NM4484NSPAXAE-3EE1 [NANYA]

MCP Specification;
NM4484NSPAXAE-3EE1
型号: NM4484NSPAXAE-3EE1
厂家: Nanya Technology Corporation.    Nanya Technology Corporation.
描述:

MCP Specification

文件: 总381页 (文件大小:12774K)
中文:  中文翻译
下载:  下载PDF数据表文档文件
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MCP Specification  
4Gb SLC NAND Flash (X8) + 4Gb LPDDR4X (X16)  
Nanya Technology Corporation  
Version 1.0  
1
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
NTC has the rights to change any specifications or product without notification.  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Ordering Information  
MCP  
NAND  
DRAM  
Density Program Erase  
Density  
(Org.)  
Part Number  
Type  
Type  
Speed RL  
3733 32  
(Org.)  
Time  
Time  
4Gb  
(512Mb X 8)  
4Gb  
(256Mb X 16)  
NM4484NSPAXAEꢀ3EE1 SLC  
300ꢁs 3.5ms LPDDR4X  
Note 1 Please confirm with NTC for the available schedule.  
Version 1.0  
12/2018  
2
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
NANYA MCP Part Numbering Guide  
NM  
4
48  
4N  
E
S
P
AX  
A
E
3
E
Grade  
E = Wide Temperature  
NANYA  
MCP  
DRAM Speed  
E = 3733Mbps @ RL=32  
Product Family  
4 =SLC NAND + LPDDR4X  
NAND Organization  
(Density, Config)  
48= 4Gb x8  
NAND Speed  
3 = 300μs  
DRAM Organization  
(Density, Config)  
4N= 4Gb x16  
Package  
E= 149ball FBGA  
Device Version  
A = 1st version  
NAND Voltage  
S= 1.8V  
Reserve Code  
AX=Default  
Interface & Power (VDD1 , VDD2 , VDDQ  
)
P = LVSTL (1.8V, 1.1V, 0.6V)  
Version 1.0  
3
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Features  
MCP  
Separate SLC NAND and LPDDR4X RAM interfaces  
Leadꢀfree (RoHS compliant) and Halogenꢀfree Package : 149ꢀball FBGA 8.00 x 9.50 x 0.80 (mm)  
Operating temperature range: –40°C to +85°C  
4Gb X8 SLC NAND  
Voltage Supply(VCC/VCCQ): 1.70V ~ 1.95V  
4Gb X16 LPDDR4X  
Speed, Addressing and Retention Specification  
Organization  
Items  
Speed Grade  
Number of Banks  
Bank Address  
Row  
256Mb x 16_1ch  
3733 / RL = 32  
8
ꢀ X8 Memory Cell Array: 4352 x 128K x 8  
ꢀ X8 Register: 4352 x 8  
ꢀ X8 Page Program: 4352 Bytes  
ꢀ X8 Block Erase: (256K + 16K) Bytes  
Modes  
BA[2:0]  
R[14:0]  
Column  
C[9:0]  
3.9  
tREFI (us)  
Read, Reset, Auto Page Program, Auto Block Erase,  
Status Read, Page Copy, Multi Page Program,  
Multi Block Erase, Multi Page Copy, Multi Page Read  
Mode control  
Basis LPDDR4X Compliant  
ꢀ Low Power Consumption  
ꢀ 16n Prefetch Architecture and BL16, BL32 (OTF)  
LVSTL Interface and Power Supply  
ꢀ Serial input/output  
ꢀ Command control  
ꢀ VDD1/VDD2/VDDQ = 1.8V/1.1V/0.6V  
Signal Integrity  
Number of valid blocks  
ꢀ Internal VREF and VREF Training  
ꢀ Configurable DS for System Compatibility  
ꢀ Configurable OnꢀDie Termination  
ꢀ ZQ Calibration for DS/ODT Impedance Accuracy Via  
External ZQ Pad (240ꢂ± 1%)  
ꢀ Min 2008 blocks  
ꢀ Max 2048 blocks  
Access time  
ꢀ Cell array to register: 25ꢁs max  
ꢀ Serial Read Cycle: 25ns min (CL=30pF)  
Program/Erase time  
ꢀ Data Bus Inversion(DBI)  
Training for Signals’ Synchronization  
ꢀ DQ Calibration Offering Specific DQ Output Patterns  
ꢀ CA Calibration  
ꢀ Auto Page Program: 300ꢁs/page typical  
ꢀ Auto Block Erase: 3.5ms/block typical  
Operating current  
ꢀ Write Leveling  
Data Integrity  
ꢀ Read (25ns cycle): 30 mA max  
ꢀ Program (avg.): 30 mA max  
ꢀ DRAM builtꢀin Temperature Sensor for Temperature  
Compensated Self Refresh (TCSR)  
ꢀ Auto Refresh and Self Refresh Modes  
ꢀ Erase (avg.): 30 mA max  
Power Saving Modes  
ꢀ Standby: 50 ꢁA max  
8 bit ECC for each 512 Bytes is required.  
ꢀ Partial Array Self Refresh (PASR)  
ꢀ Frequency Set Point(WR/OP)  
ꢀ Clockꢀstop capability  
Programmable Function  
ꢀ RON (Typical:40/48/60/80/120/240)  
ꢀ RTT (40/48/60/80/120/240)  
ꢀ RL/WL Select (Set A / Set B)  
ꢀ nWR (X16 mode / X8 Byte mode)  
ꢀ PASR (bank/segment)  
nWR (3/4/5/6/7/8)  
Version 1.0  
4
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Ball Assignment – (149b Flash X 8 + DRAM X 16)  
Part Number: NM4484NSPAXAEꢀXXX  
Top View, A1 in Top Left Corner  
1
2
3
4
5
6
7
8
9
10  
11  
12  
13  
14  
A
B
C
D
E
F
DNU  
DNU  
NC  
DNU  
NC  
DNU  
NC  
DNU  
DNU  
NC  
A
B
C
D
E
F
NC  
NC  
NC  
NC  
NC  
NC  
NC  
WE  
RE  
NC  
NC  
I/O 7  
VSS  
I/O 2  
VSS  
ODT_CA  
VSS  
VSS  
CA1  
CA4  
CA3  
CA2  
VCC  
I/O 6  
VSS  
I/O 5  
VSS  
NC  
NC  
R/  
B
VSS  
VSS  
VSS  
ALE  
VSS  
CLE  
VCC  
I/O 1  
VCC  
I/O 3  
NC  
WP  
NC  
NC  
NC  
I/O 4  
VCC  
I/O 0  
NC  
CE  
VDD2 VDD2 VDD2  
VSS VSS DQS1  
DQ10 VDD2 DQ8  
DQ9  
G
H
J
DQ11 VDDQ VDDQ VSS DQ12 VDDQ  
G
H
J
DQS1  
DMI1  
VSS VDDQ DQ14 VSS DQ15 VDDQ  
NC  
VSS  
VSS  
NC  
CK  
DQ13 VSS  
VSS  
VSS VDD2 VDD2 VDD2  
CA0  
VSS  
VSS  
VSS  
VSS  
CK  
NC  
K
L
K
L
CS  
CKE  
M
N
P
R
T
DQ3  
DQ2  
DQ1  
VSS  
VSS  
DMI0  
VSS  
VSS  
DQ5  
DQ6  
VSS  
DQ4  
VSS  
VSS  
CA5  
M
N
P
R
T
DQS0  
RESET  
NC  
DQ7 DQS0  
VSS VDD2  
DQ0 VDDQ VSS  
VDD2 VDD2 VDD1  
ZQ  
DNU VDD1 VDD2 VDDQ VDDQ VDD2 VDD1  
VDDQ VDDQ VDD1 DNU  
DNU  
1
DNU  
2
DNU  
13  
DNU  
14  
3
4
5
6
7
8
9
10  
11  
12  
Version 1.0  
5
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Package Outline Drawing (8.00mm x 9.50mm x 0.80mm)  
Unit: mm  
* BSC (Basic Spacing between Center)  
Version 1.0  
6
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Ball Description ꢀ 4Gb X8 SLC NAND  
Symbol  
Type  
Function  
Data Bus: The I/O0 to 7 pins are used as a port for transferring address, command and input/output data to  
X8: I/O[7:0]  
Input/output  
and from the device.  
Command Latch Enable: The CLE input signal is used to control loading of the operation mode command  
into the internal command register. The command is latched into the command register from the I/O port on  
the rising edge of the WE signal while CLE is High.  
CLE  
ALE  
CE  
Input  
Input  
Input  
Input  
Address Latch Enable: The ALE signal is used to control loading address information into the internal  
address register. Address information is latched into the address register from the I/O ports on the rising  
edge of WE while ALE is High.  
Chip Enable: The device goes into a lowꢀpower Standby mode when CE goes High during the device is in  
Ready state. The CE signal is ignored when device is in Busy state ( RY / BY = L), such as during a Program  
or Erase or Read operation, and will not enter Standby mode even if the CE input goes High.  
Read Enable: The RE signal controls serial data output. Data is available tREA after the falling edge of RE.  
RE  
The internal column address counter is also incremented (Address = Address +1) on this falling edge.  
WE  
WP  
Input  
Input  
Write Enable: The WE signal is used to control the acquisition of data from the I/O port.  
Write Protect: The WP signal is used to protect the device from accidental programming or erasing. The  
internal voltage regulator is reset when WP is Low. This signal is usually used protecting the data during the  
powerꢀon/off sequence when input signals are invalid.  
Ready / Busy Output: The RY / BY output signal is used to indicate the operation condition of the device.  
The RY / BY signal is in Busy state ( RY / BY =L) during the Program, Erase and Read operations and will  
return to Ready state ( RY / BY =H) afeter completion of the operation. The output buffer for this signal is an  
open drain and has to be pulledꢀup to Vccq with an appropriate resister.  
RY/BY  
Output  
If RY / BY signal is not pulledꢀup to Vccq ( “Open” state ), device operation cannot guarantee.  
VCC  
VSS  
NC  
Supply  
Supply  
Power  
Ground  
No Connect: These pins should be left unconnected.  
Version 1.0  
12/2018  
7
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Ball Description ꢀ 4Gb X16 LPDDR4X  
Symbol  
Type  
Function  
Clock: CK and CK are differential clock inputs. All address, command, and control input signals are  
sampled on the crossing of the positive edge of CK and the negative edge of CK. AC timings for CA  
parameters are referenced to CK.  
CK , CK  
Input  
Clock Enable: CKE HIGH activates and CKE LOW deactivates the internal clock circuits, input buffers,  
and output drivers. Powerꢀsaving modes are entered and exited via CKE transitions.  
CKE is part of the command code.  
CKE  
Input  
CS  
Input  
Input  
Chip Select: CS is part of the command code.  
Command/Address Inputs: CA signals provide the Command and Address inputs according to  
the Command Truth Table.  
CA[5:0]  
Data Mask Inversion: DMI is a biꢀdirectional signal which is driven HIGH when the data on the data bus is  
inverted, or driven LOW when the data is in its normal state. Data Inversion can be disabled via a mode  
register setting. Each byte of data has a DMI signal. Each channel (A & B) has its own DMI signals. This  
signal is also used along with the DQ signals to provide write data masking information to the DRAM. The  
DMI pin function ꢀ Data Inversion or Data mask ꢀ depends on Mode Register setting.  
DMI[1:0]  
DQ[15:0]  
Input/output  
Input/output  
Input/output  
Data Bus: Biꢀdirection data bus.  
Data Strobe: DQS and DQS are biꢀdirectional differential output clock signals used to strobe data during a  
READ or WRITE. The Data Strobe is generated by the DRAM for a READ and is edgeꢀaligned with Data.  
The Data Strobe is generated by the Memory Controller for a WRITE and must arrive prior to Data. Each  
byte of data has a Data Strobe signal pair..  
DQS[1:0]  
DQS[1:0]  
CA ODT Control: The ODT_CA pin is ignored by LPDDR4X devices. ODTꢀCS/CA/CK function is fully  
ODT_CA  
ZQ  
Input  
controlled through MR11 and MR22. The ODT_CA pin shall be connected to either VDD2 or VSS.  
Calibration Reference. Used to calibrate the output drive strength and the termination resistance. There  
Reference  
is one ZQ pin per die. The ZQ pin shall be connected to VDDQ through a 240ꢂ ± 1% resistor.  
Supply  
Supply  
Supply  
GND  
VDD1  
VDD2  
VDDQ  
VSS  
Power Supply 1: Core power supply  
Power Supply 2: Core power supply  
DQ Power Supply: Isolated on the die for improved noise immunity.  
Ground  
RESET  
Input  
Reset: When asserted LOW, the RESET signal resets both channels of the die.  
NOTE 1: The signal may show up in a different symbol but it indicates to the same thing. e.g., /CK = CK# = CK = CKb, /DQS =  
DQS# = DQS = DQSb, /CS = CS# = CS = CSb.  
Version 1.0  
8
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Functional Block Diagram  
VDD1 VDD2 VDDQ  
VDDQ  
Vss  
ZQ  
RZQ  
RESET  
CS  
CKE  
CK  
4Gb LPDDR4X  
CK  
CA[5:0]  
ODT_CA  
X16  
DMI[1:0]  
DQ[15:0]  
DQS[1:0]  
DQS[1:0]  
NAND  
CLE  
ALE  
CE  
4Gb NAND Flash  
RE  
X8:I/O[7:0]  
WE  
WP  
R/B  
VCC VSS  
Version 1.0  
12/2018  
9
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
4Gb(X8) SLC NAND Flash  
Version 1.0  
10  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Descriptions  
The device is a single 1.8V 4Gbit (4,563,402,752 bits) NAND Electrically Erasable and Programmable ReadꢀOnly  
2
Memory (NAND E PROM) organized as X8: (4096 +256) bytes x 64 pages x 2048 blocks.  
The device has two 4352ꢀbytes static registers which allow program and read data to be transferred between the  
register ad the memory cell array in 4352ꢀbytes increments. The Erase operation is implemented in a single block  
unit (X8=256 Kbytes + 16 Kbytes: 4352 bytes x 64 pages).  
The device is a serialꢀtype memory device which utilizes the I/O pins for both address and data input/output as well  
as for command inputs. The Erase and Program operations are automatically executed making the device most  
suitable for applications such as solidꢀstate file storage, voice recording, image file memory for still cameras and  
other systems which require highꢀdensity nonꢀvolatile memory data storage.  
Version 1.0  
11  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Function Block Diagram (X8)  
VCC  
Status register  
Address register  
Command register  
VSS  
I/O 0  
Column buffer  
Column decoder  
Data register  
I/O  
to  
Control circuit  
I/O 7  
Command register  
CE  
CLE  
ALE  
Memory cell array  
Logic control  
Control circuit  
RE  
WE  
WP  
RY/BY  
RY/BY  
HV generator  
Schematic Cell Layout and Address Assignment (X8)  
The Program operation works on page units while the Erase operation works on block units  
A page consists of 4352 bytes in which 4096 bytes are used for  
main memory storage and 256 bytes are for redundancy or for  
other uses.  
I/O 0  
I/O 7  
1 Page = 4352 Bytes  
1 Block = 4352 Bytes x 64 Pages = (256K + 16K) Bytes  
Capacity = 4352 Bytes x 64 Pages x 2048 Blocks  
1 Block = 64 Pages  
(256K + 16K) Bytes  
Array Address (X8)  
I/O 0  
I/O 1  
I/O 2  
I/O 3  
I/O 4  
I/O 5  
I/O 6  
I/O 7  
Address  
1st cycle  
2nd cycle  
3rd cycle  
4th cycle  
5th cycle  
CA0  
CA8  
PA0  
PA8  
PA16  
CA1  
CA9  
PA1  
PA9  
L
CA2  
CA10  
PA2  
PA10  
L
CA3  
CA11  
PA3  
PA11  
L
CA4  
CA12  
PA4  
PA12  
L
CA5  
L
CA6  
L
CA7  
L
Column Address  
Column Address  
Page Address  
Page Address  
Page Address  
PA5  
PA13  
L
PA6  
PA14  
L
PA7  
PA15  
L
PA6 to PA16: Block address  
PA0 to PA5: NAND address in block  
Version 1.0  
12/2018  
12  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Absolute Maximum Ratings  
Symbol  
Rating  
Value  
ꢀ0.6 to +2.5  
ꢀ0.6 to +2.5  
ꢀ0.6 to Vcc + 0.3 (2.5V)  
0.3  
Unit  
VCC  
Power Supply Voltage  
Input Voltage  
VIN  
V
VI/O  
Input / Output Voltage  
Power Dissipation  
PD  
W
oC  
oC  
TSOLDER  
TSTG  
Soldering Temperature (10 s)  
Storage Temperature  
260  
ꢀ55 to +125  
Capacitance1  
(TA=25, f=1.0MHz)  
Symbol  
CIN  
Parameter  
Input  
Test Condition  
VIN=0V  
Min  
Max  
Unit  
pF  
10  
10  
COUT  
Output  
VOUT=0V  
pF  
NOTE 1  
This parameter is periodically sampled and is not tested for every device.  
Version 1.0  
13  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Valid Blocks  
Symbol  
NVB  
Parameter  
Min  
Typ.  
Max  
2,048  
Unit  
Number of Valid Blocks  
2,008  
Blocks  
NOTE 1  
The device occasionally contains unusable blocks.  
The first block (Block 0) is guaranteed to be a valid block at the time of shipment.  
The specification for the minimum number of valid blocks is applicable over lifetime.  
The number of valid blocks is on the basis of single plane operations, and this may be decreased with two  
plane operations.  
Recommended DC Operating Conditions  
Symbol  
Parameter  
Min  
Typ.  
Max  
1.95  
Unit  
V
VCC  
Power Supply Voltage  
1.70  
VIH  
High Level Input Voltage  
Low Level Input Voltage  
VCC x 0.8  
ꢀ0.31  
VCC + 0.3  
VCC x 0.2  
V
VIL  
V
Note 1  
-2V (pulse width lower than 20 ns)  
DC and Operation Characteristics  
(Ta= ꢀ40 to 85℃, VCC=1.70 to 1.95V )  
Symbol  
IIL  
Parameter  
Input Leakage Current  
Output Leakage Current  
Serial Read Current  
Test Conditions  
Min  
Typ.  
Max  
Unit  
ꢁA  
VIN=0 to VCC  
±10  
±10  
30  
ꢁA  
ILO  
VOUT=0 to VCC  
mA  
mA  
ICCO1  
ICCO2  
ICCO3  
ICCS  
CE=VIL,IOUT= 0 mA, tcycle=25ns  
Programming Current  
Erasing Current  
30  
30  
mA  
ꢁA  
V
Standby Current  
CE = VCC ꢀ 0.2 V, WP = 0 V/VCC  
IOH = ꢀ0.1mA  
50  
VOH  
VOL  
High Level Output Voltage  
Low Level Output Voltage  
VCC ꢀ 0.2  
4
IOL = 0.1mA  
0.2  
V
IOL (RY/BY) Output Current of (RY/BY) pin  
VOL=0.2V  
mA  
Version 1.0  
12/2018  
14  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AC Characteristics and Recommended Operating Conditions  
(Ta= ꢀ40 to 85℃, VCC=1.70 to 1.95V)  
Symbol  
Parameter  
Min  
Max  
Unit  
tCLS  
tCLH  
tCS  
CLE Setup Time  
CLE Hold Time  
CE Setup Time  
CE Hold Time  
12  
5
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
20  
5
tCH  
tWP  
tALS  
tALH  
tDS  
Write Pulse Width  
ALE Setup Time  
ALE Hold Time  
Data Setup Time  
Data Hold Time  
Write Cycle Time  
WE High Hold Time  
12  
12  
5
12  
5
tDH  
tWC  
tWH  
25  
10  
AC Characteristics for Operation  
Symbol  
Parameter  
Min  
Max  
Unit  
tWW  
tRR  
WP High to WE Low  
100  
20  
20  
12  
25  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ꢁs  
ꢁs  
ꢁs  
ns  
ꢁs  
Ready to RE Falling Edge  
tRW  
Ready to WE Falling Edge  
Read Pulse Width  
tRP  
tRC  
Read Cycle Time  
tREA  
tCEA  
tCLR  
RE Access Time  
20  
CE Access Time  
25  
CLE Low to RE Low  
10  
10  
25  
5
tAR  
ALE Low to RE Low  
tRHOH  
tRLOH  
tRHZ  
tCHZ  
tCSD  
tREH  
tIR  
RE High to Output Hold Time  
RE Low to Output Hold Time  
RE High to Output High Impedance  
CE High to Output High Impedance  
CE High to ALE or CLE Don’t care  
RE High Hold Time  
60  
20  
0
10  
0
OutputꢀHighꢀimpedanceꢀtoꢀRE Falling Edge  
RE High to WE Low  
tRHW  
tWHC  
tWHR  
tR  
30  
30  
60  
WE High to CE Low  
WE High to RE Low  
25  
Memory Cell Array to Starting Address  
Data Cache Busy in Read Cache (following 31h and 3Fh)  
Data Cache Busy in Page Copy (following 3Ah)  
WE High to Busy  
tDCBSYR1  
tDCBSYR2  
tWB  
25  
30  
100  
tRST  
Device Reset Time (Ready/Read/Program/Erase)  
5/5/10/500  
NOTE 1 tCLS and tALS cannot be shorter than tWP.  
NOTE 2 tCS should be longer than tWP + 8ns.  
Version 1.0  
15  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AC Test Condition  
Condition  
VCC : 1.70 to 1.95V  
VCC – 0.2 V, 0.2 V  
3ns  
Parameter  
Input level  
Input pulse rise and fall time  
Input comparison level  
Output data comparison level  
Output Load  
Vcc / 2  
Vcc / 2  
1 TTL GATE and CL=30pF  
NOTE 1  
Busy to ready time depends on the pull-up resistor tied to the RY/BY pin.  
Programming / Erasing Characteristics  
(Ta= ꢀ40 to 85℃, VCC=1.70 to 1.95V)  
Symbol  
tPROG  
Parameter  
Average Programming Time  
Min  
Typ.  
300  
Max  
700  
10  
Unit  
ꢁs  
tDCBSYW1 Data Cache Busy Time in Write Cache (following 11h)  
tDCBSYW21 Data Cache Busy Time in Write Cache (following 15h)  
ꢁs  
700  
4
ꢁs  
N
Number of Partial Program Cycles in the Same Page  
Block Erase Time  
cycle  
ms  
tBERASE  
NOTE 1  
3.5  
10  
t
depends on the timing between internal programming time and data in time.  
DCBSYW2  
Data Output  
When tREH is long, output buffers are disabled by RE=High, and the hold time of data output depend on tRHOH  
(25ns MIN). On this condition, waveforms look like normal serial read mode.  
When tREH is short, output buffers are not disabled by RE =High, and the hold time of data output depend on  
tRLOH (5ns MIN). On this condition, output buffers are disabled by the rising edge of CLE, ALE, CE or falling  
edge of WE, and waveforms look like Extended Data Output Mode.  
Version 1.0  
16  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Operation Mode: Logic and Command Tables  
The operation modes such as Program, Erase, Read and Reset are controlled by operations shown in command  
table. Address input, command input and data input/output are controlled by the CLE, ALE, CE, WE, RE and WP  
signals, as shown in Mode Selection Table.  
Logic Table  
CLE  
ALE  
L
CE  
L
WE  
RE  
H
WP  
Mode  
Command Input  
H
L
L
L
H
H
Data Input  
L
H
L
H
Address Input  
L
L
L
H
Serial Data Output  
During Program (Busy)  
During Erase (Busy)  
H
L
H1  
H1  
H
H
During Read (Busy)  
H
L
Program, Erase Inhibit  
Stand-by  
0V/VCC  
H: VIH, L=VIL *: VIH or VIL.  
Note 1: If CE is low during read busy. WE and RE must be held High to avoid unintended command/address input to  
device or read to device. Reset or Status Read command can be input during Read Busy.  
Version 1.0  
17  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command Table  
Function  
1stCycle  
80H  
00H  
05H  
31H  
3FH  
80H  
85H  
80H  
80H  
81H  
81H  
00H  
8CH  
8CH  
60H  
90H  
70H  
71H  
FFH  
2ndCycle  
Acceptable Command during Busy  
Serial Data Input  
Read  
30H  
E0H  
Column Address Change in Serial Data Output  
Read with Data Cache  
Read Start for Last Page in Read Cycle with Data Cache  
Auto Page Program  
10H  
Column Address Change in Serial Data Input  
Auto Program with Data Cache  
15H  
11H  
15H  
10H  
3AH  
15H  
10H  
D0H  
Multi Page Program  
Read for Page Copy (2) with Data Out  
Auto Program with Data Cache during Page Copy (2)  
Auto Program for last page during Page Copy (2)  
Auto Block Erase  
ID Read  
Status Read  
O
O
O
Status Read for MultiꢀPage Program or Multi Block Erase  
Reset  
Read mode operation states  
CLE  
L
ALE  
L
CE  
L
WE  
H
RE  
L
I/O0 to I/O7  
Data output  
Power  
Active  
Active  
Output Select  
Output Deselect  
L
L
L
H
H
High impedance  
H: VIH, L=VIL  
Version 1.0  
18  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DEVICE OPERATION  
Read Mode  
Read mode is set when the "00h" and “30h” commands are issued to the Command register. Between the two  
commands, a start address for the Read mode needs to be issued. After initial power on sequence, “00h”  
command is latched into the internal command register. Therefore read operation after power on sequence is  
executed by the setting of only five address cycles and “30h” command. Refer to the figures below for the  
sequence and the block diagram (Refer to the detailed timing chart.).  
For X8 :  
A data transfer operation from the cell array to the Data Cache via Page Buffer starts on the rising edge of WE in the  
30h command input cycle (after the address information has been latched). The device will be in the Busy state  
during this transfer period.  
After the transfer period, the device returns to Ready state. Serial data can be output synchronously with the RE  
clock from the start address designated in the address input cycle.  
Version 1.0  
19  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Random Column Address Change in Read Cycle  
During the serial data output from the Data Cache, the column address can be changed by inputting a new column  
address using the 05h and E0h commands. The data is read out in serial starting at the new column address.  
Random Column Address Change operation can be done multiple times within the same page.  
Version 1.0  
20  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read Operation with Read Cahe  
The device has a Read operation with Data Cache that enables the high speed read operation shown below. When  
the block address changes, this sequence has to be started from the beginning.  
For X8 :  
If the 31h command is issued to the device, the data content of the next page is transferred to the Page Buffer  
during serial data out from the Data Cache, and therefore the tR (Data transfer from memory cell to data register)  
will be reduced.  
1 Normal read. Data is transferred from Page N to Data Cache through Page Buffer. During this time period, the device outputs Busy state for tR  
max.  
2 After the Ready/Busy returns to Ready, 31h command is issued and data is transferred to Data Cache from Page Buffer again. This data  
transfer takes tDCBSYR1 max and the completion of this time period can be detected by Ready/Busy signal.  
3 Data of Page N + 1 is transferred to Page Buffer from cell while the data of Page N in Data cache can be read out by RE clock simultaneously.  
4 The 31h command makes data of Page N + 1 transfer to Data Cache from Page Buffer after the completion of the transfer from cell to Page  
Buffer. The device outputs Busy state for tDCBSYR1 max.  
This Busy period depends on the combination of the internal data transfer time from cell to Page buffer and the serial data out time.  
5 Data of Page N + 2 is transferred to Page Buffer from cell while the data of Page N + 1 in Data cache can be read out by RE clock  
simultaneously.  
6 The 3Fh command makes the data of Page N + 2 transfer to the Data Cache from the Page Buffer after the completion of the transfer from cell  
to Page Buffer. The device outputs Busy state for tDCBSYR1 max. This Busy period depends on the combination of the internal data transfer  
time from cell to Page buffer and the serial data out time.  
7 Data of Page N + 2 in Data Cache can be read out, but since the 3Fh command does not transfer the data from the memory cell to Page Buffer,  
the device can accept new command input immediately after the completion of serial data out.  
Version 1.0  
21  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Multi Page Read Operation  
The device has a Multi Page Read operation and Multi Page Read with Data Cache operation.  
(1) Multi Page Read without Data Cache  
The sequence of command and address input is shown below.  
Same page address (PA0 to PA5) within each district has to be selected.  
For X8 :  
The data transfer operation from the cell array to the Data Cache via Page Buffer starts on the rising edge of WE in  
the 30h command input cycle (after the 2 Districts address information has been latched). The device will be in the  
Busy state during this transfer period.  
After the transfer period, the device returns to Ready state. Serial data can be output synchronously with the RE  
clock from the start address designated in the address input cycle.  
Version 1.0  
22  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(2) Multi Page Read with Data Cache  
When the block address changes (increments) this sequenced has to be started from the beginning.  
The sequence of command and address input is shown below.  
Same page address (PA0 to PA5) within each district has to be selected.  
For X8 :  
Version 1.0  
23  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(3) Notes  
(a) Internal addressing in relation with the Districts  
To use Multi Page Read operation, the internal addressing should be considered in relation with the District.  
• The device consists from 2 Districts.  
• Each District consists from 1024 erase blocks.  
• The allocation rule is follows.  
District 0: Block 0, Block 2, Block 4, Block 6,———, Block 2046  
District 1: Block 1, Block 3, Block 5, Block 7,———, Block 2047  
(b) Address input restriction for the Multi Page Read operation  
There are following restrictions in using Multi Page Read;  
(Restriction)  
Maximum one block should be selected from each District.  
Same page address (PA0 to PA5) within two districts has to be selected.  
For example;  
(60) [District 0, Page Address 0x00000] (60) [District 1, Page Address 0x00040] (30)  
(60) [District 0, Page Address 0x00001] (60) [District 1, Page Address 0x00041] (30)  
(Acceptance)  
There is no order limitation of the District for the address input.  
For example, following operation is accepted;  
(60) [District 0] (60) [District 1] (30)  
(60) [District 1] (60) [District 0] (30)  
It requires no mutual address relation between the selected blocks from each District.  
(c) WP signal  
Make sure WP is held to High level when Multi Page Read operation is performed  
Version 1.0  
12/2018  
24  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Auto Page Program Operation  
The device carries out an Automatic Page Program operation when it receives a "10h" Program command  
after the address and data have been input. The sequence of command, address and data input is shown below.  
(Refer to the detailed timing chart.)  
The data is transferred (programmed) from the Data Cache via the Page Buffer to the selected page on the rising  
edge of WE following input of the “10h” command. After programming, the programmed data is transferred back to  
the Page Buffer to be automatically verified by the device.  
If the programming does not succeed, the Program/Verify operation is repeated by the device until success is  
achieved or until the maximum loop number set in the device is reached.  
Version 1.0  
25  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Random Column Address Change in Auto Page Program Operation  
The column address can be changed by the 85h command during the data input sequence of the Auto Page  
Program operation.  
Two address input cycles after the 85h command are recognized as a new column address for the data input. After  
the new data is input to the new column address, the 10h command initiates the actual data program into the  
selected page automatically. The Random Column Address Change operation can be repeated multiple times within  
the same page.  
Version 1.0  
26  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Multi Page Program  
The device has a Multi Page Program, which enables even higher speed program operation compared to Auto Page  
Program. The sequence of command, address and data input is shown below. (Refer to the detailed timing chart.)  
Although two planes are programmed simultaneously, pass/fail is not available for each page by "70h" command  
when the program operation completes. Status bit of I/O 1 is set to “1” when any of the pages fails. Limitation in  
addressing with Multi Page Program is shown below.  
For X8 :  
NOTE: Any command between 11h and 81h is prohibited except 70h and FFh.  
Version 1.0  
27  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Auto Page Program Operation with Data Cache  
The device has an Auto Page Program with Data Cache operation enabling the high speed program operation  
shown below. When the block address changes this  
sequenced has to be started from the beginning.  
Issuing the 15h command to the device after serial data input initiates the program operation with Data Cache  
1 Data for Page N is input to Data Cache.  
2 Data is transferred to the Page Buffer by the 15h command. During the transfer the Ready/Busy outputs Busy State (tDCBSYW2).  
3 Data is programmed to the selected page while the data for page N + 1 is input to the Data Cache.  
4 By the 15h command, the data in the Data Cache is transferred to the Page Buffer after the programming of page N is completed. The device  
output busy state from the 15h command until the Data Cache becomes empty. The duration of this period depends on timing between the  
internal programming of page N and serial data input for Page N + 1 (tDCBSYW2).  
5 Data for Page N + P is input to the Data Cache while the data of the Page N + P 1 is being programmed.  
6 The programming with Data Cache is terminated by the 10h command. When the device becomes Ready, it shows that the internal  
programming of the Page N + P is completed.  
NOTE: Since the last page programming by the 10h command is initiated after the previous cache program, the tPROG during cache  
programming is given by the following;  
tPROG =tPROG for the last page +tPROG of the previous page ( command input cycle +address input cycle +data input cycle time of the last  
page)  
Version 1.0  
28  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Pass/fail status for each page programmed by the Auto Page Programming with Data Cache operation can be detected by the Status Read  
operation.  
I/O1 : Pass/fail of the current page program operation.  
I/O2 : Pass/fail of the previous page program operation.  
The Pass/Fail status on I/O1 and I/O2 are valid under the following conditions.  
Status on I/O1: Page Buffer Ready/Busy is Ready State.  
The Page Buffer Ready/Busy is output on I/O6 by Status Read operation or RY / BY pin after the 10h command  
Status on I/O2: Data Cache Read/Busy is Ready State.  
The Data Cache Ready/Busy is output on I/O7 by Status Read operation or RY / BY pin after the 15h command.  
If the Page Buffer Busy returns to Ready before the next 80h command input, and if Status Read is done during  
this Ready period, the Status Read provides pass/fail for Page 2 on I/O1 and pass/fail result for Page1 on I/O2  
Version 1.0  
29  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Multi Page Program with Data Cache  
The device has a Multi Page Program with Data Cache operation, which enables even higher speed program  
operation compared to Auto Page Program with Data Cache as shown below. When the block address changes  
(increments) this sequenced has to be started from the beginning.  
The sequence of command, address and data input is shown below. (Refer to the detailed timing chart.)  
For X8 :  
After “15h” or “10h” Program command is input to device, physical programing starts as follows. For details  
of Auto Program with Data Cache, refer to “Auto Page Program with Data Cache”.  
The data is transferred (programmed) from the page buffer to the selected page on the rising edge of WE following input of the “15h” or “10h”  
command. After programming, the programmed data is transferred back to the register to be automatically verified by the device. If the  
programming does not succeed, the Program/Verify operation is repeated by the device until success is achieved or until the maximum loop  
number set in the device is reached.  
Version 1.0  
30  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Starting the above operation from 1st page of the selected erase blocks, and then repeating the operation  
total 64 times with incrementing the page address in the blocks, and then input the last page data of the  
blocks, “10h” command executes final programming. Make sure to terminate with 81hꢀ10hꢀ command  
sequence.  
In this full sequence, the command sequence is following.  
After the “15h” or “10h” command, the results of the above operation is shown through the “71h”Status Read  
command.  
The 71H Command Status Description  
I/O  
Status  
Output  
Pass : 0 / Fail : 1  
I/O 0  
I/O 1  
I/O 2  
I/O 3  
I/O 4  
Chip Status1 : Pass / Fail  
Pass : 0 / Fail : 1  
Pass : 0 / Fail : 1  
Pass : 0 / Fail : 1  
Pass : 0 / Fail : 1  
District 0 Chip Status1 : Pass / Fail  
District 1 Chip Status2 : Pass / Fail  
District 0 Chip Status1 : Pass / Fail  
District 1 Chip Status2 : Pass / Fail  
I/O 5  
I/O 6  
I/O 7  
Ready / Busy  
Data Cache Ready / Busy  
Write Protect  
Busy : 0 / Ready : 1  
Busy : 0 / Ready : 1  
Protected : 0 / Not Protected : 1  
I/O0 describes Pass/Fail condition of district 0 and 1 (OR data of I/O1 and I/O2). If one of the districts fails during multi page  
program operation, it shows “Fail”.  
I/O1 to I/O4 shows the Pass/Fail condition of each district. For details on “Chip Status 1” and “Chip Status2” refer to  
section “Status Read”.  
Version 1.0  
31  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Internal addressing in relation with the Districts  
To use Multi Page Program operation, the internal addressing should be considered in relation with the  
District.  
The device consists from 2 Districts.  
Each District consists from 1024 erase blocks.  
The allocation rule is follows.  
District 0: Block 0, Block 2, Block 4, Block 6,———, Block 2046  
District 1: Block 1, Block 3, Block 5, Block 7,———, Block 2047  
Address input restriction for the Multi Page Program with Data Cache operation  
There are following restrictions in using Multi Page Program with Data Cache;  
(Restriction)  
Maximum one block should be selected from each District.  
Same page address (PA0 to PA5) within two districts has to be selected.  
For example;  
(80) [District 0, Page Address 0x00000] (11) (81) [District 1, Page Address 0x00040] (15 or 10)  
(80) [District 0, Page Address 0x00001] (11) (81) [District 1, Page Address 0x00041] (15 or 10)  
(Acceptance)  
There is no order limitation of the District for the address input.  
For example, following operation is accepted;  
(80) [District 0] (11) (81) [District 1] (15 or 10)  
(80) [District 1] (11) (81) [District 0] (15 or 10)  
It requires no mutual address relation between the selected blocks from each District.  
Operating restriction during the Multi Page Program with Data Cache operation  
(Restriction)  
The operation has to be terminated with “10h” command.  
Once the operation is started, no commands other than the commands shown in the timing diagram is allowed  
to be input except for Status Read command and reset command.  
Version 1.0  
32  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Page Copy (2)  
By using Page Copy (2), data in a page can be copied to another page after the data has been read out.  
When the block address changes (increments) this sequenced has to be started from the beginning.  
For X8 :  
Page  
Copy (2) operation is as following.  
1 Data for Page N is transferred to the Data Cache.  
2 Data for Page N is read out.  
3 Copy Page address M is input and if the data needs to be changed, changed data is input.  
4 Data Cache for Page M is transferred to the Page Buffer.  
5 After the Ready state, Data for Page N + P1 is output from the Data Cache while the data of Page M is being programmed.  
Version 1.0  
33  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
For X8 :  
6 Copy Page address (M + R1) is input and if the data needs to be changed, changed data is input.  
7 After programming of page M is completed, Data Cache for Page M + R1 is transferred to the Page Buffer.  
8 By the 15h command, the data in the Page Buffer is programmed to Page M + R1. Data for Page N + P2 is transferred to the Data cache.  
9 The data in the Page Buffer is programmed to Page M + Rn − 1. Data for Page N + Pn is transferred to the Data Cache.  
Version 1.0  
34  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
For X8 :  
10 Copy Page address (M + Rn) is input and if the data needs to be changed, changed data is input.  
11 By issuing the 10h command, the data in the Page Buffer is programmed to Page M + Rn.  
(*1) Since the last page programming by the 10h command is initiated after the previous cache program, the tPROG here will be expected as the  
following,  
tPROG = tPROG of the last page + tPROG of the previous page − ( command input cycle + address input cycle + data output/input cycle time of  
the last page)  
NOTE) This operation needs to be executed within Districtꢀ0 or Districtꢀ1.  
Data input is required only if previous data output needs to be altered.  
If the data has to be changed, locate the desired address with the column and page address input after the 8Ch command, and change  
only the data that needs be changed.  
If the data does not have to be changed, data input cycles are not required.  
Make sure WP is held to High level when Page Copy (2) operation is performed.  
Also make sure the Page Copy operation is terminated with 8Chꢀ10h command sequence  
Version 1.0  
35  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mutil Page Copy (2)  
By using Multi Page Copy (2), data in two pages can be copied to other pages after the data has been read out.  
When each block address changes (increments) this sequence has to be started from the beginning.  
Same page address (PA0 to PA5) within two districts has to be selected.  
For X8 :  
Version 1.0  
36  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Auto Block Erase  
The Auto Block Erase operation starts on the rising edge of WE after the Erase Start command “D0h” which  
follows the Erase Setup command “60h”. This twoꢀcycle process for Erase operations acts as an extra layer of  
protection from accidental erasure of data due to external noise. The device automatically executes the Erase  
and Verify operations.  
Multi Block Erase  
The Multi Block Erase operation starts by selecting two block addresses before D0h command as in below  
diagram. The device automatically executes the Erase and Verify operations and the result can be monitored by  
checking the status by 71h status read command. For details on 71h status read command, refer to section  
Multi Page Program with Data Cache”.  
Version 1.0  
37  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Internal addressing in relation with the Districts  
To use Multi Block Erase operation, the internal addressing should be considered in relation with the District.  
The device consists from 2 Districts.  
Each District consists from 1024 erase blocks.  
The allocation rule is follows.  
District 0: Block 0, Block 2, Block 4, Block 6,———, Block 2046  
District 1: Block 1, Block 3, Block 5, Block 7,———, Block 2047  
Address input restriction for the Multi Block Erase  
There are following restrictions in using Multi Block Erase  
(Restriction)  
Maximum one block should be selected from each District.  
For example;  
(60) [District 0] (60) [District 1] (D0)  
(Acceptance)  
There is no order limitation of the District for the address input.  
For example, following operation is accepted;  
(60) [District 1] (60) [District 0] (D0)  
It requires no mutual address relation between the selected blocks from each District.  
Make sure to terminate the operation with D0h command. If the operation needs to be terminated before D0h  
command input, input the FFh reset command to terminate the operation.  
Version 1.0  
38  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ID Read  
The device contains ID codes which can be used to identify the device type, the manufacturer, and features of the  
device. The ID codes can be read out under the following timing conditions:  
ID Definition Table (X8)  
I/O7 I/O6 I/O5 I/O4 I/O3 I/O2 I/O1 I/O0 Hex Data  
Description  
1st Data  
2nd Data  
3rd Data  
4th Data  
5th Data  
Maker Code  
Device Code  
1
1
1
0
0
0
0
0
0
1
0
1
0
1
1
1
0
1
0
1
1
1
0
0
0
0
1
0
1
1
0
0
0
1
1
0
0
0
0
0
98H  
ACH  
90H  
26H  
76H  
Chip Number, Cell Type  
Page Size, Block Size, I/O Width  
Plane Number  
Version 1.0  
39  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
3rd ID Data  
Item  
Description  
I/O7  
I/O6  
I/O5  
I/O4  
I/O3  
I/O2  
I/O1  
I/O0  
1
2
4
8
0
0
1
1
0
1
0
1
Internal Chip Number  
2 Level Cell  
4 Level Cell  
8 Level Cell  
16 Level Cell  
0
0
1
1
0
1
0
1
Cell Type  
Reserved  
1
0
0
1
4th ID Data  
Item  
Description  
I/O7  
I/O6  
I/O5  
I/O4  
I/O3  
I/O2  
I/O1  
I/O0  
1 KB  
2 KB  
4 KB  
0
0
1
1
0
1
0
1
Page Size  
(without redundant area)  
8 KB  
64 KB  
128 KB  
256 KB  
512 KB  
X8  
0
0
1
1
0
1
0
1
Block Size  
(without redundant area)  
0
1
I/O Width  
Reserved  
X16  
0
0
1
5th ID Data  
Item  
Description  
I/O7  
I/O6  
I/O5  
I/O4  
I/O3  
I/O2  
I/O1  
I/O0  
1 Plane  
2 Plane  
4 Plane  
8 Plane  
0
0
1
1
0
1
0
1
Plane Number  
Reserved  
0
1
1
1
1
0
Version 1.0  
40  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Status Read  
The device automatically implements the execution and verification of the Program and Erase operations. The  
status Read function is used to monitor the Ready/Busy status of the device, determine the result (pass/fail) of a  
Program or Erase operation, and determine whether the device is in Protect mode. The device status is output via  
the I/O port using RE after a “70h” command input. The Status Read can also be used during a Read operation to  
find out the Ready/Busy status.  
Status Register Definition for 70H Command  
I/O  
Page Program  
Block Erase  
Read  
Cache Read  
Cache Program  
Definition  
Chip Status1  
Pass : 0 / Fail : 1  
Chip Status2  
I/O 0  
Pass / Fail  
Pass / Fail  
Invalid  
Invalid  
Pass / Fail  
I/O 1  
Invalid  
Invalid  
Invalid  
Invalid  
Pass / Fail  
Pass : 0 / Fail : 1  
I/O 2  
I/O 3  
I/O 4  
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Not Used  
Not Used  
Not Used  
Page Buffer  
Busy : 0 / Ready : 1  
ctData Cache  
I/O 5  
I/O 6  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Ready / Busy  
Busy : 0 / Ready : 1  
Write Prot  
I/O 7  
Write Protect  
Write Protect  
Write Protect  
Write Protect  
Write Protect  
Protected : 0  
/ Not Protected : 1  
NOTE  
The Pass/Fail status on I/O0 and I/O1 is only valid during a Program/Erase operation when the device is in the  
Ready state.  
Chip Status 1:  
During a Auto Page Program or Auto Block Erase operation this bit indicates the pass/fail result.  
During a Auto Page Programming with Data Cache operation, this bit shows the pass/fail results of the current page  
program operation, and therefore this bit is only valid when I/O5 shows the Ready state.  
Chip Status 2:  
This bit shows the pass/fail result of the previous page program operation during Auto Page Programming with Data  
Cache. This status is valid when I/O6 shows the Ready State.  
The status output on the I/O5 is the same as that of I/O6 if the command input just before the 70h is not 15h or 31h.  
Version 1.0  
41  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
An application example with multiple devices is shown in the figure below.  
System Design Note: If the RY / BY pin signals from multiple devices are wired together as shown in the  
diagram, the Status Read function can be used to determine the status of each individual device.  
Reset  
The Reset mode stops all operations. For example, in case of a Program or Erase operation, the internally  
generated voltage is discharged to 0 volt and the device enters the Wait state.  
Reset during a Cache Program/Page Copy may not just stop the most recent page program but it may also  
stop the previous program to a page depending on when the FF reset is input.  
The response to a “FFh” Reset command input during the various device operations is as follows:  
When a Reset (FFh) command is input during programming  
Version 1.0  
42  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
When a Reset (FFh) command is input during erasing  
When a Reset (FFh) command is input during Read operation  
When a Reset (FFh) command is input during Ready  
When a Status Read command (70h) is input after a Reset  
When two or more Reset commands are input in succession  
Version 1.0  
43  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
APPLICATION NOTES AND COMMENTS  
(1)Powerꢀon/off sequence  
The timing sequence shown in the figure below is necessary for the powerꢀon/off sequence.  
The device internal initialization starts after the power supply reaches an appropriate level in the power on sequence.  
During the initialization the device Ready/Busy signal indicates the Busy state as shown in the figure below. In this  
time period, the acceptable commands are FFh or 70h.  
The WP signal is useful for protecting against data corruption at powerꢀon/off.  
(2)Do not turn off the power before write/erase operation is complete. Avoid using the device when the battery  
is low. Power shortage and/or power failure before write/erase operation is complete will cause loss of data  
and/or damage to data.  
(3)Powerꢀon Reset  
The following sequence is necessary because some input signals may not be stable at powerꢀon.  
(4)Prohibition of unspecified commands  
The operation commands are listed in Logic Table. Input of a command other than those specified in Logic Table is  
prohibited. Stored data may be corrupted if an unknown command is entered during the command cycle.  
(5)Restriction of commands while in the Busy state  
During the Busy state, do not input any command except 70h(71h) and FFh.  
(6)Acceptable commands after Serial Input command “80h”  
Once the Serial Input command “80h” has been input, do not input any command other than the Column  
Version 1.0  
44  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Address Change in Serial Data Input command “85h”, Auto Program command “10h”, Multi Page Program  
command “11h”, Auto Program with Data Cache Command “15h”, or the Reset command “FFh”.  
If a command other than “85h” , “10h” , “11h” , “15h” or “FFh” is input, the Program operation is not performed and  
the device operation is set to the mode which the input command specifies.  
(7)Addressing for Program Operation  
Within a block, the pages must be programmed consecutively from the LSB (least significant bit) page of the block to  
MSB (most significant bit) page of the block. Random page address programming is prohibited.  
Page 63  
Page 31  
Page 63  
Page 31  
(64)  
(64)  
(1)  
(32)  
Page2  
Page 1  
Page 0  
Page2  
Page 1  
Page 0  
(3)  
(32)  
(2)  
(3)  
(2)  
(1)  
Data Register  
Data Register  
From the LSB page to MSB page  
Ex.) Random page program (Prohibition)  
DATA IN: Data(1) Data (64)  
DATA IN: Data(1) Data (64)  
Version 1.0  
45  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(8)Status Read during a Read operation  
The device status can be read out by inputting the Status Read command “70h” in Read mode. Once the  
device has been set to Status Read mode by a “70h” command, the device will not return to Read mode  
unless the Read command “00h” is inputted during [A]. If the Read command “00h” is inputted during [A],  
Status Read mode is reset, and the device returns to Read mode. In this case, data output starts  
automatically from address N and address input is unnecessary  
(9)Auto programming failure  
Version 1.0  
46  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(10)RY / BY : termination for the Ready/Busy pin (RY / BY )  
A pullꢀup resistor needs to be used for termination because the RY / BY buffer consists of an open drain  
circuit.  
Version 1.0  
47  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(11)Note regarding the WP signal  
The Erase and Program operations are automatically reset when WP goes Low. The operations are  
enabled and disabled as follows:  
Enable Programming  
Disable Programming  
Enable Erasing  
Disable Erasing  
Version 1.0  
48  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(12)When six address cycles are input  
Although the device may read in a sixth address, it is ignored inside the chip.  
Read operation  
Program operation  
Version 1.0  
49  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(13)Several programming cycles on the same page (Partial Page Program)  
Each segment can be programmed individually as follows:  
1st Programming  
2nd Programming  
Data Pattern 1  
All 1 s  
All 1 s  
All 1 s  
Data Pattern 2  
4th Programming  
Result  
All 1 s  
Data Pattern 4  
Data Pattern 4  
Data Pattern 1  
Data Pattern 2  
-------------------------------------  
Version 1.0  
50  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(14)Invalid blocks (bad blocks)  
The device occasionally contains unusable blocks. Therefore, the following issues must be recognized:  
Please do not perform an erase operation to bad blocks. It may be impossible to recover  
the bad block information if the information is erased.  
Check if the device has any bad blocks after installation into the system.  
Refer to the test flow for bad block detection. Bad blocks which are detected by the test  
flow must be managed as unusable blocks by the system.  
A bad block does not affect the performance of good blocks because it is isolated from  
the bit lines by select gates.  
The number of valid blocks over the device lifetime is as follows:  
Symbol  
Min  
Typ.  
Max  
Unit  
Valid(Good) Block Number  
2,008  
2,048  
Blocks  
Bad Block Test Flow  
Regarding invalid blocks, bad block mark is in whole pages.  
Please read one column of any page in each block. It makes sure that every invalid block has Marjority “0” data at  
this column. If the data of the column is Marjority “0”, define the block as a bad block.  
Start  
Block No = 1  
Fail  
Read Check  
Pass  
Block No. = Block No. + 1  
Bad Block1  
No  
Last Block  
Yes  
End  
Note1: No erase operation is allowed to detected bad blocks.  
Version 1.0  
12/2018  
51  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(15)Failure phenomena for Program and Erase Operations  
The device may fail during a Program or Erase operation.  
The following possible failure modes should be considered when implementing a highly reliable system.  
Failure Mode  
Erase Failure  
Programming Failure  
Bit Error  
Detection and Countermeasure Sequence  
Read Status after Erase Block Replacement  
Read Status after Program Block Replacement  
ECC Correction / Block Refresh  
Block  
Page  
Read  
NOTE 1  
ECC: Error Correction Code. 8 bit correction per 512 Bytes is necessary.  
Block Replacement  
Program  
When an error happens in Block A, try to reprogram the data into another Block (Block B) by loading from an external  
buffer. Then, prevent further system accesses to Block A ( by creating a bad block table or by using anther  
appropriate scheme).  
Erase  
When an error occurs during an Erase operation, prevent future accesses to this bad block (again by creating a table  
within the system or by using another appropriate scheme).  
(16)The number of valid blocks is on the basis of single plane operations, and this may be decreased with two  
plane operations.  
Version 1.0  
52  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
(17)Reliability Guidance  
This reliability guidance is intended to notify some guidance related to using NAND flash with 8bit ECC for each 512  
bytes. For detailed reliability data, please refer to TOSHIBA’s reliability note.  
Although random bit errors may occur during use, it does not necessarily mean that a block is bad.  
Generally, a block should be marked as bad when a program status failure or erase status failure is detected.  
The other failure modes may be recovered by a block erase.  
ECC treatment for read data is mandatory due to the following Data Retention and Read Disturb failures.  
Write/Erase Endurance  
Write/Erase endurance failures may occur in a cell, page, or block, and are detected by doing a status read  
after either an auto program or auto block erase operation. The cumulative bad block count will increase  
along with the number of write/erase cycles.  
Data Retention  
The data in memory may change after a certain amount of storage time. This is due to charge loss or charge  
gain. After block erasure and reprogramming, the block may become usable again.  
Here is the combined characteristics image of Write/Erase Endurance and Data Retention.  
Read Disturb  
A read operation may disturb the data in memory. The data may change due to charge gain. Usually, bit  
errors occur on other pages in the block, not the page being read. After a large number of read cycles  
(between block erases), a tiny charge may build up and can cause a cell to be soft programmed to another  
state. After block erasure and reprogramming, the block may become usable again.  
Version 1.0  
53  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
TIMING DIAGRAMS  
Latch Timing Diagram for Command/Address/Data  
Command Input Cycle Timing Diagram  
Version 1.0  
54  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Address Input Cycle Timing Diagram  
Data Input Cycle Timing Diagram  
Version 1.0  
55  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Serial Read Cycle Timing Diagram  
Status Read Cycle Timing Diagram  
*: 70h represents the hexadecimal number  
Version 1.0  
56  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read Cycle Timing Diagram  
Read Cycle Timing Diagram: When Interrupted by CE  
Version 1.0  
57  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read Cycle with Data Cache Timing Diagram (1/2)  
*: The column address will be reset to 0 by the 31h command input.  
Read Cycle with Data Cache Timing Diagram (2/2)  
Make sure to terminate the operation with 3Fh command  
.
Version 1.0  
58  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Column Address Change in Read Cycle Timing Diagram (1/2)  
Version 1.0  
59  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Column Address Change in Read Cycle Timing Diagram (2/2)  
Version 1.0  
60  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data Output Timing Diagram  
Version 1.0  
61  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AutoꢀProgram Operation Timing Diagram  
*: M: up to 4351 (byte input data for ×8 device).  
Version 1.0  
62  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AutoꢀProgram Operation with Data Cache Timing Diagram (1/3)  
CA0 to CA12 is 0 in this diagram.  
Version 1.0  
63  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AutoꢀProgram Operation with Data Cache Timing Diagram (2/3)  
Repeat a max of 62 times  
(in order to program pages 1 to 62 of a block).  
Version 1.0  
64  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AutoꢀProgram Operation with Data Cache Timing Diagram (3/3)  
(*1)  
tPROG: Since the last page programming by 10h command is initiated after the previous cache  
program, the tPROG during cache programming is given by the following equation.  
tPROG = tPROG of the last page + tPROG of the previous page − A  
A = (command input cycle + address input cycle + data input cycle time of the last page)  
If “A” exceeds the tPROG of previous page, tPROG of the last page is tPROG max.  
NOTE : Make sure to terminate the operation with 80hꢀ10hꢀ command sequence.  
If the operation is terminated by 80hꢀ15h command sequence, monitor I/O 6 (Ready / Busy) by  
issuing Status Read command (70h) and make sure the previous page program operation is  
completed. If the page program operation is completed issue FFh reset before next operation.  
Version 1.0  
65  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MultiꢀPage Program Operation with Data Cache Timing Diagram (1/4)  
Version 1.0  
66  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MultiꢀPage Program Operation with Data Cache Timing Diagram (2/4)  
Repeat a max of 63 times  
(in order to program pages 0 to 62 of a block).  
Version 1.0  
67  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MultiꢀPage Program Operation with Data Cache Timing Diagram (3/4)  
Version 1.0  
68  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MultiꢀPage Program Operation with Data Cache Timing Diagram (4/4)  
(*1)  
tPROG: Since the last page programming by 10h command is initiated after the previous cache  
program, the tPROG during cache programming is given by the following equation.  
tPROG = tPROG of the last page + tPROG of the previous page − A  
A = (command input cycle + address input cycle + data input cycle time of the last page)  
If “A” exceeds the tPROG of previous page, tPROG of the last page is tPROG max.  
NOTE : Make sure to terminate the operation with 81hꢀ10hꢀ command sequence.  
If the operation is terminated by 81hꢀ15h command sequence, monitor I/O 6 (Ready / Busy)  
by issuing Status  
Read command (70h) and make sure the previous page program operation is completed.  
If the page program operation is completed issue FFh reset before next operation.  
Version 1.0  
69  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Auto Block Erase Timing Diagram  
Version 1.0  
70  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Multi Block Erase Timing Diagram  
Version 1.0  
71  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ID Read Operation Timing Diagram  
Table 5: ID Definition Table  
Version 1.0  
72  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
4Gb(X16) LPDDR4X  
Version 1.0  
73  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
LPDDR4X Descriptions  
LPDDR4 devices use a 2 or 4 clocks architecture on the Command/Address (CA) bus to reduce the  
number of input pins in the system. The 6ꢀbit CA bus contains command, address, and bank information.  
Each command uses 1, 2 or 4 clock cycle, during which command information is transferred on the positive  
edge of the clock. See command truth table for details.  
These devices use a double data rate architecture on the DQ pins to achieve high speed operation. The  
double data rate architecture is essentially an 16n prefetch architecture with an interface designed to  
transfer two data bits per DQ every clock cycle at the I/O pins. A single read or write access for the  
LPDDR4 SDRAM effectively consists of a single 16nꢀbit wide, one clock cycle data transfer at the internal  
DRAM core and eight corresponding nꢀbit wide, one halfꢀclockꢀcycle data transfers at the I/O pins. Read  
and write accesses to the LPDDR4 SDRAMs are burst oriented; accesses start at a selected location and  
continue for a programmed number of locations in a programmed sequence. Accesses begin with the  
registration of an Activate command, which is then followed by a Read, Write or Mask Write command.  
The address and BA bits registered coincident with the Activate command are used to select the row and  
the bank to be accessed. The address bits registered coincident with the Read, Write or Mask Write  
command are used to select the bank and the starting column location for the burst access.  
Prior to normal operation, the LPDDR4 SDRAM must be initialized. The following section provides detailed  
information covering device initialization, register definition, command description and device operation.  
Operating Frequency  
Please confirm with NTC when the operating frequency is slower than the defined frequency in the following table.  
Frequency [MHz]  
RL [nCK]  
1866  
32  
Unit  
VDDQ [V]  
0.6  
3733  
Mbps  
NM4484NSPAXAEꢀ3EE  
Notes:  
Any part number also supports functional operation at lower frequencies as shown in the table which are not subject to Production  
Tests but has been verified  
Version 1.0  
74  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Powerꢀup, Initialization and Powerꢀoff Procedure  
For powerꢀup and reset initialization, in order to prevent DRAM from functioning improperly, default values of the  
following MR settings are defined.  
MRS defaults settings  
Item  
FSP-OP/WR  
WLS  
MRS  
Default Setting  
00  
Description  
FSP-OP/WR[0] are enabled  
Write Latency Set 0 is selected  
WL = 4  
MR13 OP[7:6]  
MR2 OP[6]  
B
0
B
WL  
MR2 OP[5:3]  
MR2 OP[2:0]  
MR1 OP[6:4]  
MR3 OP[7:6]  
MR11 OP[6:4]  
MR11 OP[2:0]  
MR12 OP[6]  
MR12 OP[5:0]  
MR14 OP[6]  
MR14 OP[5:0]  
000  
000  
000  
B
B
B
RL  
RL = 6, nRTP=8  
nWR  
nWR = 6  
DBI-WR/RD  
CA ODT  
DQ ODT  
Write & Read DBI are disabled  
CA ODT is disabled  
DQ ODT is disabled  
00  
B
000  
000  
B
B
V
(CA) Setting  
1
B
V
REF  
(CA) Range[1] enabled  
REF  
V
(CA) Value  
001101  
Range1 : 50.3% of V  
DDQ  
REF  
B
B
V
(DQ) Setting  
1
B
V
REF  
(DQ) Range[1] enabled  
REF  
V
(DQ) Value  
001101  
Range1 : 50.3% of V  
DDQ  
REF  
Version 1.0  
75  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Voltage Ramp and Device Initialization  
The following sequence shall be used to power up the device. Unless specified otherwise, these steps are  
mandatory. Note that the powerꢀup sequence of all channels must proceed simultaneously.  
Power Ramp and Initialization Sequence  
Notes:  
1. Training is optional and may be done at the system architects discretion. The training sequence after ZQ_CAL Latch  
(Th, Sequence7~9) in Initialization Sequence is simplified recommendation and actual training sequence may vary depending on  
systems.  
Version 1.0  
76  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
1) While applying power (after Ta), RESET is recommended to be LOW (0.2 x VDD2) and all other inputs must be  
between VILmin and VIHmax. The device outputs remain at HighꢀZ while RESET is held LOW.  
Power supply voltage ramp requirements are provided. VDD1 must ramp at the same time or earlier than VDD2  
.
VDD2 must ramp at the same time or earlier than VDDQ  
.
Voltage Ramp Conditions  
After  
Applicable Conditions  
V
DD1  
must be greater than V  
DD2  
Ta is reached  
V
DD2  
must be greater than V  
- 200mV  
DDQ  
Notes:  
1. Ta is the point when any power supply first reaches 300mV.  
2. Voltage ramp conditions apply between Ta and power-off (controlled or uncontrolled).  
3. Tb is the point at which all supply and reference voltages are within their defined ranges.  
4. Power ramp duration tINIT0 (Tb-Ta) must not exceed 20ms.  
5. The voltage difference between any of VSS and VSSQ pins must not excess 100mV.  
2) Following the completion of the voltage ramp (Tb), RESET must be maintained LOW. DQ, DMI, DQS and DQS  
voltage levels must be between VSSQ and VDDQ during voltage ramp to avoid latchꢀup. CKE, CK, CK, CS and CA  
input levels must be between VSS and VDD2 during voltage ramp to avoid latchꢀup.  
3) Beginning at Tb, RESET must remain LOW for at least tINIT1 (Tc), after which RESET can be deꢀasserted to  
HIGH(Tc). At least 10ns before RESET deꢀassertion, CKE is required to be set LOW. All other input signals are  
"Don't Care".  
4) After RESET is deꢀasserted (Tc), wait at least tINIT3 before activating CKE. Clock (CK, CK) is required to be  
started and stabilized for tINIT4 before CKE goes active(Td). CS is required to be maintained LOW when  
controller activates CKE.  
5) After setting CKE high, wait minimum of tINIT5 to issue any MRR or MRW commands (Te). For both MRR and  
MRW commands, the clock frequency must be within the range defined for tCKb. Some AC parameters (for  
example, tDQSCK) could have relaxed timings (such as tDQSCKb) before the system is appropriately  
configured.  
6) After completing all MRW commands to set the Pullꢀup, Pullꢀdown and Rx termination values, the DRAM  
controller can issue ZQCAL Start command to the memory (Tf). This command is used to calibrate VOH level  
and output impedance over process, voltage and temperature. In systems where more than one LPDDR4 DRAM  
devices share one external ZQ resistor, the controller must not overlap the ZQ calibration sequence of each  
LPDDR4 device. ZQ calibration sequence is completed after tZQCAL (Tg) and the ZQCAL Latch command must  
be issued to update the DQ drivers and DQ+CA ODT to the calibrated values.  
7) After tZQLAT is satisfied (Th) the command bus (internal V  
(CA), CS, and CA) should be trained for  
REF  
highꢀspeed operation by issuing an MRW command (Command Bus Training Mode). This command is used to  
calibrate the device's internal VREF and align CS/CA with CK for highꢀspeed operation. The LPDDR4 device will  
Version 1.0  
77  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
powerꢀup with receivers configured for lowꢀspeed operations, and VREF(CA) set to a default factory setting.  
Normal device operation at clock speeds higher than tCKb may not be possible until command bus training has  
been completed.  
NOTE: The command bus training MRW command uses the CA bus as inputs for the calibration data stream, and outputs the results  
asynchronously on the DQ bus. See Command Bus Training, MRW for information on how to enter/exit the training mode.  
8) After command bus training, DRAM controller must perform write leveling. Write leveling mode is enabled when  
MR2 OP[7] is high (Ti). See Mode Register WriteꢀWR Leveling Mode, for detailed description of write leveling  
entry and exit sequence. In write leveling mode, the DRAM controller adjusts write DQS/DQS timing to the point  
where the LPDDR4 device recognizes the start of write DQ data burst with desired write latency.  
9) After write leveling, the DQ Bus (internal VREF(DQ), DQS, and DQ) should be trained for highꢀspeed operation  
using the MPC training commands and by issuing MRW commands to adjust VREF(DQ)(Tj). The LPDDR4 device  
will powerꢀup with receivers configured for lowꢀspeed operations and VREF(DQ) set to a default factory setting.  
Normal device operation at clock speeds higher than tCKb should not be attempted until DQ Bus training has  
been completed. The MPC Read Calibration command is used together with MPC FIFO Write/Read commands  
to train DQ bus without disturbing the memory array contents. See DQ Bus Training section for detailed DQ Bus  
Training sequence.  
10) At Tk the LPDDR4 device is ready for normal operation, and is ready to accept any valid command. Any more  
registers that have not previously been set up for normal operation should be written at this time.  
Initialization Timing Parameters  
Value  
Parameter  
Unit  
Comment  
Min  
Max  
tINIT0  
tINIT1  
tINIT2  
tINIT3  
tINIT4  
tINIT5  
tZQCAL  
tZQLAT  
tCKb  
-
20  
-
ms Maximum voltage-ramp time  
200  
us Minimum RESET LOW time after completion of voltage ramp  
ns Minimum CKE low time before RESET high  
ms Minimum CKE low time after RESET high  
tCK Minimum stable clock before first CKE high  
us Minimum idle time before first MRW/MRR command  
us ZQ calibration time  
10  
-
2
-
5
-
2
-
1
-
Max(30ns, 8tCK)  
-
ns ZQCAL latch quiet time.  
*1,2  
*1,2  
ns Clock cycle time during boot  
Note  
Note  
Notes:  
1. Min tCKb guaranteed by DRAM test is 18ns.  
2. The system may boot at a higher frequency than dictated by min tCKb. The higher boot frequency is system dependent.  
Version 1.0  
78  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Reset Initialization with Stable Power  
The following sequence is required for RESET at no power interruption initialization.  
1. Assert RESET below 0.2 x VDD2 anytime when reset is needed. RESET needs to be maintained for minimum  
tPW_RESET. CKE must be pulled LOW at least 10 ns before deꢀasserting RESET.  
2. Repeat steps 4 to 10 in "Voltage Ramp and Device Initialization" section.  
Reset Timing Parameter  
Value  
Parameter  
Unit  
Comment  
Min  
Max  
tPW_RESET  
100  
-
ns Minimum RESET low Time for Reset Initialization with stable power  
Powerꢀoff Sequence  
The following procedure is required to power off the device.  
While powering off, CKE must be held LOW (0.2 X VDD2) and all other inputs must be between VILmin and VIHmax.  
The device outputs remain at HighꢀZ while CKE is held LOW. DQ, DMI, DQS and DQS voltage levels must be  
between VSSQ and VDDQ during voltage ramp to avoid latchꢀup. RESET, CK, CK, CS and CA input levels must be  
between VSS and VDD2 during voltage ramp to avoid latchꢀup.  
Tx is the point where any power supply drops below the minimum value specified.  
Tz is the point where all power supplies are below 300mV. After TZ, the device is powered off.  
Power Supply Conditions  
After  
Applicable Conditions  
V
DD1  
must be greater than V  
DD2  
Tx and Tz  
V
DD2  
must be greater than V  
- 200mV  
DDQ  
The voltage difference between any of VSS, VSSQ pins must not exceed 100mV.  
Version 1.0  
79  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Uncontrolled PowerꢀOff Sequence  
When an uncontrolled powerꢀoff occurs, the following conditions must be met:  
At Tx, when the power supply drops below the minimum values specified, all power supplies must be turned off and  
all power supply current capacity must be at zero, except any static charge remaining in the system.  
After Tz (the point at which all power supplies first reach 300mV), the device must power off. During this period the  
relative voltage between power supplies is uncontrolled. VDD1 and VDD2 must decrease with a slope lower than  
0.5V/ꢁs between Tx and Tz.  
An uncontrolled powerꢀoff sequence can occur a maximum of 400 times over the life of the device.  
Timing Parameters Power Off  
Value  
Symbol  
Unit  
Comment  
Min  
Max  
tPOFF  
-
2
s
Maximum Power-off ramp item  
Version 1.0  
12/2018  
80  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mode Register Definition  
Mode Register Assignment and Definition in LPDDR4 SDRAM  
Each register is denoted as "R" if it can be read but not written, "W" if it can be written but not read, and "R/W" if it  
can be read and written. A Mode Register Read command is used to read a mode register. A Mode Register Write  
command is used to write a mode register.  
MR#  
0
OP[7]  
Reserved  
RPST  
OP[6]  
OP[5]  
RFU  
OP[4]  
OP[3]  
OP[2]  
RFU  
OP[1]  
OP[0]  
RFU  
RZQI  
Latency mode Refresh Mode  
1
nWR (for AP)  
RD-PRE  
WR-PRE  
BL  
2
WR Lev  
DBI-WR  
TUF  
WLS  
WL  
RL  
3
DBI-RD  
PDDS  
PPRE  
PPRP  
WR PST  
PU-CAL  
4
Thermal Offset  
SR Abort  
Refresh Rate  
5
LPDDR4 Manufacturer ID  
Revision ID-1  
6
7
Revision ID-2  
8
IO Width  
Density  
Type  
9
Vendor Specific Test Register  
10  
11  
12  
13  
14  
15  
16  
17  
18  
19  
20  
21  
RFU  
ZQ-Reset  
CBT  
RFU  
CBT Mode  
FSP-OP  
RFU  
CA ODT  
DMD  
RFU  
DQ ODT  
RPT  
VR-CA  
FSP-WR  
VR(dq)  
V
(CA)  
REF  
RRO  
VRCG  
VRO  
(DQ)  
V
REF  
Lower-Byte Invert Register for DQ Calibration  
PASR Bank Mask  
PASR Segment Mask  
DQS Oscillator Count - LSB  
DQS Oscillator Count - MSB  
Upper-Byte Invert Register for DQ Calibration  
RFU  
22 ODTD for x8 2ch(Byte) mode ODTD-CA  
ODTE-CS  
DQS interval timer run time setting  
TRR Mode BAn  
ODTE-CK  
SOC ODT  
23  
24  
TRR Mode  
Unlimited MAC  
MAC Value  
25  
PPR Resource  
26~29  
30  
RFU  
Reserved for testing – SDRAM will ignore  
RFU  
31 Byte mode Vref Selection  
32  
DQ Calibration Pattern “A” (default = 5AH)  
RFU  
33~38  
39  
40  
Reserved for testing – SDRAM will ignore  
DQ Calibration Pattern “B” (default = 3CH)  
Version 1.0  
81  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR0 Register Information (MA[5:0] = 00H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Reserved  
RFU  
RFU  
RZQI  
RFU  
Latency modeRefresh mode  
Register  
Type  
Function  
Operand  
OP[0]  
Data  
0 : Both legacy & modified refresh mode supported  
Notes  
B
Refresh mode  
Latency mode  
1 : Only modified refresh mode supported  
B
0 : Device supports normal latency  
B
OP[1]  
5,6  
1 : Device supports byte mode latency  
B
00 : RZQ Self-Test Not Supported  
B
Read-only  
01 : ZQ pin may connect to VSSQ or float  
B
RZQI  
10 : ZQ-pin may short to VDDQ  
B
OP[4:3]  
1,2,3,4  
(Built-in Self-Test for RZQ)  
11 : ZQ-pin Self-Test Completed, no error  
B
condition detected (ZQ-pin may not connect  
to VSSQ or float, nor short to VDDQ)  
Notes:  
1. RZQI MR value, if supported, will be valid after the following sequence:  
a: Completion of MPC ZQCAL Start command to either channel.  
b: Completion of MPC ZQCAL Latch command to either channel then tZQLAT is satisified.  
RZQI value will be lost after Reset.  
2. If the ZQ-pin is connected to VSSQ to set default calibration, OP[4:3] shall be set to 01B. If the ZQ-pin is not connected to VSSQ, either  
OP[4:3] = 01B or OP[4:3] = 10B might indicate might indicate a ZQ-pin assembly error. It is recommended that the assembly error is  
corrected.  
3. In the case of possible assembly error, the LPDDR4-SDRAM device will default to factory trim settings for RON, and will ignore ZQ  
Calibration commands. In either case, the device may not function as intended.  
4. If ZQ Self-Test returns OP[4:3] = 11B, the device has detected a resistor connected to the ZQ-pin. However, this result cannot be used to  
validate the ZQ resistor value or that the ZQ resistor tolerance meets the specified limits (i.e. 240Ω ± 1%).  
5. See byte mode addendum spec for byte mode latency details.  
6. Byte mode latency for 2Ch. x16 device is only allowed when it is stacked in a same package with byte mode device.  
Version 1.0  
82  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR1 Register Information (MA[5:0] = 01H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
RPST  
nWR (for AP)  
RD-PRE  
WR-PRE  
BL  
Register  
Type  
Function  
Operand  
Data  
00 : BL=16 Sequential (default)  
Notes  
B
BL  
01 : BL=32 Sequential  
B
OP[1:0]  
1,7  
(Burst Length)  
10 : BL=16 or 32 Sequential (on-the-fly)  
B
All Others: Reserved  
WR-PRE  
(WR Pre-amble Length)  
RD-PRE  
0 : Reserved  
B
OP[2]  
OP[3]  
5,6  
1 : WR Pre-amble = 2*tCK  
B
0 : RD Pre-amble = Static (default)  
B
3,5,6  
(RD Pre-amble Type)  
1 : RD Pre-amble = Toggle  
B
For x16 mode  
000 : nWR = 6 (default)  
B
001 : nWR = 10  
B
010 : nWR = 16  
B
011 : nWR = 20  
B
100 : nWR = 24  
B
Write-only  
101 : nWR = 30  
B
110 : nWR = 34  
B
nWR  
111 : nWR = 40  
B
(Write-Recovery for Auto-  
Pre-charge commands)  
OP[6:4]  
2,5,6  
For Byte (x8) mode  
000 : nWR = 6 (default)  
B
001 : nWR = 12  
B
010 : nWR = 16  
B
011 : nWR = 22  
B
100 : nWR = 28  
B
101 : nWR = 32  
B
110 : nWR = 38  
B
111 : nWR = 44  
B
RPST  
0 : RD Post-amble = 0.5*tCK (default)  
B
OP[7]  
4,5,6  
(RD Post-Amble Length)  
1 : RD Post-amble = 1.5*tCK  
B
Notes:  
1. Burst length on-the-fly can be set to either BL=16 or BL=32 by setting the “BL” bit in the command operands. See the Command Truth  
Table.  
Version 1.0  
83  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
2. The programmed value of nWR is the number of clock cycles the LPDDR4-SDRAM device uses to determine the starting point of an  
internal Pre-charge operation after a Write burst with AP (auto-precharge) enabled.  
(See section “Read and Write Latencies.)  
3. For Read operations this bit must be set to select between a "toggling" pre-amble and a "Non-toggling"  
Pre-amble.  
(See Read Preamble and Postamble, for a drawing of each type of pre-amble.)  
4. OP[7] provides an optional READ post-amble with an additional rising and falling edge of DQS. The optional postamble cycle is provided  
for the benefit of certain memory controllers.  
5. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1.  
Only the registers for the set point determined by the state of the FSP-WR bit (MR13 OP[6]) will be read from  
with an MRR command to this MR address.  
6. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1.  
The device will operate only according to the values stored in the registers for the active set point, i.e., the set  
point determined by the state of the FSP-OP bit (MR13 OP[7]). The values in the registers for the inactive set  
point will be ignored by the device, and may be changed without affecting device operation.  
7. Supporting the two physical registers for Burst Length: MR1 OP[1:0] as optional feature. Applications requiring  
support of both vendor options shall assure that both FSP-OP[0] and FSP-OP[1] are set to the same code. Refer  
to vendor datasheets for detail.  
Version 1.0  
84  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst Sequence for READ  
Burst Burst  
Burst Cycle Number and Burst Address Sequence  
C4 C3 C2 C1 C0  
Length Type  
1
0
4
8
C
0
4
8
C
2
1
5
9
D
1
5
9
D
3
2
6
A
E
2
6
A
E
4
3
7
B
F
5
4
8
C
0
4
8
C
0
6
5
9
D
1
5
9
D
1
7
6
A
E
2
6
A
E
2
8
7
B
F
9
8
C
0
4
8
C
0
4
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30  
31 32  
V
V
V
V
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
9
D
1
5
9
D
1
5
A
E
2
6
A
E
2
6
B
F
C
0
4
8
C
0
4
8
D
1
5
9
D
1
5
9
E
2
6
A
E
2
6
A
F
3
7
B
F
16  
SEQ  
3
7
B
F
3
7
B
F
3
7
B
F
10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F  
14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 10 11 12 13  
18 19 1A 1B 1C 1D 1E 1F 10 11 12 13 14 15 16 17  
1C 1D 1E 1F 10 11 12 13 14 15 16 17 18 19 1A 1B  
3
7
B
3
7
3
32  
SEQ  
10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F  
14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 10 11 12 13  
18 19 1A 1B 1C 1D 1E 1F 10 11 12 13 14 15 16 17  
1C 1D 1E 1F 10 11 12 13 14 15 16 17 18 19 1A 1B  
0
4
8
C
1
5
9
D
2
6
A
E
3
7
B
F
4
8
C
0
5
9
D
1
6
A
E
2
7
B
F
8
C
0
4
9
D
1
5
A
E
2
6
B
F
C
0
4
8
D
1
5
9
E
2
6
A
F
3
7
B
3
7
3
Notes:  
1. C0-C1 are assumed to be '0', and are not transmitted on the command bus.  
2. The starting burst address is on 64-bit (4n) boundaries.  
Burst Sequence for Write  
Burst Burst  
Burst Cycle Number and Burst Address Sequence  
C4 C3 C2 C1 C0  
Length Type  
1
0
0
2
1
1
3
2
2
4
3
3
5
4
4
6
5
5
7
6
6
8
7
7
9
8
8
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32  
16  
32  
SEQ  
SEQ  
V
0
0
0
0
0
0
0
0
0
9
9
A
A
B
B
C
C
D
D
E
E
F
F
10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F  
Notes:  
1. C0-C1 are assumed to be '0', and are not transmitted on the command bus.  
2. The starting address is on 256-bit(16n) boundaries for Burst length 16.  
3. The starting address is on 512-bit(32n) boundaries for Burst length 32.  
4. C2-C3 shall be set to '0' for all Write operations.  
Version 1.0  
85  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR2 Register Information (MA[5:0] = 02H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
WR Lev  
WLS  
WL  
RL  
Register  
Type  
Function  
Operand  
Data  
Notes  
For x16 mode  
RL & nRTP for DBI-RD Disabled (MR3 OP[6]=0 )  
B
000 : RL=6, nRTP = 8 (Default)  
B
001 : RL=10, nRTP = 8  
B
010 : RL=14, nRTP = 8  
B
011 : RL=20, nRTP = 8  
B
100 : RL=24, nRTP = 10  
B
101 : RL=28, nRTP = 12  
B
110 : RL=32, nRTP = 14  
B
111 : RL=36, nRTP = 16  
B
RL & nRTP for DBI-RD Enabled (MR3 OP[6]=1 )  
B
000 : RL=6, nRTP = 8  
B
001 : RL=12, nRTP = 8  
B
010 : RL=16, nRTP = 8  
B
011 : RL=22, nRTP = 8  
B
100 : RL=28, nRTP = 10  
B
101 : RL=32, nRTP = 12  
B
110 : RL=36, nRTP = 14  
B
RL  
(Read latency)  
111 : RL=40, nRTP = 16  
B
Write-only  
OP[2:0]  
1,3,4  
For Byte (x8) mode  
RL & nRTP for DBI-RD Disabled (MR3 OP[6]=0 )  
B
000 : RL=6, nRTP = 8 (Default)  
B
001 : RL=10, nRTP = 8  
B
010 : RL=16, nRTP = 8  
B
011 : RL=22, nRTP = 8  
B
100 : RL=26, nRTP = 10  
B
101 : RL=32, nRTP = 12  
B
110 : RL=36, nRTP = 14  
B
111 : RL=40, nRTP = 16  
B
RL & nRTP for DBI-RD Enabled (MR3 OP[6]=1 )  
B
000 : RL=6, nRTP = 8  
B
001 : RL=12, nRTP = 8  
B
010 : RL=18, nRTP = 8  
B
011 : RL=24, nRTP = 8  
B
100 : RL=30, nRTP = 10  
B
101 : RL=36, nRTP = 12  
B
110 : RL=40, nRTP = 14  
B
111 : RL=44, nRTP = 16  
B
Version 1.0  
12/2018  
86  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Register  
Function  
Operand  
Data  
Notes  
Type  
For x16 mode  
WL Set "A”  
(MR2 OP[6]=0 )  
For Byte (x8) mode  
WL Set "A”  
(MR2 OP[6]=0 )  
B
B
000 : WL=4 (Default)  
000 : WL=4 (Default)  
B
B
001 : WL=6  
001 : WL=6  
B
B
010 : WL=8  
010 : WL=8  
B
B
011 : WL=10  
011 : WL=10  
B
B
100 : WL=12  
100 : WL=12  
B
B
101 : WL=14  
101 : WL=14  
B
B
110 : WL=16  
B
110 : WL=16  
B
WL  
OP[5:3]  
1,3,4  
111 : WL=18  
B
111 : WL=18  
B
(Write latency)  
WL Set "B"  
WL Set "B"  
(MR2 OP[6]=1 )  
B
(MR2 OP[6]=1 )  
B
000 : WL=4  
B
000 : WL=4  
B
001 : WL=8  
B
001 : WL=8  
B
010 : WL=12  
B
010 : WL=12  
B
011 : WL=18  
B
011 : WL=18  
B
100 : WL=22  
B
100 : WL=22  
B
101 : WL=26  
B
101 : WL=26  
B
110 : WL=30  
B
110 : WL=30  
B
111 : WL=34  
111 : WL=34  
B
B
WLS  
0 : WL Set "A" (default)  
B
OP[6]  
OP[7]  
1,3,4  
2
(Write Latency Set)  
WR LEV  
1 : WL Set "B"  
B
0 : Disabled (default)  
B
(Write Leveling)  
1 : Enabled  
B
Notes:  
1. See “Read and Write Latencies” for detail.  
2. After a MRW to set the Write Leveling Enable bit (OP[7]=1B), the LPDDR4-SDRAM device remains in the MRW state until another MRW  
command clears the bit (OP[7]=0 B). No other commands are allowed until the Write Leveling Enable bit is cleared.  
3. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. Only the registers for  
the set point determined by the state of the FSP-WR bit (MR13 OP[6]) will be written to with an MRW command to this MR address, or  
read from with an MRR command to this address.  
4. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. The device will  
operate only according to the values stored in the registers for the active set point, i.e., the set point determined by the state of the  
FSP-OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed without  
affecting device operation.  
Version 1.0  
87  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR3 Register Information (MA[5:0] = 03H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
DBI-WR  
DBI-RD  
PDDS  
PPRP  
WR PST  
PU-CAL  
Register  
Type  
Function  
Operand  
OP[0]  
Data  
Notes  
1,4  
PU-Cal  
(Pull-up Calibration Point)  
WR PST  
0 : VDDQ*0.6  
B
1 : VDDQ*0.5 (default)  
B
0 : WR Post-amble = 0.5*tCK (default)  
B
OP[1]  
2,3,5  
(WR Post-Amble Length)  
Post Package Repair  
Protection  
1 : WR Post-amble = 1.5*tCK (Vendor specific function)  
B
0 : PPR protection disabled (default)  
B
OP[2]  
6
1 : PPR protection enabled  
B
000 : RFU  
B
001 : RZQ/1  
B
010 : RZQ/2  
B
Write-only  
PDDS  
011 : RZQ/3  
B
OP[5:3]  
1,2,3  
100 : RZQ/4  
B
(Pull-Down Drive Strength)  
101 : RZQ/5  
B
110 : RZQ/6 (default)  
B
111 : Reserved  
B
DBI-RD  
(DBI-Read Enable)  
DBI-WR  
0 : Disabled (default)  
B
OP[6]  
OP[7]  
2,3  
2,3  
1 : Enabled  
B
0 : Disabled (default)  
B
(DBI-Write Enable)  
Notes:  
1B: Enabled  
1. All values are "typical". The actual value after calibration will be within the specified tolerance for a given volt-age and temperature.  
Re-calibration may be required as voltage and temperature vary.  
2. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. Only the registers for  
the set point determined by the state of the FSP-WR bit (MR13 OP[6]) will be written to with an MRW command to this MR address, or  
read from with an MRR command to this address  
3. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. The device will  
operate only according to the values stored in the registers for the active set point, i.e., the set point determined by the state of the  
FSP-OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed  
without affecting device operation.  
4. For dual channel devices, PU-CAL setting is required as the same value for both Ch.A and Ch.B before issuing ZQ Cal start command.  
5. Refer to the supplier data sheet for vender specific function. 1.5*tCK apply > 1.6GHz clock..  
6. If MR3 OP[2] is set to 1b then PPR protection mode is enabled. The PPR Protection bit is a sticky bit and can only be set to 0b by a  
power on reset. MR4 OP[4] controls entry to PPR Mode. If PPR protection is enabled then DRAM will not allow writing of 1 to MR4  
OP[4].  
Version 1.0  
88  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR4 Register Information (MA[5:0] = 04H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
TUF  
Thermal Offset  
PPRE  
SR Abort  
Refresh Rate  
Register  
Type  
Function  
Operand  
Data  
000 : SDRAM Low temperature operating limit exceeded  
Notes  
B
001 : 4x refresh  
B
010 : 2x refresh  
B
1,2,3,4,  
7,8,9  
011 : 1x refresh (default)  
B
Refresh Rate  
Read  
OP[2:0]  
100 : 0.5x refresh  
B
101 : 0.25x refresh, no de-rating  
B
110 : 0.25x refresh, with de-rating  
B
111 : SDRAM High temperature operating limit exceeded  
B
SR Abort  
(Self Refresh Abort)  
PPRE  
0 : Disable (default)  
B
Write  
Write  
OP[3]  
OP[4]  
9,11  
5,9  
1 : Enable  
B
0 : Exit PPR mode (default)  
B
(Post-package repair entry/exit)  
1 : Enter PPR mode  
B
00 : No offset, 0~5°C gradient (default)  
B
Thermal Offset  
01 : 5°C offset, 5~10°C gradient  
B
Write  
Read  
OP[6:5]  
OP[7]  
10  
(Vender Specific Function)  
10 : 10°C offset, 10~15°C gradient  
B
11 : Reserved  
B
TUF  
(Temperature Update Flag)  
Notes:  
0 : No change in OP[2:0] since last MR4 read (default)  
B
6,7,8  
1 : Change in OP[2:0] since last MR4 read  
B
1. The refresh rate for each MR4-OP[2:0] setting applies to tREFI, tREFIpb, and tREFW. OP[2:0]=011B corresponds to a device temperature  
of 85°C. Other values require either a longer (2x, 4x) refresh interval at lower temperatures, or a shorter (0.5x, 0.25x) refresh interval at  
higher temperatures. If OP[2]=1B, the device temperature is greater than 85°C.  
2. At higher temperatures (>85°C), AC timing derating may be required. If derating is required the LPDDR4-SDRAM will set OP[2:0]=110B.  
See derating timing requirements.  
3. DRAM vendors may or may not report all of the possible settings over the operating temperature range of the device. Each vendor  
guarantees that their device will work at any temperature within the range using the refresh interval requested by their device.  
4. The device may not operate properly when OP[2:0]=000B or 111B.  
5. Post-package repair can be entered or exited by writing to OP[4].  
6. When OP[7]=1, the refresh rate reported in OP[2:0] has changed since the last MR4 read. A mode register read from MR4 will reset  
OP[7] to '0'.  
7. OP[7]=0 at power-up. OP[2:0] bits are valid after initialization sequence(Te)  
8. See the section on "Temperature Sensor" for information on the recommended frequency of reading MR4.  
Version 1.0  
89  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
9. OP[6:3] bits that can be written in this register. All other bits will be ignored by the DRAM during a MRW to this register.  
10. Refer to the supplier data sheet for vender specific function.  
11. Self refresh abort feature is available for higher density devices starting with 12Gb device.  
Version 1.0  
90  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR5 Register Information (MA[5:0] = 05H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Manufacturer ID  
Register  
Function  
Operand  
Data  
Notes  
Type  
0000 0101B: Nanya  
All Others: Reserved  
Manufacturer ID  
Read-only  
OP[7:0]  
MR6 Register Information (MA[5:0] = 06H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Revision ID-1  
Register  
Function  
Operand  
Data  
Notes  
Type  
0000 0000B: A-die  
Revision ID-1  
Read-only  
OP[7:0]  
1
All Others: Reserved  
Notes:  
1. MR6 is vendor specific.  
MR7 Register Information (MA[5:0] = 07H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Revision ID-2  
Register  
Function  
Operand  
Data  
Notes  
Type  
0100 0000B: A Version  
All Others: Reserved  
Revision ID-2  
Read-only  
OP[7:0]  
1
Notes:  
1. MR7 is vendor specific.  
Version 1.0  
91  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR8 Register Information (MA[5:0] = 08H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
IO Width  
Density  
Type  
Register  
Function  
Operand  
Data  
Notes  
Type  
00 : S16 SDRAM (16n pre-fetch)  
B
Type  
OP[1:0]  
OP[5:2]  
OP[7:6]  
All Others: Reserved  
0000 : 4Gb dual channel die/2Gb single channel die  
B
0001 : 6Gb dual channel die/3Gb single channel die  
B
0010 : 8Gb dual channel die/4Gb single channel die  
B
0011 : 12Gb dual channel die/6Gb single channel die  
B
Density  
Read-only  
0100 : 16Gb dual channel die/8Gb single channel die  
B
0101 : 24Gb dual channel die/12Gb single channel die  
B
0110 : 32Gb dual channel die/16Gb single channel die  
B
1100 : 2Gb dual channel die/1Gb single channel die  
B
All Others: Reserved  
00 : x16 (per channel)  
B
IO Width  
All Others: Reserved  
MR9 Register Information (MA[5:0] = 09H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Vendor Specific Test Register  
Notes:  
1. Only 00H should be written to this register.  
MR10 Register Information (MA[5:0] = 0AH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
RFU  
ZQ-Reset  
Register  
Function  
Operand  
Data  
0 : Normal Operation (Default)  
Notes  
Type  
B
ZQ-Reset  
Write-only  
OP[0]  
1,2  
1 : ZQ Reset  
B
Notes:  
1. ZQCal Timing Parameters for calibration latency and timing.  
2. If the ZQ-pin is connected to VDDQ through RZQ, either the ZQ calibration function or default calibration (via ZQ-Reset) is supported. If  
the ZQ-pin is connected to VSS, the device operates with default calibration, and ZQ calibration commands are ignored. In both cases,  
the ZQ connection shall not change after power is applied to the device.  
Version 1.0  
92  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR11 Register Information (MA[5:0] = 0BH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
RFU  
CA ODT  
RFU  
DQ ODT  
Register  
Function  
Operand  
Data  
000 : Disable (Default)  
Notes  
Type  
B
001 : RZQ/1  
B
010 : RZQ/2  
B
DQ ODT  
011 : RZQ/3  
B
(DQ Bus Receiver  
OP[2:0]  
1,2,3  
100 : RZQ/4  
B
On-Die-Termination)  
101 : RZQ/5  
B
110 : RZQ/6  
B
111 : RFU  
B
Write-only  
000 : Disable (Default)  
B
001 : RZQ/1  
B
010 : RZQ/2  
B
CA ODT  
011 : RZQ/3  
B
(CA Bus Receiver  
On-Die-Termination)  
OP[6:4]  
1,2,3  
100 : RZQ/4  
B
101 : RZQ/5  
B
110 : RZQ/6  
B
111 : RFU  
B
Notes:  
1. All values are "typical". The actual value after calibration will be within the specified tolerance for a given voltage and temperature.  
Re-calibration may be required as voltage and temperature vary.  
2. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. Only the registers for  
the set point determined by the state of the FSP-WR bit (MR13 OP[6]) will be written to with an MRW command to this MR address, or  
read from with an MRR command to this address.  
3. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. The device will  
operate only according to the values stored in the registers for the active set point, i.e., the set point determined by the state of the  
FSP-OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed  
without affecting device operation.  
Version 1.0  
93  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR12 Register Information (MA[5:0] = 0CH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
CBT Mode  
VR-CA  
VREF(CA)  
Register  
Function  
Operand  
Data  
Notes  
Type  
000000 :  
B
VREF(CA)  
-- Thru --  
1,2,3,  
5,6  
OP[5:0]  
(VREF(CA) Setting)  
110010 : See table below  
B
All Others: Reserved  
Read/  
Write  
VR-CA  
1,2,4,  
5,6  
0 : VREF(CA) Range[0] enabled  
B
OP[6]  
OP[7]  
(VREF(CA) Range)  
1 : VREF(CA) Range[1] enabled (default)  
B
0 : Mode1(Default)  
B
CBT Mode  
7
1 : Mode2  
B
Notes:  
1. This register controls the VREF(CA) levels. Refer to VREF Settings for Range[0] and Range[1] for actual voltage of VREF(CA).  
2. A read to this register places the contents of OP[7:0] on DQ[7:0]. Any RFU bits and unused DQ's shall be set to '0'. See the section on  
MRR Operation.  
3. A write to OP[5:0] sets the internal VREF(CA) level for FSP[0] when MR13 OP[6]=0B, or sets the internal VREF(CA) level for FSP[1] when  
MR13 OP[6]=1B. The time required for VREF(CA) to reach the set level depends on the step size from the current level to the new level.  
See the section on VREF(CA) training for more information.  
4. A write to OP[6] switches the LPDDR4-SDRAM between two internal VREF(CA) ranges. The range (Range[0] or Range[1]) must be  
selected when setting the VREF(CA) register. The value, once set, will be retained until overwritten, or until the next power-on or RESET  
event.  
5. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. Only the registers for  
the set point determined by the state of the FSP-WR bit (MR13 OP[6]) will be written to with an MRW command to this MR address, or  
read from with an MRR command to this address.  
6. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. The device will  
operate only according to the values stored in the registers for the active set point, i.e., the set point determined by the state of the  
FSP-OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed  
without affecting device operation.  
7. This field can be activated in only Byte Mode: x8. Device.  
Version 1.0  
94  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
VREF Settings for Range[0] and Range[1]  
Function Operand  
Notes  
Range[0] Values (% of V  
)
Range[1] Values (% of V  
)
DDQ  
DDQ  
000000 : 15.0%  
011010 : 30.5%  
000000 : 32.9%  
011010 : 48.5%  
B
B
B
B
000001 : 15.6%  
011011 : 31.1%  
000001 : 33.5%  
011011 : 49.1%  
B
B
B
B
000010 : 16.2%  
011100 : 31.7%  
000010 : 34.1%  
011100 : 49.7%  
B
B
B
B
011101 : 50.3%  
B
000011 : 16.8%  
011101 : 32.3%  
000011 : 34.7%  
B
B
B
(Default)  
000100 : 17.4%  
011110 : 32.9%  
000100 : 35.3%  
011110 : 50.9%  
B
B
B
B
000101 : 18.0%  
011111 : 33.5%  
000101 : 35.9%  
011111 : 51.5%  
B
B
B
B
000110 : 18.6%  
B
100000 : 34.1%  
B
000110 : 36.5%  
B
100000 : 52.1%  
B
000111 : 19.2%  
B
100001 : 34.7%  
B
000111 : 37.1%  
B
100001 : 52.7%  
B
001000 : 19.8%  
B
100010 : 35.3%  
B
001000 : 37.7%  
B
100010 : 53.3%  
B
001001 : 20.4%  
B
100011 : 35.9%  
B
001001 : 38.3%  
B
100011 : 53.9%  
B
001010 : 21.0%  
B
100100 : 36.5%  
B
001010 : 38.9%  
B
100100 : 54.5%  
B
V
REF  
001011 : 21.6%  
B
100101 : 37.1%  
B
001011 : 39.5%  
B
100101 : 55.1%  
B
Settings  
for  
OP[5:0]  
1,2,3  
001100 : 22.2%  
B
100110 : 37.7%  
B
001100 : 40.1%  
B
100110 : 55.7%  
B
001101 : 22.8%  
B
100111 : 38.3%  
B
001101 : 40.7%  
B
100111 : 56.3%  
B
MR12  
001110 : 23.4%  
B
101000 : 38.9%  
B
001110 : 41.3%  
B
101000 : 56.9%  
B
001111 : 24.0%  
B
101001 : 39.5%  
B
001111 : 41.9%  
B
101001 : 57.5%  
B
010000 : 24.6%  
B
101010 : 40.1%  
B
010000 : 42.5%  
B
101010 : 58.1%  
B
010001 : 25.1%  
B
101011 : 40.7%  
B
010001 : 43.1%  
B
101011 : 58.7%  
B
010010 : 25.7%  
B
101100 : 41.3%  
B
010010 : 43.7%  
B
101100 : 59.3%  
B
010011 : 26.3%  
B
101101 : 41.9%  
B
010011 : 44.3%  
B
101101 : 59.9%  
B
010100 : 26.9%  
B
101110 : 42.5%  
B
010100 : 44.9%  
B
101110 : 60.5%  
B
010101 : 27.5%  
101111 : 43.1%  
010101 : 45.5%  
101111 : 61.1%  
B
B
B
B
010110 : 28.1%  
110000 : 43.7%  
010110 : 46.1%  
110000 : 61.7%  
B
B
B
B
010111 : 28.7%  
110001 : 44.3%  
010111 : 46.7%  
110001 : 62.3%  
B
B
B
B
011000 : 29.3%  
110010 : 44.9%  
011000 : 47.3%  
110010 : 62.9%  
B
B
B
B
All Others: Reserved  
All Others: Reserved  
011001 : 29.9%  
011001 : 47.9%  
B
B
Notes:  
1. These values may be used for MR12 OP[5:0] to set the VREF(CA) levels in the LPDDR4-SDRAM.  
2. The range may be selected in the MR12 register by setting OP[6] appropriately.  
3. The MR12 registers represents either FSP[0] or FSP[1]. Two frequency-set-points each for CA and DQ are provided to allow for faster  
switching between terminated and un-terminated operation, or between different high-frequency setting which may use different  
terminations values.  
Version 1.0  
95  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR13 Register Information (MA[5:0] = 0DH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
FSP-OP  
FSP-WR  
DMD  
RRO  
VRCG  
VRO  
RPT  
CBT  
Register  
Function  
CBT  
Operand  
Data  
0 : Normal Operation (default)  
Notes  
Type  
B
OP[0]  
OP[1]  
1
(Command Bus Training)  
RPT  
1 : Command Bus Training Mode Enabled  
B
0 : Disable (default)  
B
(Read Preamble Training Mode)  
1 : Enable  
B
0 : Normal operation (default)  
B
VRO  
OP[2]  
2
1 : Output the V (CA) and V (DQ) values  
REF  
B
REF  
(VREF Output)  
on DQ bits  
0 : Normal Operation (default)  
VRCG  
(VREF Current Generator)  
RRO  
B
OP[3]  
OP[4]  
OP[5]  
OP[6]  
3
4, 5  
6
1 : V Fast Response (high current) mode  
B REF  
Write-only  
0 : Disable codes 001 and 010 in MR4 OP[2:0]  
B
(Refresh rate option)  
DMD  
1 : Enable all codes in MR4 OP[2:0]  
B
0 : Data Mask Operation Enabled (default)  
B
(Data Mask Disable)  
1 : Data Mask Operation Disabled  
B
FSP-WR  
0 : Frequency-Set-Point[0] (default)  
B
7
(Frequency Set Point Write Enable)  
1 : Frequency-Set-Point [1]  
B
FSP-OP  
0 : Frequency-Set-Point[0] (default)  
B
OP[7]  
8
(Frequency Set Point Operation Mode)  
1 : Frequency-Set-Point [1]  
B
Notes:  
1. A write to set OP[0]=1 causes the LPDDR4-SDRAM to enter the Command Bus Training mode. When OP[0]=1 and CKE goes LOW,  
commands are ignored and the contents of CA[5:0] are mapped to the DQ bus. CKE must be brought HIGH before doing a MRW to clear  
this bit (OP[0]=0) and return to normal operation. See the Command Bus Training section for more information.  
2. When set, the LPDDR4-SDRAM will output the VREF(CA) and VREF (DQ) voltages on DQ pins. Only the “active” frequency-set-point, as  
defined by MR13 OP[7], will be output on the DQ pins. This function allows an external test system to measure the internal VREF levels.  
The DQ pins used for VREF output are vendor specific.  
3. When OP[3]=1, the VREF circuit uses a high-current mode to improve VREF settling time.  
4. MR13 OP4 RRO bit is valid only when MR0 OP0 = 1. For LPDDR4 devices with MR0 OP0 = 0, MR4 OP[2:0] bits are not dependent on  
MR13 OP4.  
5. When OP[4] = 0, only 001B and 010B in MR4 OP[2:0] are disabled. LPDDR4 devices must report 011B instead of 001B or 010B in this case.  
Controller should follow the refresh mode reported by MR4 OP[2:0], regardless of RRO setting. TCSR function does not depend on RRO  
setting.  
Version 1.0  
96  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
6. When enabled (OP[5]=0B) data masking is enabled for the device. When disabled (OP[5]=1B), masked write command is illegal. See  
LPDDR4 Data Mask (DM) and Data Bus Inversion (DBIdc) Function.  
7. FSP-WR determines which frequency-set-point registers are accessed with MRW commands for the following functions such as VREF(CA)  
Setting, VREF(CA) Range, VREF(DQ) Setting, VREF(DQ) Range. For more information, refer to “Frequency Set Point.  
8. FSP-OP determines which frequency-set-point register values are currently used to specify device operation for the following functions  
such as VREF(CA) Setting, VREF(CA) Range, VREF(DQ) Setting, VREF(DQ) Range. For more information, refer to “Frequency Set Point  
section.  
Version 1.0  
97  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR14 Register Information (MA[5:0] = 0EH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
RFU  
VR(DQ)  
VREF(DQ)  
Register  
Function  
Operand  
Data  
Notes  
Type  
000000 :  
B
-- Thru --  
1,2,3,  
5,6  
V
(DQ)  
REF  
OP[5:0]  
OP[6]  
Read/  
Write  
(V (DQ) Setting)  
REF  
110010 : See table below  
B
All Others: Reserved  
VR(dq)  
1,2,4,  
5,6  
0 : V (DQ) Range[0] enabled  
B REF  
(V (DQ) Range)  
REF  
1 : V (DQ) Range[1] enabled (default)  
B REF  
Notes:  
1. This register controls the VREF(DQ) levels for Frequency-Set-Point[1:0]. Values from either VR(DQ)[0] or VR(dq)[1] may be selected by  
setting OP[6] appropriately.  
2. A read (MRR) to this register places the contents of OP[7:0] on DQ[7:0]. Any RFU bits and unused DQ's shall be set to‘0’. See the section  
on MRR Operation.  
3. A write to OP[5:0] sets the internal VREF(DQ) level for FSP[0] when MR13 OP[6]=0B, or sets FSP[1] when MR13 OP[6]=1B. The time  
required for VREF(DQ) to reach the set level depends on the step size from the cur-rent level to the new level. See the section on  
VREF(DQ) training for more information.  
4. A write to OP[6] switches the LPDDR4-SDRAM between two internal VREF(DQ) ranges. The range (Range[0] or Range[1]) must be  
selected when setting the VREF(DQ) register. The value, once set, will be retained until overwritten, or until the next power-on or RESET  
event.  
5. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. Only the registers for  
the set point determined by the state of the FSP-WR bit (MR13 OP[6]) will be written to with an MRW command to this MR address, or  
read from with an MRR command to this address.  
6. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. The device will  
operate only according to the values stored in the registers for the active set point, i.e., the set point determined by the state of the  
FSP-OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed  
without affecting device operation.  
Version 1.0  
98  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
VREF Settings for Range[0] and Range[1]  
Function Operand  
Notes  
Range[0] Values (% of V  
)
Range[1] Values (% of V  
)
DDQ  
DDQ  
000000 : 15.0%  
011010 : 30.5%  
000000 : 32.9%  
011010 : 48.5%  
B
B
B
B
000001 : 15.6%  
011011 : 31.1%  
000001 : 33.5%  
011011 : 49.1%  
B
B
B
B
000010 : 16.2%  
011100 : 31.7%  
000010 : 34.1%  
011100 : 49.7%  
B
B
B
B
011101 : 50.3%  
B
000011 : 16.8%  
011101 : 32.3%  
000011 : 34.7%  
B
B
B
(Default)  
000100 : 17.4%  
011110 : 32.9%  
000100 : 35.3%  
011110 : 50.9%  
B
B
B
B
000101 : 18.0%  
011111 : 33.5%  
000101 : 35.9%  
011111 : 51.5%  
B
B
B
B
000110 : 18.6%  
B
100000 : 34.1%  
B
000110 : 36.5%  
B
100000 : 52.1%  
B
000111 : 19.2%  
B
100001 : 34.7%  
B
000111 : 37.1%  
B
100001 : 52.7%  
B
001000 : 19.8%  
B
100010 : 35.3%  
B
001000 : 37.7%  
B
100010 : 53.3%  
B
001001 : 20.4%  
B
100011 : 35.9%  
B
001001 : 38.3%  
B
100011 : 53.9%  
B
001010 : 21.0%  
B
100100 : 36.5%  
B
001010 : 38.9%  
B
100100 : 54.5%  
B
V
REF  
001011 : 21.6%  
B
100101 : 37.1%  
B
001011 : 39.5%  
B
100101 : 55.1%  
B
Settings  
for  
OP[5:0]  
1,2,3  
001100 : 22.2%  
B
100110 : 37.7%  
B
001100 : 40.1%  
B
100110 : 55.7%  
B
001101 : 22.8%  
B
100111 : 38.3%  
B
001101 : 40.7%  
B
100111 : 56.3%  
B
MR14  
001110 : 23.4%  
B
101000 : 38.9%  
B
001110 : 41.3%  
B
101000 : 56.9%  
B
001111 : 24.0%  
B
101001 : 39.5%  
B
001111 : 41.9%  
B
101001 : 57.5%  
B
010000 : 24.6%  
B
101010 : 40.1%  
B
010000 : 42.5%  
B
101010 : 58.1%  
B
010001 : 25.1%  
B
101011 : 40.7%  
B
010001 : 43.1%  
B
101011 : 58.7%  
B
010010 : 25.7%  
B
101100 : 41.3%  
B
010010 : 43.7%  
B
101100 : 59.3%  
B
010011 : 26.3%  
B
101101 : 41.9%  
B
010011 : 44.3%  
B
101101 : 59.9%  
B
010100 : 26.9%  
B
101110 : 42.5%  
B
010100 : 44.9%  
B
101110 : 60.5%  
B
010101 : 27.5%  
101111 : 43.1%  
010101 : 45.5%  
101111 : 61.1%  
B
B
B
B
010110 : 28.1%  
110000 : 43.7%  
010110 : 46.1%  
110000 : 61.7%  
B
B
B
B
010111 : 28.7%  
110001 : 44.3%  
010111 : 46.7%  
110001 : 62.3%  
B
B
B
B
011000 : 29.3%  
110010 : 44.9%  
011000 : 47.3%  
110010 : 62.9%  
B
B
B
B
All Others: Reserved  
All Others: Reserved  
011001 : 29.9%  
011001 : 47.9%  
B
B
Notes:  
1. These values may be used for MR14 OP[5:0] to set the VREF(DQ) levels in the LPDDR4-SDRAM.  
2. The range may be selected in the MR14 register by setting OP[6] appropriately.  
3. The MR14 registers represents either FSP[0] or FSP[1]. Two frequency-set-points each for CA and DQ are pro-vided to allow for faster  
switching between terminated and un-terminated operation, or between different high frequency setting which may use different  
terminations values.  
Version 1.0  
99  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR15 Register Information (MA[5:0] = 0FH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Lower-Byte Invert Register for DQ Calibration  
Register  
Operand  
Type  
Function  
Data  
Notes  
The following values may be written for any  
operand OP[7:0], and will be applied to the  
corresponding DQ locations DQ[7:0] within a  
byte lane:  
Lower-Byte Invert  
for DQ Calibration  
Write-only  
OP[7:0]  
1,2,3  
0 : Do not invert  
B
1 : Invert the DQ Calibration patterns in MR32  
B
and MR40  
Default value for OP[7:0]=55  
H
Notes:  
1. This register will invert the DQ Calibration pattern found in MR32 and MR40 for any single DQ, or any combination of DQ's. Example: If  
MR15 OP[7:0]=00010101B, then the DQ Calibration patterns transmitted on DQ[7,6,5,3,1] will not be inverted, but the DQ Calibration  
patterns transmitted on DQ[4,2,0] will be inverted.  
2. DMI[0] is not inverted, and always transmits the “true” data contained in MR32/MR40.  
3. No Data Bus Inversion (DBI) function is enacted during DQ Read Calibration, even if DBI is enabled in MR3-OP[6].  
MR15 Invert Register Pin Mapping  
PIN  
DQ0  
DQ1  
DQ2  
DQ3  
DMI0  
DQ4  
DQ5  
DQ6  
DQ7  
MR15  
OP0  
OP1  
OP2  
OP3  
NO-Invert  
OP4  
OP5  
OP6  
OP7  
Version 1.0  
100  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR16 Register Information (MA[5:0] = 10H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
PASR Bank Mask  
Register  
Function  
Operand  
Data  
0 : Bank Refresh enabled (default) : Unmasked  
Notes  
Type  
B
Bank[7:0] Mask  
Write-only  
OP[7:0]  
1
1 : Bank Refresh disabled : Masked  
B
OP[n]  
Bank Mask  
8-Bank SDRAM  
Bank 0  
0
1
2
3
4
5
6
7
xxxxxxx1  
xxxxxx1x  
xxxxx1xx  
xxxx1xxx  
xxx1xxxx  
xx1xxxxx  
x1xxxxxx  
1xxxxxxx  
Bank 1  
Bank 2  
Bank 3  
Bank 4  
Bank 5  
Bank 6  
Bank 7  
Notes:  
1. When a mask bit is asserted (OP[n]=1), refresh to that bank is disabled.  
2. PASR bank-masking is on a per-channel basis. The two channels on the die may have different bank masking in dual channel devices.  
Version 1.0  
101  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR17 Register Information (MA[5:0] = 11H) for x16 mode  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
PASR Segment Mask  
Register  
Function  
Operand  
Data  
0 : Segment Refresh enabled (default)  
Notes  
Type  
B
PASR Segment Mask  
Write-only  
OP[7:0]  
1 : Segment Refresh disabled  
B
1Gb  
per  
2Gb  
per  
3Gb  
per  
4Gb  
per  
6Gb  
per  
8Gb  
per  
12Gb  
per  
16Gb  
per  
Segment  
Mask  
Segment OP[n]  
channel channel channel channel channel channel channel channel  
R12:R10 R13:R11 R14:R12 R14:R12 R15:R13 R15:R13 R16:R14 R16:R14  
0
1
0
1
xxxxxxx1  
xxxxxx1x  
000  
001  
B
B
2
2
3
4
5
6
7
xxxxx1xx  
xxxx1xxx  
xxx1xxxx  
xx1xxxxx  
x1xxxxxx  
1xxxxxxx  
010  
011  
100  
101  
B
B
B
B
3
4
5
6
110  
111  
110  
111  
110  
111  
110  
111  
110  
111  
B
B
B
B
B
Not  
Allowed  
Not  
Allowed  
Not  
Allowed  
7
B
B
B
B
B
Notes:  
1. This table indicates the range of row addresses in each masked segment. "X" is don't care for a particular segment.  
2. PASR segment-masking is on a per-channel basis. The two channels on the die may have different segment masking in dual channel  
devices.  
3. For 3Gb, 6Gb, and 12Gb densities, OP[7:6] must always be LOW (=00B).  
Version 1.0  
102  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR17 Register Information (MA[5:0] = 11H) for Byte mode (x8_2ch)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
PASR Segment Mask  
Register  
Function  
Operand  
Data  
0 : Segment Refresh enabled (default)  
Notes  
Type  
B
PASR Segment Mask  
Write-only  
OP[7:0]  
1 : Segment Refresh disabled  
B
1Gb  
per  
2Gb  
per  
3Gb  
per  
4Gb  
per  
6Gb  
per  
8Gb  
per  
12Gb  
per  
16Gb  
per  
Segment  
Mask  
Segment OP[n]  
channel channel channel channel channel channel channel channel  
R13:R11 R14:R12 R15:R13 R15:R13 R16:R14 R16:R14 R17:R15 R17:R15  
0
1
0
1
xxxxxxx1  
xxxxxx1x  
000  
001  
B
B
2
2
3
4
5
6
7
xxxxx1xx  
xxxx1xxx  
xxx1xxxx  
xx1xxxxx  
x1xxxxxx  
1xxxxxxx  
010  
011  
100  
101  
B
B
B
B
3
4
5
6
110  
111  
110  
111  
110  
111  
110  
111  
110  
111  
B
B
B
B
B
Not  
Allowed  
Not  
Allowed  
Not  
Allowed  
7
B
B
B
B
B
Notes:  
1. This table indicates the range of row addresses in each masked segment. "X" is don't care for a particular segment.  
2. PASR segment-masking is on a per-channel basis. The two channels on the die may have different segment masking.  
3. For 3Gb, 6Gb, and 12Gb densities, OP[7:6] must always be LOW (=00B).  
Version 1.0  
103  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR18 Register Information (MA[5:0] = 12H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
DQS Oscillator Count - LSB  
Register  
Function  
Operand  
Data  
Notes  
Type  
DQS Oscillator  
Read-only  
OP[7:0]  
0 - 255 LSB DRAM DQS Oscillator Count  
1,2,3  
(WR Training DQS Oscillator)  
Notes:  
1. MR18 reports the LSB bits of the DRAM DQS Oscillator count. The DRAM DQS Oscillator count value is used to train DQS to the DQ data  
valid window. The value reported by the DRAM in this mode register can be used by the memory controller to periodically adjust the  
phase of DQS relative to DQ.  
2. Both MR18 and MR19 must be read (MRR) and combined to get the value of the DQS Oscillator count.  
3. A new MPC [Start DQS Oscillator] should be issued to reset the contents of MR18/MR19.  
Version 1.0  
104  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR19 Register Information (MA[5:0] = 13H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
DQS Oscillator Count - MSB  
Register  
Type  
Function  
Operand  
Data  
Notes  
DQS Oscillator  
Read-only  
OP[7:0]  
0-255 MSB DRAM DQS Oscillator Count  
1,2,3  
(WR Training DQS Oscillator)  
Notes:  
1. MR19 reports the MSB bits of the DRAM DQS Oscillator count. The DRAM DQS Oscillator count value is used to train DQS to the DQ  
data valid window. The value reported by the DRAM in this mode register can be used by the memory controller to periodically adjust  
the phase of DQS relative to DQ.  
2. Both MR18 and MR19 must be read (MRR) and combined to get the value of the DQS Oscillator count.  
3. A new MPC [Start DQS Oscillator] should be issued to reset the contents of MR18/MR19.  
Version 1.0  
105  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR20 Register Information (MA[5:0] = 14H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Upper-Byte Invert Register for DQ Calibration  
Register  
Operand  
Type  
Function  
Data  
Notes  
The following values may be written for any  
operand OP[7:0], and will be applied to the  
corresponding DQ locations DQ[15:8] within a  
byte lane:  
Upper-Byte Invert  
for DQ Calibration  
Write-only  
OP[7:0]  
1,2  
0 : Do not invert  
B
1 : Invert the DQ Calibration patterns in MR32  
B
and MR40  
Default value for OP[7:0] = 55  
H
Notes:  
1. This register will invert the DQ Calibration pattern found in MR32 and MR40 for any single DQ, or any combination of DQ's. Example: If  
MR20 OP[7:0]=00010101B, then the DQ Calibration patterns transmitted on DQ[15,14,13,11,9] will not be inverted, but the DQ  
Calibration patterns transmitted on DQ[12,10,8] will be inverted.  
2. DMI[1] is not inverted, and always transmits the "true" data contained in MR32/MR40.  
3. No Data Bus Inversion (DBI) function is enacted during DQ Read Calibration, even if DBI is enabled in MR3-OP[6].  
MR20 Invert Register Pin Mapping  
PIN  
DQ8  
DQ9  
DQ10  
DQ11  
DMI1  
DQ12  
DQ13  
DQ14  
DQ15  
MR20  
OP0  
OP1  
OP2  
OP3  
NO-Invert  
OP4  
OP5  
OP6  
OP7  
Version 1.0  
106  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR22 Register Information (MA[5:0] = 16H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
ODTD for x8_2ch(Byte) mode  
ODTD-CA  
ODTE-CS  
ODTE-CK  
SOC ODT  
Register  
Function  
Operand  
Data  
000 : Disable (Default)  
Notes  
Type  
B
001 : RZQ/1(illegal if MR3 OP[0] = 0 )  
B
B
010 : RZQ/2  
B
SoC ODT  
011 : RZQ/3(illegal if MR3 OP[0] = 0 )  
B
B
(Controller ODT Value for  
VOH calibration)  
OP[2:0]  
1,2,3  
100 : RZQ/4  
B
101 : RZQ/5(illegal if MR3 OP[0] = 0 )  
B
B
110 : RZQ/6(illegal if MR3 OP[0] = 0 )  
B
B
111 : RFU  
B
ODTE-CK  
(CK ODT enabled for non  
terminating rank)  
ODTE-CS  
ODT bond PAD is ignored  
OP[3]  
OP[4]  
OP[5]  
2,3,4  
2,3,4  
2,3,4  
0 : ODT-CK Enabled (Default)  
B
1 : ODT-CK Disabled  
B
ODT bond PAD is ignored  
(CS ODT enable for non  
terminating rank)  
0 : ODT-CS Enabled (Default)  
B
Write-only  
1 : ODT-CS Disabled  
B
ODT bond PAD is ignored  
ODTD-CA  
0 : ODT-CA Enabled (Default)  
B
(CA ODT termination disable)  
1 : ODT-CA Disabled  
B
Byte mode device x8 2ch only, lower [7:0]  
Byte selected Device  
X8ODTD[7:0]  
(CA/CK ODT termination disable,  
[7:0] Byte select)  
OP[6]  
4
0 : ODT-CS/CA/CLK follows MR11 OP[6:4] and  
B
MR22 OP[5:3] (default)  
1 : ODT-CS/CA/CLK Disabled  
B
Byte mode device x8 2ch only, upper [15:8]  
Byte selected Device  
X8ODTD[15:8]  
(CA/CK ODT termination disable,  
[15:8] Byte select)  
OP[7]  
4
0 : ODT-CS/CA/CLK follows MR11 OP[6:4] and  
B
MR22 OP[5:3] (default)  
1 : ODT-CS/CA/CLK Disabled  
B
Notes:  
1. All values are “typical".  
2. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. Only the registers for  
the set point determined by the state of the FSP-WR bit (MR13 OP[6]) will be written to with an MRW command to this MR address, or  
read from with an MRR command to this address.  
Version 1.0  
107  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
3. There are two physical registers assigned to each bit of this MR parameter, designated set point 0 and set point 1. The device will  
operate only according to the values stored in the registers for the active set point, i.e., the set point determined by the state of the  
FSP-OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed  
without affecting device operation.  
4. The ODT_CA pin is ignored by LPDDR4X devices. The ODT_CA pin shall be connected to either VDD2 or VSS. CA/ CS/ CK ODT is fully  
controlled through MR11 and MR22. Before enabling CA termination via MR11, all ranks should have appropriate MR22 termination  
settings programmed.  
LPDDR4X Byte Mode Device (MR11 OP[6:4] 000 Case)  
B
ODT PAD Ignore  
CS  
ODTD  
Byte mode  
ODT  
CA  
ODT  
CS  
ODT  
CK  
CA  
CK  
MR22  
Lower  
Byte  
Upper  
Bye  
Lower  
Byte  
Upper  
Bye  
Lower  
Byte  
T
Upper  
Bye  
T
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
0
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
T
T
T
T
T
T
T
T
T
T
T
T
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
LP4X  
T
T
T
T
T
T
T
T
T
T
T
T
T
T
Notes:  
1. T means “terminated” condition. Blank is “unterminated”  
Version 1.0  
108  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR23 Register Information (MA[5:0] = 17H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
DQS interval timer run time setting  
Register  
Type  
Function  
Operand  
Data  
Notes  
00000000 : DQS interval timer stop via  
B
MPC Command (Default)  
00000001 : DQS timer stops automatically  
B
at 16th clocks after timer start  
00000010 : DQS timer stops automatically  
B
at 32nd clocks after timer start  
00000011 : DQS timer stops automatically  
B
at 48th clocks after timer start  
00000100 : DQS timer stops automatically  
B
DQS interval timer run time Write-only  
OP[7:0]  
at 64th clocks after timer start  
-------------- Thru --------------  
1, 2  
00111111 : DQS timer stops automatically  
B
at (63X16)th clocks after timer start  
01XXXXXX : DQS timer stops automatically  
B
at 2048th clocks after timer start  
10XXXXXX : DQS timer stops automatically  
B
at 4096th clocks after timer start  
11XXXXXX : DQS timer stops automatically  
B
at 8192nd clocks after timer start  
Notes:  
1. MPC command with OP[6:0]=1001101B (Stop DQS Interval Oscillator) stops DQS interval timer in case of MR23 OP[7:0] = 00000000B.  
2. MPC command with OP[6:0]=1001101B (Stop DQS Interval Oscillator) is illegal with non-zero values in MR23 OP[7:0].  
Version 1.0  
109  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR24 Register Information (MA[5:0] = 18H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Unlimited MAC  
TRR Mode  
TRR Mode BAn  
MAC Value  
Register  
Type  
Function  
Operand  
Data  
Notes  
000 : Unknown when bit OP3=0 (Note 1)  
B
Unlimited when bit OP3=1 (Note 2)  
001 : 700K  
B
010 : 600K  
B
MAC Value  
OP[2:0]  
011 : 500K  
B
Read-only  
100 : 400K  
B
101 : 300K  
B
110 : 200K  
B
111 : Reserved  
B
0 : OP[2:0] define MAC value  
B
Unlimited MAC  
TRR Mode BAn  
TRR Mode  
OP[3]  
OP[6:4]  
OP[7]  
1 : Unlimited MAC value (Note 2, Note 3)  
B
000 : Bank 0  
B
001 : Bank 1  
B
010 : Bank 2  
B
011 : Bank 3  
B
100 : Bank 4  
B
Write-only  
101 : Bank 5  
B
110 : Bank 6  
B
111 : Bank 7  
B
0 : Disabled (default)  
B
1 : Enabled  
B
Notes:  
1. Unknown means that the device is not tested for tMAC and pass/fail value in unknown.  
2. There is no restriction to number of activates.  
3. MR24 OP [2:0] is set to zero.  
Version 1.0  
110  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR25 Register Information (MA[5:0] = 19H)  
Mode Register 25 contains one bit of readout per bank indicating that at least one resource is available for Post  
Package Repair programming.  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Bank7  
Bank6  
Bank5  
Bank4  
Bank3  
Bank2  
Bank1  
Bank0  
Register  
Function  
Operand  
Data  
0 : PPR Resource is not available  
Notes  
Type  
B
PPR Resource  
Read-only  
OP[7:0]  
1 : PPR Resource is available  
B
Version 1.0  
12/2018  
111  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR30 Register Information (MA[5:0] = 1EH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Valid 0 or 1  
Register  
Function  
Operand  
Data  
Notes  
Type  
SDRAM will ignore  
Write-only  
OP[7:0]  
Don’t care  
Notes:  
1. This register is reserved for testing purposes. The logical data values written to OP[7:0] shall have no effect on SDRAM operation,  
however timings need to be observed as for any other MR access command..  
MR31 Register Information (MA[5:0] = 1FH)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Bytemode Vref Selection  
RFU  
Register  
Function  
Operand  
Data  
Notes  
Type  
0B: x16 device and no Byte mode  
selection(Default)  
Bytemode Vref Selection-  
Lower Byte  
OP[6]  
1,2,3  
1B: Disable to update MR12/MR14 for lower  
byte  
Write-only  
0B: x16 device and no Byte mode  
selection(Default)  
Bytemode Vref Selection-  
Upper Byte  
OP[7]  
1,2,3  
1B: Disable to update MR12/MR14 for upperr  
byte  
Notes:  
1. The byte mode Vref selecion is optional. Please consult with vendors for the availability to support feature.  
2. When Byte mode Vref selection is applied, the non-targeted byte is required to disable to update VrefCA and VrefDQ setting,  
assigned in MR12 and MR14 OP[6:0], for the other targeted byte.  
- In order to update MR12/MR14 setting only for upper byte, it is required to disable byte mode selection on  
lower byte, as applying MR31 OP[7:6] = 01B.  
- In order to update MR12/MR14 setting only for lower byte, it is required to disable byte mode selection on  
upper byte, as applying MR31 OP[7:6] = 10B.  
- When OP[7:6] = 00B is applied, both lower byte and upper byte will be updated.  
3. When the configuration is not composed of byte mode device, MR31 OP[7:6] shall be the default value, 00B.  
Version 1.0  
112  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR32 Register Information (MA[5:0] = 20H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
DQ Calibration Pattern “A” (default = 5A )  
H
Register  
Operand  
Type  
Function  
Data  
Notes  
X : An MPC command with OP[6:0]= 1000011  
B
B
causes the device to return the DQ Calibration  
Pattern contained in this register and (followed  
by) the contents of MR40. A default pattern  
Return DQ Calibration  
Pattern MR32 + MR40  
Write  
OP[7:0]  
1,2,3  
“5A ”is loaded at power-up or RESET, or the  
H
pattern may be overwritten with a MRW to this  
register. The contents of MR15 and MR20 will  
invert the data pattern for a given DQ  
(See MR15 for more information)  
MR39 Register Information (MA[5:0] = 27H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
Valid 0 or 1  
Register  
Function  
Operand  
Data  
Notes  
Type  
SDRAM will ignore  
Write-only  
OP[7:0]  
Don’t care  
1
Notes:  
1. This register is reserved for testing purposes. The logical data values written to OP[7:0] shall have no effect on SDRAM operation,  
however timings need to be observed as for any other MR access command.  
Version 1.0  
113  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR40 Register Information (MA[5:0] = 28H)  
OP[7]  
OP[6]  
OP[5]  
OP[4]  
OP[3]  
OP[2]  
OP[1]  
OP[0]  
DQ Calibration Pattern “B” (default = 3C )  
H
Register  
Operand  
Type  
Function  
Data  
X : A default pattern “3C ” is loaded at  
Notes  
B
H
Return DQ Calibration  
Pattern MR32 + MR40  
Write  
only  
power-up or RESET, or the pattern may be  
overwritten with a MRW to this register.  
See MR32 for more information.  
OP[7:0]  
1,2,3  
Notes:  
1. The pattern contained in MR40 is concatenated to the end of MR32 and transmitted on DQ[15:0] and DMI[1:0] when DQ Read  
Calibration is initiated via a MPC command. The pattern transmitted serially on each data lane, organized “little endian” such that the  
low-order bit in a byte is transmitted first. If the data pattern in MR40 is 27H , then the first bit transmitted with be a '1', followed by '1',  
'1', '0', '0', '1', '0', and '0'. The bit stream will be 00100111B.  
2. MR15 and MR20 may be used to invert the MR32/MR40 data patterns on the DQ pins. See MR15 and MR20 for more information. Data  
is never inverted on the DMI[1:0] pins..  
3. The data pattern is not transmitted on the DMI[1:0] pins if DBI-RD is disabled via MR3-OP[6].  
4. No Data Bus Inversion (DBI) function is enacted during DQ Read Calibration, even if DBI is enabled in MR3-OP[6].  
Version 1.0  
114  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Core Timing  
Core AC Timing for x16 mode  
Min/  
Max  
Parameter  
Symbol  
Data Rate  
Unit  
Note  
Core Parameters  
ACTIVATE-to-ACTIVATE  
command period (same bank)  
Minimum Self-Refresh Time  
(Entry to Exit)  
3733  
tRAS + tRPab (with all-bank precharge)  
tRAS + tRPpb (with per-bank precharge)  
tRC  
tSR  
Min  
Min  
Min  
ns  
ns  
ns  
max(15ns, 3nCK)  
SELF REFRESH exit to  
next valid command delay  
Exit Power-Down to next  
valid command delay  
CAS-to-CAS delay  
tXSR  
max(tRFCab + 7.5ns, 2nCK)  
tXP  
tCCD  
tRTP  
tRCD  
tRPpb  
Min  
Min  
Min  
Min  
Min  
max(7.5ns, 5nCK)  
8
ns  
tCK(avg)  
ns  
3
Internal READ to  
max(7.5ns, 8nCK)  
max(18ns, 4nCK)  
max(18ns, 4nCK)  
PRECHARGE command delay  
RAS-to-CAS delay  
ns  
Row precharge time  
ns  
(single bank)  
Row precharge time (all banks) tRPab  
Min  
max(21ns, 4nCK)  
ns  
Min  
Max  
Min  
Min  
Min  
Min  
Min  
max(42ns, 3nCK)  
ns  
us  
Row active time  
tRAS  
Min(9 * tREFI * Refresh Rate, 70.2 us)  
4
WRITE recovery time  
WRITE-to-READ delay  
tWR  
tWTR  
tRRD  
tPPD  
tFAW  
max(18ns, 6nCK)  
ns  
max(10ns, 8nCK)  
ns  
Active bank-A to active bank-B  
Precharge to Precharge Delay4  
Four-bank ACTIVATE window  
Notes:  
max(10ns, 4nCK)  
ns  
2
1
2
4
tCK(avg)  
ns  
40  
1. Precharge to precharge timing restriction does not apply to Auto-Precharge commands.  
2. Devices supporting 4267 Mbps specification shall support these timings at lower data rates.  
3. The value is based on BL16. For BL32 need additional 8 tCK(avg) delay.  
4. Refresh Rate is specified by MR4, OP[2:0]  
Version 1.0  
115  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Core AC Timing for Byte (x8) mode  
Min/  
Max  
Parameter  
Symbol  
Data Rate  
Unit  
Note  
Core Parameters  
3733  
WRITE recovery time  
WRITE-to-READ delay  
tWR  
Min  
Min  
max(20ns, 6nCK)  
max(12ns, 8nCK)  
ns  
ns  
tWTR  
Notes:  
1. The rest of the Core AC timing is the same as x16 mode.  
Version 1.0  
116  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tLZ(DQS), tLZ(DQ), tHZ(DQS), tHZ(DQ) Calculation  
tHZ and tLZ transitions occur in the same time window as valid data transitions. These parameters are referenced  
to a specific voltage level that specifies when the device output is no longer driving tHZ(DQS) and tHZ(DQ), or  
begins driving tLZ(DQS), tLZ(DQ).  
This section shows a method to calculate the point when the device is no longer driving tHZ(DQS) and tHZ(DQ), or  
begins driving tLZ(DQS), tLZ(DQ), by measuring the signal at two different voltages. The actual voltage  
measurement points are not critical as long as the calculation is consistent. The parameters tLZ(DQS), tLZ(DQ),  
tHZ(DQS), and tHZ(DQ) are defined as single ended.  
Version 1.0  
117  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tLZ(DQS) and tHZ(DQS) Calculation for ATE(Automatic Test Equipment)  
CKꢀCK corssing at 2nd CASꢀ2 of Read Command  
tLZ(DQS) method for calculating transitions and end point  
Notes:  
1. Conditions for Calibration: Pull Down Driver Ron = 40ohm, VOH = VDDQ/3  
2. Termination condition for DQS and DQS = 50ohm to VSSQ  
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances.  
Use the actual VOH value for tHZ and tLZ measurements.  
Version 1.0  
118  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CKꢀCK corssing at 2nd CASꢀ2 of Read Command  
tHZ(DQS) method for calculating transitions and end point  
Notes:  
1. Conditions for Calibration: Pull Down Driver Ron = 40ohm, VOH = VDDQ/3  
2. Termination condition for DQS and DQS = 50ohm to VSSQ  
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances.  
Use the actual VOH value for tHZ and tLZ measurements.  
Reference Voltage for tLZ(DQS), tHZ(DQS) Timing Measurements  
Measured Parameter  
Symbol  
tLZ(DQS)  
tHZ(DQS)  
Vsw1[V]  
0.4 x VOH  
0.4 x VOH  
Vsw2[V]  
Notes  
DQS low-impedance time from CK, CK  
DQS high impedance time from CK, CK  
0.6 x VOH  
0.6 x VOH  
Version 1.0  
12/2018  
119  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tLZ(DQ) and tHZ(DQ) Calculation for ATE(Automatic Test Equipment)  
CKꢀCK corssing at 2nd CASꢀ2 of Read Command  
tLZ(DQ) method for calculating transitions and end point  
Notes:  
1. Conditions for Calibration: Pull Down Driver Ron = 40ohm, VOH = VDDQ/3  
2. Termination condition for DQS and DMI = 50ohm to VSSQ  
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances.  
Use the actual VOH value for tHZ and tLZ measurements.  
Version 1.0  
120  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CKꢀCK corssing at 2nd CASꢀ2 of Read Command  
tHZ(DQ) method for calculating transitions and end point  
Notes:  
1. Conditions for Calibration: Pull Down Driver Ron = 40ohm, VOH = VDDQ/3  
2. Termination condition for DQS and DMI = 50ohm to VSSQ  
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances.  
Use the actual VOH value for tHZ and tLZ measurements.  
Reference Voltage for tLZ(DQ), tHZ(DQ) Timing Measurements  
Measured Parameter  
Symbol  
tLZ(DQ)  
tHZ(DQ)  
Vsw1[V]  
0.4 x VOH  
0.4 x VOH  
Vsw2[V]  
Notes  
DQ Low-impedance time from CK, CK  
DQ high impedance time from CK, CK  
0.6 x VOH  
0.6 x VOH  
Version 1.0  
12/2018  
121  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tRPRE Calculation for ATE(Automatic Test Equipment)  
The method for calculating differential pulse widths for tRPRE is shown in figure below.  
Method for calculating tRPRE transitions and endpoints  
Notes:  
1. Conditions for Calibration: Pull Down Driver Ron = 40ohm, VOH = VDDQ/3  
2. Termination condition for DQS, DQS, DQ and DMI = 50ohm to VSSQ  
3. Preamble = Static  
Method for calculating tRPRE transitions and endpoints  
Notes:  
1. Conditions for Calibration: Pull Down Driver Ron = 40ohm, VOH = VDDQ/3  
2. Termination condition for DQS, DQS, DQ and DMI = 50ohm to VSSQ  
3. Preamble = Toggle  
Reference Voltage for tRPRE Timing Measurements  
Measured Parameter  
Symbol  
Vsw1[V]  
Vsw2[V]  
-(0.7 x VOH)  
Notes  
DQS, DQS Differential Read Preamble  
tRPRE  
-(0.3 x VOH)  
Version 1.0  
12/2018  
122  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tRPST Calculation for ATE(Automatic Test Equipment)  
The method for calculating differential pulse widths for tRPST is shown in figure below.  
.
Method for calculating tRPST transitions and endpoints  
Notes:  
1. Conditions for Calibration: Pull Down Driver Ron = 40ohm, VOH = VDDQ/3  
2. Termination condition for DQS, DQS, DQ and DMI = 50ohm to VSSQ  
3. Read Postamble: 0.5tCK  
4. The method for calculating differential pulse widths for 1.5tCK Postamble is same as 0.5tCK Postamble.  
Reference Voltage for tRPST Timing Measurements  
Measured Parameter  
Symbol  
Vsw1[V]  
Vsw2[V]  
-(0.3 x VOH)  
Notes  
DQS, DQS Differential Read Postamble  
tRPST  
-(0.7 x VOH)  
Version 1.0  
12/2018  
123  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read AC Timing  
Parameter  
Read Timing  
Min/  
Symbol  
Data Rate  
Max  
Unit  
3733  
1.8  
READ preamble  
tRPRE  
tRPST  
tRPST  
Min  
Min  
Min  
tCK(avg)  
tCK(avg)  
tCK(avg)  
0.5 tCK READ postamble  
1.5 tCK READ postamble  
0.4  
1.4  
DQ low-impedance time from CK, CK  
DQ high impedance time from CK, CK  
tLZ(DQ)  
Min  
(RL x tCK) + tDQSCK(Min) - 200ps  
ps  
ps  
(RL x tCK) + tDQSCK(Max) + tDQSQ(Max)  
+ (BL/2 x tCK) - 100ps  
tHZ(DQ) Max  
tLZ(DQS) Min  
(RL x tCK) + tDQSCK(Min)  
- (tPRE(Max) x tCK) - 200ps  
(RL x tCK) + tDQSCK(Max) + (BL/2 x tCK)  
+ (RPST(Max) x tCK) - 100ps  
0.18  
DQS low-impedance time from CK, CK  
ps  
DQS high impedance time from CK, CK tHZ(DQS) Max  
DQS-DQ skew tDQSQ Max  
ps  
UI  
Version 1.0  
124  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tDQSCK Timing Table  
Parameter  
Symbol  
Min  
Max  
3.5  
4
Unit Notes  
DQS Output Access Time from CK/CK  
tDQSCK  
1.5  
ns  
1
2
3
DQS Output Access Time from CK/CK - Temperature Variation tDQSCK_temp  
DQS Output Access Time from CK/CK - Voltage Variation tDQSCK_volt  
Notes:  
-
-
ps/°C  
ps/mV  
7
1. Includes DRAM process, voltage and temperature variation. It includes the AC noise impact for frequencies > 20 MHz and max voltage  
of 45 mV pk-pk from DC-20 MHz at a fixed temperature on the package. The voltage supply noise must comply to the component  
Min-Max DC Operating conditions.  
2. tDQSCK_temp max delay variation as a function of Temperature.  
3. tDQSCK_volt max delay variation as a function of DC voltage variation for VDDQ and VDD2. tDQSCK_volt should be used to calculate  
timing variation due to VDDQ and VDD2 noise < 20 MHz. Host controller do not need to account for any variation due to VDDQ and VDD2  
noise > 20 MHz. The voltage supply noise must comply to the component Min-Max DC Operating conditions. The voltage variation is  
defined as the Max[abs{tDQSCKmin@V1- tDQSCKmax@V2}, abs{tDQSCKmax@V1-tDQSCKmin@V2}]/abs{V1-V2}. For tester  
measurement VDDQ = VDD2 is assumed.  
CK to DQS Rank to Rank variation  
tDQSCK_rank2rank Timing Table  
Min/  
Parameter  
Symbol  
Data Rate  
Max  
Unit Notes  
Read Timing  
CK to DQS Rank to Rank variation  
Notes:  
3733  
tDQSCK_rank2rank  
Max  
1.0  
ns  
1,2  
1. The same voltage and temperature are applied to tDQS2CK_rank2rank.  
2. tDQSCK_rank2rank parameter is applied to multi-ranks per byte lane within a package consisting of the same design dies.  
Version 1.0  
125  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write Preamble and Postamble  
The DQS strobe for the LPDDR4ꢀSDRAM requires a preꢀamble prior to the first latching edge (the rising edge of  
DQS with DATA "valid"), and it requires a postꢀamble after the last latching edge. The preꢀamble and postꢀamble  
lengths are set via mode register writes (MRW).  
For WRITE operations, a 2*tCK preꢀamble is required at all operating frequencies.  
LPDDR4 will have a DQS Write postꢀamble of 0.5*tCK or extended to 1.5*tCK. Standard DQS postꢀamble will be  
0.5*tCK driven by the memory controller for Writes. A mode register setting instructs the DRAM to drive an  
additional (extended) one cycle DQS Write postꢀamble. The drawings below show examples of DQS Write  
postꢀamble for both standard (tWPST) and extended (tWPSTE) postꢀamble operation.  
DQS Write Preamble and Postamble: 0.5nCK Postamble  
Notes:  
1. BL = 16, Postamble = 0.5nCK  
2. DQS and DQ terminated VSSQ  
3. DQS/DQS is “don’t care” prior to the start of tWPRE  
No transition of DQS is implied, as DQS/DQS can be HIGH, LOW or HI-Z prior to tWPRE  
Version 1.0  
126  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQS Write Preamble and Postamble: 1.5nCK Postamble  
Notes:  
1. BL = 16, Postamble = 1.5nCK  
2. DQS and DQ terminated VSSQ  
3. DQS/DQS is “don’t care” prior to the start of tWPRE  
No transition of DQS is implied, as DQS/DQS can be HIGH, LOW or HI-Z prior to tWPRE  
Version 1.0  
127  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst Write Operation  
A burst WRITE command is initiated with CS, and CA[5:0] asserted to the proper state at the rising edge of CK, as  
defined by the Command Truth Table. Column addresses C[3:2] should be driven LOW for Burst WRITE  
commands, and column addresses C[1:0] are not transmitted on the CA bus (and are assumed to be zero), so that  
the starting column burst address is always aligned with a 32B boundary. The write latency (WL) is defined from  
the last rising edge of the clock that completes a write command (Ex: the second rising edge of the CASꢀ2  
command) to the rising edge of the clock from which tDQSS is measured. The first valid “latching” edge of DQS  
must be driven WL * tCK + tDQSS after the rising edge of Clock that completes a write command.  
The LPDDR4ꢀSDRAM uses an unꢀmatched DQSꢀDQ path for lower power, so the DQSꢀstrobe must arrive at the  
SDRAM ball prior to the DQ signal by the amount of tDQS2DQ. The DQSꢀstrobe output is driven tWPRE before  
the first valid rising strobe edge. The tWPRE preꢀamble is required to be 2 x tCK. The DQSꢀstrobe must be trained  
to arrive at the DQ pad centerꢀaligned with the DQꢀdata. The DQꢀdata must be held for tDIVW (data input valid  
window) and the DQS must be periodically trained to stay centered in the tDIVW window to compensate for timing  
changes due to temperature and voltage variation. Burst data is captured by the SDRAM on successive edges of  
DQS until the 16 or 32 bit data burst is complete. The DQSꢀstrobe must remain active (toggling) for tWPST  
(WRITE postꢀamble) after the completion of the burst WRITE. After a burst WRITE operation, tWR must be  
satisfied before a PRECHARGE command to the same bank can be issued. Pin input timings are measured  
relative to the crosspoint of DQS and DQS.  
Version 1.0  
128  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst Write Operation  
Notes:  
1. BL = 16, Write Postamble = 0.5nCK, DQ/DQS: VSSQ termination  
2. Din n = data-in to columnm n  
3. The minimum number of clock cycles from the burst write command to burst read command for any bank is [WL +1 +BL/2 +  
RU(tWR/tCK)]  
4. tWR starts at the rising edge of CK after the last latching edge of DQS  
5. DES commands are shown for ease of illustration; other commands may be valid at these times  
Burst Write Followed by Burst Read  
Notes:  
1. BL = 16, Write Postamble = 0.5nCK, DQ/DQS: VSSQ termination  
2. Din n = data-in to columnm n  
3. The minimum number of clock cycles from the burst write command to burst read command for any bank is [WL +1 +BL/2 +  
RU(tWTR/tCK)].  
4. tWTR starts at the rising edge of CK after the last latching edge of DQS.  
5. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
129  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write Timing  
Notes:  
1. BL = 16, Write Postamble = 0.5nCK  
2. Din n = data-in to columnm.n  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
130  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tWPRE Calculation for ATE(Automatic Test Equipment)  
The method for calculating differential pulse widths for tWPRE is shown in figure below.  
Method for calculating tWPRE transitions and endpoints  
Notes:  
1. Termination condition for DQS, DQS DQ and DMI = 50ohm to VSSQ.  
Reference Voltage for tWPRE Timing Measurements  
Measured Parameter  
Symbol  
Vsw1[V]  
Vsw2[V]  
VIHL_AC x 0.7  
Notes  
DQS, DQS differential WRITE Preamble  
tWPRE  
VIHL_AC x 0.3  
Version 1.0  
12/2018  
131  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
tWPST Calculation for ATE(Automatic Test Equipment)  
The method for calculating differential pulse widths for tWPRE is shown in figure below.  
Method for calculating tWPST transitions and endpoints  
Notes:  
1. Termination condition for DQS, DQS DQ and DMI = 50ohm to VSSQ.  
2. Wrtie Postamble: 0.5tCK  
3. The method for calculating differential pulse widths for 1.5tCK Postamble is same as 0.5tCK Postamble.  
Reference Voltage for tWPST Timing Measurements  
Measured Parameter  
Symbol  
Vsw1[V]  
Vsw2[V]  
- (VIHL_AC x 0.3)  
Notes  
DQS, DQS differential Write Postamble  
tWPST  
- (VIHL_AC x 0.7)  
Version 1.0  
12/2018  
132  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write AC Timing  
Min/  
Parameter  
Write Timing  
Symbol  
Data Rate  
Max  
Unit  
Notes  
3733  
0.75  
1.25  
0.4  
Min  
Max  
Min  
Min  
Min  
Min  
Min  
Min  
Min  
Write command to 1st DQS latching transition  
tDQSS  
tCK(avg)  
DQS input high-level width  
DQS input low-level width  
DQS falling edge to CK setup time  
DQS falling edge hold time from CK  
Write preamble  
tDQSH  
tDQSL  
tDSS  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
0.4  
0.2  
tDSH  
0.2  
tWPRE  
tWPST  
tWPST  
1.8  
0.5 tCK Write postamble  
0.4  
1
1
1.5 tCK Write postamble  
1.4  
Notes:  
1. The length of Write Postamble depends on MR3 OP1 setting.  
Version 1.0  
133  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read and Write Latencies  
Read and Write Latencies for x16 mode  
Read Latency Write Latency  
Upper Clock Frequency  
Limit [MHz] ()  
Lower Clock Frequency  
Limit [MHz] (>)  
nWR nRTP  
No DBI  
w/ DBI  
Set A  
Set B  
6
6
4
4
6
8
8
10  
266  
533  
10  
14  
20  
24  
28  
32  
36  
12  
16  
22  
28  
32  
36  
40  
6
8
10  
16  
20  
24  
30  
34  
40  
266  
8
12  
18  
22  
26  
30  
34  
8
533  
800  
10  
12  
14  
16  
18  
8
800  
1066  
1333  
1600  
1866  
2133  
10  
12  
14  
16  
1066  
1333  
1600  
1866  
Notes:  
1. The LPDDR4 SDRAM device should not be operated at a frequency above the Upper Frequency Limit, or below the Lower Frequency  
Limit, shown for each RL, WL, nRTP, or nWR value.  
2. DBI for Read operations is enabled in MR3 OP[6]. When MR3 OP[6]=0, then the "No DBI" column should be used for Read Latency.  
When MR3 OP[6]=1, then the "w/DBI" column should be used for Read Latency.  
3. Write Latency Set "A" and Set "B" is determined by MR2 OP[6]. When MR2 OP[6]=0, then Write Latency Set "A" should be used. When  
MR2 OP[6]=1, then Write Latency Set "B" should be used.  
4. The programmed value of nWR is the number of clock cycles the LPDDR4 SDRAM device uses to determine the starting point of an  
internal Precharge operation after a Write burst with AP (Auto Pre-charge). It is determined by RU(tWR/tCK).  
5. The programmed value of nRTP is the number of clock cycles the LPDDR4 SDRAM device uses to determine the starting point of an  
internal Precharge operation after a Read burst with AP (Auto-Pre-charge). It is determined by RU(tRTP/tCK).  
6. nRTP shown in this table is valid for BL16 only. For BL32, the SDRAM will add 8 clocks to the nRTP value before starting a precharge.  
7. Clock Frequency herewith is a reference base on JEDEC's. Precise setting needs to follow where defined on speed compatible table in  
section “Operating Frequency, exceptional setting please confirm with NTC.  
Version 1.0  
134  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read and Write Latencies  
Read and Write Latencies for Byte (x8) mode  
Read Latency Write Latency  
Upper Clock Frequency  
Limit [MHz] ()  
Lower Clock Frequency  
Limit [MHz] (>)  
nWR nRTP  
No DBI  
w/ DBI  
Set A  
Set B  
6
6
4
4
6
8
8
10  
266  
533  
10  
16  
22  
26  
32  
36  
40  
12  
18  
24  
30  
36  
40  
44  
6
8
12  
16  
22  
28  
32  
38  
44  
266  
8
12  
18  
22  
26  
30  
34  
8
533  
800  
10  
12  
14  
16  
18  
8
800  
1066  
1333  
1600  
1866  
2133  
10  
12  
14  
16  
1066  
1333  
1600  
1866  
Notes:  
1. The LPDDR4 SDRAM device should not be operated at a frequency above the Upper Frequency Limit, or below the Lower Frequency  
Limit, shown for each RL, WL, nRTP, or nWR value.  
2. DBI for Read operations is enabled in MR3 OP[6]. When MR3 OP[6]=0, then the "No DBI" column should be used for Read Latency.  
When MR3 OP[6]=1, then the "w/DBI" column should be used for Read Latency.  
3. Write Latency Set "A" and Set "B" is determined by MR2 OP[6]. When MR2 OP[6]=0, then Write Latency Set "A" should be used. When  
MR2 OP[6]=1, then Write Latency Set "B" should be used.  
4. The programmed value of nWR is the number of clock cycles the LPDDR4 SDRAM device uses to determine the starting point of an  
internal Precharge operation after a Write burst with AP (Auto Pre-charge). It is determined by RU(tWR/tCK).  
5. The programmed value of nRTP is the number of clock cycles the LPDDR4 SDRAM device uses to determine the starting point of an  
internal Precharge operation after a Read burst with AP (Auto Pre-charge). It is determined by RU(tRTP/tCK).  
6. nRTP shown in this table is valid for BL16 only. For BL32, the SDRAM will add 8 clocks to the nRTP value before starting a precharge.  
7. Clock Frequency herewith is a reference base on JEDEC's. Precise setting needs to follow where defined on speed compatible table in  
section “Operating Frequency, exceptional setting please confirm with NTC.  
Version 1.0  
135  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write and Masked Write operation DQS controls (WDQS Control)  
LPDDR4ꢀSDRAMs support write and masked write operations with the following DQS controls. Before and  
after Write and Masked Write operations are issued, DQS/DQS is required to have a sufficient voltage  
gap to make sure the write buffers operating normally without any risk of metastability.  
The LPDDR4ꢀSDRAM is supported by either of two WDQS control modes.  
Mode 1: Read Based Control  
Mode 2: WDQS_on / WDQS_off definition based control  
Regardless of ODT enable/disable, WDQS related timing described in ‘WDQS Control’ does not allow any change  
of existing command timing constraints for all read/write operations. In case of any conflict or ambiguity on the  
command timing constraints caused by what is specified in ‘WDQS Control’, the specifications defined in ‘MPC’,  
‘Timing Constraints for Training Commands’ should have higher priority than WDQS control requirements.  
Some legacy products may not provide WDQS control described below. However, in order to prevent the  
write preamble related failure, it is strongly recommended to support either of two WDQS controls to  
LPDDR4ꢀSDRAMs. In the case of legacy SoC which may not provide WDQS control modes, it is required  
to consult DRAM vendors to guarantee the write / masked write operation appropriately.  
Version 1.0  
136  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
WDQS Control Mode 1 ꢀ Read Based Control  
controlled as described below. WDQS control requirements here can be ignored while differential read  
DQS is operated or while DQS hands over from Read to Write and vice versa.  
1.At the time a write / masked write command is issued, SoC makes the transition from driving DQS  
high to driving differential DQS/DQS, followed by normal differential burst on DQS pins.  
2.At the end of post amble of write /masked write burst, SoC resumes driving DQS high through the  
subsequent states except for DQS toggling and DQS turn around time of WTꢀRD and RDꢀWT as long  
as CKE is high.  
3.When CKE is low, the state of DQS and DQS is allowed to be “Don’t Care”.  
Version 1.0  
137  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
WDQS Control Mode 2 ꢀ WDQS_on/off  
After write / masked write command is issued, DQS and DQS required to be differential from WDQS_on, and DQS  
and DQS can be “Don’t Care” status from WDQS_off of write / masked write  
command. When ODT is enabled, WDQS_on and WDQS_off timing is located in the middle of the  
operations. When host disables ODT, WDQS_on and WDQS_off constraints conflict with tRTW. The timing  
does not conflict when ODT is enabled because WDQS_on and WDQS_off timing is covered in ODTLon  
and ODTLoff. However, regardless of ODT on/off, WDQS_on/off timing below does not change any  
command timing constraints for all read and write operations. In order to prevent the conflict, WDQS_on/off  
requirement can be ignored when WDQS_on/off timing is overlapped with read operation period including  
Read burst period and tRPST or overlapped with turnꢀaround time (RDꢀWT or WTꢀRD). In addition, the  
period during DQS toggling caused by Read and Write can be counted as WDQS_on/off.  
Parameters  
WDQS_on: the max delay from write / masked write command to differential DQS and DQS.  
WDQS_off : the min delay for DQS and DQS differential input after the last write / masked write  
command.  
WDQS_Exception : the period where WDQS_on and WDQS_off timing is overlapped with read  
operation or with DQS turn around (RDꢀWT, WTꢀRD).  
ꢀ WDQS_Exception @ ODT disable = max (WLꢀWDQS_on+tDQSTAꢀ tWPRE ꢀ n*tCK, 0 tCK)  
where RD to WT command gap = tRTW(min)@ODT disable + n*tCK  
ꢀ WDQS_Exception @ ODT enable = tDQSTA  
WDQS_on / WDQS_off Definition  
WDQS_on  
(max)  
WDQS_off  
(min)  
Write Latency  
Lower Clock Frequency Upper Clock Frequency  
nWR  
nRTP  
Limit [MHz] (>)  
Limit [MHz] ()  
Set A  
Set B  
Set A  
Set B Set A  
15  
18  
Set B  
15  
4
4
6
12  
16  
20  
24  
30  
34  
40  
nCK  
8
8
0
0
10  
266  
533  
6
8
8
12  
18  
22  
26  
30  
34  
nCK  
0
0
0
20  
25  
266  
8
6
12  
14  
18  
20  
24  
nCK  
21  
24  
533  
800  
10  
8
4
32  
800  
1066  
1333  
1600  
1866  
2133  
MHz  
12  
10  
12  
14  
16  
nCK  
4
27  
37  
1066  
1333  
1600  
1866  
MHz  
14  
6
30  
42  
16  
6
33  
47  
18  
8
36  
52  
nCK  
Notes:  
nCK  
nCK  
nCK  
1. WDQS_on/off requirement can be ignored when WDQS_on/off timing is overlapped with read operation period including Read burst  
period and tRPST or overlapped with turn-around time (RD-WT or WT-RD).  
2. The period during which DQS is toggling because of a Read or Write can be counted as part of the WDQS_on/off requirement.  
3. Clock Frequency herewith is a reference base on JEDEC's. Precise setting needs to follow where defined on speed compatible table in  
section “Operating Frequency, exceptional setting please confirm with NTC.  
Version 1.0  
138  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
WDQS_on / WDQS_off Allowable Variation Range  
Min  
Max  
Unit  
WDQS_On  
WDQS_Off  
-0.25  
0.25  
tCK(avg)  
-0.25  
0.25  
tCK(avg)  
Burst Write Operation  
Burst Write Operation  
Notes:  
1. BL=16, Write Postamble = 0.5nCK, DQ/DQS: VSSQ termination  
2. Din n = data-in to columnm n  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
4. DRAM RTT is only applied when ODT is enabled (MR11 OP[2:0] is not 000B)  
Version 1.0  
139  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst Read followed by Burst Write or Burst Mask Write (ODT Disable)  
Notes:  
1. BL=16, Read Preamble = Toggle, Read Postamble = 0.5nCK, Write Preamble = 2nCK, Write Postamble = 0.5nCK  
2. Dout n = data-out from column n and Din n = data-in to columnm n  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
4. WDQS_on and WDQS_off requirement can be ignored where WDQS_on/off timing is overlapped with read operation period including  
Read burst period and tRPST or overlapped with turn-around time (RD-WT or WT-RD)  
Version 1.0  
140  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst Read followed by Burst Write or Burst Mask Write (ODT Enable)  
Notes:  
1. BL=16, Read Preamble = Toggle, Read Postamble = 0.5nCK, Write Preamble = 2nCK, Write Postamble = 0.5nCK, DQ/DQS: VSSQ  
termination  
2. Dout n = data-out from column n and Din n = data-in to columnm n  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
4. WDQS_on and WDQS_off requirement can be ignored where WDQS_on/off timing is overlapped with read operation period including  
Read burst period and tRPST or overlapped with turn-around time (RD-WT or WT-RD)  
Version 1.0  
141  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Postamble and Preamble merging behavior  
The DQS strobe for the device requires a preamble prior to the first latching edge (the rising edge of DQS with data  
valid), and it requires a postamble after the last latching edge. The preamble and postamble options are set via  
Mode Register Write commands.  
In Read to Read or Write to Write operations with tCCD=BL/2, postamble for 1st command and preamble for 2nd  
command will disappear to create consecutive DQS latching edge for seamless burst operations. But in the case of  
Read to Read or Write to Write operations with command interval of tCCD+1,tCCD+2, etc., they will not completely  
disappear because it’s not seamless burst operations.  
Timing diagrams in this material describe Postamble and Preamble merging behavior in Read to Read or Write to  
Write operations with tCCD+n.  
Version 1.0  
142  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read to Read Operation  
Seamless Reads Operation: tCCD = Min, Preamble = Toggle, 1.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Toggle, Postamble = 1.5nCK2. Dout n = data-out from column n and Din n =  
data-in to columnm n  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Reads Operation: tCCD = Min +1, Preamble = Toggle, 1.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Toggle, Postamble = 1.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
143  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Reads Operation: tCCD = Min +1, Preamble = Toggle, 0.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Toggle, Postamble = 0.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Reads Operation: tCCD = Min +1, Preamble = Static, 1.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Static, Postamble = 1.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
144  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Reads Operation: tCCD = Min +1, Preamble = Static, 0.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Static, Postamble = 0.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Reads Operation: tCCD = Min +2, Preamble = Toggle, 1.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Toggle, Postamble = 1.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
145  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Reads Operation: tCCD = Min +2, Preamble = Toggle, 0.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Toggle, Postamble = 0.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Reads Operation: tCCD = Min +2, Preamble = Static, 1.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Static, Postamble = 1.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
146  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Reads Operation: tCCD = Min +2, Preamble = Static, 0.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Static, Postamble = 0.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Reads Operation: tCCD = Min +3, Preamble = Toggle, 1.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Toggle, Postamble = 1.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
147  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Reads Operation: tCCD = Min +3, Preamble = Toggle, 0.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Toggle, Postamble = 0.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Reads Operation: tCCD = Min +3, Preamble = Static, 1.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Static, Postamble = 1.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
148  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Reads Operation: tCCD = Min +3, Preamble = Static, 0.5nCK Postamble  
Notes:  
1. BL = 16 for column n and column m, RL = 6, Preamble = Static, Postamble = 0.5nCK  
2. Dout n/m = data-out from column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
149  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write to Write Operation  
Seamless Writes Operation: tCCD = Min, 0.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 0.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. The minimum number of clock cycles from the burst write command to the burst write command for any bank is BL/2  
4. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Seamless Writes Operation: tCCD = Min, 1.5nCK Postamble, 533MHz < Clock Freq. 800MHz,  
ODT Worst Timing Case  
Notes:  
1. Clock Frequency = 800MHz, tCK(AVG) = 1.25ns  
2. BL=16, Write Postamble = 1.5nCK  
3. Dout n/m = data-in to column n and column m.  
4. The minimum number of clock cycles from the burst write command to the burst write command for any bank is BL/2  
5. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
150  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Seamless Writes Operation: tCCD = Min, 1.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 1.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. The minimum number of clock cycles from the burst write command to the burst write command for any bank is BL/2  
4. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Writes Operation: tCCD = Min + 1, 0.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 0.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
151  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Writes Operation: tCCD = Min + 1, 1.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 1.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Writes Operation: tCCD = Min + 2, 0.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 0.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
152  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Writes Operation: tCCD = Min + 2, 1.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 1.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Writes Operation: tCCD = Min + 3, 0.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 0.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
153  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive Writes Operation: tCCD = Min + 3, 1.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 1.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Consecutive Writes Operation: tCCD = Min + 4, 1.5nCK Postamble  
Notes:  
1. BL=16, Write Postamble = 1.5nCK  
2. Dout n/m = data-in to column n and column m.  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
154  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MASKED WRITE OPERATION  
The LPDDR4ꢀSDRAM requires that Write operations which include a byte mask anywhere in the burst sequence  
must use the Masked Write command. This allows the DRAM to implement efficient data protection schemes  
based on larger data blocks. The Masked Writeꢀ1 command is used to begin the operation, followed by a CASꢀ2  
command. A Masked Write command to the same bank cannot be issued until tCCDMW later, to allow the  
LPDDR4ꢀSDRAM to finish the internal ReadꢀModifyꢀWrite. One Data MaskꢀInvert (DMI) pin is provided per byte  
lane, and the Data MaskꢀInvert timings match data bit (DQ) timing. See the section on "Data Mask Invert" for more  
information on the use of the DMI signal.  
Masked Write Command ꢀ Same Bank  
Notes:  
1. BL=16, Write Postamble = 0.5nCK, DQ/DQS: VSSQ termination  
2. Din n = data-in to columnm n  
3. Mask-Write supports only BL16 operations. For BL32 configuration, the system needs to insert only 16 bit wide data for masked write  
operation.  
4. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
155  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Masked Write Command ꢀ Different Bank  
Notes:  
1. BL=16, DQ/DQS/DMI: VSSQ termination  
2. Din n = data-in to columnm n  
3. Mask-Write supports only BL16 operations. For BL32 configuration, the system needs to insert only 16 bit wide data for masked write  
operation.  
4. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
156  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Masked Write Timing constraints for BL16  
Timing constraints for Same bank: DQ ODT is Disabled  
Next CMD  
Current CMD  
Active  
Read  
Write  
(BL=16 or 32)  
Active  
illegal  
illegal  
Masked Write  
Precharge  
(BL=16 or 32)  
RU(tRCD/tCK)  
RU(tRCD/tCK)  
RU(tRCD/tCK)  
RU(tRAS/tCK)  
BL/2+  
Read with  
BL = 16  
RL+RU(tDQSCK(max)/tCK)  
RL+RU(tDQSCK(max)/tCK)  
81  
+BL/2-WL+tWPRE+RD(tRPST) +BL/2-WL+tWPRE+RD(tRPST) max{(8,RU(tRTP/tCK)}-8  
RL+RU(tDQSCK(max)/tCK) RL+RU(tDQSCK(max)/tCK) BL/2+  
Read with  
BL = 32  
illegal  
illegal  
illegal  
illegal  
162  
+BL/2-WL+tWPRE+RD(tRPST) +BL/2-WL+tWPRE+RD(tRPST) max{(8,RU(tRTP/tCK)}-8  
Write with  
BL = 16  
WL+1+BL/2  
+RU(tWTR/tCK)  
WL+1+BL/2  
WL+ 1 +  
81  
tCCDMW3  
tCCDMW + 84  
tCCDMW3  
illegal  
BL/2+RU(tWR/tCK)  
WL+ 1 +  
Write with  
BL = 32  
162  
+RU(tWTR/tCK)  
WL+1+BL/2  
BL/2+RU(tWR/tCK)  
WL+ 1 +  
Masked Write  
tCCD  
illegal  
+RU(tWTR/tCK)  
BL/2+RU(tWR/tCK)  
RU(tRP/tCK),  
Precharge  
Notes:  
illegal  
4
RU(tRPab/tCK)  
1. In the case of BL = 16, tCCD is 8*tCK.  
2. In the case of BL = 32, tCCD is 16*tCK.  
3. tCCDMW = 32*tCK (4*tCCD at BL=16)  
4. Write with BL=32 operation has 8*tCK longer than BL =16.  
5. tRPST values depend on MR1-OP[7] respectively.  
Timing constraints for Same bank: DQ ODT is Enabled  
Next CMD  
Read  
Write  
Active  
Masked Write  
Precharge  
(BL=16 or 32)  
(BL=16 or 32)  
Current CMD  
RL+RU(tDQSCK(max)/tCK) RL+RU(tDQSCK(max)/tCK)  
+BL/2+RD(tRPST)-ODTLon +BL/2+RD(tRPST)-ODTLon  
-RD(tODTon,min/tCK)+1 -RD(tODTon,min/tCK)+1  
RL+RU(tDQSCK(max)/tCK) RL+RU(tDQSCK(max)/tCK)  
+BL/2+RD(tRPST)-ODTLon +BL/2+RD(tRPST)-ODTLon  
-RD(tODTon,min/tCK)+1 -RD(tODTon,min/tCK)+1  
Read with  
BL = 16  
BL/2+  
illegal  
81  
max{(8,RU(tRTP/tCK)}-8  
Read with  
BL = 32  
BL/2+  
illegal  
162  
max{(8,RU(tRTP/tCK)}-8  
Notes:  
1. In the case of BL = 16, tCCD is 8*tCK.  
2. In the case of BL = 32, tCCD is 16*tCK.  
3. The rest of the timing is same as DQ ODT is Disable case.  
4. tRPST values depend on MR1-OP[7] respectively.  
Version 1.0  
157  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing constraints for Different bank: DQ ODT is Disabled  
Next CMD  
Current CMD  
Active  
Read  
(BL=16 or 32)  
4
Write  
Active  
RU(tRRD/tCK)  
4
Masked Write  
Precharge  
(BL=16 or 32)  
4
4
4
4
Read with  
BL = 16  
RL+RU(tDQSCK(max)/tCK)  
+BL/2-WL+tWPRE+RD(tRPST)  
RL+RU(tDQSCK(max)/tCK)  
+BL/2-WL+tWPRE+RD(tRPST)  
RL+RU(tDQSCK(max)/tCK)  
+BL/2-WL+tWPRE+RD(tRPST)  
RL+RU(tDQSCK(max)/tCK)  
+BL/2-WL+tWPRE+RD(tRPST)  
81  
Read with  
BL = 32  
4
4
4
162  
4
4
4
Write with  
BL = 16  
WL+1+BL/2  
+RU(tWTR/tCK)  
WL+1+BL/2  
+RU(tWTR/tCK)  
WL+1+BL/2  
+RU(tWTR/tCK)  
4
81  
81  
Write with  
BL = 32  
162  
162  
Masked Write  
4
4
81  
4
81  
4
4
4
Precharge  
Notes:  
1. In the case of BL = 16, tCCD is 8*tCK.  
2. In the case of BL = 32, tCCD is 16*tCK.  
3. tRPST values depend MR1-OP[7] respectively  
Timing constraints for Different bank: DQ ODT is Enabled  
Next CMD  
Read  
Write  
Active  
Masked Write  
Precharge  
(BL=16 or 32)  
(BL=16 or 32)  
Current CMD  
RL+RU(tDQSCK(max)/tCK)+BL/2 RL+RU(tDQSCK(max)/tCK)+BL/2  
Read with  
BL = 16  
4
81  
+RD(tRPST)-ODTLon  
+RD(tRPST)-ODTLon  
2
-RD(tODTon,min/tCK)+1  
-RD(tODTon,min/tCK)+1  
RL+RU(tDQSCK(max)/tCK)+BL/2 RL+RU(tDQSCK(max)/tCK)+BL/2  
Read with  
BL = 32  
4
162  
+RD(tRPST)-ODTLon  
+RD(tRPST)-ODTLon  
2
-RD(tODTon,min/tCK)+1  
-RD(tODTon,min/tCK)+1  
Notes:  
1. In the case of BL = 16, tCCD is 8*tCK.  
2. In the case of BL = 32, tCCD is 16*tCK.  
3. The rest of the timing is same as DQ ODT is Disable case.  
4. tRPST values depend MR1-OP[7] respectively  
Version 1.0  
158  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
LPDDR4 Data Mask (DM) and Data Bus Inversion (DBIdc) Function  
LPDDR4 SDRAM supports the function of Data Mask and Data Bus inversion. Details are shown below:  
• LPDDR4 device supports Data Mask (DM) function for Write operation.  
• LPDDR4 device supports Data Bus Inversion (DBIdc) function for Write and Read operation.  
• LPDDR4 supports DM and DBIdc function with a byte granularity.  
• DBIdc function during Write or Masked Write can be enabled or disabled through MR3 OP[7].  
• DBIdc function during Read can be enabled or disabled through MR3 OP[6].  
• DM function during Masked Write can be enabled or disabled through MR13 OP[5].  
• LPDDR4 device has one Data Mask Inversion (DMI) signal pin per byte; total of 2 DMI signals per channel.  
• DMI signal is a biꢀdirectional DDR signal and is sampled along with the DQ signals for Read and Write or  
Masked Write operation.  
There are eight possible combinations for LPDDR4 device with DM and DBIdc function.  
Version 1.0  
159  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Function Behavior of DMI Signal During Write, Masked Write and Read Operation  
Signal  
DMI Signal  
DMI Signal  
DMI Signal DMI Signal  
Write  
DBIdc  
during DMI Signal  
during DMI Signal  
DM  
Read DBIdc during  
during  
during  
Masked  
Write  
During  
Read  
MPC DQ  
Read  
During  
MRR  
Function  
Function  
Write  
MPC WR MPC RD  
Function  
Command  
FIFO  
FIFO  
Command  
Training  
Disable  
Disable  
Disable  
Disable  
Enable  
Enable  
Enable  
Enable  
Notes:  
Disable  
Enable  
Disable  
Enable  
Disable  
Enable  
Disable  
Enable  
Disable  
Disable  
Enable  
Enable  
Disable  
Disable  
Enable  
Enable  
Notes: 1 Notes: 1, 3 Notes: 2  
Notes: 4 Notes: 3 Notes: 2  
Notes: 1 Notes: 3 Notes: 5  
Notes: 4 Notes: 3 Notes: 5  
Notes: 6 Notes: 7 Notes: 2  
Notes: 4 Notes: 8 Notes: 2  
Notes: 6 Notes: 7 Notes: 5  
Notes: 4 Notes: 8 Notes: 5  
Note: 1  
Note: 9  
Note: 9  
Note: 9  
Note: 9  
Note: 9  
Note: 9  
Note: 9  
Note: 2  
Note: 2  
Notes: 2  
Note: 10 Note: 11 Notes: 2  
Note: 10 Note: 11 Notes: 12  
Note: 10 Note: 11 Notes: 12  
Note: 10 Note: 11 Notes: 2  
Note: 10 Note: 11 Notes: 2  
Note: 10 Note: 11 Notes: 12  
Note: 10 Note: 11 Notes: 12  
1. DMI input signal is a don't care. DMI input receivers are turned OFF.  
2. DMI output drivers are turned OFF.  
3. Masked Write Command is not allowed and is considered an illegal command as DM function is disabled.  
4. DMI signal is treated as DBI signal and it indicates whether DRAM needs to invert the Write data received on DQs within a byte. The  
LPDDR4 device inverts Write data received on the DQ inputs in case DMI was sampled HIGH, or leaves the Write data non-inverted in  
case DMI was sampled LOW.  
5. The LPDDR4 DRAM inverts Read data on its DQ outputs associated within a byte and drives DMI signal HIGH when the number of ‘1’  
data bits within a given byte lane is greater than four; otherwise the DRAM does not invert the read data and drives DMI signal LOW.  
6. The LPDDR4 DRAM does not perform any mask operation when it receives Write command. During the Write burst associated with  
Write command, DMI signal must be driven LOW.  
7. The LPDDR4 DRAM requires an explicit Masked Write command for all masked write operations. DMI signal is treated as DM signal and  
it indicates which bit time within the burst is to be masked. When DMI signal is HIGH, DRAM masks that bit time across all DQs  
associated within a byte. All DQ input signals within a byte are don't care (either HIGH or LOW) when DMI signal is HIGH. When DMI  
signal is LOW, the LPDDR4 DRAM does not perform mask operation and data received on DQ input is written to the array.  
8. The LPDDR4 DRAM requires an explicit Masked Write command for all masked write operations. The LPDDR4 device masks the Write  
data received on the DQ inputs if the total count of '1' data bits on DQ[2:7] or DQ[10:15] (for Lower Byte or Upper Byte respectively) is  
equal to or greater than five and DMI signal is LOW. Otherwise the LPDDR4 DRAM does not perform mask operation and treats it as a  
legal DBI pattern; DMI signal is treated as DBI signal and data received on DQ input is written to the array.  
9. DMI signal is treated as a training pattern. The LPDDR4 DRAM does not perform any mask operation and does not invert Write data  
received on the DQ inputs.  
10. DMI signal is treated as a training pattern. The LPDDR4 DRAM returns DMI pattern written in WR FIFO.  
11. DMI signal is treated as a training pattern. For more details, see ‘RD DQ Calibration’.  
Version 1.0  
160  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
12. DBI may apply or may not apply during normal MRR. It's vendor specific.  
If read DBI is enable with MRS and vendor cannot support the DBI during MRR, DBI pin status should be low.  
If read DBI is enable with MRS and vendor can support the DBI during MRR, the LPDDR4 DRAM inverts Mode Register Read data on its  
DQ outputs associated within a byte and drives DMI signal HIGH when the number of ‘1’ data bits within a given byte lane is greater  
than four; otherwise the DRAM does not invert the read data and drives DMI signal LOW.  
Version 1.0  
161  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Masked Write Command w/ Write DBI Enabled; DM Enabled  
Notes:  
1. Data Mask (DM) is Enable: MR13 OP [5] = 0, Data BUS Inversion (DBI) Write is Enable: MR3 OP[7] = 1  
Version 1.0  
162  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write Command w/ Write DBI Enabled; DM Enabled  
Notes:  
1. Data Mask (DM) is Disable: MR13 OP [5] = 1, Data BUS Inversion (DBI) Write is Enable: MR3 OP[7] = 1  
Version 1.0  
163  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
PreꢀCharge Operation  
The PRECHARGE command is used to precharge or close a bank that has been activated. The PRECHARGE  
command is initiated with CS, and CA[5:0] in the proper state as defined by the Command Truth Table.  
The PRECHARGE command can be used to precharge each bank independently or all banks simultaneously.  
The AB flag and the bank address bit are used to determine which bank(s) to precharge. The precharged bank(s)  
will be available for subsequent row access tRPab after an allꢀbank PRECHARGE command is issued, or tRPpb  
after a singleꢀbank PRECHARGE command is issued.  
To ensure that LPDDR4 devices can meet the instantaneous current demands, the rowꢀprecharge time for an  
allꢀbank PRECHARGE (tRPab) is longer than the perbank precharge time (tRPpb).  
Precharge Bank Selection  
AB  
BA2  
BA1  
BA0  
Precharged  
Bank(s)  
(CA[5], R1)  
(CA[2], R2)  
(CA[1], R2)  
(CA[0], R2)  
0
0
0
0
0
0
0
0
1
0
0
0
Bank 0 only  
Bank 1 only  
Bank 2 only  
Bank 3 only  
Bank 4 only  
Bank 5 only  
Bank 6 only  
Bank 7 only  
All Banks  
0
0
1
0
1
0
0
1
1
1
1
0
0
0
1
1
1
0
1
1
1
Valid  
Valid  
Valid  
Version 1.0  
12/2018  
164  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst Read Operation Followed by a PRECHARGE  
The PRECHARGE command can be issued as early as BL/2 clock cycles after a READ command, but  
PRECHARGE cannot be issued until after tRAS is satisfied. A new bank ACTIVATE command can be issued to  
the same bank after the row PRECHARGE time (tRP) has elapsed. The minimum READꢀtoꢀPRECHARGE time  
must also satisfy a minimum analog time from the 2nd rising clock edge of the CASꢀ2 command. tRTP begins BL/2  
ꢀ 8 clock cycles after the READ command. For LPDDR4 READꢀtoꢀPRECHARGE timings see table below.  
Burst READ followed by PRECHARGE (Shown with BL16, 2tCK preꢀamble)  
Burst READ followed by PRECHARGE (Shown with BL32, 2tCK preꢀamble)  
Version 1.0  
165  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst WRITE Followed by PRECHARGE  
A Write Recovery time (tWR) must be provided before a PRECHARGE command may be issued. This delay is  
referenced from the next rising edge of CK after the last latching DQS clock of the burst.  
LPDDR4ꢀSDRAM devices write data to the memory array in prefetch multiples (prefetch=16). An internal WRITE  
operation can only begin after a prefetch group has been clocked, so tWR starts at the prefetch boundaries. The  
minimum WRITEꢀtoꢀPRECHARGE time for commands to the same bank is WL + BL/2 + 1 + RU(tWR/tCK) clock  
cycles.  
Burst WRITE Followed by PRECHARGE (Shown with BL16, 2tCK preꢀamble)  
AutoꢀPRECHARGE Operation  
Before a new row can be opened in an active bank, the active bank must be precharged using either the  
PRECHARGE command or the AutoꢀPRECHARGE function. When a READ, a WRITE or Masked Write command  
is issued to the device, the AP bit (CA5) can be set to enable the active bank to automatically begin precharge at  
the earliest possible moment during the burst READ, WRITE or Masked Write cycle.  
If AP is LOW when the READ or WRITE command is issued, then the normal READ, WRITE or Masked Write  
burst operation is executed and the bank remains active at the completion of the burst.  
If AP is HIGH when the READ, WRITE or Masked Write command is issued, the AutoꢀPRECHARGE function is  
engaged. This feature enables the PRECHARGE operation to be partially or completely hidden during burst READ  
cycles (dependent upon READ or WRITE latency), thus improving system performance for random data access.  
Version 1.0  
166  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst READ with AutoꢀPRECHARGE  
If AP is HIGH when a READ command is issued, the READ with AutoꢀPRECHARGE function is engaged. An  
internal precharge procedure starts a following delay time after the READ command. And this delay time depends  
on BL setting.  
BL = 16: nRTP  
BL = 32: 8nCK + nRTP  
For LPDDR4 AutoꢀPRECHARGE calculations. Following an AutoꢀPRECHARGE operation, an ACTIVATE  
command can be issued to the same bank if the following two conditions are both satisfied:  
a. The RAS precharge time (tRP) has been satisfied from the clock at which the AutoꢀPRECHARGE began, or  
b. The RAS cycle time (tRC) from the previous bank activation has been satisfied.  
Burst READ with AutoꢀPRECHARGE (Shown with BL16, 2tCK preꢀamble)  
Burst READ with AutoꢀPRECHARGE (Shown with BL32, 2tCK preꢀamble)  
Version 1.0  
167  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst WRITE with AutoꢀPRECHARGE  
If AP is HIGH when a WRITE command is issued, the WRITE with AutoꢀPRECHARGE function is engaged. The  
device starts an AutoꢀPRECHARGE on the rising edge tWR cycles after the completion of the Burst WRITE.  
Following a WRITE with AutoꢀPRECHARGE, an ACTIVATE command can be issued to the same bank if the  
following conditions are met:  
a. The RAS precharge time (tRP) has been satisfied from the clock at which the AutoꢀPRECHARGE began, and  
b. The RAS cycle time (tRC) from the previous bank activation has been satisfied.  
Burst WRITE with AutoꢀPRECHARGE (Shown with BL16, 2tCK preꢀamble)  
Version 1.0  
168  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AutoꢀPrecharge Operation  
Before a new row can be opened in an active bank, the active bank must be precharged using either the  
PRECHARGE command or the AutoꢀPrecharge function. When a READ, a WRITE or Masked Write  
command is issued to the device, the AP bit (CA5) can be set to enable the active bank to automatically  
begin precharge at the earliest possible moment during the burst READ, WRITE or Masked Write cycle.  
If AP is LOW when the READ or WRITE command is issued, then the normal READ, WRITE or Masked  
Write burst operation is executed and the bank remains active at the completion of the burst.  
If AP is HIGH when the READ, WRITE or Masked Write command is issued, the AutoꢀPrecharge function  
is engaged. This feature enables the PRECHARGE operation to be partially or completely hidden during  
burst READ cycles (dependent upon READ or WRITE latency), thus improving system performance for  
random data access.  
Read with AutoꢀPrecharge or Write/Mask Write with AutoꢀPrecharge commands may be issued after tRCD  
has been satisfied. The LPDDR4 SDRAM RAS Lockout feature will schedule the internal precharge to  
assure that tRAS is satisfied.  
tRC needs to be satisfied prior to issuing subsequent Activate commands to the same bank.  
The following figure shows example of RAS lock function.  
Command Input Timing with RAS lock  
Notes:  
1. tCK(AVG) = 0.938ns, Data Rate = 2133Mbps, tRCD(Min) = Max(18ns, 4nCK), tRAS(Min) = Max(42ns, 3nCK), nRTP = 8nCK, BL = 32  
2. tRCD = 20nCK comes from Roundup(18ns/0.938ns)  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
169  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Delay time from Write to Read with AutoꢀPrecharge  
In the case of write command followed by read with AutoꢀPrecharge, controller must satisfy tWR for the  
write command before initiating the DRAM internal AutoꢀPrecharge. It means that (tWTR + nRTP) should  
be equal or longer than (tWR) when BL setting is 16, as well as (tWTR + nRTP +8nCK) should be equal or  
longer than (tWR) when BL setting is 32. Refer to the following figure for details.  
Delay time from Write to Read with AutoꢀPrecharge  
Notes:  
1. Burst Length at Read = 16  
2. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
170  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing Between Commands (PRECHARGE and AutoꢀPRECHARGE): DQ ODT is Disable  
From  
Minimum Delay between "From Command"  
and "To Command"  
To Command  
Unit  
Notes  
Command  
PRECHARGE  
(to same bank as Read)  
PRECHARGE All  
tRTP  
tRTP  
tCK  
tCK  
tCK  
tCK  
tCK  
tCK  
tCK  
1,6  
1,6  
READ  
BL=16  
PRECHARGE  
8tCK + tRTP  
8tCK + tRTP  
nRTP  
1,6  
READ  
(to same bank as Read)  
PRECHARGE All  
BL=32  
1,6  
PRECHARGE  
1,10  
1,10  
1,8,10  
(to same bank as READ w/AP)  
PRECHARGE All  
nRTP  
Activate  
nRTP + tRPpb  
(to same bank as READ w/AP)  
WRITE or WRITE w/AP  
(same bank)  
Illegal  
Illegal  
-
MASK-WR or MASK-WR w/AP  
(same bank)  
READ w/AP  
BL=16  
-
WRITE or WRITE w/AP  
(different bank)  
RL+RU(tDQSCK(max)/tCK)+BL/2+  
RD(tRPST)-WL+tWPRE  
tCK  
tCK  
-
3,4,5  
3,4,5  
MASK-WR or MASK-WR w/AP  
(different bank)  
RL+RU(tDQSCK(max)/tCK)+BL/2+  
RD(tRPST)-WL+tWPRE  
READ or READ w/AP  
(same bank)  
Illegal  
BL/2  
READ or READ w/AP  
(different bank)  
tCK  
3
Version 1.0  
171  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing Between Commands (PRECHARGE and AutoꢀPRECHARGE): DQ ODT is Disable (cont’d)  
From  
Minimum Delay between "From Command"  
and "To Command"  
To Command  
Unit  
Notes  
Command  
PRECHARGE  
(to same bank as READ w/AP)  
PRECHARGE All  
8tCK + nRTP  
8tCK + nRTP  
tCK  
tCK  
tCK  
1,10  
1,10  
Activate  
8tCK + nRTP +tRPpb  
1,8,10  
(to same bank as READ w/AP)  
WRITE or WRITE w/AP  
(same bank)  
Illegal  
Illegal  
-
-
MASK-WR or MASK-WR w/AP  
(same bank)  
READ w/AP  
BL=32  
WRITE or WRITE w/AP  
(different bank)  
RL+RU(tDQSCK(max)/tCK)+BL/2+RD(tRPST)-WL+tWPRE tCK  
RL+RU(tDQSCK(max)/tCK)+BL/2+RD(tRPST)-WL+tWPRE tCK  
3,4,5  
3,4,5  
MASK-WR or MASK-WR w/AP  
(different bank)  
READ or READ w/AP  
(same bank)  
Illegal  
BL/2  
-
READ or READ w/AP  
(different bank)  
tCK  
3
PRECHARGE  
WL + BL/2 + tWR + 1  
WL + BL/2 + tWR + 1  
WL + BL/2 + tWR + 1  
WL + BL/2 + tWR + 1  
tCK  
tCK  
tCK  
tCK  
1,7  
1,7  
1,7  
1,7  
WRITE  
(to same bank as WRITE)  
PRECHARGE All  
BL=16 & 32  
PRECHARGE  
MASK-WR  
BL=16  
(to same bank as MASK-WR)  
PRECHARGE All  
Version 1.0  
172  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing Between Commands (PRECHARGE and AutoꢀPRECHARGE): DQ ODT is Disable (cont’d)  
From  
Minimum Delay between "From Command"  
and "To Command"  
To Command  
Unit  
Notes  
Command  
PRECHARGE  
(to same bank as WRITE w/AP)  
PRECHARGE All  
WL + BL/2 + nWR + 1  
WL + BL/2 + nWR + 1  
tCK  
tCK  
tCK  
1,11  
1,11  
ACTIVATE  
WL + BL/2 +nWR + 1 + tRPpb  
1,8,11  
(to same bank as WRITE w/AP)  
WRITE or WRITE w/AP  
(same bank)  
Illegal  
-
WRITE w/AP  
BL=16 & 32  
READ or READ w/AP  
(same bank)  
Illegal  
BL/2  
-
WRITE or WRITE w/AP  
(different bank)  
tCK  
tCK  
tCK  
3
3
MASK-WR or MASK-WR w/AP  
(different bank)  
BL/2  
READ or READ w/AP  
(different bank)  
WL + BL/2 + tWTR + 1  
3,9  
Version 1.0  
173  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing Between Commands (PRECHARGE and AutoꢀPRECHARGE): DQ ODT is Disable (cont’d)  
From  
Minimum Delay between "From Command"  
and "To Command"  
To Command  
Unit  
Notes  
Command  
PRECHARGE  
(to same bank as MASK-WR  
w/AP)  
WL + BL/2 + nWR + 1  
WL + BL/2 +nWR + 1  
tCK  
tCK  
tCK  
1,11  
1,11  
PRECHARGE All  
ACTIVATE  
(to same bank as MASK-WR  
w/AP)  
WL + BL/2 + nWR + 1 + tRPpb  
1,8,11  
WRITE or WRITE w/AP  
(same bank)  
Illegal  
-
-
3
3
MASK-WR  
w/AP  
MASK-WR or MASK-WR w/AP  
(same bank)  
Illegal  
BL=16  
WRITE or WRITE w/AP  
(different bank)  
BL/2  
BL/2  
tCK  
tCK  
-
3
MASK-WR or MASK-WR w/AP  
(different bank)  
3
READ or READ w/AP  
(same bank)  
Illegal  
3
READ or READ w/AP  
(different bank)  
WL + BL/2 + tWTR + 1  
tCK  
3,9  
Version 1.0  
174  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing Between Commands (PRECHARGE and AutoꢀPRECHARGE): DQ ODT is Disable (cont’d)  
From  
Minimum Delay between "From Command"  
and "To Command"  
To Command  
Unit  
Notes  
Command  
PRECHARGE  
(to same bank as PRECHARGE)  
PRECHARGE All  
4
tCK  
1
PRECHARGE  
4
4
4
tCK  
tCK  
tCK  
1
1
1
PRECHARGE  
PRECHARGE  
All  
PRECHARGE All  
Notes:  
1. For a given bank, the precharge period should be counted from the latest precharge command, whether per-bank or all-bank, issued to  
that bank. The precharge period is satisfied tRP after that latest precharge command.  
2. Any command issued during the minimum delay time as specified in the table above is illegal.  
3. After READ w/AP, seamless read operations to different banks are supported. After WRITE w/AP or MASK-WR w/AP, seamless write  
operations to different banks are supported. READ, WRITE, and MASK-WR operations may not be truncated or interrupted.  
4. tRPST values depend on MR1-OP[7] respectively.  
5. tWPRE values depend on MR1-OP[2] respectively.  
6. Minimum Delay between "From Command" and "To Command" in clock cycle is calculated by dividing tRTP(in ns) by tCK(in ns) and  
rounding up to the next integer: Minimum Delay[cycles] = Roundup(tRTP[ns] / tCK[ns])  
7. Minimum Delay between "From Command" and "To Command" in clock cycle is calculated by dividing tWR(in ns) by tCK(in ns) and  
rounding up to the next integer: Minimum Delay[cycles] = Roundup(tWR[ns] / tCK[ns])  
8. Minimum Delay between "From Command" and "To Command" in clock cycle is calculated by dividing tRPpb(in ns) by tCK(in ns) and  
rounding up to the next integer: Minimum Delay[cycles] = Roundup(tRPpb[ns] / tCK[ns])  
9. Minimum Delay between "From Command" and "To Command" in clock cycle is calculated by dividing tWTR(in ns) by tCK(in ns) and  
rounding up to the next integer: Minimum Delay[cycles] = Roundup(tWTR[ns] / tCK[ns])  
10. For Read w/AP the value is nRTP which is defined in Mode Register 2.  
11. For Write w/AP the value is nWR which is defined in Mode Register 1.  
Version 1.0  
175  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing Between Commands (read w/ AP and write command): DQ ODT is Enabled  
From  
Minimum Delay between "From Command"  
and "To Command"  
To Command  
Unit  
tCK  
tCK  
tCK  
tCK  
Notes  
2, 3  
Command  
WRITE or WRITE w/AP  
(different bank)  
RL+RU(tDQSCK(max)/tCK)+BL/2  
+RD(tRPST)-ODTLon-RD(tODTon,min/tCK)+1  
RL+RU(tDQSCK(max)/tCK)+BL/2  
READ w/AP  
BL=16  
MASK-WR or MASK-WR w/AP  
(different bank)  
2, 3  
+RD(tRPST)-ODTLon-RD(tODTon,min/tCK)+1  
RL+RU(tDQSCK(max)/tCK)+BL/2  
WRITE or WRITE w/AP  
(different bank)  
2, 3  
+RD(tRPST)-ODTLon-RD(tODTon,min/tCK)+1  
RL+RU(tDQSCK(max)/tCK)+BL/2  
READ w/AP  
BL=32  
MASK-WR or MASK-WR w/AP  
(different bank)  
2, 3  
+RD(tRPST)-ODTLon-RD(tODTon,min/tCK)+1  
Notes:  
1. The rest of the timing about prechage and Auto-Precharge is same as DQ ODT is Disable case.  
2. After READ w/AP, seamless read operations to different banks are supported. READ, WRITE, and MASK-WR operations may not be  
truncated or interrupted.  
3. tRPST values depend on MR1-OP[7] respectively.  
Version 1.0  
176  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Refresh command  
The REFRESH command is initiated with CS HIGH, CA0 LOW, CA1 LOW, CA2 LOW, CA3 HIGH and CA4 LOW at  
the first rising edge of the clock. Perꢀbank REFRESH is initiated with CA5 LOW at the first rising edge of the clock.  
Allꢀbank REFRESH is initiated with CA5 HIGH at the first rising edge of the clock.  
A perꢀbank REFRESH command (REFpb) is performed to the bank address as transferred on CA0, CA1 and CA2 at  
the second rising edge of the clock. Bank address BA0 is transferred on CA0, bank address BA1 is transferred on  
CA1 and bank address BA2 is transferred on CA2. A perꢀbank REFRESH command (REFpb) to the eight banks can  
be issued in any order. e.g. REFpb commands are issued in the following order: 1ꢀ3ꢀ0ꢀ2ꢀ4ꢀ7ꢀ5ꢀ6. After the eight  
banks have been refreshed using the perꢀbank REFRESH command the controller can send another set of perꢀbank  
REFRESH commands in the same order or a different order. e.g. REFpb commands are issued in the following  
order that is different from the previous order: 7ꢀ1ꢀ3ꢀ5ꢀ0ꢀ4ꢀ2ꢀ6. One of the possible order can also be a sequential  
round robin: 0ꢀ1ꢀ2ꢀ3ꢀ4ꢀ5ꢀ6ꢀ7. It is illegal to send a perꢀbank REFRESH command to the same bank unless all eight  
banks have been refreshed using the perꢀbank REFRESH command. The count of eight REFpb commands starts  
with the first REFpb command after a synchronization event.  
The bank count is synchronized between the controller and the SDRAM by resetting the bank count to zero.  
Synchronization can occur upon issuing a RESET command or at every exit from self refresh. REFab command  
also synchronizes the counter between the controller and SDRAM to zero. The SDRAM device can be placed in  
selfꢀrefresh or a REFab command can be issued at any time without cycling through all eight banks using perꢀbank  
REFRESH command. After the bank count is synchronized to zero the controller can issue perꢀbank REFRESH  
commands in any order as described in the previous paragraph.  
A REFab command issued when the bank counter is not zero will reset the bank counter to zero and the DRAM will  
perform refreshes to all banks as indicated by the row counter.  
If another refresh command (REFab or REFpb) is  
issued after the REFab command then it uses an incremented value of the row counter.  
Version 1.0  
177  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Bank and Refresh counter increment behavior  
The table below shows examples of both bank and refresh counter increment behavior.  
Refresh  
Bank#  
Bank  
Counter #  
To 0  
Ref Counter #  
#
Sub# Command  
BA0  
BA1  
BA2  
(Row Address #)  
0
1
0
Reset, SRX or REFab  
-
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
0
1
2
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFpb  
REFab  
REFpb  
REFpb  
0
0
0
0
1
1
1
1
1
1
0
0
1
0
0
1
0
0
0
V
1
1
0
0
1
1
0
0
1
1
1
1
0
1
0
1
0
0
0
0
1
V
1
1
0
1
0
1
0
1
0
1
0
1
1
1
1
0
0
0
0
1
0
V
0
1
0
1
0 to 1  
1 to 2  
2 to 3  
3 to 4  
4 to 5  
5 to 6  
6 to 7  
7 to 0  
0 to 1  
1 to 2  
2 to 3  
3 to 4  
4 to 5  
5 to 6  
6 to 7  
7 to 0  
0 to 1  
1 to 2  
2 to 3  
To 0  
2
3
2
4
3
n
5
4
6
5
7
6
8
7
9
6
10  
11  
12  
13  
14  
15  
16  
17  
18  
19  
20  
21  
22  
7
1
3
n + 1  
5
2
0
4
0
1
n + 2  
2
0~7  
6
n + 2  
0 to 1  
1 to 2  
n + 3  
7
Snip  
Version 1.0  
178  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
A bank must be idle before it can be refreshed. The controller must track the bank being refreshed by the perꢀbank  
REFRESH command.  
The REFpb command must not be issued to the device until the following conditions are met:  
ꢀ tRFCab has been satisfied after the prior REFab command.  
ꢀ tRFCpb has been satisfied after the prior REFpb command.  
ꢀ tRP has been satisfied after the prior PRECHARGE command to that bank.  
ꢀ tRRD has been satisfied after the prior ACTIVATE command (if applicable, for example after activating a row in  
a different bank than the one affected by the REFpb command).  
The target bank is inaccessible during perꢀbank REFRESH cycle time (tRFCpb), however, other banks within the  
device are accessible and can be addressed during the cycle. During the REFpb operation, any of the banks other  
than the one being refreshed can be maintained in an active state or accessed by a READ or a WRITE command.  
When the perꢀbank REFRESH cycle has completed, the affected bank will be in the idle state.  
After issuing REFpb, these conditions must be met:  
ꢀ tRFCpb must be satisfied before issuing a REFab command.  
ꢀ tRFCpb must be satisfied before issuing an ACTIVATE command to the same bank.  
ꢀ tRRD must be satisfied before issuing an ACTIVATE command to a different bank.  
ꢀ tRFCpb must be satisfied before issuing another REFpb command.  
An allꢀbank REFRESH command (REFab) issues a REFRESH command to all banks. All banks must be idle when  
REFab is issued (for instance, by issuing a PRECHARGEꢀall command prior to issuing an allꢀbank REFRESH  
command). REFab also synchronizes the bank count between the controller and the SDRAM to zero. The REFab  
command must not be issued to the device until the following conditions have been met:  
ꢀ tRFCab has been satisfied following the prior REFab command.  
ꢀ tRFCpb has been satisfied following the prior REFpb command.  
ꢀ tRP has been satisfied following the prior PRECHARGE commands.  
When an allꢀbank refresh cycle has completed, all banks will be idle. After issuing REFab:  
ꢀ tRFCab latency must be satisfied before issuing an ACTIVATE command.  
ꢀ tRFCab latency must be satisfied before issuing a REFab or REFpb command.  
Version 1.0  
179  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
REFRESH Command Scheduling Separation requirements  
Symbol  
Minimum Delay From  
To  
Notes  
REFab  
tRFCab  
REFab  
Activate command to any bank  
REFpb  
REFab  
tRFCpb  
tRRD  
REFpb  
Activate command to same bank as REFpb  
REFpb  
REFpb  
Activate command to different bank than REFpb  
REFpb  
1
Activate  
Activate command to different bank than prior Activate command  
Notes:  
1. A bank must be in the idle state before it is refreshed, so following an ACTIVATE command REFab is prohibited; REFpb is supported only if  
it affects a bank that is in the idle state.  
Version 1.0  
180  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AllꢀBank Refresh Operation  
Notes:  
1. DES commands are shown for ease of illustration; other commands may be valid at these times.  
2. Activate Command is shown as an example. Other commands may be valid provided the timing specification is satisfied.  
Per Bank Refresh to a different bank Operation  
Notes:  
1. DES commands are shown for ease of illustration; other commands may be valid at these times.  
2. In the beginning of this example, the REFpb bank is pointing to bank 0.  
3. Operations to banks other than the bank being refreshed are supported during the tpbR2pbR period.  
4. Activate Command is shown as an example. Other commands may be valid provided the timing specification is satisfied.  
Version 1.0  
181  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Per Bank Refresh to the same bank Operation  
Notes:  
1. DES commands are shown for ease of illustration; other commands may be valid at these times.  
2. In the beginning of this example, the REFpb bank is pointing to bank 0.  
3. Operations to banks other than the bank being refreshed are supported during the tRFCpb period.  
4. Activate Command is shown as an example. Other commands may be valid provided the timing specification is satisfied.  
Version 1.0  
182  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
In general, a Refresh command needs to be issued to the LPDDR4 SDRAM regularly every tREFI interval. To allow  
for improved efficiency in scheduling and switching between tasks, some flexibility in the absolute refresh interval is  
provided. A maximum of 8 Refresh commands can be postponed during operation of the LPDDR4 SDRAM,  
meaning that at no point in time more than a total of 8 Refresh commands are allowed to be postponed and  
maximum number of pulledꢀin or postponed REF command is dependent on refresh rate. It is described in the table  
below. In case that 8 Refresh commands are postponed in a row, the resulting maximum interval between the  
surrounding Refresh commands is limited to 9 × tREFI. A maximum of 8 additional Refresh commands can be  
issued in advance (“pulled in”), with each one reducing the number of regular Refresh commands required later by  
one. Note that pulling in more than 8 Refresh commands in advance does not further reduce the number of regular  
Refresh commands required later, so that the resulting maximum interval between two surrounding Refresh  
commands is limited to 9 × tREFI.  
Version 1.0  
183  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
At any given time, a maximum of 16 REF commands can be issued within 2 x tREFI. SelfꢀRefresh Mode may be  
entered with a maximum of eight Refresh commands being postponed. After exiting SelfꢀRefresh Mode with one or  
more Refresh commands postponed, additional Refresh commands may be postponed to the extent that the total  
number of postponed Refresh commands (before and after the SelfꢀRefresh) will never exceed eight. During  
SelfꢀRefresh Mode, the number of postponed or pulledꢀin REF commands does not change.  
And for per bank refresh, a maximum 8 x 8 per bank refresh commands can be postponed or pulled in for  
scheduling efficiency. At any given time, a maximum of 2 x 8 x 8 per bank refresh commands can be  
issued within 2 x tREFI.  
Legacy Refresh Command Timing Constraints  
Max. No. of  
pulled-in or  
postponed REFab  
Max. Interval  
between two  
REFab  
Max. No. of REFab within  
max(2xtREFI x refresh rate Per-bank Refresh  
multiplier, 16xtRFC)  
MR4  
OP[2:0]  
Refresh rate  
000B  
001B  
010B  
011B  
100B  
101B  
110B  
111B  
Low Temp. Limit  
4x tREFI  
N/A  
8
N/A  
N/A  
16  
N/A  
9 x 4 x tREFI  
9 x 2 x tREFI  
9 x tREFI  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
N/A  
2x tREFI  
8
16  
1x tREFI  
8
16  
0.5x tREFI  
8
9 x 0.5 x tREFI  
9 x 0.25 x tREFI  
9 x 0.25 x tREFI  
N/A  
16  
0.25x tREFI  
0.25x tREFI  
High Temp. Limit  
8
16  
8
16  
N/A  
N/A  
Modified REFRESH Command Timing Constraints  
Max. No. of  
pulled-in or  
postponed REFab  
Max. Interval  
between two  
REFab  
Max. No. of REFab within  
max(2xtREFI x refresh rate Per-bank Refresh  
multiplier, 16xtRFC)  
MR4  
OP[2:0]  
Refresh rate  
000B  
001B  
010B  
011B  
100B  
101B  
110B  
Low Temp. Limit  
4x tREFI  
N/A  
2
N/A  
N/A  
4
N/A  
3 x 4 x tREFI  
5 x 2 x tREFI  
9 x tREFI  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
1/8 of REFab  
N/A  
2x tREFI  
4
8
1x tREFI  
8
16  
16  
16  
16  
N/A  
0.5x tREFI  
8
9 x 0.5 x tREFI  
9 x 0.25 x tREFI  
9 x 0.25 x tREFI  
N/A  
0.25x tREFI  
0.25x tREFI  
High Temp. Limit  
8
8
111B  
N/A  
Notes:  
1. For any thermal transition phase where Refresh mode is transitioned to either 2x tREFIor 4x tREFI, DRAM will support the previous  
postponed refresh requirement provided the number of postponed refreshes is monotonically reduced to meet the new requirement.  
However, the pulled-in refresh commands in previous thermal phase are not applied in new thermal phase. Entering new thermal phase  
the controller must count the number of pulled-in refresh commands as zero, regardless of remaining pulled-in refresh commands in  
previous thermal phase.  
2. LPDDR4 devices are refreshed properly if memory controller issues refresh commands with same or shorter refresh period than reported  
by MR4 OP[2:0]. If shorter refresh period is applied, the corresponding requirements from Table apply. For example, when MR4  
OP[2:0]=001B, controller can be in any refresh rate from 4xtREFI to 0.25x tREFI. When MR4 OP[2:0]=010B, the only prohibited refresh  
rate is 4x tREFI.  
Version 1.0  
184  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Postponing Refresh Commands (Example)  
Pullingꢀin Refresh Commands (Example)  
Version 1.0  
185  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Burst Read operation followed by Per Bank Refresh  
The Per Bank Refresh command can be issued after tRTP + tRPpb from Read command.  
Burst Read operation followed by Per Bank Refresh  
Notes:  
1. BL = 16, Preamble = Toggle, Postamble = 0.5nCK, DQ/DQS: VSSQ termination  
2. Dout n = data-out from column n.  
3. In case of BL = 32, Delay time from Read to Per Bank Precharge is 8nCK + tRTP.  
4. DES commands are shown for ease of illustration; other commands may be valid at these times.  
The Per Bank Refresh command can be issued after tRC from Read with Auto Precharge command.  
Burst Read with AutoꢀPrecharge operation followed by Per Bank Refresh  
Notes:  
1. BL = 16, Preamble = Toggle, Postamble = 0.5nCK, DQ/DQS: VSSQ termination  
2. Dout n = data-out from column n.  
3. tRC needs to be satisfied prior to issuing subsequent Per Bank Refresh command.  
4. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
186  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Refresh Requirement  
Refresh Requirement Parameters per die for Single Channel SDRAM devices  
Refresh Requirements  
Density per Channel  
Symbol  
4Gb  
4Gb  
8
Unit  
Number of banks per channel  
Refresh Window (tREFW)  
tREFW  
32  
ms  
(TCASE 85°C)  
Refresh Window (tREFW)  
(1/2 Rate Refresh)  
tREFW  
tREFW  
R
16  
8
ms  
ms  
-
Refresh Window (tREFW)  
(1/4 Rate Refresh)  
Required Number of REFRESH  
Commands in a tREFW window  
8192  
REFAB  
Average Refresh Interval  
REFPB  
tREFI  
3.904  
488  
180  
90  
us  
ns  
ns  
ns  
tREFIpb  
tRFCab  
tRFCpb  
Refresh Cycle Time (All Banks)  
Refresh Cycle Time (Per Bank)  
Per-bank Refresh to Per-bank  
tpbR2pbR  
90  
ns  
Refresh different bank Time  
Notes:  
1. Self Refresh abort feature is available for higher density devices starting with12 Gb dual channel device and 6 Gb single channel device  
and tXSR_abort(min) is defined as tRFCpb + 17.5ns.  
Version 1.0  
187  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Self Refresh Operation  
Self Refresh Entry and Exit  
The Self Refresh command can be used to retain data in the LPDDR4 SDRAM, the SDRAM retains data without  
external Refresh command. The device has a builtꢀin timer to accommodate Self Refresh operation. The Self  
Refresh is entered by Self Refresh Entry Command defined by having CS High, CA0 Low, CA1 Low, CA2 Low;  
CA3 High; CA4 High, CA5 Valid (Valid that means it is Logic Level, High or Low) for the first rising edge and CS  
Low, CA0 Valid, CA1 Valid, CA2 Valid, CA3 Valid, CA4 Valid, CA5 Valid at the second rising edge of the clock. Self  
Refresh command is only allowed when read data burst is completed and SDRAM is idle state.  
During Self Refresh mode, external clock input is needed and all input pin of SDRAM are activated. SDRAM can  
accept the following commands, MRRꢀ1, CASꢀ2, DES, SRX, MPC, MRWꢀ1, and MRWꢀ2 except PASR  
Bank/Segment setting.  
LPDDR4 SDRAM can operate in Self Refresh in both the standard or elevated temperature ranges. SDRAM will  
also manage Self Refresh power consumption when the operating temperature changes, lower at low temperature  
and higher at high temperatures.  
For proper Self Refresh operation, power supply pins (VDD1, VDD2 and VDDQ) must be at valid levels. However  
VDDQ may be turned off during SelfꢀRefresh with Power Down after tESCKE is satisfied.  
Prior to exiting SelfꢀRefresh with Power Down, VDDQ must be within specified limits. The minimum time that the  
SDRAM must remain in Self Refresh model is tSR,min. Once Self Refresh Exit is registered, only MRRꢀ1, CASꢀ2,  
DES, MPC, MRWꢀ1 and MRWꢀ2 except PASR Bank/Segment setting are allowed until tXSR is satisfied.  
The use of Self Refresh mode introduces the possibility that an internally timed refresh event can be missed when  
Self Refresh Exit is registered. Upon exit from Self Refresh, it is required that at least one REFRESH command (8  
perꢀbank or 1 allꢀbank) is issued before entry into a subsequent Self Refresh. This REFRESH command is not  
included in the count of regular refresh commands required by the tREFI interval, and does not modify the  
postponed or pulledꢀin refresh counts; the REFRESH command does count toward the maximum refreshes  
permitted within 2 X tREFI.  
Self Refresh Entry/Exit Timing  
Notes:  
1. MRR-1, CAS-2, DES, SRX, MPC, MRW-1 and MRW-2 except PASR Bank/Segment and SR Abort setting is allowed during Self Refresh.  
2. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
188  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Power Down Entry and Exit during Self Refresh  
Entering/Exiting Power Down Mode is allowed during Self Refresh mode in SDRAM. The related timing parameters  
between Self Refresh Entry/Exit and Power Down Entry/Exit are shown in figure below.  
Self Refresh Entry/Exit Timing with Power Down Entry/Exit  
Notes:  
1. MRR-1, CAS-2, DES, SRX, MPC, MRW-1 and MRW-2 except PASR Bank/Segment and SR Abort setting is allowed during Self Refresh.  
2. Input clock frequency can be changed or the input clock can be stopped or floated after tCKELCK satisfied and during power-down,  
provided that upon exiting power-down, the clock is stable and within specified limits for a minimum of tCKCKEH of stable clock prior to  
power-down exit and the clock frequency is between the minimum and maximum specified frequency for the speed grade in use.  
3. 2 Clock command for example.  
Version 1.0  
189  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command input Timing after Power Down Exit  
Command input timings after Power Down Exit during Self Refresh  
Notes:  
1. MRR-1, CAS-2, DES, SRX, MPC, MRW-1 and MRW-2 except PASR Bank/Segment setting is allowed during Self Refresh.  
2. Input clock frequency can be changed or the input clock can be stopped or floated after tCKELCK satisfied and during power-down,  
provided that upon exiting power-down, the clock is stable and within specified limits for a minimum of tCKCKEH of stable clock prior to  
power-down exit and the clock frequency is between the minimum and maximum specified frequency for the speed grade in use.  
3. 2 Clock command for example.  
Version 1.0  
190  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AC Timing Table  
Min/  
Data Rate  
Max  
Parameter  
Symbol  
Unit Notes  
Self Refresh Timing  
Delay from SRE command to CKE Input low  
Minimum Self Refresh Time  
Exit Self Refresh to Valid commands  
Notes:  
tESCKE Min  
Max(1.75ns, 3tCK)  
Max(15ns, 3tCK)  
ns  
ns  
ns  
1
1
tSR  
Min  
Min  
tXSR  
Max(tRFCab + 7.5ns, 2tCK)  
1,2  
1. Delay time has to satisfy both analog time(ns) and clock count(tCK).  
It means that tESCKE will not expire until CK has toggled through at least 3 full cycles (3 *tCK) and 1.75ns has transpired.  
The case which 3tCK is applied to is shown below.  
tESCKE Timing  
2. MRR-1, CAS-2, DES, MPC, MRW-1 and MRW-2 except PASR Bank/Segment setting are only allowed during this period.  
Self Refresh Abort  
If MR4 OP[3] is enabled then DRAM aborts any ongoing refresh during Self Refresh exit and does not increment the  
internal refresh counter. Controller can issue a valid command after a delay of tXSR_abort instead of tXSR.  
The value of tXSR_abort(min) is defined as tRFCpb + 17.5 ns.  
Upon exit from Self Refresh mode, the LPDDR4 SDRAM requires a minimum of one extra refresh (8 per bank or 1  
all bank) before entry into a subsequent Self Refresh mode. This requirement remains the same irrespective of the  
setting of the MR bit for self refresh abort.  
Self refresh abort feature is available for higher density devices starting with12 Gb dual channel device and 6Gb  
single channel device.  
Version 1.0  
191  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MRR, MRW, MPC Command during tXSR, tRFC  
Mode Register Read (MRR), Mode Register Write (MRW) and Multi Purpose Command (MPC) can be issued  
during tXSR period.  
MRR, MRW and MPC Commands Issuing Timing during tXSR  
Notes:  
1. MPC and MRW command are shown in figure at this time, Any combination of MRR, MRW and MPC is allowed during tXSR period.  
2. Any command also includes MRR, MRW and all MPC command.  
Mode Register Read (MRR), Mode Register Write (MRW) and Multi Purpose Command (MPC) can be issued during  
tRFC period.  
MRR, MRW and MPC Commands Issuing Timing during tRFC  
Notes:  
1. MPC and MRW command are shown in figure at this time, Any combination of MRR, MRW and MPC is allowed during tRFCab or  
tRFCpb period.  
2. Refresh cycle time depends on Refresh command. In case of REF per Bank command issued, Refresh cycle time will be tRFCpb.  
3. Any command also includes MRR, MRW and all MPC command.  
Version 1.0  
192  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MODE REGISTER READ(MRR)  
The Mode Register Read (MRR) command is used to read configuration and status data from the  
LPDDR4ꢀSDRAM registers.The MRR command is initiated with CS and CA[5:0] in the proper state as defined by  
the Command Truth Table. The mode register address operands (MA[5:0]) allow the user to select one of 64  
registers. The mode register contents are available on the first 4UI's data bits of DQ[7:0] after RL x tCK + tDQSCK  
+ tDQSQ following the MRR command. Subsequent data bits contain valid but undefined content. DQS is toggled  
for the duration of the Mode Register READ burst. The MRR has a command burst length 16.  
MRR operation must not be interrupted.  
DQ output mapping  
BL  
DQ0  
0
1
2
3
4
5
6
7
8
9
10  
11  
12  
13  
14  
15  
OP0  
OP1  
OP2  
OP3  
OP4  
OP5  
OP6  
OP7  
V
V
V
V
V
V
V
V
DQ1  
DQ2  
DQ3  
DQ4  
DQ5  
DQ6  
DQ7  
DQ8-15  
DMI0-1  
V
V
Notes:  
1. MRR data are extended to first 4 UI’s for DRAM controller to sample data easily.  
2. DBI may apply or may not apply during normal MRR. It’s vendor specific. If read DBI is enable with MRS and vendor cannot support the  
DBI during MRR, DMI pin status should be low.  
3. The read pre-amble and post-amble of MRR are same as normal read.  
Version 1.0  
193  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mode Register Read Operation  
Notes:  
1. Only BL=16 is supported  
2. Only DES is allowed during tMRR period  
3. There are some exceptions about issuing commands after tMRR. Refer to MRR/MRW Timing Constraints Table for detail.  
4. DBI is Disable mode.  
5. DES commands except tMRR period are shown for ease of illustration; other commands may be valid at these times.  
6. DQ/DQS: VSSQ termination  
Version 1.0  
194  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MRR after Read and Write command  
After a prior READ command, the MRR command must not be issued earlier than BL/2 clock cycles, in a similar way  
WL + BL/2 + 1 + RU(tWTR/tCK) clock cycles after a prior Write, Write with AP, Mask Write, Mask Write with AP and  
MPC Write FIFO command in order to avoid the collision of Read and Write burst data on SDRAM’s internal Data  
bus.  
READ to MRR Timing  
Notes:  
1. The minimum number of clock cycles from the burst READ command to the MRR command is BL/2.  
2. Read BL = 32, MRR BL = 16, RL = 14, Preamble = Toggle, Postamble = 0.5nCK, DBI = Disable, DQ/DQS: VSSQ termination  
3. DES commands except tMRR period are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
195  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write to MRR Timing  
Notes:  
1. Write BL=16, Write Postamble = 0.5nCK, DQ/DQS: VSSQ termination.  
2. Only DES is allowed during tMRR period.  
3. Din n = data-in to columnm n.  
4. The minimum number of clock cycles from the burst write command to MRR command is WL + BL/2 + 1 + RU(tWTR/tCK).  
5. tWTR starts at the rising edge of CK after the last latching edge of DQS.  
6. DES commands except tMRR period are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
196  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MRR after PowerꢀDown Exit  
Following the powerꢀdown state, an additional time, tMRRI, is required prior to issuing the mode register read (MRR)  
command. This additional time (equivalent to tRCD) is required in order to be able to maximize powerꢀdown current  
savings by allowing more powerꢀup time for the MRR data path after exit from powerꢀdown mode.  
MRR Following PowerꢀDown  
Notes:  
1. Only DES is allowed during tMRR period.  
2. DES commands except tMRR period are shown for ease of illustration; other commands may be valid at these times.  
Mode Register Read/Write AC timing  
Min/  
Parameter  
Symbol  
Data Rate  
Unit Notes  
Max  
Mode Register Read/Write Timing  
Additional time after tXP has expired until  
MRR command may be issued  
tMRRI  
Min  
Min  
tRCD + 3nCK  
8
-
MODE REGISTER READ command period  
MODE REGISTER WRITE command period  
Mode register set command delay  
tMRR  
tMRW  
tMRD  
nCK  
Min MAX(10ns, 10nCK)  
Min max(14ns, 10nCK)  
-
-
Version 1.0  
197  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mode Register Write (MRW) Operation  
The Mode Register Write (MRW) command is used to write configuration data to the mode registers. The MRW  
command is initiated by setting CKE, CS, and CA[5:0] to valid levels at a rising edge of the clock (see Command  
Truth Table). The mode register address and the data written to the mode registers is contained in CA[5:0]  
according to the Command Truth Table. The MRW command period is defined by tMRW. Mode register Writes to  
readꢀonly registers have no impact on the functionality of the device.  
Mode Register Write Timing  
Notes:  
1. Only Deselect command is allowed during tMRW and tMRD periods.  
Mode Register Write  
MRW can be issued from either a BankꢀIdle or BankꢀActive state. Certain restrictions may apply for MRW from an  
Active state.  
Truth Table for Mode Register Read (MRR) and Mode Register Write (MRW)  
Current State  
SDRAM  
Intermediate State  
SDRAM  
Next State  
SDRAM  
Command  
MRR  
Mode Register Reading  
(All Banks Idle)  
All Banks Idle  
All Banks Idle  
All Banks Idle  
Bank(s) Active  
Mode Register Writing  
(All Banks Idle)  
MRW  
MRR  
Mode Register Reading  
Mode Register Writing  
Bank(s) Active  
Bank(s) Active  
MRW  
Version 1.0  
198  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MRR/MRW Timing Constraints: DQ ODT is Disable  
Minimum Delay between "From Command"  
From Command  
To Command  
Unit Notes  
and "To Command"  
tMRR  
MRR  
RD/RDA  
-
-
tMRR  
MRR  
WR/WRA/  
MWR/MWRA  
MRW  
RL+RU(tDQSCK(max)/tCK)+BL/2  
-WL+tWPRE+RD(tRPST)  
RL+RU(tDQSCK(max)/tCK)+BL/2+3  
BL/2  
nCK  
nCK  
nCK  
RD/RDA  
WR/WRA/  
WL+1+BL/2+RU(tWTR/tCK)  
nCK  
MWR/MWRA  
MRW  
MRR  
tMRD  
tXP+tMRRI  
tMRD  
-
-
-
Power Down Exit  
RD/RDA  
WR/WRA/  
MWR/MWRA  
MRW  
MRW  
tMRD  
-
-
tMRW  
RD/  
RD FIFO/  
RL+BL/2+RU(tDQSCKmax/tCK)+RD(tRPST)  
+max(RU(7.5ns/tCK),8nCK)  
nCK  
nCK  
nCK  
nCK  
RD DQ CAL  
RD with  
RL+BL/2+RU(tDQSCKmax/tCK)+RD(tRPST)  
+max(RU(7.5ns/tCK),8nCK)+nRTP-8  
Auto-Precharge  
WR/  
MRW  
MWR/  
WL+1+BL/2+max(RU(7.5ns/tCK),8nCK)  
WR FIFO  
WR/MWR with  
Auto-Precharge  
WL+1+BL/2+max(RU(7.5ns/tCK),8nCK)+nWR  
Version 1.0  
12/2018  
199  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MRR/MRW Timing Constraints: DQ ODT is Enable  
Minimum Delay between "From Command"  
and "To Command"  
From Command  
To Command  
Unit Notes  
MRR  
RD/RDA  
Same as ODT Disable Case  
-
MRR  
WR/WRA/  
MWR/MWRA  
MRW  
RL+RU(tDQSCK(max)/tCK)+BL/2-ODTLon  
-RD(tODTon(min)/tCK)+RD(tRPST)+1  
Same as ODT Disable Case  
nCK  
-
RD/RDA  
WR/WRA/  
MWR/MWRA  
MRW  
MRR  
Same as ODT Disable Case  
-
-
Powe Down Exit  
RD/RDA  
WR/WRA/  
MWR/MWRA  
MRW  
MRW  
Same as ODT Disable Case  
RD/  
RD FIFO/  
RD DQ CAL  
RD with  
Auto-Precharge  
WR/  
MRW  
Same as ODT Disable Case  
-
MWR/  
WR FIFO  
WR/MWR with  
Auto-Precharge  
Version 1.0  
12/2018  
200  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
VREF Current Generator (VRCG)  
LPDDR4 SDRAM VREF current generators (VRCG) incorporate a high current mode to reduce the settling time of the  
internal VREF(DQ) and VREF(CA) levels during training and when changing frequency set points during operation. The  
high current mode is enabled by setting MR13[OP3] = 1. Only Deselect commands may be issued until  
tVRCG_ENABLE is satisfied. tVRCG_ENABLE timing is shown in figure below.  
VRCG Enable timing  
VRCG high current mode is disabled by setting MR13[OP3] = 0. Only Deselect commands may be issued until  
tVRCG_DISABLE is satisfied. tVRCG_DISABLE timing is shown in figure below.  
VRCG Disable timing  
Note that LPDDR4 SDRAM devices support VREF(CA) and VREF(DQ) range and value changes without enabling  
VRCG high current mode.  
VRCG Enable/Disable Timing  
Speed  
3733  
Unit Notes  
Parameter  
Symbol  
Min  
Max  
200  
100  
VREF high current mode enable time  
VREF high current mode disable time  
tVRCG_ENABLE  
tVRCG_DISABLE  
-
-
ns  
ns  
Version 1.0  
12/2018  
201  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CA VREFTraining  
The DRAM internal CA VREF specification parameters are voltage operating range, stepsize, VREF set tolerance, VREF  
step time and VREF valid level.  
The voltage operating range specifies the minimum required VREF setting range for LPDDR4 DRAM devices. The  
minimum range is defined by VREFmax and VREFmin as depicted in figure below.  
V
REF  
operating range (V  
min, V  
max)  
REF  
REF  
Version 1.0  
202  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The V  
stepsize is defined as the stepsize between adjacent steps. However, for a given design, DRAM has one  
REF  
value for V  
step size that falls within the range.  
REF  
The V  
set tolerance is the variation in the V  
voltage from the ideal setting. This accounts for accumulated  
REF  
REF  
error over multiple steps. There are two ranges for V  
uncertainty is a function of number of steps n.  
set tolerance uncertainty. The range of V  
set tolerance  
REF  
REF  
The V  
set tolerance is measured with respect to the ideal line which is based on the two endpoints. Where the  
REF  
endpoints are at the min and max V  
values for a specified range. An illustration depicting an example of the  
REF  
stepsize and V  
set tolerance is below.  
REF  
Example of V  
set tolerance (max case only shown) and step size  
REF  
Version 1.0  
203  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The V  
increment/decrement step times are define by V  
_timeꢀshort, Middle and long. The VREF_timeꢀshort,  
REF  
REF  
V
REF  
_timeꢀMiddle and V  
_timeꢀlong is defined from TS to TE as shown in the figure below where TE is  
REF  
referenced to when the V  
voltage is at the final DC level within the V  
valid tolerance(V  
_val_tol).  
REF  
REF  
REF  
The V  
valid level is defined by V  
_val tolerance to qualify the step time TE as shown in figure below. This  
REF  
REF  
parameter is used to insure an adequate RC time constant behavior of the voltage level change after any V  
increment/decrement adjustment. This parameter is only applicable for DRAM component level  
validation/characerization.  
REF  
V
V
_timeꢀShort is for a single stepsize increment/decrement change in VREF voltage.  
_timeꢀMiddle is at least 2 stepsizes increment/decrement change within the same V  
REF  
CA range in V  
REF  
REF  
REF  
voltage.  
V
V
_timeꢀLong is the time including up to V  
min to V max or V max to V  
REF REF REF  
min change across the  
REF  
REF  
CA Range in V  
voltage.  
REF  
REF  
TS ꢀ is referenced to MRS command clock  
TE ꢀ is referenced to the V  
_val_tol  
REF  
V
REF  
_time for Short, Middle and Long Timing Diagram  
The MRW command to the mode register bits are as follows.  
MR12 OP[5:0] : V  
(CA) Setting  
REF  
MR12 OP[6] : V  
(CA) Range  
REF  
Version 1.0  
12/2018  
204  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The minimum time required between two V  
MRS commands is V  
_timeꢀshort for single step and  
REF  
REF  
V
REF  
_timeꢀMiddle for a full voltage range step.  
V
REF  
step single stepsize increment case  
V
REF  
step single stepsize decrement case  
Version 1.0  
205  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
V
REF  
full step from V  
min to V  
max case  
REF  
REF  
V
REF  
full step from V  
max to V  
min case  
REF  
REF  
Version 1.0  
206  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CA Internal V  
Specifications  
REF  
The table below contains the CA internal V  
compliance.  
specifications that will be characterized at the component level for  
REF  
Parameter  
Symbol  
_max_R0  
Min  
Typ  
Max  
44.9%  
-
Unit  
Notes  
1,11  
1,11  
1,11  
1,11  
2
-
-
V
REF  
Max operating point Range0  
Min operating point Range0  
Max operating point Range1  
Min operating point Range1  
V
V
REF  
DDQ  
DDQ  
DDQ  
DDQ  
DDQ  
15%  
-
V
V
_min_R0  
V
V
V
V
REF  
REF  
REF  
-
-
62.9%  
-
V
V
_max_R1  
REF  
32.9%  
-
V
REF  
V
_min_R1  
REF  
0.50%  
0.60%  
0.70%  
11  
V
REF  
Stepsize  
V
_step  
REF  
-11  
0
mV  
mV  
ns  
3,4,6  
3,5,7  
8
V
Set Tolerance  
V
_set_tol  
REF  
REF  
-1.1  
0
1.1  
-
-
100  
200  
250  
1
V
REF  
_time-Short  
-
-
ns  
12  
V
_time_Middle  
REF  
V
Step Time  
REF  
-
-
-
ns  
9
V
_time-Long  
REF  
-
ms  
13,14  
10  
V
_time_weak  
REF  
-0.10%  
0.00%  
0.10%  
V
REF  
Valid tolerance  
V
_val_tol  
V
DDQ  
REF  
Notes:  
1. VREF DC voltage referenced to VDD2_DC.  
2. VREF stepsize increment/decrement range. VREF at DC level.  
3. VREF_new = VREF_old + n*VREF_step; n= number of steps; if increment use "+"; If decrement use "-".  
4. The minimum value of VREF setting tolerance = VREF_new - 11mV. The maximum value of VREF setting tolerance = VREF_new + 11mV. For  
n>4.  
5. The minimum value of VREF setting tolerance = VREF_new - 1.1mV. The maximum value of VREF setting tolerance = VREF_new + 1.1mV. For  
n4.  
6. Measured by recording the min and max values of the VREF output over the range, drawing a straight line between those points and  
comparing all other VREF output settings to that line.  
7. Measured by recording the min and max values of the VREF output across 4 consecutive steps(n=4), drawing a straight line between  
those points and comparing all other VREF output settings to that line.  
8. Time from MRS command to increment or decrement one step size for VREF  
.
9. Time from MRS command to increment or decrement VREFmin to VREFmax or VREFmax to VREFmin change across the VREFCA Range in  
VREF voltage.  
10. Only applicable for DRAM component level test/characterization purpose. Not applicable for normal mode of operation. VREF valid is to  
qualify the step times which will be characterized at the component level.  
11. DRAM range 0 or 1 set by MR12 OP[6].  
12. Time from MRS command to increment or decrement more than one step size up to a full range of VREF voltage withiin the same VREFCA  
range.  
13. Applies when VRCG high current mode is not enabled, specified by MR13[OP3] = 0.  
14. VREF_time_weak covers all VREF(CA) Range and Value change conditions are applied to VREF_time_Short/Middle/Long.  
Version 1.0  
207  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ VREF Training  
The DRAM internal DQ V  
specification parameters are voltage operating range, stepsize, V  
set tolerance,  
REF  
REF  
V
REF  
step time and V  
valid level.  
REF  
The voltage operating range specifies the minimum required V  
setting range for LPDDR4 DRAM devices. The  
REF  
minimum range is defined by V  
max and V min as depicted in figure below.  
REF  
REF  
VREF operating range (V  
min, V  
max)  
REF  
REF  
Version 1.0  
208  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The V  
stepsize is defined as the stepsize between adjacent steps. However, for a given design, DRAM has one  
REF  
value for V  
step size that falls within the range.  
REF  
The V  
set tolerance is the variation in the V  
voltage from the ideal setting. This accounts for accumulated  
REF  
REF  
error over multiple steps. There are two ranges for V  
set tolerance  
REF  
uncertainty. The range of V  
set tolerance uncertainty is a function of number of steps n.  
REF  
The V  
set tolerance is measured with respect to the ideal line which is based on the two endpoints. Where the  
REF  
endpoints are at the min and max V  
values for a specified range. An illustration depicting an example of the  
REF  
stepsize and V  
set tolerance is shown in below.  
REF  
Example of V  
set tolerance (max case only shown) and stepsize  
REF  
Version 1.0  
209  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The V  
increment/decrement step times are define by V  
_timeꢀshort, Middle and long. The V  
_timeꢀshort,  
REF  
REF  
REF  
V
REF  
_timeꢀMiddle and V  
_timeꢀlong is defined from TS to TE as shown in the figure below where TE is  
REF  
referenced to when the V  
voltage is at the final DC level within the V  
valid tolerance(V  
_val_tol).  
REF  
REF  
REF  
The V  
valid level is defined by V  
_val tolerance to qualify the step time TE as shown in figure below. This  
REF  
REF  
parameter is used to insure an adequate RC time constant behavior of the voltage level change after any V  
increment/decrement adjustment. This parameter is only applicable for DRAM component level  
validation/characerization.  
REF  
V
V
_timeꢀShort is for a single stepsize increment/decrement change in V  
voltage.  
REF  
REF  
_timeꢀMiddle is at least 2 stepsizes increment/decrement change within the same V  
DQ range in V  
REF  
REF  
REF  
voltage.  
V
V
_timeꢀLong is the time including up to V  
min to V max or V max to V  
REF REF REF  
min change across the  
REF  
REF  
DQ Range in V  
voltage.  
REF  
REF  
TS ꢀ is referenced to MRS command clock  
TE ꢀ is referenced to the V  
_val_tol  
REF  
V
REF  
_time for Short, Middle and Long Timing Diagram  
The MRW command to the mode register bits are as follows.  
MR14 OP[5:0] : V  
(DQ) Setting  
(DQ) Range  
REF  
MR14 OP[6] : V  
REF  
Version 1.0  
12/2018  
210  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The minimum time required between two V  
MRS commands is V  
_timeꢀshort for single step and  
REF  
REF  
V
REF  
_timeꢀMiddle for a full voltage range step.  
V
REF  
step single stepsize increment case  
V
REF  
step single stepsize decrement case  
Version 1.0  
211  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
V
REF  
full step from V  
min to V  
max case  
REF  
REF  
V
REF  
full step from V  
max to V  
min case  
REF  
REF  
Version 1.0  
212  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ Internal V  
Specifications  
REF  
The table below contains the DQ internal V  
compliance.  
specifications that will be characterized at the component level for  
REF  
Parameter  
Symbol  
_max_R0  
Min  
Typ  
Max  
44.9%  
-
Unit  
Notes  
1,11  
1,11  
1,11  
1,11  
2
-
-
V
REF  
Max operating point Range0  
Min operating point Range0  
Max operating point Range1  
Min operating point Range1  
V
V
REF  
DDQ  
DDQ  
DDQ  
DDQ  
DDQ  
15%  
-
V
V
_min_R0  
V
V
V
V
REF  
REF  
REF  
-
-
62.9%  
-
V
V
_max_R1  
REF  
32.9%  
-
V
REF  
V
_min_R1  
REF  
0.50%  
0.60%  
0.70%  
11  
V
REF  
Stepsize  
V
_step  
REF  
-11  
0
mV  
mV  
ns  
3,4,6  
3,5,7  
8
V
Set Tolerance  
V
_set_tol  
REF  
REF  
-1.1  
0
1.1  
-
-
100  
200  
250  
1
V
REF  
_time-Short  
-
-
ns  
12  
V
_time_Middle  
REF  
V
Step Time  
REF  
-
-
-
ns  
9
V
_time-Long  
REF  
-
ms  
13,14  
10  
V
_time_weak  
REF  
-0.10%  
0.00%  
0.10%  
V
REF  
Valid tolerance  
V
_val_tol  
V
DDQ  
REF  
Notes:  
1. VREF DC voltage referenced to VDDQ_DC.  
2. VREF stepsize increment/decrement range. VREF at DC level.  
3. VREF_new = VREF_old + n*VREF_step; n= number of steps; if increment use "+"; If decrement use "-".  
4. The minimum value of VREF setting tolerance = VREF_new - 11mV. The maximum value of VREF setting tolerance = VREF_new + 11mV. For  
n>4.  
5. The minimum value of VREF setting tolerance = VREF_new - 1.1mV. The maximum value of VREF setting tolerance = VREF_new + 1.1mV. For  
n<4.  
6. Measured by recording the min and max values of the VREF output over the range, drawing a straight line between those points and  
comparing all other VREF output settings to that line.  
7. Measured by recording the min and max values of the VREF output across 4 consectuive steps(n=4), drawing a straight line between  
those points and comparing all other VREF output settings to that line.  
8. Time from MRS command to increment or decrement one step size for VREF  
.
9. Time from MRS command to increment or decrement VREFmin to VREFmax or VREFfmax to VREFmin change across the VREFDQ Range in  
VREF voltage.  
10.Only applicable for DRAM component level test/characterization purpose. Not applicable for normal mode of operation. VREF valid is to  
qualify the step times which will be characterized at the component level.  
11. DRAM range 0 or 1 set by MR14 OP[6].  
12. Time from MRS command to increment or decrement more than one step size up to a full range of VREF voltage withiin the same  
VREFDQ range.  
13. Applies when VRCG high current mode is not enabled, specified by MR13[OP3] = 0.  
14. VREF_time_weak covers all VREF(DQ) Range and Value change conditions are applied to VREF_time_Short/Middle/Long.  
Version 1.0  
213  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command Bus Training  
Command Bus Trainging for x16 mode  
The LPDDR4ꢀSDRAM command bus must be trained before enabling termination for highꢀfrequency operation.  
LPDDR4 provides an internal VREF(CA) that defaults to a level suitable for unꢀterminated, lowꢀfrequency operation,  
but the VREF(CA) must be trained to achieve suitable receiver voltage margin for terminated, highꢀfrequency  
operation. The training mode described here centers the internal VREF(CA) in the CAdata eye and at the same time  
allows for timing adjustments of the CS and CA signals to meet setup/hold requirements. Because it can be difficult  
to capture commands prior to training the CA inputs, the training mode described here uses a minimum of external  
commands to enter, train, and exit the Command Bus Training mode.  
NOTES: it is up to the system designer to determine what constitutes “low-frequency” and “high-frequency” based on the capabilities of  
the system. Low-frequency should then be defined as an operating frequency in which the system can reliably communicate with  
the SDRAM before Command Bus Training is executed.  
The LPDDR4ꢀSDRAM die has a bondꢀpad (ODTꢀCA) for multiꢀrank operation. In a multiꢀrank system, the  
terminating rank should be trained first, followed by the nonterminating rank(s).  
The LPDDR4ꢀSDRAM uses Frequency SetꢀPoints to enable multiple operating settings for the die. The  
LPDDR4ꢀSDRAM defaults to FSPꢀOP[0] at powerꢀup, which has the default settings to operate in unꢀterminated,  
lowꢀfrequency environments. Prior to training, the mode register settings should be configured by setting MR13  
OP[6]=1B (FSPꢀWR[1]) and setting all other mode register bits for FSPꢀOP[1] to the desired settings for  
highꢀfrequency operation. Prior to entering Command Bus Training, the SDRAM will be operating from FSPꢀOP[x].  
Upon Command Bus Training entry when CKE is driven LOW, the LPDDR4ꢀSDRAM will automatically switch to the  
alternate FSP register set (FSPꢀOP[y]) and use the alternate register settings during training. Upon training exit  
when CKE is driven HIGH, the LPDDR4ꢀSDRAM will automatically switch back to the original FSP register set  
(FSPꢀOP[x]), returning to the “knownꢀgood” state that was operating prior to training. The training values for  
V
REF  
(CA) are not retained by the DRAM in FSPꢀOP[y] registers, and must be written to the registers after training  
exit.  
Version 1.0  
214  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
1. To enter Command Bus Training mode, issue a MRWꢀ1 command followed by a MRWꢀ2 command to set MR13  
OP[0]=1 (Command Bus Training Mode Enabled).  
B
2. After time tMRD, CKE may be set LOW, causing the LPDDR4ꢀSDRAM to switch from FSPꢀOP[x] to FSPꢀOP[y],  
and completing the entry into Command Bus Training mode.  
A status of DQS, DQS, DQ and DMI are as follows, and DQ ODT state of DQS, DQS, DQ and DMI will be followed  
by MR11 OP[2:0]: DQ ODT and MR13 OP[7]: FSPꢀOP except output pins.  
ꢀ DQS[0], DQS[0] become input pins for capturing DQ[6:0] levels by its toggling.  
ꢀ DQ[5:0] become input pins for setting V  
ꢀ DQ[6] becomes a input pin for setting V  
(CA) Level.  
REF  
(CA) Range.  
REF  
ꢀ DQ[7] and DMI[0] become input pins and their input level is Valid level or floating, either way is fine.  
ꢀ DQ[13:8] become output pins to feedback its capturing value via command bus by CS signal.  
ꢀ DQS[1], DQS[1],DMI[1] and DQ[15:14] become output pins or disable, it means that SDRAM may drive to a valid  
level or left floating.  
3. At time tCAENT later, LPDDR4 SDRAM can accept to change its VFREF(ca) Range and Value using input  
signals of DQS[0], DQS[0] and DQ[6:0] from existing value that’s setting via MR12 OP[6:0]. The mapping  
between MR12 OP code and DQs is shown in the following table. At least one V  
proceed to next training steps.  
CA setting is required before  
REF  
Mapping of MR12 OP Code and DQ Numbers  
Mapping  
MR12 OP Code  
DQ Number  
OP6  
DQ6  
OP5  
DQ5  
OP4  
DQ4  
OP3  
DQ3  
OP2  
DQ2  
OP1  
DQ1  
OP0  
DQ0  
4. The new V  
(CA) value must “settle” for time tV  
_LONG before attempting to latch CA information.  
REF  
REF  
5. To verify that the receiver has the correct V  
(CA) setting and to further train the CA eye relative to clock (CK),  
REF  
values latched at the receiver on the CA bus are asynchronously output to the DQ bus.  
6. To exit Command Bus Training mode, drive CKE HIGH, and after time tVREF_LONG issue the MRWꢀ1  
command followed by the MRWꢀ2 command to set MR13 OP[0]=0 . After time tMRW the LPDDR4ꢀSDRAM is  
B
ready for normal operation. After training exit the LPDDR4ꢀSDRAM will automatically switch back to the FSPꢀOP  
registers that were in use prior to training.  
Command Bus Training may executed from IDLE or Self Refresh states. When executing CBT within the Self  
Refresh state, the SDRAM must not be a power down state (i.e. CKE must be HIGH prior to training entry).  
Command Bus Training entry and exit is the same, regardless of the SDRAM state from which CBT is initiated.  
Version 1.0  
215  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Sequence for singleꢀrank systems  
Note that a example shown here is assuming an initial lowꢀfrequency, noꢀterminating operating point, training a  
highꢀfrequency, terminating operating point. The green text is lowꢀfrequency, magenta text is highꢀfrequency. Any  
operating point may be trained from any known good operating point.  
1. Set MR13 OP[6]=1 to enable writing to Frequency Set Point ‘y’ (FSPꢀWR[y]) (or FSPꢀOP[x], See note).  
B
2. Write FSPꢀWR[y] (or FSPꢀWR[x]) registers for all channels to set up highꢀfrequency operating parameters.  
3. Issue MRWꢀ1 and MRWꢀ2 commands to enter Command Bus Training mode.  
4. Drive CKE LOW, and change CK frequency to the highꢀfrequency operating point.  
5. Perform Command Bus Training (V  
CA, CS, and CA).  
REF  
6. Exit training, a change CK frequency to the lowꢀfrequency operating point prior to driving CKE HIGH, then issue  
MRWꢀ1 and MRWꢀ2 commands. When CKE is driven HIGH, the SDRAM will automatically switch back to the  
FSPꢀOP registers that were in use prior to training (i.e. trained values are not retained by the SDRAM).  
7. Write the trained values to FSPꢀWR[y] (or FSPꢀWR[x]) by issuing MRWꢀ1 and MRWꢀ2 commands to the SDRAM  
and setting all applicable mode register parameters.  
8. Issue MRWꢀ1 and MRWꢀ2 commands to switch to FSPꢀOP[y] (or FSPꢀOP[x]), to turn on termination, and change  
CK frequency to the high frequency operating point. At this point the Command Bus is trained and you may  
proceed to other training or normal operation.  
Version 1.0  
216  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Sequence for multiꢀrank systems  
Note that a example shown here is assuming an initial lowꢀfrequency operating point, training a highꢀfrequency  
operating point. The green text is lowꢀfrequency, magenta text is highꢀfrequency. Any operating point may be trained  
from any known good operating point.  
1. Set MR13 OP[6]=1 to enable writing to Frequency Set Point ‘y’ (FSPꢀWR[y]) (or FSPꢀWR[x], See Note).  
B
2. Write FSPꢀWR[y] (or FSPꢀWR[x]) registers for all channels and ranks to set up high frequency operating  
parameters.  
3. Read MR0 OP[7] on all channels and ranks to determine which die are terminating, signified by  
MR0 OP[7]=1 .  
B
4. Issue MRWꢀ1 and MRWꢀ2 commands to enter Command Bus Training mode on the terminating rank.  
5. Drive CKE LOW on the terminating rank (or all ranks), and change CK frequency to the highꢀfrequency operating  
point.  
6. Perform Command Bus Training on the terminating rank (V  
CA, CS, and CA).  
REF  
7. Exit training by driving CKE HIGH, change CK frequency to the lowꢀfrequency operating point, and issue MRWꢀ1  
and MRWꢀ2 commands to write the trained values to FSPꢀWR[y] (or FSPꢀWR[x]). When CKE is driven HIGH, the  
SDRAM will automatically switch back to the FSPꢀOP registers that were in use prior to training (i.e. trained  
values are not retained by the SDRAM).  
8. Issue MRWꢀ1 and MRWꢀ2 command to enter training mode on the nonꢀterminating rank (but keep CKE HIGH)  
9. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[y] (or FSPꢀOP[x]), to turn on  
termination, and change CK frequency to the high frequency operating point.  
10. Drive CKE LOW on the nonꢀterminating (or all) ranks. The nonꢀterminating rank(s) will now be using FSPꢀOP[y]  
(or FSPꢀOP[x]).  
11. Perform Command Bus Training on the nonꢀterminating rank (V  
CA, CS, and CA).  
REF  
12. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[x] (or FSPꢀOP[y]) to turn off  
termination.  
13. Exit training by driving CKE HIGH on the nonꢀterminating rank, change CK frequency to the lowꢀfrequency  
operating point, and issue MRWꢀ1 and MRWꢀ2 commands. When CKE is driven HIGH, the SDRAM will  
automatically switch back to the FSPꢀOP registers that were in use prior to training (i.e. trained values are not  
retained by the SDRAM).  
14. Write the trained values to FSPꢀWR[y] (or FSPꢀWR[x]) by issuing MRWꢀ1 and MRWꢀ2 commands to the  
SDRAM and setting all applicable mode register parameters.  
15. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[y] (or FSPꢀOP[x]), to turn on  
termination, and change CK frequency to the high frequency operating point. At this point the Command Bus is  
trained for both ranks and you may proceed to other training or normal operation.  
Version 1.0  
217  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Relation between CA input pin DQ output pin.  
Mapping of CA Input pin and DQ Output pin.  
Mapping  
CA Number  
DQ Number  
CA5  
CA4  
CA3  
CA2  
CA1  
CA0  
DQ13  
DQ12  
DQ11  
DQ10  
DQ9  
DQ8  
Timing Diagram  
The basic Timing diagrams of Command Bus Training are shown in the following figures.  
Entering Command Bus Training Mode and CA Training Pattern Input  
and Output with VREFCA Value Update  
Notes:  
1. After tCKELCK clock can be stopped or frequency changed any time.  
2. The input clock condition should be satisfied tCKPRECS and tCKPSTCS.  
3. Continue to Drive CK and Hold CS pins low until tCKELCK after CKE is low (which disables command decoding).  
4. DRAM may or may not capture first rising/falling edge of DQS/DQS due to an unstable first rising edge. Hence provide at least  
consecutive 2 pulses of DQS signal input is required in every DQS input signal at capturing DQ[6:0] signals. The captured value of DQ6:0  
signal level by each DQS edges are overwritten at any time and the DRAM updates its VREFca setting of MR12 temporary after time  
tVREFca_Long.  
5. tVREF_LONG may be reduced to tVREF_SHORT if the following conditions are met: 1) The new Vref setting is a single step above or  
below the old Vref setting, and 2) The DQS pulses a single time, or the new Vref setting value on DQ[6:0] is static and meets  
tDSTRAIN/tDHTRAIN for every DQS pulse applied.  
6. When CKE is driven LOW, the SDRAM will switch its FSP-OP registers to use the alternate (i.e. non-active) set. Example: If the SDRAM is  
currently using FSP-OP[0], then it will switch to FSP-OP[1] when CKE is driven LOW. All operating parameters should be written to the  
alternate mode registers before entering Command Bus Training to ensure that ODT settings, RL/WL/nWR setting, etc., are set to the  
correct values. If the alternate FSP-OP has ODT_CA disabled then termination will not enable in CA Bus Training mode. If the ODT_CA  
pad is bonded to Vss, ODT_CA termination will never enable for that die.  
7. When CKE is driven low in Command Bus Training mode, the LPDDR4-SDRAM will change operation to the alternate FSP, i.e. the inverse  
of the FSP programmed in the FSP-OP mode register.  
Version 1.0  
218  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive VREFCA Value Update  
Notes:  
1. After tCKELCK clock can be stopped or frequency changed any time.  
2. The input clock condition should be satisfied tCKPRECS.  
3. Continue to Drive CK and Hold CS pins low until tCKELCK after CKE is low (which disables command decoding).  
4. DRAM may or may not capture first rising/falling edge of DQS_t/c due to an unstable first rising edge. Hence provide at least consecutive  
2 pulses of DQS signal input is required in every DQS input signal at capturing DQ6:0 signals.  
The captured value of DQ6:0 signal level by each DQS edges are overwritten at any time and the DRAM updates its VREFca setting of  
MR12 temporary after time tVREFca_Long.  
5. tVREF_LONG may be reduced to tVREF_SHORT if the following conditions are met: 1) The new Vref setting is a single step above or  
below the old Vref setting, and 2) The DQS pulses a single time, or the new Vref setting value on DQ[6:0] is static and meets  
tDSTRAIN/tDHTRAIN for every DQS pulse applied.  
6. When CKE is driven LOW, the SDRAM will switch its FSP-OP registers to use the alternate (i.e. non-active) set. Example: If the SDRAM is  
currently using FSP-OP[0], then it will switch to FSP-OP[1] when CKE is driven LOW. All operating parameters should be written to the  
alternate mode registers before entering Command Bus Training to ensure that ODT settings, RL/WL/nWR setting, etc., are set to the  
correct values. If the alternate FSP-OP has ODT_CA disabled then termination will not enable in CA Bus Training mode. If the ODT_CA  
pad is bonded to Vss, ODT_CA termination will never enable for that die.  
7. When CKE is driven low in Command Bus Training mode, the LPDDR4-SDRAM will change operation to the alternate FSP, i.e. the inverse  
of the FSP programmed in the FSP-OP mode register.  
Version 1.0  
219  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Exiting Command Bus Training Mode with Valid Command  
Notes:  
1. Clock can be stopped or frequency changed any time before tCKCKEH. CK must meet tCKCKEH before CKE is driven high.  
When CKE is driven high the clock frequency must be returned to the original frequency (the frequency corresponding to the FSP at  
which Command Bus Training mode was entered)  
2. CS must be Deselect (low) tCKCKEH before CKE is driven high.  
3. When CKE is driven high, the SDRAM’s ODT_CA will revert to the state/value defined by FSP-OP prior to Command Bus Training mode  
entry, i.e. the original frequency set point (FSP-OP, MR13-OP[7]).  
Example: If the SDRAM was using FSP-OP[1] for training, then it will switch to FSP-OP[0] when CKE is driven HIGH.  
4. Training values are not retained by the SDRAM, and must be written to the FSP-OP register set before returning to operation at the  
trained frequency.  
Example: VREF(ca) will return to the value programmed in the original set point.  
5. When CKE is driven high the LPDDR4-SDRAM will revert to the FSP in operation when Command Bus Training mode was entered.  
Version 1.0  
220  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Exiting Command Bus Training Mode with Power Down Entry  
Notes:  
1. Clock can be stopped or frequency changed any time before tCKCKEH. CK must meet tCKCKEH before CKE is driven high. When CKE is  
driven high the clock frequency must be returned to the original frequency (the frequency corresponding to the FSP at which Command  
Bus Training mode was entered)  
2. CS must be Deselect (low) tCKCKEH before CKE is driven high.  
3. When CKE is driven high, the SDRAM’s ODT_CA will revert to the state/value defined by FSP-OP prior to Command Bus Training mode  
entry, i.e. the original frequency set point (FSP-OP, MR13-OP[7]).  
Example: If the SDRAM was using FSP-OP[1] for training, then it will switch to FSP-OP[0] when CKE is driven HIGH.  
4. Training values are not retained by the SDRAM, and must be written to the FSP-OP register set before returning to operation at the  
trained frequency. Example: VREF(ca) will return to the value programmed in the original set point.  
5. When CKE is driven high the LPDDR4-SDRAM will revert to the FSP in operation when Command Bus Training mode was entered.  
Version 1.0  
221  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command Bus Training AC Timing Table  
Data Rate  
3733  
Min/  
Max  
Parameter  
Symbol  
Unit Notes  
Command Bus Training Timing  
Valid Clock Requirement  
tCKELCK  
Min  
max(5ns, 5nCK)  
tCK  
after CKE Input low  
Data Setup for VREF Training Mode  
Data Hold forVREF Training Mode  
Asynchronous Data Read  
tDStrain  
tDHtrain  
tADR  
Min  
Min  
Max  
2
2
ns  
ns  
ns  
20  
CA Bus Training Command  
tCACD  
tDQSCKE  
tCAENT  
Min  
Min  
Min  
RU(tADR/tCK)  
tCK  
ns  
2
1
to CA Bus Training Command Delay  
Valid Strobe Requirement before CKE Low  
First CA Bus Training  
10  
250  
ns  
Command Following CKE Low  
VREF Step Time – multiple steps  
VREF Step Time – one step  
tVREFCA_LONG Max  
tVREFCA_SHORT Max  
250  
80  
ns  
ns  
-
Valid Clock Requirement before CS High  
Valid Clock Requirement after CS High  
Minimum delay from  
tCKPRECS  
tCKPSTCS  
Min 2tCK + tXP (tXP = max(7.5ns, 5nCK))  
Min  
max(7.5ns, 5nCK))  
-
tCS_VREF  
Min  
2
tCK  
ns  
CS to DQS toggle in command bus training  
Minimum delay from  
tCKEHDQS  
10  
CKE High to Strobe High Impedance  
Clock and Command Valid before CKE High  
CA Bus Training CKE High to DQ Tri-state  
ODT turn-on Latency from CKE  
ODT turn-off Latency from CKE  
tCKCKEH  
tMRZ  
Min  
Min  
max(1.75ns, 3nCK)  
tCK  
ns  
ns  
ns  
-
1.5  
20  
tCKELODTon Min  
tCKELODToff Min  
tXCBT_Short Min  
tXCBT_Middle Min  
tXCBT_Long Min  
20  
max(5nCK, 200ns)  
max(5nCK, 200ns)  
max(5nCK, 250ns)  
3
3
3
Exit Command Bus Training Mode  
to next vaild command delay  
-
-
Notes:  
1. DQS has to retain a low level during tDQSCKE period, as well as DQS has to retain a high level.  
2. If tCACD is violated, the data for samples which violate tCACD will not be available, except for the last sample (where tCACD after this  
sample is met). Valid data for the last sample will be available after tADR.  
3. Exit Command Bus Training Mode to next valid command delay Time depends on value of VREF(CA) setting:  
MR12 OP[5:0] and VREF(CA) Range: MR12 OP[6] of FSP-OP 0 and 1. The details are shown in tFC value mapping.  
Additionally exit Command Bus Training Mode to next valid command delay Time may affect VREF(DQ) setting.  
Settling time of VREF(DQ) level is same as VREF(CA) level.  
Version 1.0  
222  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command Bus Trainging for Byte (x8) mode  
The LPDDR4ꢀSDRAM command bus must be trained before enabling termination for highꢀfrequency  
operation. LPDDR4 provides an internal VREF(ca) that defaults to a level suitable for unꢀterminated, lowfrequency  
operation, but the VREF(ca) must be trained to achieve suitable receiver voltage margin for  
terminated, highꢀfrequency operation. The training methodology described here centers the internal  
VREF(ca) in the CAdata eye and at the same time allows for timing adjustments of the CS and CA signals  
to meet setup/hold requirements. Because it can be difficult to capture commands prior to training the CA  
inputs, the training methodology described here uses a minimum of external commands to enter, train, and  
exit the Command Bus Training methodology.  
NOTES: it is up to the system designer to determine what constitutes “low-frequency” and “high-frequency” based on the capabilities of  
the system. Low-frequency should then be defined as an operating frequency in which the system can reliably communicate with  
the SDRAM before Command Bus Training is executed.  
The Byte mode LPDDR4ꢀSDRAM (x8 2ch.) is supported two Command Bus Training (CBT) modes and  
their feature is as follows.  
Mode1: DQ[6:0] only uses as output and VrefCA input procedure removes from CBT function of x16 2ch. device.  
Mode2: The status (Input or Output) of DQ[6:0] is controlled by DQ[7] pin.  
Aboveꢀmentioned CBT mode is selected by MRx [OPy].  
The LPDDR4ꢀSDRAM die has a bondꢀpad (ODTꢀCA) for multiꢀrank operation. In a multiꢀrank system, the  
terminating rank should be trained first, followed by the nonterminating rank(s). See the ODT section for  
more information.  
The corresponding DQ pins in this definition depends on the package configuration. DQ0 becomes DQ8 in  
some cases, as well as DQ1 to DQ6.  
Version 1.0  
223  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Mode 1  
The LPDDR4ꢀSDRAM uses Frequency SetꢀPoints to enable multiple operating settings for the die. The  
LPDDR4ꢀSDRAM defaults to FSPꢀOP[0] at powerꢀup, which has the default settings to operate in unterminated,  
lowꢀfrequency environments. Prior to training, the mode register settings should be configured  
by setting MR13 OP[6]=1B (FSPꢀWR[1]) and setting all other mode register bits including MR12 OP[6:0]  
(VREF(CA) Range and Setting) for FSPꢀOP[1] to the desired settings for highꢀfrequency operation. Prior to  
entering Command Bus Training, the SDRAM will be operating from FSPꢀOP[x]. Upon Command Bus  
Training entry when CKE is driven LOW, the LPDDR4ꢀSDRAM will automatically switch to the alternate  
FSP register set (FSPꢀOP[y]) and use the alternate register settings during training. Upon training exit when CKE  
is driven HIGH, the LPDDR4ꢀSDRAM will automatically switch back to the original FSP register set (FSPꢀOP[x]),  
returning to the “knownꢀgood” state that was operating prior to training.  
1. To set MRx OP[y] = 0: CBT Training Mode 1  
2. To enter Command Bus Training mode, issue a MRWꢀ1 command followed by a MRWꢀ2 command to  
set MR13 OP[0]=1 (Command Bus Training Mode Enabled).  
B
3. After time tMRD, CKE may be set LOW, causing the LPDDR4ꢀSDRAM to switch from FSPꢀOP[x] to  
FSPꢀOP[y], and completing the entry into Command Bus Training mode.  
A status of DQS, DQS, DQ and DMI are as follows, and DQ ODT state will be followed Frequency  
Set Point function except output pins.  
4. At time tCAENT later, LPDDR4 SDRAM can accept to input CA training pattern via CA bus.  
5. To verify that the receiver has the correct VREF(ca) setting and to further train the CA eye relative to  
clock (CK), values latched at the receiver on the CA bus are asynchronously output to the DQ bus.  
6. To exit Command Bus Training mode, drive CKE HIGH, and after time tXCBT issue the MRWꢀ1  
command followed by the MRWꢀ2 command to set MR13 OP[0]=0 . After time tMRW the LPDDR4ꢀ  
B
SDRAM is ready for normal operation. After training exit the LPDDR4ꢀSDRAM will automatically  
switch back to the FSPꢀOP registers that were in use prior to training.  
Command Bus Training may executed from IDLE or Self Refresh states. When executing CBT within the  
Self Refresh state, the SDRAM must not be in a power down state (i.e. CKE must be HIGH prior to training  
entry). Command Bus Training entry and exit is the same, regardless of the SDRAM state from which CBT  
is initiated.  
Version 1.0  
224  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Sequence of mode 1 for singleꢀrank systems  
Note that a example shown here is assuming an initial lowꢀfrequency, noꢀterminating operating point, training  
a highꢀfrequency, terminating operating point. The green text is lowꢀfrequency, magenta text is highꢀfrequency.  
Any operating point may be trained from any known good operating point.  
1. Set MR13 OP[6]=1 to enable writing to Frequency Set Point ‘y’ (FSPꢀWR[y]) (or FSPꢀOP[x], See  
B
note).  
2. Write FSPꢀWR[y] (or FSPꢀWR[x]) registers for all channels to set up highꢀfrequency operating  
parameters including VREF(CA) Range and Setting.  
3. Issue MRWꢀ1 and MRWꢀ2 commands to enter Command Bus Training mode.  
4. Drive CKE LOW, and change CK frequency to the highꢀfrequency operating point.  
5. Perform Command Bus Training (CS, and CA).  
6. Exit training, a change CK frequency to the lowꢀfrequency operating point prior to driving CKE HIGH,  
then issue MRWꢀ1 and MRWꢀ2 commands. When CKE is driven HIGH, the SDRAM will automatically  
switch back to the FSPꢀOP registers that were in use prior to training (i.e. trained values are not  
retained by the SDRAM).  
7. Issue MRWꢀ1 and MRWꢀ2 commands to switch to FSPꢀOP[y] (or FSPꢀOP[x]), to turn on termination,  
and change CK frequency to the high frequency operating point. At this point the Command Bus is  
trained and you may proceed to other training or normal operation.  
Note: Repeat steps 1 through 2 (Table- Timing constraints for Same bank: DQ ODT is Enabled) until the proper VREFCA level is  
established.  
Command Bus Training Steps.  
Step  
Mode  
1
Normal  
Low  
0
2
3(1)  
Normal  
Low  
0
4(2)  
CBT  
CBT  
Operation Frequency  
FSP-OP  
High  
1
High  
1
FSP-WR  
1
1
1
1
Training Pattern Input  
then comparison  
between output Data  
and expected data.  
Training Pattern Input  
then comparison  
between output Data  
and expected data.  
VREFCA Range/Value  
Setting via MRW  
VREFCA Range/Value  
Setting via MRW  
Operation  
Version 1.0  
225  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Sequence of mode 1 for multiꢀrank systems  
Note that a example shown here is assuming an initial lowꢀfrequency operating point, training a highꢀfrequency  
operating point. The green text is lowꢀfrequency, magenta text is highꢀfrequency. Any operating  
point may be trained from any known good operating point.  
1. Set MR13 OP[6]=1 to enable writing to Frequency Set Point ‘y’ (FSPꢀWR[y]) (or FSPꢀWR[x], See  
B
Note).  
2. Write FSPꢀWR[y] (or FSPꢀWR[x]) registers for all channels and ranks to set up high frequency  
operating parameters including VREF(CA) Range and Setting.  
3. Read MR0 OP[7] on all channels and ranks to determine which die are terminating, signified by  
MR0 OP[7]=1 .  
B
4. Issue MRWꢀ1 and MRWꢀ2 commands to enter Command Bus Training mode on the terminating rank.  
5. Drive CKE LOW on the terminating rank (or all ranks), and change CK frequency to the highfrequency  
operating point.  
6. Perform Command Bus Training on the terminating rank (CS, and CA).  
7. Exit training by driving CKE HIGH, change CK frequency to the lowꢀfrequency operating point, and  
issue MRWꢀ1 and MRWꢀ2 commands to write the trained values to FSPꢀWR[y] (or FSPꢀWR[x]). When  
CKE is driven HIGH, the SDRAM will automatically switch back to the FSPꢀOP registers that were in  
use prior to training.  
8. Issue MRWꢀ1 and MRWꢀ2 command to enter training mode on the nonꢀterminating rank (but keep  
CKE HIGH)  
9. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[y] (or FSPꢀOP[x]), to  
turn on termination, and change CK frequency to the high frequency operating point.  
10. Drive CKE LOW on the nonꢀterminating (or all) ranks. The nonꢀterminating rank(s) will now be using  
FSPꢀOP[y] (or FSPꢀOP[x]).  
11. Perform Command Bus Training on the nonꢀterminating rank (CS, and CA).  
12. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[x] (or FSPꢀOP[y]) to  
turn off termination.  
13. Exit training by driving CKE HIGH on the nonꢀterminating rank, change CK frequency to the lowfrequency  
operating point, and issue MRWꢀ1 and MRWꢀ2 commands. When CKE is driven HIGH, the  
SDRAM will automatically switch back to the FSPꢀOP registers that were in use prior to training.  
14. Write the trained values to FSPꢀWR[y] (or FSPꢀWR[x]) by issuing MRWꢀ1 and MRWꢀ2 commands to  
the SDRAM and setting all applicable mode register parameters.  
15. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[y] (or FSPꢀOP[x]), to  
turn on termination, and change CK frequency to the high frequency operating point. At this point the  
Command Bus is trained for both ranks and you may proceed to other training or normal operation.  
Version 1.0  
226  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Relation between CA input pin DQ output pin for mode 1  
Mapping of CA Input pin and DQ Output pin.  
Mapping  
CA Number  
DQ Number  
CA5  
CA4  
CA3  
CA2  
CA1  
CA0  
DQ5  
DQ4  
DQ3  
DQ2  
DQ1  
DQ0  
Timing Diagram for mode 1  
The basic Timing diagrams of Command Bus Training are shown in the following figures.  
Entering Command Bus Training Mode and CA Training Pattern Input and Output  
Notes:  
1. After tCKELCK clock can be stopped or frequency changed any time.  
2. The input clock condition should be satisfied tCKPRECS and tCKPSTCS.  
3. Continue to Drive CK and Hold CA & CS pins low until tCKELCK after CKE is low (which disables command decoding).  
4. When CKE is driven LOW, the SDRAM will switch its FSP-OP registers to use the alternate (i.e. non-active) set. Example: If the SDRAM is  
currently using FSP-OP[0], then it will switch to FSP-OP[1] when CKE is driven LOW. All operating parameters should be written to the  
alternate mode registers before entering Command Bus Training to ensure that ODT settings, RL/WL/nWR setting, etc., are set to the  
correct values. If the alternate FSP-OP has ODT_CA disabled then termination will not enable in CA Bus Training mode. If the ODT_CA  
pad is bonded to Vss or floating, ODT_CA termination will never enable for that die.  
5. When CKE is driven low in Command Bus Training mode, the LPDDR4-SDRAM will change operation to the alternate FSP, i.e. non-active  
FSP programmed in the FSP-OP mode register.  
Version 1.0  
227  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Exiting Command Bus Training Mode with Valid Command  
Notes:  
1. CK must meet tCKCKEH before CKE is driven high.  
When CKE is driven high the clock frequency must be returned to the original frequency (the frequency corresponding to the FSP at  
which Command Bus Training mode was entered)  
2. CS and CA[5:0] must be Deselect (all low) tCKCKEH before CKE is driven high.  
3. When CKE is driven high, the SDRAM’s ODT_CA will revert to the state/value defined by FSP-OP prior to Command Bus Training mode  
entry, i.e. the original frequency set point (FSP-OP, MR13-OP[7]).  
Example: If the SDRAM was using FSP-OP[1] for training, then it will switch to FSP-OP[0] when CKE is driven HIGH.  
4. When CKE is driven high the LPDDR4-SDRAM will revert to the FSP in operation when Command Bus Training mode was entered.  
Version 1.0  
228  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Exiting Command Bus Training Mode with Power Down Entry  
Notes:  
1. Clock can be stopped or frequency changed any time before tCKCKEH. CK must meet tCKCKEH before CKE is driven high.  
When CKE is driven high the clock frequency must be returned to the original frequency (the frequency corresponding to the FSP at  
which Command Bus Training mode was entered)  
2. CS and CA[5:0] must be Deselect (all low) tCKCKEH before CKE is driven high.  
3. When CKE is driven high, the SDRAM’s ODT_CA will revert to the state/value defined by FSP-OP prior to Command Bus Training mode  
entry, i.e. the original frequency set point (FSP-OP, MR13-OP[7]).  
Example: If the SDRAM was using FSP-OP[1] for training, then it will switch to FSP-OP[0] when CKE is driven HIGH.  
4. When CKE is driven high the LPDDR4-SDRAM will revert to the FSP in operation when Command Bus Training mode was entered.  
Version 1.0  
229  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command Bus Training AC Timing Table for Mode 1  
Data Rate  
3733  
Min/  
Parameter  
Symbol  
Unit Notes  
Max  
Command Bus Training Timing  
Clock and Command Valid after CKE Low  
Asynchronous Data Read  
tCKELCK  
tADR  
Min  
Max  
max(7.5ns, 3nCK)  
20  
tCK  
ns  
CA Bus Training Command  
tCACD  
Min  
Min  
RU(tADR/tCK )  
250  
tCK  
ns  
2
to CA Bus Training Command Delay  
First CA Bus Training  
tCAENT  
Command Following CKE Low  
Valid Clock Requirement before CS High  
Valid Clock Requirement after CS High  
Clock and Command Valid before CKE High  
CA Bus Training CKE High to DQ Tri-state  
ODT turn-on Latency from CKE  
ODT turn-off Latency from CKE  
tCKPRECS  
tCKPSTCS  
tCKCKEH  
tMRZ  
Min 2tCK + tXP (tXP = max(7.5ns, 5nCK))  
-
-
Min  
Min  
Min  
max(7.5ns, 5nCK))  
2
tCK  
ns  
ns  
ns  
-
1.5  
20  
tCKELODTon Min  
tCKELODToff Min  
tXCBT_Short Min  
tXCBT_Middle Min  
tXCBT_Long Min  
20  
max(5nCK, 200ns)  
max(5nCK, 200ns)  
max(5nCK, 250ns)  
3
3
3
Exit Command Bus Training Mode  
to next vaild command delay  
-
-
Notes:  
1. If tCACD is violated, the data for samples which violate tCACD will not be available, except for the last sample (where tCACD after this  
sample is met). Valid data for the last sample will be available after tADR.  
2. Exit Command Bus Training Mode to next valid command delay Time depends on value of VREF(CA) setting:  
MR12 OP[5:0] and VREF(CA) Range: MR12 OP[6] of FSP-OP 0 and 1. The details are shown in tFC value mapping.  
Additionally exit Command Bus Training Mode to next valid command delay Time may affect VREF(DQ) setting.  
Settling time of VREF(DQ) level is same as VREF(CA) level.  
Version 1.0  
230  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Mode 2  
The LPDDR4ꢀSDRAM uses Frequency SetꢀPoints to enable multiple operating settings for the die. The  
LPDDR4ꢀSDRAM defaults to FSPꢀOP[0] at powerꢀup, which has the default settings to operate in unterminated,  
lowꢀfrequency environments. Prior to training, the mode register settings should be configured by  
setting MR13 OP[6]=1 (FSPꢀWR[1]) and setting all other mode register bits for FSPꢀOP[1] to the desired  
B
settings for highꢀfrequency operation. Prior to entering Command Bus Training, the SDRAM will be operating  
from FSPꢀOP[x]. Upon Command Bus Training entry when CKE is driven LOW, the LPDDR4ꢀSDRAM  
will automatically switch to the alternate FSP register set (FSPꢀOP[y]) and use the alternate register settings  
during training. Upon training exit when CKE is driven HIGH, the LPDDR4ꢀSDRAM will automatically switch back  
to the original FSP register set (FSPꢀOP[x]), returning to the “knownꢀgood” state that was operating prior to  
training. The training values for VREFCA are not retained by the DRAM in FSPꢀOP[y] registers, and must be  
written to the registers after training exit.  
1. To set MR12 OP[7] = 1: CBT Training Mode 2  
2. To enter Command Bus Training mode, issue a MRWꢀ1 command followed by a MRWꢀ2 command to  
set MR13 OP[0]=1 (Command Bus Training Mode Enabled).  
B
3. After time tMRD, CKE may be set LOW, causing the LPDDR4ꢀSDRAM to switch from FSPꢀOP[x] to  
FSPꢀOP[y], and completing the entry into Command Bus Training mode.  
A status of DQS, DQS, DQ and DMI are as follows, and ODT state of DQS, DQS, DQ and  
DMI will be followed by MR11 OP[2:0]: DQ ODT and MR13 OP[7]: FSPꢀOP except when pin is output  
or transition state.  
ꢀ DQS, DQS become input pins for capturing DQ[6:0] levels by its toggling. The ODT for the DQS, DQS is  
always enabled during CBT Mode 2. The DQS, DQS ODT use the value specified by MR11 OP[2:0]: DQ  
ODT and MR13 OP[7]: FSPꢀOP.  
ꢀ DQ[5:0] become input pins for setting VREFCA Level during tDStrain + tDQSICYC + tDHtrain period.  
ꢀ DQ[5:0] become output pins to feedback its capturing value via command bus by CS signal during tADVW  
period.  
ꢀ DQ[6] becomes a input pin for setting VREFCA Range during tDStrain + tDQSICYC + tDHtrain period.  
ꢀ DQ[6] becomes an output pin during tADVW period and the output data is meaningless.  
ꢀ DQ[7] becomes an output pin to indicate the meaningful data output by its toggling during tADVW period.  
The meaningful data is its capturing value via command bus by CS signal. DQ[7] status except tADVW  
period becomes input or disable, this state is vendor specific, as well as ODT behavior.  
ꢀ DMI become Input, output or disable, The DMI state is vendor specific.  
4. At time tCAENT later, LPDDR4 SDRAM can accept to change its V  
CA Range and Value using input  
REF  
signals of DQS, DQS and DQ[6:0] from existing value that’s setting via MR12 OP[6:0]. The mapping between  
MR12 OP code and DQs is shown in following table. At least one V  
next training steps.  
CA setting is required before proceed to  
REF  
Version 1.0  
12/2018  
231  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mapping of CA Input pin and DQ Output pin.  
Mapping  
MR12 OP Code  
DQ Number  
OP6  
DQ6  
OP5  
DQ5  
OP4  
DQ4  
OP3  
DQ3  
OP2  
DQ2  
OP1  
DQ1  
OP0  
DQ0  
5. The new VREFCA value must “settle” for time tVREF_LONG before attempting to latch CA information.  
6. To verify that the receiver has the correct VREFCA setting and to further train the CA eye relative to clock (CK),  
values latched at the receiver on the CA bus are asynchronously output to the DQ bus.  
7. Command followed by the MRWꢀ2 command to set MR13 OP[0]=0B. After time tMRW the LPDDR4ꢀSDRAM is  
ready for normal operation. After training exit the LPDDR4ꢀSDRAM will automatically switch back to the  
FSPꢀOP registers that were in use prior to training.  
Command Bus Training may executed from IDLE or Self Refresh states. When executing CBT within the  
Self Refresh state, the SDRAM must not be a power down state (i.e. CKE must be HIGH prior to training  
entry). Command Bus Training entry and exit is the same, regardless of the SDRAM state from which CBT  
is initiated.  
Version 1.0  
232  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Sequence of mode 2 for singleꢀrank systems  
Note that a example shown here is assuming an initial lowꢀfrequency, noꢀterminating operating point, training  
a highꢀfrequency, terminating operating point. The green text is lowꢀfrequency, magenta text is highꢀfrequency.  
Any operating point may be trained from any known good operating point. This example is  
assuming on the following condition. Frequency Set Point ‘x’ for low frequency operation and Frequency  
Set Point ‘y’ for High frequency operation.  
1. Set MR13 OP[6]=1 to enable writing to Frequency Set Point ‘y’ (FSPꢀWR[y]).  
B
2. Write FSPꢀWR[y] registers for all channels to set up highꢀfrequency operating parameters.  
3. Issue MRWꢀ1 and MRWꢀ2 commands to enter Command Bus Training mode.  
4. Drive CKE LOW, then change CK frequency to the highꢀfrequency operating point.  
5. Perform Command Bus Training (VREFCA, CS, and CA).  
6. Exit training, a change CK frequency to the lowꢀfrequency operating point prior to driving CKE HIGH,  
then issue MRWꢀ1 and MRWꢀ2 commands to exit Command Bus Training mode. When CKE is driven  
HIGH, the SDRAM will automatically switch back to the FSPꢀOP registers that were in use prior to  
training (i.e. trained values are not retained by the SDRAM).  
7. Write the trained values to FSPꢀWR[y] by issuing MRWꢀ1 and MRWꢀ2 commands to the SDRAM and  
setting all applicable mode register parameters.  
8. Issue MRWꢀ1 and MRWꢀ2 commands to switch to FSPꢀOP[y] to turn on termination, and change CK  
frequency to the high frequency operating point. At this point the Command Bus is trained and you  
may proceed to other training or normal operation.  
Version 1.0  
233  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Training Sequence of mode 2 for multiꢀrank systems  
Note that a example shown here is assuming an initial lowꢀfrequency operating point, training a highꢀfrequency  
operating point. The green text is lowꢀfrequency, magenta text is highꢀfrequency. Any operating  
point may be trained from any known good operating point. This example is assuming on the following  
condition. Frequency Set Point ‘x’ for low frequency operation and Frequency Set Point ‘y’ for High frequency  
operation.  
1. Set MR13 OP[6]=1 to enable writing to Frequency Set Point ‘y’ (FSPꢀWR[y]).  
B
2. Write FSPꢀWR[y] registers for all channels and ranks to set up high frequency operating parameters.  
3. Read MR0 OP[7] on all channels and ranks to determine which die are terminating, signified by MR0  
OP[7]=1B.  
4. Issue MRWꢀ1 and MRWꢀ2 commands to enter Command Bus Training mode on the terminating rank.  
5. Drive CKE LOW on the terminating rank (or all ranks), and change CK frequency to the high  
frequency operating point.  
6. Perform Command Bus Training on the terminating rank (VREFCA, CS, and CA).  
7. Exit training by driving CKE HIGH, change CK frequency to the lowꢀfrequency operating point, and  
issue MRWꢀ1 and MRWꢀ2 commands to exit Command Bus Training mode. When CKE is driven  
HIGH, the SDRAM will automatically switch back to the FSPꢀOP registers that were in use prior to  
training (i.e. trained values are not retained by the SDRAM).  
8. Write the trained values to FSPꢀWR[y] by issuing MRWꢀ1 and MRWꢀ2 commands to the SDRAM and  
setting all applicable mode register parameters.  
9. Issue MRWꢀ1 and MRWꢀ2 command to enter training mode on the nonꢀterminating rank (but keep  
CKE HIGH).  
10. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[y], to turn on  
termination, and change CK frequency to the high frequency operating point.  
11. Drive CKE LOW on the nonꢀterminating (or all) ranks. The nonꢀterminating rank(s) will now be using  
FSPꢀOP[y].  
12. Perform Command Bus Training on the nonꢀterminating rank (VREFCA, CS, and CA).  
13. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[x] to turn off termination.  
14. Exit training by driving CKE HIGH on the nonꢀterminating rank, change CK frequency to the low  
frequency operating point, and issue MRWꢀ1 and MRWꢀ2 commands to exit Command Bus Training  
mode. When CKE is driven HIGH, the SDRAM will automatically switch back to the FSPꢀOP registers  
that were in use prior to training (i.e. trained values are not retained by the SDRAM).  
15. Write the trained values to FSPꢀWR[y] by issuing MRWꢀ1 and MRWꢀ2 commands to the SDRAM and  
setting all applicable mode register parameters.  
16. Issue MRWꢀ1 and MRWꢀ2 commands to switch the terminating rank to FSPꢀOP[y], to turn on  
termination, and change CK frequency to the high frequency operating point. At this point the  
Command Bus is trained for both ranks and you may proceed to other training or normal operation.  
Version 1.0  
234  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Relation between CA input pin DQ output pin for mode 2  
Mapping of CA Input pin and DQ Output pin.  
Mapping  
CA Number  
DQ Number  
CA5  
CA4  
CA3  
CA2  
CA1  
CA0  
DQ5  
DQ4  
DQ3  
DQ2  
DQ1  
DQ0  
Timing Diagram for mode 2  
The basic Timing diagrams of Command Bus Training are shown in the following figures.  
Entering Command Bus Training Mode and CA Training Pattern Input with VrefCA Value Update  
Notes:  
1. After tCKELCK clock can be stopped or frequency changed any time.  
2. The input clock condition should be satisfied tCKPRECS and tCKPSTCS.  
3. Continue to Drive CK and Hold CS pins low until tCKELCK after CKE is low (which disables command decoding).  
4. The DRAM may or may not capture the first rising/falling edge of DQS_t/c due to an unstable first rising edge. At least 2 consecutive  
pulses of DQS signal input are required for every DQS input signal when capturing DQ[6:0] signals.  
The captured value of the DQ[6:0] signal level by each DQS edge is overwritten at any time. The DRAM updates its VREFca setting of  
MR12 temporary, after time tVREFca_Long.  
5. tVREFca_Long may be reduced to tVREFca_Middle or tVREFca_Short.  
6. When CKE is driven LOW, the SDRAM will switch its FSP-OP registers to use the alternate (i.e. non-active) set. Example: If the SDRAM is  
currently using FSP-OP[0], then it will switch to FSP-OP[1] when CKE is driven LOW. All operating parameters should be written to the  
alternate mode registers before entering Command Bus Training to ensure that ODT settings, RL/WL/nWR setting, etc., are set to the  
correct values. If the alternate FSP-OP has ODT_CA disabled then termination will not enable in CA Bus Training mode. If the ODT_CA  
pad is bonded to Vss, ODT_CA termination will never enable for that die.  
7. When CKE is driven low in Command Bus Training mode, the LPDDR4-SDRAM will change operation to the alternate FSP, i.e. non-active  
FSP programmed in the FSP-OP mode register  
8. tDStrai+ tDQSICYC + tDHtrain period on DQ7 become Input or disable, this state during CBT Mode 2 is vendor specific.  
9. DMI become Input, output or disable, The DMI state during CBT Mode 2 is vendor specific  
Version 1.0  
235  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CA pattern Input/Output to Vref setting Input  
Notes:  
1. The DRAM may or may not capture the first rising/falling edge of DQS_t/c due to an unstable first rising edge. At least 2 consecutive  
pulses of DQS signal input are required for every DQS input signal when capturing DQ[6:0] signals.  
The captured value of the DQ[6:0] signal level by each DQS edge is overwritten at any time. The DRAM updates its VREFca setting of  
MR12 temporary, after time tVREFca_Long.  
2. tVREFca_Long may be reduced to tVREFca_Middle or tVREFca_Short.  
3. tDStrain + tDQSICYC + tDHtrain period on DQ7 become Input or disable, this state during CBT Mode 2 is vendor specific.  
4. DMI become Input, output or disable, The DMI state during CBT Mode 2 is vendor specific.  
Version 1.0  
236  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Consecutive CA training pattern Input/Output  
Notes:  
1. DMI become Input, output or disable, The DMI state during CBT Mode 2 is vendor specific.  
Version 1.0  
237  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Exiting Command Bus Training Mode  
Notes:  
1. CK must meet tCKCKEH before CKE is driven high.  
When CKE is driven high the clock frequency must be returned to the original frequency (the frequency corresponding to the FSP at  
which Command Bus Training mode was entered)  
2. CS and CA[5:0] must be all low tCKCKEH before CKE is driven high.  
3. CKE must be held low from when CS transitions high to when tCBTRTW is satisfied. Exiting CBT mode is prohibited during this period.  
4. When CKE is driven high, the SDRAM’s ODT_CA will revert to the state/value defined by FSP-OP prior to Command Bus Training mode  
entry, i.e. the original frequency set point (FSP-OP, MR13-OP[7]).  
Example: If the SDRAM was using FSP-OP[1] for training, then it will switch to FSP-OP[0] when CKE is driven HIGH.  
5. Training values are not retained by the SDRAM, and must be written to the FSP-OP register set before returning to operation at the  
trained frequency. Example: VREF(ca) will return to the value programmed in the original set point.  
6. When CKE is driven high the LPDDR4-SDRAM will revert to the FSP in operation when Command Bus Training mode was entered.  
7. DMI become Input, output or disable, The DMI state during CBT Mode 2 is vendor specific.  
Version 1.0  
238  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQS ODT Timing during Command Bus Training Mode 2  
Notes:  
1. After tCKELCK clock can be stopped or frequency changed any time.  
2. Continue to Drive CK and Hold CS pins low until tCKELCK after CKE is low (which disables command decoding).  
3. When CKE is driven low in Command Bus Training mode, the LPDDR4-SDRAM will change operation to the alternate FSP, i.e non-active  
FSP programmed in the FSP-OP mode register.  
4. CK must meet tCKCKEH before CKE is driven high.  
When CKE is driven high the clock frequency must be returned to the original frequency (the frequency corresponding to the FSP at  
which Command Bus Training mode was entered)  
5. CS and CA[5:0] must be all low tCKCKEH before CKE is driven high.  
6. When CKE is driven high the LPDDR4-SDRAM will revert to the FSP in operation when Command Bus Training mode was entered.  
Version 1.0  
239  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command Bus Training AC Timing Table for Mode 2  
Min/  
Data Rate  
3733  
Parameter  
Symbol  
Unit Notes  
Max  
Command Bus Training Timing  
Clock and Command Valid after CKE Low  
Valid Clock Requirement before CS High  
Valid Clock Requirement after CS High  
Valid Strobe Requirement before CKE Low  
First CA Bus Training  
tCKELCK  
tCKPRECS  
tCKPSTCS  
tDQSCKE  
Min  
max(5ns, 5nCK)  
tCK  
Min 2tck + tXP (tXP = max(7.5ns, 5nCK))  
-
-
Min  
Min  
max(7.5ns, 5nCK))  
10  
ns  
1
tCAENT  
Min  
250  
ns  
Command Following CKE Low  
VREF Step Time – Long  
Max  
Max  
Max  
Min  
Min  
Min  
Max  
Min  
Max  
Max  
Min  
Min  
Max  
Min  
Max  
Min  
Max  
250  
200  
100  
2
ns  
ns  
ns  
ns  
ns  
Ns  
ns  
Ns  
ns  
ns  
2
3
4
tVREFCA_LONG  
tVREFCA_Middle  
tVREFCA_SHORT  
tDStrain  
VREF Step Time – Middle  
VREF Step Time – Short  
Data Setup for VREF Training Mode  
Data Hold forVREF Training Mode  
tDHtrain  
2
16  
80  
5
Asynchronous Data Read Valid Window  
DQS Input period at CBT mode  
tADVW  
tDQSICYC  
100  
20  
0
Asynchronous Data Read  
tADR  
DQS high impedance time from CS High  
tHZCBT  
3
Asynchronous Data Read to DQ7 toggle  
DQ7sample hold time  
tAD2DQ7  
tDQ7SH  
tADSPW  
10  
10  
60  
3
Asynchronous Data Read Pulse Width  
Hi-Z to asynchronous VrefCA valid data  
10  
tHZ2VREF  
tCBTRTW  
Min  
Min  
Max(10ns, 5nCK)  
tCK  
ns  
Read to Write Delay at CBT mode  
CA Bus Training Command to CA Bus  
Training Command Delay  
tCACD  
Min  
Max(110ns, 4nCK)  
10  
Minimum delay from  
tCKEHDQS  
tCKCKEH  
Min  
Min  
ns  
CKE High to Strobe High Impedance  
Clock and Command Valid before CKE High  
ODT turn-on Latency from CKE  
max(1.75ns, 3nCK)  
20  
tCK  
ns  
tCKELODTon Max  
Version 1.0  
240  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data Rate  
3733  
Min/  
Max  
Parameter  
Symbol  
Unit Notes  
Command Bus Training Timing  
ODT turn-off Latency from CKE for ODT_CA tCKELODToff Max  
20  
20  
ns  
ns  
ODT turn-off Latency from CKE  
tCKEHODToff Max  
for ODT_DQ and DQS  
ODT_DQ turn-off Latency from  
tODToffCBT Max  
20  
ns  
-
CS high during CB Training  
ODT_DQ turn-on Latencyfrom  
tODTonCBT Max  
the end of Valid Data out  
Max(10ns, 5nCK)  
tXCBT_Short Min  
Exit Command Bus Training Mode  
tXCBT_Middle Min  
max(5nCK, 200ns)  
max(5nCK, 200ns)  
max(5nCK, 250ns)  
-
-
-
5
5
5
to next vaild command delay  
tXCBT_Long Min  
Notes:  
1. DQS has to retain a low level during tDQSCKE period, as well as DQS has to retain a high level.  
2. VREFCA_Long is the time including up to VREFmin to VREFmax or VREFmax to VREFmin change across the VREFDQ Range in VREF  
voltage.  
3. VREF_Middle is at least 2 stepsizes increment/decrement change within the same VREFDQ range in VREF voltage.  
4. VREF_Short is for a single stepsize increment/decrement change in VREF voltage.  
5. Exit Command Bus Training Mode to next valid command delay Time depends on value of VREF(CA) setting: MR12 OP[5:0] and  
VREF(CA) Range: MR12 OP[6] of FSP-OP 0 and 1. The details are shown in Table ‘tFC value mapping’.  
Additionally exit Command Bus Training Mode to next valid command delay Time may affect VREF(DQ) setting. Settling time of VREF(DQ)  
level is same as VREF(CA) level.  
Version 1.0  
241  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Frequency Set Point  
Frequency SetꢀPoints allow the LPDDR4ꢀSDRAM CA Bus to be switched between two differing operating  
frequencies, with changes in voltage swings and termination values, without ever being in an unꢀtrained state which  
could result in a loss of communication to the DRAM. This is accomplished by duplicating all CA Bus mode register  
parameters, as well as other mode register parameters commonly changed with operating frequency. These  
duplicated registers form two sets that use the same mode register addresses, with read/write access controlled by  
MR bit FSPꢀWR (Frequency SetꢀPoint Write/Read) and the DRAM operating point controlled by another MR bit  
FSPOP (Frequency SetꢀPoint Operation). Changing the FSPꢀWR bit allows MR parameters to be changed for an  
alternate Frequency SetꢀPoint without affecting the LPDDR4ꢀSDRAM's current operation. Once all necessary  
parameters have been written to the alternate SetꢀPoint, changing the FSPꢀOP bit will switch operation to use all of  
the new parameters simultaneously (within tFC), eliminating the possibility of a loss of communication that could be  
caused by a partial configuration change.  
Version 1.0  
242  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mode Register Function with two physical registers  
Parameters which have two physical registers controlled by FSPꢀWR and FSPꢀOP include:  
MR#  
Operand  
OP[2]  
Function  
WR-PRE (WR Pre-amble Length)  
Notes  
OP[3]  
RD-PRE (RD Pre-amble Type)  
MR1  
1
OP[6:4]  
OP[7]  
nWR (Write-Recovery for Auto-Pre-charge commands)  
PST (RD Post-Amble Length)  
OP[2:0]  
OP[5:3]  
OP[6]  
RL (Read latency)  
MR2  
MR3  
WL (Write latency)  
WLS (Write Latency Set)  
OP[0]  
PU-Cal (Pull-up Calibration Point)  
WR PST(WR Post-Amble Length)  
PDDS (Pull-Down Drive Strength)  
DBI-RD (DBI-Read Enable)  
2
OP[1]  
OP[5:3]  
OP[6]  
OP[7]  
DBI-WR (DBI-Write Enable)  
OP[2:0]  
OP[6:4]  
OP[5:0]  
OP[6]  
DQ ODT (DQ Bus Receiver On-Die-Termination)  
CA ODT (CA Bus Receiver On-Die-Termination)  
VREF(ca) (VREF(ca) Setting)  
MR11  
MR12  
MR14  
VR-CA (VREF(ca) Range)  
OP[5:0]  
OP[6]  
VREF(dq) (VREF(dq) Setting)  
VR(dq) (VREF(dq) Range)  
OP[2:0]  
OP[3]  
SoC ODT (Controller ODT Value for VOH calibration)  
ODTE-CK (CK ODT enabled for nonterminating rank)  
ODTE-CS (CS ODT enable for non terminating rank)  
ODTD-CA (CA ODT termination disable)  
MR22  
OP[4]  
OP[5]  
Notes:  
1. Supporting the two physical registers for Burst Length: MR1 OP[1:0] is optional.  
Applications requiring support of both vendor options shall assure that both FSP-OP[0] and FSP-OP[1] are set to the  
same code. Refer to vendor datasheets for detail.  
2. For dual channel devices, PU-CAL setting is required as the same value for both Ch.A and Ch.B before issuing ZQ Cal start command.  
See Mode Register Definition for more details.  
Version 1.0  
243  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Relation between MR Setting and DRAM Operation  
The following table shows how the two mode registers for each of the parameters above can be modified by setting  
the appropriate FSPꢀWR value, and how device operation can be switched between operating points by setting the  
appropriate FSPꢀOP value. The FSPꢀWR and FSPꢀOP functions operate completely independently.  
Function  
MR# & Operand  
Data  
Operation  
Notes  
Data write to Mode Register N for FSP-OP[0]  
by MRW Command  
0 (Default)  
Data read from Mode Register N for FSP-OP[0] by MRR  
Command.  
FSP-WR  
MR13 OP[6]  
1
Data write to Mode Register N for FSP-OP[1]  
by MRW Command.  
1
Data read from Mode Register N for FSP-OP[1] by MRR  
Command.  
DRAM operates with Mode Register N for FSP-OP[0]  
setting.  
0 (Default)  
1
FSP-OP  
MR13 OP[7]  
2
DRAM operates with Mode Register N for FSP-OP[1]  
setting.  
Notes:  
1. FSP-WR stands for Frequency Set Point Write/Read.  
2. FSP-OP stands for Frequency Set Point Operating Point.  
Version 1.0  
244  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Frequency set point update Timing  
The Frequency set point update timing is shown in following figure. When changing the frequency set point via  
MR13 OP[7], the VRCG setting: MR13 OP[3] have to be changed into VREF Fast Response (high current) mode at  
the same time. After Frequency change time(tFC) is satisfied. VRCG can be changed into Normal Operation mode  
via MR13 OP[3].  
Frequency Set Point Switching Timing  
Notes:  
1. The definition that is Clock frequency change during CKE HIGH should be followed at the frequency change operation.  
For more information, refer to Section 4.49 Input Clock Stop and Frequency Change.  
Version 1.0  
245  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AC Timing Table  
Data Rate  
3733  
Min/  
Max  
Parameter  
Symbol  
Unit Notes  
Frequency Set Point parameters  
tFC_Short  
tFC_Middle  
tFC_Long  
Min  
Min  
Min  
200  
200  
250  
ns  
ns  
ns  
1
1
1
Frequency Set Point Switching Time  
Valid Clock Requirement  
tCKFSPE  
tCKFSPX  
Min  
Min  
max(7.5ns, 4nCK)  
max(7.5ns, 4nCK)  
-
-
after Entering FSP Change  
Valid Clock Requirement before 1st Valid  
Command after FSP change  
NOTE1. Frequency Set Point Switching Time depends on value of Vref(ca) setting: MR12 OP[5:0] and Vref(ca) Range: MR12 OP[6] of FSP-OP  
0 and 1. The details are shown in following table.  
tFC value mapping  
Additionally change of Frequency Set Point may affect Vref(dq) setting. Settling time of Vref(dq) level is same as Vref(ca) level.  
Step Size  
Range  
Application  
From  
FSP -OP0  
Base  
To  
From  
FSP -OP0  
Base  
To  
FSP-OP1  
A single step size increment/decrement  
Two or more step size increment/decrement  
-
FSP-OP1  
No Change  
No Change  
Change  
tFC_Short  
tFC_Middle  
tFC_Long  
Base  
Base  
-
Base  
NOTE1. As well as change from FSP-OP1 to FSP-OP0.  
tFC value mapping example  
The following table provides an example of tFC value mapping when FSP-OP moves from OP0 to OP1.  
FSP-OP:  
VREF(ca) Setting:  
MR12: OP[5:0]  
001100  
VREF(ca) Range:  
Case  
From/To  
Application  
Notes  
MR13 OP[7]  
MR12 OP[6]  
From  
To  
0
1
0
1
0
1
0
0
0
0
0
1
1
2
3
tFC_Short  
tFC_Middle  
tFC_Long  
1
2
3
001101  
From  
To  
001100  
001110  
From  
To  
Don’t Care  
Don’t Care  
Notes:  
1. A single step size increment/decrement for VREF(CA) Setting Value.  
2. Two or more step size increment/decrement for VREF(CA) Setting Value.  
3. VREF(CA) Range is changed. In tis case changing VREF(CA) Setting doesn’t affect tFC value.  
Version 1.0  
246  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The LPDDR4ꢀSDRAM defaults to FSPꢀOP[0] at powerꢀup. Both SetꢀPoints default to settings needed to operate in  
unꢀterminated, lowꢀfrequency environments. To enable the LPDDR4ꢀSDRAM to operate at higher frequencies,  
Command Bus Training mode should be utilized to train the alternate Frequency SetꢀPoint. See the section  
Command Bus Training for more details on this training mode.  
Prepare for CA Bus  
Training of FSP[1] for  
High Frequency  
FSP-OP=0  
Powerup/  
Initialization  
FSP-OP=0  
FSP-WR=0  
Freq = Boot  
CA Bus Training,  
FSP-OP[1]  
FSP-WR=1  
Freq = high  
CKE highlow  
FSP-WR=1  
Freq = Boot  
CKE lowhigh  
Prepare for CA Bus  
Training of FSP[0]  
for Med. Frequency  
FSP-OP=1  
Exit CA Bus  
Training  
FSP-OP=0  
FSP-WR=1  
Freq = Boot  
Switch to  
High-Speed Mode  
FSP-OP=1  
FSP-WR=1  
Freq = High  
FSP-WR=0  
Freq = high  
CKE highlow  
Exit CA Bus  
Training  
FSP-OP=1  
FSP-WR=0  
Freq = High  
CA Bus Training,  
FSP-OP[0]  
FSP-WR=0  
Operate at  
High speed  
CKE lowhigh  
Freq = Med  
Training Two Frequency SetꢀPoints  
Version 1.0  
247  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Once both Frequency SetꢀPoints have been trained, switching between points can be performed by a  
single MRW followed by waiting for tFC.  
State n-1: FSP-OP=1  
MWR Command  
State n: FSP-OP=0  
Operate at  
Medium speed  
Operate at  
High speed  
tFC  
State n-1: FSP-OP=1  
MWR Command  
State n: FSP-OP=0  
Operate at  
High speed  
tFC  
Switching Between Two Trained Frequency SetꢀPoints  
Switching to a third (or more) SetꢀPoint can be accomplished if the memory controller has stored the  
previouslyꢀtrained values (in particular the VrefꢀCA calibration value) and reꢀwrites these to the alternate SetꢀPoint  
before switching FSPꢀOP.  
MRW Vref-CA  
CA-ODT, DQ-ODT,RL,  
WL, Vref-DQ,...etc.  
State n-1: FSP-OP=1  
State n: FSP-OP=0  
Operate at  
High speed  
tFC  
Operate at  
Third  
State n-1: FSP-OP=1  
State n: FSP-OP=0  
speed  
tFC  
Switching to a Third Trained Frequency SetꢀPoint  
Version 1.0  
248  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mode Reister WriteꢀWR Leveling Mode  
To improve signalꢀintegrity performance, the LPDDR4 SDRAM provides a writeꢀleveling feature to compensate  
CKꢀtoꢀDQS timing skew affecting timing parameters such as tDQSS, tDSS, and tDSH. The DRAM samples the  
clock state with the rising edge of DQS signals, and asynchronously feeds back to the memory controller. The  
memory controller references this feedback to adjust the clockꢀtoꢀdata strobe signal relationship for each DQS/DQS  
signal pair.  
All data bits (DQ[7:0] for DQS/DQS[0], and DQ[15:8] for DQS/DQS [1]) carry the training feedback to the controller.  
Both DQS signals in each channel must be leveled independently. Writeꢀleveling entry/exit is independent between  
channels.  
The LPDDR4 SDRAM enters into writeꢀleveling mode when mode register MR2ꢀOP[7] is set HIGH. When entering  
writeꢀleveling mode, the state of the DQ pins is undefined. During writeꢀleveling mode, only DESELECT commands  
are allowed, or a MRW command to exit the writeꢀleveling operation. Depending on the absolute values of tDQSL  
and tDQSH in the application, the value of tDQSS may have to be better than the limits provided in the Write AC  
Timing Table1 in order to satisfy the tDSS and tDSH specifications. Upon completion of the writeꢀleveling operation,  
the DRAM exits from writeꢀleveling mode when MR2ꢀOP[7] is reset LOW.  
Write Leveling should be performed before Write Training (DQS2DQ Training).  
NOTE1. As of publication of this document, under discussion by the formulating committee.  
Write Leveling Procedure  
1. Enter into Writeꢀleveling mode by setting MR2ꢀOP[7]=1.  
2. Once entered into Writeꢀleveling mode, DQS must be driven LOW and DQS HIGH after a delay of tWLDQSEN.  
3. Wait for a time tWLMRD before providing the first DQS signal input. The delay time tWLMRD(MAX) is controller  
dependent.  
4. DRAM may or may not capture first rising edge of DQS due to an unstable first risign edge. Hence provide at least  
consecutive 2 pulses of DQS signal input is required in every DQS input signal during Write Training Mode.  
The captured clock level by each DQS edges are overwritten at any time and the DRAM provides asynchronous  
feedback on all the DQ bits after time tWLO.  
5. The feedback provided by the DRAM is referenced by the controller to increment or decrement the DQS and/or  
DQS delay settings.  
6. Repeat step 4 through step 5 until the proper DQS/DQS delay is established.  
7. Exit from Writeꢀleveling mode by setting MR2ꢀOP[7]=0.  
Version 1.0  
249  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
A Write Leveling timing example is shown in figure below.  
Write Leveling Timing, tDQSL(max)  
Write Leveling Timing, tDQSL(min)  
Version 1.0  
250  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Input Clock Frequency Stop and Change  
The input clock frequency can be stopped or changed from one stable clock rate to another stable clock rate during  
Write Leveling mode.  
The Frequency stop or change timing is shown in following figure.  
Clock Stop and Timing during Write Leveling  
Notes:  
1. CK is held LOW and CK is held HIGH during clock stop.  
2. CS shall be held LOW during clock clock stop.  
Write Leveling Timing Parameters  
Min/  
Parameter  
Symbol  
tWLDQSEN  
tWLWPRE  
tWLMRD  
tWLO  
Value  
Unit Notes  
Max  
Min  
Max  
Min  
Max  
Min  
Max  
Min  
Max  
Min  
Max  
Min  
Max  
Min  
Max  
20  
DQS/DQS delay  
tCK  
tCK  
tCK  
ns  
ns  
-
after write leveling mode is programmed  
-
20  
Write preamble for Write Leveling  
-
40  
First DQS/DQS edge  
after write leveling mode is programmed  
-
0
Write leveling output delay  
20  
max(14ns, 10nCK)  
Mode register set command delay  
Valid Clock Requirement before DQS Toggle  
Valid Clock Requirement after DQS Toggle  
tMRD  
-
max(7.5ns, 4nCK)  
tCKPRDQS  
tCKPSTDQS  
-
max(7.5ns, 4nCK)  
-
-
Version 1.0  
251  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write Leveling Setup and Hold Time  
Data Rate  
3733  
Min/  
Max  
Parameter  
Symbol  
Unit  
Write Leveling Parameters  
Write leveling hold time  
tWLH  
tWLS  
Min  
Min  
Min  
60  
60  
ps  
ps  
ps  
Write leveling setup time  
Write leveling input valid window  
tWLIVW  
100  
Notes:  
1. In addition to the traditional setup and hold time specifications above, there is value in a input valid window based specification for  
write-leveling training. As the training is based on each device, worst case process skews for setup and hold do not make sense to close  
timing between CK and DQS.  
2. tWLIVW is defined in a similar manner to tdIVW_Total, except that here it is a DQS input valid window with respect to CK. This would  
need to account for all VT (voltage and temperature) drift terms between CK and DQS within the DRAM that affect the write-leveling  
input valid window.  
The DQS input mask for timing with respect to CK is shown in following figure. The "total" mask (tWLIVW) defines  
the time the input signal must not encroach in order for the DQS input to be successfully captured by CK with a BER  
of lower than tbd. The mask is a receiver property and it is not the valid dataꢀeye.  
DQS/DQS and CK/CK at DRAM Latch  
DQS/DQS to CK/CK timings at the DRAM pins referenced from the internal latch  
Version 1.0  
252  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
RD DQ Calibration  
RD DQ Calibration for x16 mode  
LPDDR4 devices feature a RD DQ Calibration training function that outputs a 16ꢀbit userꢀdefined pattern on the DQ  
pins. RD DQ Calibration is initiated by issuing a MPCꢀ1 [RD DQ Calibration] command followed by a CASꢀ2  
command, cause the LPDDR4ꢀSDRAM to drive the contents of MR32 followed by the contents of MR40 on each of  
DQ[15:0] and DMI[1:0]. The pattern can be inverted on selected DQ pins according to userꢀdefined invert masks  
written to MR15 and MR20.  
RD DQ Calibration Training Procedure  
The procedure for executing RD DQ Calibration is:  
• Issue MRW commands to write MR32 (first eight bits), MR40 (second eight bits), MR15 (eightꢀbit invert mask  
for byte 0), and MR20 (eightꢀbit invert mask for byte 1).  
• Optionally this step could be skipped to use the default patterns  
ꢀ MR32 default = 5Ah  
ꢀ MR40 default = 3Ch  
ꢀ MR15 default = 55h  
ꢀ MR20 default = 55h  
•Issue an MPCꢀ1 [RD DQ Calibration] command followed immediately by a CASꢀ2 command.  
• Each time an MPCꢀ1 [RD DQ Calibration] command followed by a CASꢀ2 is received by the LPDDR4  
SDRAM, a 16ꢀbit data burst will, after the currently set RL, drive the eight bits programmed in MR32  
followed by the eight bits programmed in MR40 on all I/O pins.  
• The data pattern will be inverted for I/O pins with a '1' programmed in the corresponding invert mask mode  
register bit.  
• Note that the pattern is driven on the DMI pins, but no data bus inversion function is enabled, even if Read  
DBI is enabled in the DRAM mode register.  
• The MPCꢀ1 [RD DQ Calibration] command can be issued every tCCD seamlessly, and tRTRRD delay is  
required between Array Read command and the MPCꢀ1 [RD DQ Calibration] command as well the delay  
required between the MPCꢀ1 [RD DQ Calibration] command and an array read.  
• The operands received with the CASꢀ2 command must be driven LOW.  
• DQ Read Training can be performed with any or no banks active, during Refresh, or during SREF with CKE  
high.  
Invert Mask Assignments  
DQ Pin  
0
1
2
3
DMI0  
4
5
6
7
MR15 bit  
0
1
2
3
N/A  
4
5
6
7
DQ Pin  
8
9
10  
11  
DMI1  
12  
13  
14  
15  
MR20 bit  
0
1
2
3
N/A  
4
5
6
7
Version 1.0  
12/2018  
253  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ Read Training Timing: Read to Read DQ Calibration  
Notes:  
1. Read-1 to MPC [RD DQ Calibration] Operation is shown as an example of command-to-command timing.  
Timing from Read-1 to MPC [RD DQ Calibration] command is tRTRRD.  
2. MPC [RD DQ Calibration] uses the same command-to-data timing relationship (RL, tDQSCK, tDQSQ) as a Read-1 command.  
3. BL = 16, Read Preamble: Toggle, Read Postamble: 0.5nCK.  
4. DES commands are shown for ease of illustration; other commands may be valid at these times.  
DQ Read Training Timing: Read DQ Cal. to Read DQ Cal. / Read  
Notes:  
1. MPC [RD DQ Calibration] to MPC [RD DQ Calibration] Operation is shown as an example of command-to-command timing.  
2. MPC [RD DQ Calibration] to Read-1 Operation is shown as an example of command-to-command timing.  
3. MPC [RD DQ Calibration] uses the same command-to-data timing relationship (RL, tDQSCK, tDQSQ) as a Read-1 command.  
4. Seamless MPC [RD DQ Calibration] commands may be executed by repeating the command every tCCD time.  
5. Timing from MPC [RD DQ Calibration] command to Read-1 is tRTRRD.  
6. BL = 16, Read Preamble: Toggle, Read Postamble: 0.5nCK.  
7. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
254  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ Read Training Example  
An example of DQ Read Training output is shown in table below. This shows the 16ꢀbit data pattern that will be  
driven on each DQ in byte 0 when one DQ Read Training command is executed. This output assumes the following  
mode register values are used:  
ꢀ MR32 = 1CH  
ꢀ MR40 = 59H  
ꢀ MR15 = 55H  
ꢀ MR20 = 55H  
DQ Read Calibration Bit Ordering and Inversion Example  
Bit Sequence →  
Pin  
Invert  
15 14 13 12 11 10  
9
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
8
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
7
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
6
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
5
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
4
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
3
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
2
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
1
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
0
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
DQ0  
DQ1  
Yes  
No  
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
0
1
0
1
1
0
1
0
1
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
1
0
DQ2  
Yes  
No  
DQ3  
DMI0  
DQ4  
Never  
Yes  
No  
DQ5  
DQ6  
Yes  
No  
DQ7  
DQ8  
Yes  
No  
DQ9  
DQ10  
DQ11  
DMI1  
DQ12  
DQ13  
DQ14  
DQ15  
Yes  
No  
Never  
Yes  
No  
Yes  
No  
Notes:  
1. The patterns contained in MR32 and MR40 are transmitted on DQ[15:0] and DMI[1:0] when RD DQ Calibration is initiated via a MPC-1  
[RD DQ Calibration] command. The pattern transmitted serially on each data lane, organized "little endian" such that the low-order bit  
in a byte is transmitted first. If the data pattern is 27H, then the first bit transmitted with be a '1', followed by '1', '1', '0', '0', '1', '0', and  
'0'. The bit stream will be 00100111 →.  
2. MR15 and MR20 may be used to invert the MR32/MR40 data pattern on the DQ pins. See MR15 and MR20 for more information. Data is  
never inverted on the DMI[1:0] pins.  
3. DMI [1:0] outputs status follows table.  
Version 1.0  
255  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MR Setting vs. DMI Status  
DM Function  
MR13 OP[5]  
1: Disable  
1: Disable  
1: Disable  
1: Disable  
0: Enable  
Write DBIdc Function  
MR3 OP[7]  
0: Disable  
Read DBIdc Function  
DMI Status  
Hi-Z  
MR3 OP[6]  
0: Disable  
0: Disable  
1: Enable  
1: Enable  
0: Disable  
0: Disable  
1: Enable  
1: Enable  
1: Enable  
The data pattern is transmitted  
The data pattern is transmitted  
The data pattern is transmitted  
The data pattern is transmitted  
The data pattern is transmitted  
The data pattern is transmitted  
The data pattern is transmitted  
0: Disable  
1: Enable  
0: Disable  
0: Enable  
1: Enable  
0: Enable  
0: Disable  
0: Enable  
1: Enable  
4. No Data Bus Inversion (DBI) function is enacted during RD DQ Calibration, even if DBI is enabled in MR3-OP[6].  
MPC of Read DQ Calibration after PowerꢀDown Exit  
Following the powerꢀdown state, an additional time, tMRRI, is required prior to issuing the MPC of Read DQ  
Calibration command. This additional time (equivalent to tRCD) is required in order to be able to maximize  
powerꢀdown current savings by allowing more powerꢀup time for the Read DQ data in MR32 and MR40 data path  
after exit from standby, powerꢀdown mode.  
MPC Read DQ Calibration Following PowerꢀDown State  
Version 1.0  
256  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
RD DQ Calibration for Byte (x8) mode  
LPDDR4 devices feature a RD DQ Calibration training function that outputs a 8ꢀbit userꢀdefined pattern on  
the DQ pins. RD DQ Calibration is initiated by issuing a MPCꢀ1 [RD DQ Calibration] command followed by  
a CASꢀ2 command, cause the LPDDR4ꢀSDRAM to drive the contents of MR32 followed by the contents of  
MR40 on each of DQ[7:0] and DMI[0]. The pattern can be inverted on selected DQ pins according to userdefined  
invert masks written to MR15 and MR20.  
RD DQ Calibration Training Procedure  
Issue MRW commands to write MR32 (first eight bits), MR40 (second eight bits),  
MR15 (eightꢀbit invert mask for byte 0 : DQ[7:0] ) and MR20 (eightꢀbit invert mask for byte 1 : DQ[15:8] )  
Optionally this step could be skipped to use the default patterns  
ꢀ MR32 default = 5Ah  
ꢀ MR40 default = 3Ch  
ꢀ MR15 default = 55h  
ꢀ MR20 default = 55h  
Issue an MPCꢀ1 [RD DQ Calibration] command followed immediately by a CASꢀ2 command  
Each time an MPCꢀ1 [RD DQ Calibration] command followed by a CASꢀ2 is received by the  
LPDDR4 SDRAM, a 16ꢀbit data burst will, after the currently set RL, drive the eight bits programmed  
in MR32 followed by the eight bits programmed in MR40 on all I/O pins.  
The data pattern will be inverted for I/O pins with a ‘1’ programmed in the corresponding invert mask  
mode register bit (see Table ‘Invert Mask Assignments’).  
Note that the pattern is driven on the DMI pins, but no data bus inversion function is enabled, even if  
Read DBI is enabled in the DRAM mode register.  
This command can be issued every tCCD seamlessly, and can be issued seamlessly with array  
Read commands.  
The operands received with the CASꢀ2 command must be driven LOW.  
DQ Read Training can be performed with any or no banks active, during Refresh, or during SREF with  
CKE high.  
Invert Mask Assignments  
DQ Pin  
0
1
2
3
DMI0  
4
5
6
7
MR 15 bit  
0
1
2
3
N/A  
4
5
6
7
DQ Pin  
8
9
10  
11 DMI1 12  
N/A  
13  
14  
15  
MR 20 bit  
0
1
2
3
4
5
6
7
Version 1.0  
12/2018  
257  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ Read Training Example  
An example of DQ Read Training output is shown in table below. This shows the 16ꢀbit data pattern that will be  
driven on each DQ in byte 0 when one DQ Read Training command is executed. This output assumes the following  
mode register values are used:  
ꢀ MR32 = 1CH  
ꢀ MR40 = 59H  
ꢀ MR15 = 55H  
ꢀ MR20 = 55H  
DQ Read Calibration Bit Ordering and Inversion Example  
Bit Sequence →  
Pin  
Invert  
Yes  
15 14 13 12 11 10  
9
8
7
6
5
4
3
2
1
0
DQ0  
(DQ8)  
DQ1  
1
0
1
0
1
0
1
0
1
0
1
0
0
1
0
1
0
1
1
0
1
1
1
0
0
0
1
1
No  
0
1
1
0
0
1
0
1
0
1
1
0
1
0
1
0
0
1
0
1
(DQ9)  
DQ2  
Yes  
(DQ10)  
DQ3  
No  
Never  
Yes  
0
0
1
1
1
0
0
0
1
1
1
0
1
1
0
0
0
1
0
0
1
1
1
0
0
0
1
0
0
1
0
0
1
1
1
0
1
1
0
1
1
0
0
0
1
0
0
1
(DQ11)  
DMI0  
DQ4  
(DQ12)  
DQ5  
No  
Yes  
No  
0
1
0
1
0
1
0
1
0
1
0
1
1
0
1
0
1
0
0
1
0
1
0
1
0
1
0
0
1
0
0
1
0
1
0
1
1
0
1
1
0
1
0
1
0
0
1
0
(DQ13)  
DQ6  
(DQ14)  
DQ7  
(DQ15)  
Notes:  
1. The patterns contained in MR32 and MR40 are transmitted on lower byte select : DQ[7:0] or upper byte select : DQ[15:8], DMI[0] or  
DMI[1] when RD DQ Calibration is initiated via a MPC-1 [RD DQ Calibration] command. The pattern transmitted serially on each data  
lane, organized “little endian” such that the low-order bit in a byte is transmitted first. If the data pattern is 27H, then the first bit  
transmitted with be a ‘1’, followed by ‘1’, ‘1’, ‘0’, ‘0’, ‘1’, ‘0’, and ‘0’. The bit stream will be 00100111.  
2. MR15 and MR20 may be used to invert the MR32/MR40 data pattern on the DQ pins. See MR15 for more information. Data is never  
inverted on the DMI[0] pins.  
3. The data pattern is not transmitted on the DMI[0] or DMI[1] pins if DBI-RD is disabled via MR3 OP[6].  
4. No Data Bus Inversion (DBI) function is enacted during RD DQ Calibration, even if DBI is enabled in MR3 OP[6].  
Version 1.0  
258  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQSꢀDQ Training  
The LPDDR4ꢀSDRAM uses an unꢀmatched DQSꢀDQ path to enable high speed performance and save power in the  
DRAM. As a result, the DQS strobe must be trained to arrive at the DQ latch centerꢀaligned with the Data eye. The  
SDRAM DQ receiver is located at the DQ pad, and has a shorter internal delay in the SDRAM than does the DQS  
signal. The SDRAM DQ receiver will latch the data present on the DQ bus when DQS reaches the latch, and training  
is accomplished by delaying the DQ signals relative to DQS such that the Data eye arrives at the receiver latch  
centered on the DQS transition.  
Two modes of training are available in LPDDR4:  
• Commandꢀbased FIFO WR/RD with user patterns  
• A internal DQS clockꢀtree oscillator, to determine the need for, and the magnitude of required training.  
The commandꢀbased FIFO WR/RD uses the MPC command with operands to enable this special mode of operation.  
When issuing the MPC command, if OP6 is set LOW then the DRAM will perform a NOP command. When OP6 is  
set HIGH, then OP5:0 enable training functions or are reserved for future use (RFU). MPC commands that initiate a  
Read FIFO, READ DQ Calibration or Write FIFO to the SDRAM must be followed immediately by a CASꢀ2  
command. See "Multi Purpose Command (MPC) Definition" for more information.  
To perform Write Training, the controller can issue a MPC [Write DQ FIFO] command with OP[6:0] set as described  
in the MPC Definition section, followed immediately by a CASꢀ2 command (CASꢀ2 operands should be driven LOW)  
to initiate a Write DQ FIFO. Timings for MPC [Write DQ FIFO] are identical to a Write command, with WL (Write  
Latency) timed from the 2nd rising clock edge of the CASꢀ2 command. Up to 5 consecutive MPC [Write DQ FIFO]  
commands with user defined patterns may be issued to the SDRAM to store up to 80 values (BL16 x5) per pin that  
can be read back via the MPC [Read DQ FIFO] command. Write/Read FIFO Pointer operation is described later in  
this section.  
After writing data to the SDRAM with the MPC [Write DQ FIFO] command, the data can be read back with the MPC  
[Read DQ FIFO] command and results compared with “expect” data to see if further training (DQ delay) is needed.  
MPC [Read DQ FIFO] is initiated by issuing a MPC command with OP[6:0] set as described in the MPC Definition  
section, followed immediately by a CASꢀ2 command (CASꢀ2 operands must be driven LOW). Timings for the MPC  
[Read DQ FIFO] command are identical to a Read command, with RL (Read Latency) timed from the 2nd rising  
clock edge of the CASꢀ2 command.  
Read DQ FIFO is nonꢀdestructive to the data captured in the FIFO, so data may be read continuously until it is either  
overwritten by a Write DQ FIFO command or disturbed by CKE LOW or any of the following commands; Write,  
Masked Write, Read, Read DQ Calibration and a MRR. If fewer than 5 Write DQ FIFO commands were executed,  
then unwritten registers will have unꢀdefined (but valid) data when read back.  
Version 1.0  
259  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The following command about MRW is only allowed from MPC [Write DQ FIFO] command to MPC [Read DQ FIFO].  
Allowing MRW command is for OP[7]:FSPꢀOP, OP[6]:FSPꢀWR and OP[3]:VRCG of MR13 and MR14. And the rest  
of MRW command is prohibited.  
For example: If 5 Write DQ FIFO commands are executed sequentially, then a series of Read DQ FIFO commands  
will read valid data from FIFO[0], FIFO[1]ꢃ.FIFO[4], and will then wrap back to FIFO[0] on the next Read DQ FIFO.  
On the other hand, if fewer than 5 Write DQ FIFO commands are executed sequentially (example=3), then a series  
of Read DQ FIFO commands will return valid data for FIFO[0], FIFO[1], and FIFO[2], but the next two Read DQ  
FIFO commands will return unꢀdefined data for FIFO[3] and FIFO[4] before wrapping back to the valid data in  
FIFO[0].  
FIFO Pointer Reset and Synchronism  
The Read and Write DQ FIFO pointers are reset under the following conditions:  
• Powerꢀup initialization  
RESET asserted  
• Powerꢀdown entry  
• Self Refresh PowerꢀDown entry  
The MPC [Write DQ FIFO] command advances the WRꢀFIFO pointer, and the MPC [Read DQ FIFO] advances the  
RDꢀFIFO pointer. Also any normal (nonꢀFIFO) Read Operation (RD, RDA) advances both  
WRꢀFIFO pointer and RDꢀFIFO pointer. Issuing (nonꢀFIFO) Read Operation command is inhibited during Write  
training period. To keep the pointers aligned, the SoC memory controller must adhere to the following restriction at  
the end of Write training period:  
b = a + (n x c)  
Where:  
‘a’ is the number of MPC [Write DQ FIFO] commands  
‘b’ is the number of MPC [Read DQ FIFO] commands  
‘c’ is the FIFO depth (=5 for LPDDR4)  
‘n’ is a positive integer, 0  
Version 1.0  
260  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write to MPC [Write FIFO] Operation Timing  
Notes:  
1. MPC [WR FIFO] can be executed with a single bank or multiple banks active, during Refresh, or during SREF with CKE HIGH.  
2. Write-1 to MPC is shown as an example of command-to-command timing for MPC. Timing from Write-1 to MPC [WR-FIFO] is tWRWTR.  
3. Seamless MPC [WR-FIFO] commands may be executed by repeating the command every tCCD time.  
4. MPC [WR-FIFO] uses the same command-to-data timing relationship (WL, tDQSS, tDQS2DQ) as a Write-1 command.  
5. A maximum of 5 MPC [WR-FIFO] commands may be executed consecutively without corrupting FIFO data.  
The 6th MPC [WR-FIFO] command will overwrite the FIFO data from the first command. If fewer than 5 MPC [WR-FIFO] commands are  
executed, then the remaining FIFO locations will contain undefined data.  
6. For the CAS-2 command following a MPC command, the CAS-2 operands must be driven “LOW.”  
7. To avoid corrupting the FIFO contents, MPC [RD-FIFO] must immediately follow MPC [WR-FIFO]/CAS-2 without any other command  
disturbing  
FIFO pointers in-between. FIFO pointers are disturbed by CKE Low, Write, Masked Write, Read, Read DQ Calibration and MRR.  
8. BL = 16, Write Postamble = 0.5nCK  
9. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
261  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MPC [Write FIFO] to MPC [Read FIFO] Timing  
Notes:  
1. MPC [WR FIFO] can be executed with a single bank or multiple banks active, during Refresh, or during SREF with CKE HIGH.  
2. MPC [WR-FIFO] to MPC [RD-FIFO] is shown as an example of command-to-command timing for MPC.  
Timing from MPC [WR-FIFO] to MPC [RD-FIFO] is specified in the command-to-command timing table.  
3. Seamless MPC [RD-FIFO] commands may be executed by repeating the command every tCCD time.  
4. MPC [RD-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK, tDQSQ) as a Read-1 command.  
5. Data may be continuously read from the FIFO without any data corruption. After 5 MPC [RD-FIFO] commands the FIFO pointer will wrap  
back to the 1st FIFO and continue advancing. If fewer than 5 MPC [WR-FIFO] commands were executed, then the MPC [RD-FIFO]  
commands to those FIFO locations will return undefined data. See the Write Training section for more information on the FIFO pointer  
behavior.  
6. For the CAS-2 command immediately following a MPC command, the CAS-2 operands must be driven “LOW.”  
7. DMI[1:0] signals will be driven if any of WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training section for more  
information on DMI behavior.  
8. BL = 16, Write Postamble = 0.5nCK, Read Preamble: Toggle, Read Postamble: 0.5nCK  
9. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
262  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MPC [Read FIFO] to Read Timing  
Notes:  
1. MPC [WR FIFO] can be executed with a single bank or multiple banks active, during Refresh, or during SREF with CKE HIGH.  
2. MPC [RD-FIFO] to Read-1 Operation is shown as an example of command-to-command timing for MPC. Timing from MPC [RD-FIFO]  
command to Read is tRTRRD.  
3. Seamless MPC [RD-FIFO] commands may be executed by repeating the command every tCCD time.  
4. MPC [RD-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK, tDQSQ) as a Read-1 command.  
5. Data may be continuously read from the FIFO without any data corruption. After 5 MPC [RD-FIFO] commands the FIFO pointer will wrap  
back to the 1st FIFO and continue advancing. If fewer than 5 MPC [WR-FIFO] commands were executed, then the MPC [RD-FIFO]  
commands to those FIFO locations will return undefined data. See the Write Training section for more information on the FIFO pointer  
behavior.  
6. For the CAS-2 command immediately following a MPC command, the CAS-2 operands must be driven “LOW.”  
7. DMI[1:0] signals will be driven if any of WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training section for more  
information on DMI behavior.  
8. BL = 16, Read Preamble: Toggle, Read Postamble: 0.5nCK  
9. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
263  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MPC [Write FIFO] with DQ ODT Timing  
Notes:  
1. MPC [WR FIFO] can be executed with a single bank or multiple banks active, during Refresh, or during SREF with CKE HIGH.  
2. MPC [WR-FIFO] uses the same command-to-data/ODT timing relationship (WL, tDQSS, tDQS2DQ, ODTLon, ODTLoff, tODTon, tODToff) as  
a Write-1 command.  
3. For the CAS-2 command immediately following a MPC command, the CAS-2 operands must be driven “LOW.”  
4. BL = 16, Write Postamble = 0.5nCK  
5. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
264  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Power Down Exit to MPC [Write FIFO] Timing  
Notes:  
1. Any commands except MPC WR FIFO and other exception commands defined other section in this document (i.e. MPC Read DQ Cal).  
2. DES commands are shown for ease of illustration; other commands may be valid at these times.  
MPC [Write FIFO] AC Timing  
Min/  
Parameter  
Symbol  
Data Rate  
Unit Notes  
Max  
MPC Write FIFO Timing  
Additional time after tXP has expired until  
MPC [Write FIFO] command may be issued  
tMPCWR  
Min  
tRCD + 3nCK  
Version 1.0  
265  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQS Interval Oscillator  
As voltage and temperature change on the SDRAM die, the DQS clock tree delay will shift and may require  
reꢀtraining. The LPDDR4ꢀSDRAM includes an internal DQS clockꢀtree oscillator to measure the amount of delay  
over a given time interval (determined by the controller), allowing the controller to compare the trained delay value to  
the delay value seen at a later time. The DQS Oscillator will provide the controller with important information  
regarding the need to reꢀtrain, and the magnitude of potential error.  
The DQS Interval Oscillator is started by issuing a MPC [Start DQS Osc] command with OP[6:0] set as described in  
the MPC Operation section, which will start an internal ring oscillator that counts the number of time a signal  
propagates through a copy of the DQS clock tree.  
The DQS Oscillator may be stopped by issuing a MPC [Stop DQS Osc] command with OP[6:0] set as described in  
the MPC Operation section, or the controller may instruct the SDRAM to count for a specific number of clocks and  
then stop automatically (See MR23 for more information). If MR23 is set to automatically stop the DQS Oscillator,  
then the MPC [Stop DQS Osc] command should not be used (illegal). When the DQS Oscillator is stopped by either  
method, the result of the oscillator counter is automatically stored in MR18 and MR19.  
The controller may adjust the accuracy of the result by running the DQS Interval Oscillator for shorter (less accurate)  
or longer (more accurate) duration. The accuracy of the result for a given temperature and voltage is determined by  
the following equation:  
2 * (DQS delay)  
=
DQS Oscillator Granularity Error  
Run Time  
Where:  
Run Time = total time between start and stop commands  
DQS delay = the value of the DQS clock tree delay (tDQS2DQ min/max)  
Additional matching error must be included, which is the difference between DQS training circuit and the actual DQS  
clock tree across voltage and temperature. The matching error is vendor specific.  
Therefore, the total accuracy of the DQS Oscillator counter is given by:  
DQS Oscillator Accuracy = 1 - Granularity Error - Matching Error  
Version 1.0  
266  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Example: If the total time between start and stop commands is 100ns, and the maximum DQS clock tree delay is  
800ps (tDQS2DQ max), then the DQS Oscillator Granularity Error is:  
2 * (0.8ns)  
100ns  
DQS Oscillator Granularity Error =  
= 1.6%  
This equates to a granularity timing error of 12.8ps.  
Assuming a circuit Matching Error of 5.5ps across voltage and temperature, then the accuracy is:  
12.8 + 5.5  
DQS Oscillator Accuracy = 1-  
= 97.7%  
800  
Example: Running the DQS Oscillator for a longer period improves the accuracy. If the total time between start and  
stop commands is 500ns, and the maximum DQS clock tree delay is 800ps (tDQS2DQ max), then the DQS  
Oscillator Granularity Error is:  
2 * (0.8ns)  
500ns  
DQS Oscillator Granularity Error =  
= 0.32%  
This equates to a granularity timing error or 2.56ps.  
Assuming a circuit Matching Error of 5.5ps across voltage and temperature, then the accuracy is:  
2.56 + 5.5  
DQS Oscillator Accuracy = 1-  
= 99.0%  
800  
The result of the DQS Interval Oscillator is defined as the number of DQS Clock Tree Delays that can be counted  
within the “run time,” determined by the controller. The result is stored in MR18ꢀOP[7:0] and MR19ꢀOP[7:0]. MR18  
contains the least significant bits (LSB) of the result, and MR19 contains the most significant bits (MSB) of the result.  
MR18 and MR19 are overwritten by the SDRAM when a MPCꢀ1 [Stop DQS Osc] command is received. The SDRAM  
counter will count to its maximum value (=2^16) and stop. If the maximum value is read from the mode registers,  
then the memory controller must assume that the counter overflowed the register and discard the result. The longest  
“run time” for the oscillator that will not overflow the counter registers can be calculated as follows:  
Longest Run Time Interval = 216 * tDQS2DQ(min) = 216 * 0.2ns = 13.1us  
Version 1.0  
267  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Interval Oscillator matching error  
The interval oscillator matching error is defined as the difference between the DQS training ckt(interval oscillator)  
and the actual DQS clock tree across voltage and temperature.  
• Parameters:  
ꢀ tDQS2DQ: Actual DQS clock tree delay  
ꢀ tDQSOSC: Training ckt(interval oscillator) delay  
ꢀ OSCOffset: Average delay difference over voltage and temp.  
ꢀ OSCMatch: DQS oscillator matching error  
Interval oscillator offset OSCoffset  
• OSCMatch  
:
OSC Match = [ tDQS2DQ(V,T) – tDQSOSC(V,T) – OSCoffset  
]
• tDQSOSC  
:
Runtime  
tDQS OSC(V,T)  
=
2 * Count  
Version 1.0  
12/2018  
268  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQS Oscillator Matching Error Specification  
Parameter  
Symbol  
OSC  
Min  
-20  
Max  
20  
Unit  
Notes  
1,2,3,4,5,6,7  
2,4,7  
DQS Oscillator Matching Error  
DQS Oscillator Offset  
ps  
ps  
Match  
-100  
100  
OSC  
offset  
Notes:  
1. The OSCMatch is the matching error per between the actual DQS and DQS interval oscillator over voltage and temp.  
2. This parameter will be characterized or guaranteed by design.  
3. The OSCMatch is defined as the following:  
OSC Match = [ tDQS2DQ(V,T) – tDQSOSC(V,T) – OSCoffset  
]
Where tDQS2DQ(V,T) and tDQSOSC(V,T) are determined over the same voltage and temp conditions.  
4. The runtime of the oscillator must be at least 200ns for determining tDQSOSC(V,T)  
Runtime  
tDQS OSC(V,T)  
=
2 * Count  
5. The input stimulus for tDQS2DQ will be consistent over voltage and temp conditions.  
6. The OSCoffset is the average difference of the endpoints across voltage and temp.  
7. These parameters are defined per channel.  
8. tDQS2DQ(V,T) delay will be the average of DQS to DQ delay over the runtime period.  
Version 1.0  
269  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQS Interval Oscillator Readout Timing  
OSC Stop to its counting value readout timing is shown in figure below.  
In case of DQS Interval Oscillator is stopped by MPC Command  
Notes:  
1. DQS interval timer run time setting : MR23 OP[7:0] = 00000000  
2. DES commands are shown for ease of illustration; other commands may be valid at these times.  
In case of DQS Interval Oscillator is stopped by DQS interval timer  
Notes:  
1. DQS interval timer run time setting : MR23 OP[7:0]00000000  
2. Setting counts of MR23  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
DQS Interval Oscillator AC Timing  
Min/  
Parameter  
Symbol  
Value  
Unit Notes  
Max  
Delay time from OSC stop to Mode Register Readout  
Notes:  
tOSCO  
Min  
Max(40ns,8nCK)  
ns  
1. Start DQS OSC command is prohibited until tOSCO(Min) is satisfied.  
Version 1.0  
270  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
READ Preamble Training  
LPDDR4 READ Preamble Training is supported through the MPC function.  
This mode can be used to train or read level the DQS receivers. Once READ Preamble Training is enabled by  
MR13[OP1] = 1, the LPDDR4 DRAM will drive DQS LOW, DQS HIGH within tSDO and remain at these levels until  
an MPC DQ READ Calibration command is issued.  
During READ Preamble Training the DQS preamble provided during normal operation will not be driven by the  
DRAM. Once the MPC DQ READ Calibration command is issued, the DRAM will drive DQS/DQS and DQ like a  
normal READ burst after RL and tDQSCK. Prior to the MPC DQ READ Calibration command, the DRAM may or  
may not drive DQ[15:0] in this mode.  
While in READ Preamble Training Mode, only READ DQ Calibration commands may be issued.  
•Issue an MPC [RD DQ Calibration] command followed immediately by a CASꢀ2 command.  
• Each time an MPC [RD DQ Calibration] command followed by a CASꢀ2 is received by the LPDDR4 SDRAM,  
a 16ꢀbit data burst will, after the currently set RL, drive the eight bits programmed in MR32 followed by the  
eight bits programmed in MR40 on all I/O pins.  
• The data pattern will be inverted for I/O pins with a '1' programmed in the corresponding invert mask mode  
register bit.  
• Note that the pattern is driven on the DMI pins, but no data bus inversion function is enabled, even if Read  
DBI is enabled in the DRAM mode register.  
• This command can be issued every tCCD seamlessly.  
• The operands received with the CASꢀ2 command must be driven LOW.  
READ Preamble Training is exited within tSDO after setting MR13[OP1] = 0.  
LPDDR4 supports the READ Preamble Training as optional feature. Refer to vendor specific datasheets.  
Read Preamble Training  
Notes:  
1. Read DQ Calibration supports only BL16 operation  
Timing Parameters  
Parameter  
Symbol  
Min  
Max  
Unit Notes  
Delay from MRW command to DQS Driven Out  
tSDO  
-
Max(12nCK, 20ns)  
-
Version 1.0  
12/2018  
271  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MultiꢀPurpose Command (MPC)  
LPDDR4ꢀSDRAMs use the MPC command to issue a NOP and to access various training modes. The MPC  
command is initiated with CS, and CA[5:0] asserted to the proper state at the rising edge of CK, as defined by the  
Command Truth Table. The MPC command has seven operands (OP[6:0]) that are decoded to execute specific  
commands in the SDRAM. OP[6] is a special bit that is decoded on the first rising CK edge of the MPC command.  
When OP[6]=0 then the SDRAM executes a NOP (no operation) command, and when OP[6]=1 then the SDRAM  
further decodes one of several training commands.  
When OP[6]=1 and when the training command includes a Read or Write operation, the MPC command must be  
followed immediately by a CASꢀ2 command. For training commands that Read or Write the SDRAM, read latency  
(RL) and write latency (WL) are counted from the second rising CK edge of the CASꢀ2 command with the same  
timing relationship as any normal Read or Write command. The operands of the CASꢀ2 command following a MPC  
Read/Write command must be driven LOW.  
The following MPC commands must be followed by a CASꢀ2 command:  
ꢀ Write FIFO  
ꢀ Read FIFO  
ꢀ Read DQ Calibration  
All other MPCꢀ1 commands do not require a CASꢀ2 command, including:  
ꢀ NOP  
ꢀ Start DQS Interval Oscillator  
ꢀ Stop DQS Interval Oscillator  
ꢀ Start ZQ Calibration  
ꢀ Latch ZQ Calibration  
Version 1.0  
272  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MPC Command Definition  
SDR Command Pins  
CKE  
SDR CA Pins  
SDRAM  
CK  
Notes  
EDGE  
Command  
CS  
CA0  
CA1  
CA2  
CA3  
CA4  
CA5  
CK(n-1)  
CK(n)  
H
L
L
L
L
L
L
OP6  
OP5  
R1  
MPC  
H
H
1, 2  
R2  
(Train, NOP)  
OP0  
OP1  
OP2  
OP3  
OP4  
MPC Command Definition for OP[6:0]  
Function  
Operand  
Data  
Notes  
0XXXXXX : NOP  
B
1000001 : RD FIFO: RD FIFO supports only BL16 operation  
B
1000011 : RD DQ Calibration (MR32/MR40)  
B
1000101 : RFU  
B
1000111 : WR FIFO: WR FIFO supports only BL16 operation  
B
Training Modes  
OP[6:0]  
1, 2, 3  
1001001 : RFU  
B
1001011 : Start DQS Osc  
B
1001101 : Stop DQS Osc  
B
1001111 : ZQCal Start  
B
1010001 : ZQCal Latch  
B
All Others: Reserved  
Notes:  
1. See command truth table for more information.  
2. MPC commands for Read or Write training operations must be immediately followed by CAS-2 command consecutively without any  
other commands in-between. MPC command must be issued first before issuing the CAS-2 command.  
3. Write FIFO and Read FIFO commands will only operate as BL16, ignoring the burst length selected by MR1 OP[1:0].  
Version 1.0  
273  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MPC [WRITE FIFO] Operation : tWPRE=2nCK, tWPST=0.5nCK  
Notes:  
1. MPC [WR FIFO] can be executed with a single bank or multiple banks active, during Refresh, or during SREF with CKE HIGH.  
2. Write-1 to MPC is shown as an example of command-to-command timing for MPC. Timing from Write-1 to MPC [WR-FIFO] is tWRWTR.  
3. Seamless MPC [WR-FIFO] commands may be executed by repeating the command every tCCD time.  
4. MPC [WR-FIFO] uses the same command-to-data timing relationship (WL, tDQSS, tDQS2DQ) as a Write-1 command.  
5. A maximum of 5 MPC [WR-FIFO] commands may be executed consecutively without corrupting FIFO data.  
The 6th MPC [WR-FIFO] command will overwrite the FIFO data from the first command. If fewer than 5 MPC [WR-FIFO] commands  
are executed, then the remaining FIFO locations will contain undefined data.  
6. For the CAS-2 command following a MPC command, the CAS-2 operands must be driven “LOW.”  
7. To avoid corrupting the FIFO contents, MPC-1 [RD-FIFO] must immediately follow MPC-1 [WR-FIFO]/CAS-2 without any other command  
disturbing FIFO pointers in-between. FIFO pointers are disturbed by CKE Low, Write, Masked Write, Read, Read DQ Calibration and MRR.  
See Write Training session for more information on FIFO pointer behavior.  
MPC [RD FIFO] Read Operation :  
tWPRE=2nCK, tWPST=0.5nCK, tRPRE=toggling, tRPST=1.5nCK  
Notes:  
1. MPC [WR FIFO] can be executed with a single bank or multiple banks active, during Refresh, or during SREF with CKE HIGH.  
2. Write-1 to MPC is shown as an example of command-to-command timing for MPC. Timing from Write-1 to MPC-1 [WR-FIFO] is tWRWTR.  
3. Seamless MPC [RD-FIFO] commands may be executed by repeating the command every tCCD time.  
4. MPC [RD-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK) as a Read-1 command.  
5. Data may be continuously read from the FIFO without any data corruption. After 5 MPC [RD-FIFO] commands the FIFO pointer will wrap  
back to the 1st FIFO and continue advancing. If fewer than 5 MPC [WR-FIFO] commands were executed, then the MPC [RD-FIFO]  
commands to those FIFO locations will return undefined data. See the Write Training section for more information on the FIFO pointer  
behavior.  
6. For the CAS-2 command immediately following a MPC command, the CAS-2 operands must be driven “LOW.”  
7. DMI[1:0] signals will be driven if any of WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training section for more  
information on DMI behavior.  
Version 1.0  
274  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MPC [RD FIFO] Operation : tRPRE=toggling, tRPST=1.5nCK  
Notes:  
1. MPC [WR FIFO] can be executed with a single bank or multiple banks active, during Refresh, or during SREF with CKE HIGH.  
2. MPC [RD-FIFO] to Read-1 Operation is shown as an example of command-to-command timing for MPC. Timing from MPC-1 [RD-FIFO]  
command to Read is tRTRRD.  
3. Seamless MPC [RD-FIFO] commands may be executed by repeating the command every tCCD time.  
4. MPC [RD-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK) as a Read-1 command.  
5. Data may be continuously read from the FIFO without any data corruption. After 5 MPC [RD-FIFO] commands the FIFO pointer will wrap  
back tothe 1st FIFO and continue advancing. If fewer than 5 MPC [WR-FIFO] commands were executed, then the MPC [RD-FIFO]  
commands to thoseFIFO locations will return undefined data. See the Write Training section for more information on the FIFO pointer  
behavior.  
6. For the CAS-2 command immediately following a MPC command, the CAS-2 operands must be driven “LOW.”  
7. DMI[1:0] signals will be driven if any of WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training section for more  
information on DMI behavior.  
Version 1.0  
275  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Timing Constraints for Training Commands  
Previous  
Next Command  
Command  
Minimum Delay  
Unit Notes  
MPC [WR FIFO]  
tWRWTR  
nCK  
-
1
2
MPC [RD FIFO]  
Not Allowed  
WR/MWR  
MPC  
WL+RU(tDQSS(max)/tCK) +BL/2+RU(tWTR/tCK)  
nCK  
[RD DQ Calibration]  
MPC [WR FIFO]  
tRTW  
nCK  
-
3
2
MPC [RD FIFO]  
Not Allowed  
RD/MRR  
MPC  
tRTRRD  
nCK  
3
2
[RD DQ Calibration]  
WR/MWR  
Not Allowed  
-
MPC [WR FIFO]  
tCCD  
Not Allowed  
nCK  
-
MPC  
RD/MRR  
MPC [RD FIFO]  
MPC  
2
[WR FIFO]  
WL+RU(tDQSS(max)/tCK) +BL/2+RU(tWTR/tCK)  
nCK  
Not Allowed  
-
2
[RD DQ Calibration]  
WR/MWR  
tRTW  
tRTW  
nCK  
nCK  
nCK  
nCK  
3
4
3
MPC [WR FIFO]  
RD/MRR  
MPC  
tRTRRD  
tCCD  
[RD FIFO]  
MPC [RD FIFO]  
MPC  
tRTRRD  
nCK  
3
[RD DQ Calibration]  
WR/MWR  
tRTW  
tRTW  
nCK  
nCK  
nCK  
-
4
4
3
2
MPC [WR FIFO]  
RD/MRR  
MPC  
tRTRRD  
[RD DQ Calibration]  
MPC [RD FIFO]  
MPC  
Not Allowed  
tCCD  
nCK  
[RD DQ Calibration]  
Notes:  
1. tWRWTR = WL + BL/2 + RU(tDQSS(max)/tCK) + max(RU(7.5ns/tCK),8nCK)  
2. No commands are allowed between MPC [WR FIFO] and MPC-1 [RD FIFO] except MRW commands related to training parameters.  
3. tRTRRD = RL + RU(tDQSCK(max)/tCK) + BL/2 + RD(tRPST) + max(RU(7.5ns/tCK),8nCK)  
4. tRTW : In Case of DQ ODT Disable MR11 OP[2:0] = 000B:  
RL+RU(tDQSCK(max)/tCK) + BL/2 - WL+tWPRE + RD(tRPST)  
In Case of DQ ODT Enable MR11 OP[2:0] 000B:  
RL + RU(tDQSCK(max)/tCK) + BL/2 + RD(tRPST) - ODTLon - RD(tODTon,min/tCK) + 1  
Version 1.0  
276  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Thermal Offset  
Because of their tight thermal coupling with the LPDDR4 device, hot spots on an SOC can induce thermal gradients  
across the LPDDR4 device. As these hot spots may not be located near the device thermal sensor, the devices’  
temperature compensated selfꢀrefresh circuit may not generate enough refresh cycles to guarantee memory  
retention. To address this shortcoming, the controller can provide a thermal offset that the memory uses to adjust its  
TCSR circuit to ensure reliable operation.  
This offset is provided through MR4(6:5) to either or to both the channels. This temperature offset may modify  
refresh behavior for the channel to which the offset is provided. It will take a max of 200us to have the change  
reflected in MR4(2:0) for the channel to which the offset is provided. If the induced thermal gradient from the device  
temperature sensor location to the hot spot location of the controller is larger than 15 degrees C, then selfꢀrefresh  
mode will not reliably maintain memory contents.  
To accurately determine the temperature gradient between the memory thermal sensor and the induced hot spot,  
the memory thermal sensor location must be provided to the LPDDR4 memory controller.  
Support of thermal offset function is optional. Please refer to vendor datasheet to figure out if the function is  
supported or not.  
Version 1.0  
277  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Temperature Sensor  
LPDDR4 devices feature a temperature sensor whose status can be read from MR4. This sensor can be used to  
determine an appropriate refresh rate, determine whether AC timing deꢀrating is required in the elevated  
temperature range, and/or monitor the operating temperature. Either the temperature sensor or the device TOPER  
may be used to determine whether operating temperature requirements are being met.  
LPDDR4 devices shall monitor device temperature and update MR4 according to tTSI. Upon assertion of CKE (Low  
to High transition), the device temperature status bits shall be no older than tTSI. MR4 will be updated even when  
device is in self refresh state with CKE HIGH.  
When using the temperature sensor, the actual device case temperature may be higher than the TOPER  
specification that applies for the standard or elevated temperature ranges. For example, TCASE may be above  
85°C when MR4[2:0] equals ‘b011. LPDDR4 devices shall allow for 2°C temperature margin between the point at  
which the device updates the MR4 value and the point at which the controller reconfigures the system accordingly.  
In the case of tight thermal coupling of the memory device to external hot spots, the maximum device temperature  
might be higher than what is indicated by MR4.  
To assure proper operation using the temperature sensor, applications should consider the following factors:  
• TempGradient is the maximum temperature gradient experienced by the memory device at the temperature of  
interest over a range of 2°C.  
• ReadInterval is the time period between MR4 reads from the system.  
• TempSensorInterval (tTSI) is maximum delay between internal updates of MR4.  
• SysRespDelay is the maximum time between a read of MR4 and the response by the system.  
In order to determine the required frequency of polling MR4, the system shall use the maximum TempGradient and  
the maximum response time of the system using the following equation:  
TempGradient x (ReadInterval + tTSI + SysRespDelay) 2°C  
Temperature Sensor  
Max/  
Parameter  
Symbol  
Value  
Unit Notes  
Min  
Max  
Max  
Max  
Max  
Max  
System Temperature Gradient  
MR4 Read Interval  
TempGradient  
ReadInterval  
tTSI  
System Dependent  
°C/s  
ms  
ms  
ms  
°C  
System Dependent  
Temperature Sensor Interval  
System Response Delay  
Device Temperature Margin  
32  
SysRespDelay  
TempMargin  
System Dependent  
2
For example, if TempGradient is 10°C/s and the SysRespDelay is 1 ms:  
(10°C/s) x (ReadInterval + 32ms + 1ms) 2°C  
In this case, ReadInterval shall be no greater than 167 ms.  
Version 1.0  
12/2018  
278  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Temp Sensor Timing  
Version 1.0  
279  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ZQ Calibration  
The MPC command is used to initiate ZQ Calibration, which calibrates the output driver impedance across process,  
temperature, and voltage. ZQ Calibration occurs in the background of device operation, and is designed to eliminate  
any need for coordination between channels (i.e. it allows for channel independence).  
There are two ZQ Calibration modes initiated with the MPC command: ZQCal Start, and ZQCal Latch. ZQCal Start  
initiates the SDRAM’s calibration procedure, and ZQCal Latch captures the result and loads it into the SDRAM's  
drivers.  
A ZQCal Start command may be issued anytime the LPDDR4ꢀSDRAM is not in a powerꢀdown state. A ZQCal Latch  
Command may be issued anytime outside of powerꢀdown after tZQCAL has expired and all DQ bus operations have  
completed. The CA Bus must maintain a Deselect state during tZQLAT to allow CA ODT calibration settings to be  
updated. The following mode register fields that modify I/O parameters cannot be changed following a ZQCal Start  
command and before tZQCAL has expired:  
• PUꢀCal (Pullꢀup Calibration VOH Point)  
• PDDS (Pull Down Drive Strength and Rx Termination)  
• DQꢀODT (DQ ODT Value)  
• CAꢀODT (CA ODT Value)  
Version 1.0  
280  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ZQCal Reset  
The ZQCal Reset command resets the output impedance calibration to a default accuracy of +/ꢀ 30% across  
process, voltage, and temperature. This command is used to ensure output impedance accuracy to +/ꢀ 30% when  
ZQCal Start and ZQCal Latch commands are not used.  
The ZQCal Reset command is executed by writing MR10 OP[0]=1B.  
ZQCal Timing Parameters  
Parameter  
Symbol  
tZQCAL  
Min/Max  
Min  
Value  
1
Unit  
us  
ZQ Calibration Time  
ZQ Calibration Latch Time  
ZQ Calibration Reset Time  
tZQLAT  
Min  
max(30ns,8nCK)  
max(50ns,3nCK)  
ns  
tZQRESET  
Min  
ns  
ZQCal Timing  
Notes:  
1. Write and Precharge operations shown for illustrative purposes.  
Any single or multiple valid commands may be executed within the tZQCAL time and prior to latching the results.  
2. Before the ZQ-Latch command can be executed, any prior commands utilizing the DQ bus must have completed.  
Write commands with DQ Termination must be given enough time to turn off the DQ-ODT before issuing the ZQ-Latch command.  
See the ODT section for ODT timing.  
Version 1.0  
281  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
MultiꢀChannel Considerations for Dual Channel Devices  
The LPDDR4ꢀSDRAM includes a single ZQ pin and associated ZQ Calibration circuitry. Calibration values from this  
circuit will be used by both channels according to the following protocol:  
1. ZQCal Start commands may be issued to either or both channels.  
2. ZQCal Start commands may be issued when either or both channels are executing other commands and  
other commands may be issued during tZQCAL.  
3. ZQCal Start commands may be issued to both channels simultaneously.  
4. The ZQCal Start command will begin the calibration unless a previously requested ZQ calibration is in  
progress.  
5. If a ZQCal Start command is received while a ZQ calibration is in progress on the SDRAM, the ZQCal Start  
command will be ignored and the inꢀprogress calibration will not be interrupted.  
6. ZQCal Latch commands are required for each channel.  
7. ZQCal Latch commands may be issued to both channels simultaneously.  
8. ZQCal Latch commands will latch results of the most recent ZQCal Start command provided tZQCAL has  
been met.  
9. ZQCal Latch commands which do not meet tZQCAL will latch the results of the most recently completed ZQ  
calibration.  
10. ZQ Reset MRW commands will only reset the calibration values for the channel issuing the command.  
In compliance with complete channel independence, either channel may issue ZQCal Start and ZQCal Latch  
commands as needed without regard to the state of the other channel.  
ZQ External Resistor, Tolerance, and Capacitive Loading  
To use the ZQ calibration function, a 240 ohm +/ꢀ 1% tolerance external resistor must be connected between the ZQ  
pin and VDDQ  
.
If the system configuration shares the CA bus to form a x32 (or wider) channel, the ZQ pin of each die’s x16 channel  
shall use a separate ZQCal resistor.  
If the system configuration has more than one rank, and if the ZQ pins of both ranks are attached to a single resistor,  
then the SDRAM controller must ensure that the ZQCal’s don’t overlap.  
The total capacitive loading on the ZQ pin must be limited to 25pF.  
Example: If a system configuration shares a CA bus between ‘n’ channels to form a n * 16 wide bus, and no means  
are available to control the ZQCal separately for each channel (i.e. separate CS, CKE, or CK), then each x16  
channel must have a separate ZQCal resistor.  
Example: For a x32, two rank system, each x16 channel must have its own ZQCal resistor, but the ZQCal resistor  
can be shared between ranks on each x16 channel. In this configuration, the CS signal can be used to ensure that  
the ZQCal commands for Rank[0] and Rank[1] don't overlap.  
Version 1.0  
282  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Pull Up/Pull Down Driver Characteristics and Calibration  
Pullꢀdown Driver Characteristics, with ZQ Calibration  
Resistor  
Min  
0.9  
0.9  
0.9  
0.9  
0.9  
0.9  
Nom  
Max  
1.1  
1.1  
1.1  
1.1  
1.1  
1.1  
Unit  
R
,nom  
ONPD  
40 Ohm  
48 Ohm  
60 Ohm  
80 Ohm  
1
1
1
1
1
1
RZQ/6  
RZQ/5  
RZQ/4  
RZQ/3  
RZQ/2  
RZQ/1  
R
R
R
R
ON40PD  
ON48PD  
ON60PD  
ON80PD  
120 Ohm  
240 Ohm  
R
R
ON120PD  
ON240PD  
Notes:  
1. All value are after ZQ Calibration. Without ZQ Calibration RONPD values are ± 30%.  
PullꢀUp Characteristics, with ZQ Calibration  
VOH,nom(mV)  
Min  
0.9  
Nom  
Max  
1.1  
Unit  
VOH ,nom  
PU  
300  
360  
1
1
VOH,nom  
VOH,nom  
V
*0.5  
DDQ  
DDQ  
0.9  
1.1  
V
*0.6  
Notes:  
1. All values are after ZQ Calibration. Without ZQ Calibration VOH(nom) values are ± 30%.  
2. VOH,nom (mV) values are based on a nominal VDDQ =0.6V.  
Terminated Valid Calibration Points  
SOCODT Value  
VOHPU,nom  
*0.5  
240  
VALID  
DNU  
120  
80  
60  
48  
40  
VALID  
VALID  
VALID  
DNU  
VALID  
VALID  
VALID  
DNU  
VALID  
DNU  
V
DDQ  
V 0.6  
DDQ*  
Notes:  
1. Once the output is calibrated for a given VOH(nom) calibration point, the ODT value may be changed without recalibration.  
2. If the VOH(nom) calibration point is changed, then re-calibration is required.  
3. DNU = Do Not Use  
Version 1.0  
283  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
On Die Termination for Command/Address Bus  
ODT (OnꢀDie Termination) is a feature of the LPDDR4 SDRAM that allows the SDRAM to turn on/off termination  
resistance for CK, CK, CS and CA[5:0] signals without the ODT control pin.  
The ODT feature is designed to improve signal integrity of the memory channel by allowing the DRAM controller to  
turn on and off termination resistance for any target DRAM devices via Mode Register setting.  
A simple functional representation of the DRAM ODT feature is shown in following figure.  
Functional Representation of CA ODT  
Version 1.0  
284  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT Mode Register and ODT State Table  
ODT termination values are set and enabled via MR11. The CA bus (CK, CK, CS, CA[5:0]) ODT resistance values  
are set by MR11 OP[6:4]. The default state for the CA is ODT disabled.  
ODT is applied on the CA bus to the CK, CK, CS and CA[5:0] signals. Generally, only one termination load will be  
present even if multiple devices are sharing the command signals. In contrast to LPDDR4 where the ODT_CA input  
is used in combination with mode registers, LPDDR4X uses mode registers exclusively to enable CA termination.  
Before enabling CA termination via MR11, all ranks should have appropriate MR22 termination settings  
programmed. In a multi rank system, the terminating rank should be trained first, followed by the nonterminating  
rank(s).  
Command Bus ODT State  
ODTE-CA  
MR11[6:4]  
Disabled1  
Valid2  
ODTD-CA  
ODTF-CK  
ODTF-CS  
ODT State  
for CA  
Off  
ODT State  
for CK/CK  
Off  
ODT State  
for CS  
Off  
MR22[5]  
MR22[3]  
MR22[4]  
Valid2  
Valid2  
Valid2  
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
On  
On  
On  
Valid2  
On  
On  
Off  
Valid2  
On  
Off  
On  
Valid2  
On  
Off  
Off  
Valid2  
Off  
On  
On  
Valid2  
Off  
On  
Off  
Valid2  
Off  
Off  
On  
Valid2  
Off  
Off  
Off  
Notes:  
1. Default Value  
2. “Valid” means “0 or 1”  
Version 1.0  
285  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT Mode Register and ODT Characteristics  
Vout  
RTT =  
Iout|  
On Die Termination for CA  
Version 1.0  
286  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT DC Electrical Characteristics, assuming RZQ = 240+/ꢀ1% over the entire operating temperature  
range after a proper ZQ calibration  
MR11[6:4]  
RTT  
Vout  
Min  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
Nom  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
Max  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
Unit  
Notes  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
RZQ  
VOLdc= 0.20 * V  
DDQ  
001  
RZQ  
240ꢀ  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
RZQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
RZQ/2  
RZQ/2  
RZQ/2  
RZQ/3  
RZQ/3  
RZQ/3  
RZQ/4  
RZQ/4  
RZQ/4  
RZQ/5  
RZQ/5  
RZQ/5  
RZQ/6  
RZQ/6  
RZQ/6  
010  
011  
100  
101  
110  
120ꢀ  
80ꢀ  
60ꢀ  
48ꢀ  
40ꢀ  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
VOHdc= 0.75 * V  
Mismatch CA-CA  
within clk group  
Notes:  
-
2
%
1,2,3  
0.50* V  
DDQ  
1. The tolerance limits are specified after calibration with stable voltage and temperature. For the behavior of the tolerance limits if  
temperature or voltage changes after calibration, see following section on voltage and temperature sensitivity.  
2. Pull-dn ODT resistors are recommended to be calibrated at 0.50*VDDQ. Other calibration schemes may be used to achieve the linearity  
spec shown above, e.g. calibration at 0.75*VDDQ and 0.2*VDDQ  
.
3. CA to CA mismatch within clock group (CA,CS) variation for a given component including CK and CK (characterized).  
RODT (max) – RODT(min)  
CA- CA Mismatch  
=
RODT(avg)  
Version 1.0  
287  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT DC Electrical Characteristics, assuming RZQ = 240+/ꢀ1% over the entire operating temperature  
range after a proper ZQ calibration  
MR11[6:4]  
RTT  
Vout  
Min  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
Nom  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
Max  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
Unit  
RZQ  
Notes  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
VOLdc= 0.20 * V  
DDQ  
001  
RZQ  
240ꢀ  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
RZQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
RZQ/2  
RZQ/2  
RZQ/2  
RZQ/3  
RZQ/3  
RZQ/3  
RZQ/4  
010  
011  
100  
120ꢀ  
80ꢀ  
60ꢀ  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
0.9  
1.0  
1.1  
RZQ/4  
VOMdc= 0.50 * V  
DDQ  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
RZQ/4  
RZQ/5  
RZQ/5  
RZQ/5  
RZQ/6  
RZQ/6  
RZQ/6  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
VOHdc= 0.75 * V  
DDQ  
DDQ  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
101  
110  
48ꢀ  
40ꢀ  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
VOHdc= 0.75 * V  
Mismatch CA-CA  
within clk group  
Notes:  
-
2
%
1,2,3  
0.50* V  
DDQ  
1. The tolerance limits are specified after calibration with stable voltage and temperature. For the behavior of the tolerance limits if  
temperature or voltage changes after calibration, see following section on voltage and temperature sensitivity.  
2. Pull-dn ODT resistors are recommended to be calibrated at 0.50*VDDQ. Other calibration schemes may be used to achieve the linearity  
spec shown above, e.g. calibration at 0.75*VDD2 and 0.2*VDDQ  
.
4. CA to CA mismatch within clock group (CA,CS) variation for a given component including CK and CK (characterized).  
RODT (max) – RODT(min)  
CA- CA Mismatch  
=
RODT(avg)  
Version 1.0  
288  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT for Command/Address update time  
ODT for Command/Address update time after Mode Register set are shown in following figure.  
ODT for Command/Address setting update timing in 4 Clock Cycle Command  
Notes:  
1. 4 Clock Cycle Command  
ODT CA AC Timing  
Speed  
3733  
Unit Notes  
Parameter  
ODT CA Value Update Time  
Symbol  
Min  
RU(20ns/tCK(avg))  
Max  
tODTUP  
-
Version 1.0  
12/2018  
289  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
OnꢀDie Termination  
ODT (OnꢀDie Termination) is a feature of the LPDDR4 SDRAM that allows the DRAM to turn on/off termination  
resistance for each DQ, DQS, DQS and DMI signals without the ODT control pin. The ODT feature is designed to  
improve signal integrity of the memory channel by allowing the DRAM controller to turn on and off termination  
resistance for any target DRAM devices during Write or Mask Write operation.  
The ODT feature is off and cannot be supported in Power Down and SelfꢀRefresh modes.  
A simple functional representation of the DRAM ODT feature is shown in following figure.  
Functional Representation of ODT  
The switch is enabled by the internal ODT control logic, which uses the Writeꢀ1 or Mask Writeꢀ1 command and other  
mode register control information. The value of RTT is determined by the settings of Mode Register bits.  
ODT Mode Register  
The ODT Mode is enabled if MR11 OP[3:0] are non zero. In this case, the value of RTT is determined by the settings  
of those bits. The ODT Mode is disabled if MR11 OP[3] = 0.  
Version 1.0  
290  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Asynchronous ODT  
When ODT Mode is enabled in MR11 OP[3:0], DRAM ODT is always HiꢀZ. DRAM ODT feature is automatically  
turned ON asynchronously based on the Writeꢀ1 or Mask Writeꢀ1 command that DRAM samples. After the write  
burst is complete, DRAM ODT featured is automatically turned OFF asynchronously.  
Following timing parameters apply when DRAM ODT mode is enabled:  
• ODTLon, tODTon,min, tODTon,max  
• ODTLoff, tODToff,min, tODToff,max  
ODTLon is a synchronous parameter and it is the latency from CASꢀ2 command to tODTon reference. ODTLon  
latency is a fixed latency value for each speed bin. Each speed bin has a different ODTLon latency.  
Minimum RTT turnꢀon time (tODTon,min) is the point in time when the device termination circuit leaves high  
impedance state and ODT resistance begins to turn on.  
Maximum RTT turn on time (tODTon,max) is the point in time when the ODT resistance is fully on.  
tODTon,min and tODTon,max are measured once ODTLon latency is satisfied from CASꢀ2 command.  
ODTLoff is a synchronous parameter and it is the latency from CASꢀ2 command to tODToff reference. ODTLoff  
latency is a fixed latency value for each speed bin. Each speed bin has a different ODTLoff latency.  
Minimum RTT turnꢀoff time (tODToff,min) is the point in time when the device termination circuit starts to turn off the  
ODT resistance.  
Maximum ODT turn off time (tODToff,max) is the point in time when the onꢀdie termination has reached high  
impedance.  
tODToff,min and tODToff,max are measured once ODTLoff latency is satisfied from CASꢀ2 command.  
Version 1.0  
291  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODTLon and ODTLoff Latency Values  
1
ODTLon Latency  
2
Lower Clock  
Upper Clock  
ODTLoff Latency  
tWPRE = 2 tCK  
Frequency Limit(>)  
Frequency Limit()  
WL Set "A"  
WL Set "B"  
WL Set "A"  
N/A  
N/A  
N/A  
20  
WL Set "B"  
N/A  
N/A  
N/A  
4
N/A  
N/A  
6
N/A  
N/A  
22  
10  
266  
533  
266  
533  
800  
12  
28  
800  
1066  
1333  
1600  
1866  
2133  
MHz  
4
14  
22  
32  
1066  
1333  
1600  
1866  
MHz  
6
18  
24  
36  
6
20  
26  
40  
8
24  
28  
44  
nCK  
nCK  
nCK  
nCK  
Notes:  
1. ODTLon is referenced from CAS-2 command.  
2. ODTLoff as shown in table assumes BL=16. For BL32, 8 tCK should be added.  
3. Clock Frequency herewith is a reference base on JEDEC's. Precise setting needs to follow where defined on speed compatible table in  
section “Operating Frequency, exceptional setting please confirm with NTC.  
Asynchronous ODT Turn On and Turn Off Timing  
Parameter  
tODTon, min  
tODTon, max  
tODToff, min  
tODToff, max  
800 - 2133 MHz  
Unit  
ns  
1.5  
3.5  
1.5  
3.5  
ns  
ns  
ns  
Version 1.0  
292  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Asynchronous ODTon/ODToff Timing  
Notes:  
1. BL=16, Write Postamble = 0.5nCK, DQ/DQS: VSSQ termination  
2. Din n = data-in to columnm n  
3. DES commands are shown for ease of illustration; other commands may be valid at these times.  
Version 1.0  
293  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT during Write Leveling  
If ODT is enabled in MR11 OP[3:0], in Write Leveling mode, DRAM always provides the termination on DQS/DQS  
signals. DQ termination is always off in Write Leveling mode regardless.  
DRAM Termination Function in Write Leveling Mode  
ODT Enabled in MR11  
Disabled  
DQS/DQS termination  
DQ termination  
OFF  
ON  
OFF  
OFF  
Enabled  
Version 1.0  
12/2018  
294  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
On Die Termination for DQ, DQS and DMI  
OnꢀDie Termination effective resistance RTT is defined by MR11 OP[2:0].  
ODT is applied to the DQ, DMI, DQS and DQS pins.  
A functional representation of the onꢀdie termination is shown in the figure below.  
Vout  
RTT =  
Iout|  
On Die Termination  
Version 1.0  
295  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT DC Electrical Characteristics, assuming RZQ = 240ꢂ +/ꢀ1% over the entire operating temperature range  
after a proper ZQ calibration  
MR11  
RTT  
Vout  
Min  
Nom  
Max  
Unit  
Notes  
OP[2:0]  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.0  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
RZQ  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
VOLdc= 0.20 * V  
DDQ  
001  
240Ω  
RZQ  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
RZQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
RZQ/2  
RZQ/2  
RZQ/2  
RZQ/3  
RZQ/3  
RZQ/3  
RZQ/4  
RZQ/4  
RZQ/4  
RZQ/5  
RZQ/5  
RZQ/5  
RZQ/6  
RZQ/6  
RZQ/6  
010  
011  
100  
101  
110  
120Ω  
80Ω  
60Ω  
48Ω  
40Ω  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
VOHdc= 0.75 * V  
Mismatch DQ-DQ  
within byte  
-
2
%
1,2,3  
0.50* V  
DDQ  
Notes:  
1. The tolerance limits are specified after calibration with stable voltage and temperature. For the behavior of the tolerance limits if  
temperature or voltage changes after calibration, see following section on voltage and temperature sensitivity.  
2. Pull-dn Pull-dn ODT resistors are recommended to be calibrated at 0.75*VDDQ and 0.2*VDDQ. Other calibration schemes may be used to  
achieve the linearity spec shown above, e.g., calibration at 0.75*VDDQ and 0.1*VDDQ  
.
3. DQ to DQ mismatch within byte variation for a given component including DQS and DQS (characterized).  
RODT (max) – RODT(min)  
DQ - DQ Mismatch  
=
RODT(avg)  
Version 1.0  
12/2018  
296  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
ODT DC Electrical Characteristics, assuming RZQ = 240ꢂ +/ꢀ1% over the entire operating temperature range  
after a proper ZQ calibration  
MR11  
RTT  
Vout  
Min  
Nom  
Max  
Unit  
Notes  
OP[2:0]  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
0.8  
0.9  
0.9  
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
1.1  
1.1  
1.3  
RZQ  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
1,2  
VOLdc= 0.20 * V  
DDQ  
001  
240Ω  
RZQ  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
RZQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
RZQ/2  
RZQ/2  
RZQ/2  
RZQ/3  
RZQ/3  
RZQ/3  
RZQ/4  
RZQ/4  
RZQ/4  
RZQ/5  
RZQ/5  
RZQ/5  
RZQ/6  
RZQ/6  
RZQ/6  
010  
011  
100  
101  
110  
120Ω  
80Ω  
60Ω  
48Ω  
40Ω  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
DDQ  
VOHdc= 0.75 * V  
VOLdc= 0.20 * V  
VOMdc= 0.50 * V  
DDQ  
DDQ  
VOHdc= 0.75 * V  
Mismatch DQ-DQ  
within byte  
-
2
%
1,2,3  
0.50* V  
DDQ  
Notes:  
1. The tolerance limits are specified after calibration with stable voltage and temperature. For the behavior of the tolerance limits if  
temperature or voltage changes after calibration, see following section on voltage and temperature sensitivity1.  
2. Pull-dn ODT resistors are recommended to be calibrated at 0.75*VDDQ and 0.2*VDDQ. Other calibration schemes may be used to achieve  
the linearity spec shown above, e.g., calibration at 0.75*VDDQ and 0.1*VDDQ  
.
3. DQ to DQ mismatch within byte variation for a given component including DQS and DQS (characterized).  
RODT (max) – RODT(min)  
DQ - DQ Mismatch  
=
RODT(avg)  
NOTE1. As of publication of this document, under discussion by the formulating committee.  
Version 1.0  
297  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Output Driver and Termination Register Temperature and Voltage Sensitivity  
If temperature and/or voltage change after calibration, the tolerance limits widen according to the tables shown  
below.  
Output Driver and Termination Register Sensitivity Definition  
Definition  
Resistor  
Min  
Max  
Unit Notes  
Point  
0.50 x  
VDDQ  
0.50 x  
VDDQ  
0.50 x  
VDDQ  
0.50 x  
VDDQ  
%
%
%
%
1,2  
R
90-(dR dT x |T|)-(dR dV x |V|) 110+(dR dT x |T|)+(dR dV x |V|)  
on on on on  
ONPD  
1,2,5  
1,2,3  
1,2,4  
90-(dVOHdT x |T|)-(dVOHdV x |V|) 110+(dVOHdT x |T|)+(dVOHdV x|V|)  
90-(dR dT x |T|)-(dR dV x |V|) 110+(dR dT x |T|)+(dR dV x |V|)  
VOH  
PU  
R
)
TT(I/O  
on  
on  
on  
on  
R
90-(dR dT x |T|)-(dR dV x |V|) 110+(dR dT x |T|)+(dR dV x |V|)  
on on on on  
TT(In)  
Notes:  
1. T = T - T(@ Calibration), V = V - V(@ Calibration)  
2. dRONdT, dRONdV, dVOHdT, dVOHdV, dRTTdV, and dRTTdT are not subject to production test but are verified.  
3. This parameter applies to Input/Output pin such as DQS, DQ and DMI.  
4. This parameter applies to Input pin such as CK, CA and CS.  
5. Refer to Pull Up/Pull Down Driver Characteristics for VOHPU.  
Output Driver and Termination Register Temperature and Voltage Sensitivity  
Symbol  
dR dT  
Parameter  
Min  
0.00  
0.00  
0.00  
0.00  
0.00  
0.00  
Max  
0.75  
0.20  
0.75  
0.35  
0.75  
0.20  
Unit  
%/°C  
%/mV  
%/°C  
%/mV  
%/°C  
%/mV  
R
Temperature Sensitivity  
ON  
ON  
dR dV  
ON  
R
Voltage Sensitivity  
ON  
dVOHdT  
dVOHdV  
VOH Temperature Sensitivity  
VOH Voltage Sensitivity  
dR dT  
TT  
R
Temperature Sensitivity  
TT  
dR dV  
TT  
R
Voltage Sensitivity  
TT  
Version 1.0  
12/2018  
298  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
PowerꢀDown Mode  
PowerꢀDown Entry and Exit  
Powerꢀdown is asynchronously entered when CKE is driven LOW. CKE must not go LOW while the following  
operations are in progress:  
• Mode Register Read  
• Mode Register Write  
• Read  
• Write  
• VREF(CA) Range and Value setting via MRW  
• VREF(DQ) Range and Value setting via MRW  
• Command Bus Training mode Entering/Exiting via MRW  
• VRCG High Current mode Entering/Exiting via MRW  
And the LPDDR4 DRAM cannot be placed in powerꢀdown state during “Start DQS Interval Oscillator” operation.  
CKE can go LOW while any other operations such as row activation, Precharge, Auto Precharge, or Refresh are in  
progress. The powerꢀdown IDD specification will not be applied until such operations are complete.  
Entering powerꢀdown deactivates the input and output buffers, excluding CKE and RESET. To ensure that there is  
enough time to account for internal delay on the CKE signal path, CS input is required stable Low level and CA input  
level is don’t care after CKE is driven LOW, this timing period is defined as tCKELCS. Clock input is required after  
CKE is driven LOW, this timing period is defined as tCKELCK. CKE LOW will result in deactivation of all input  
receivers except RESET after tCKELCK has expired. In powerꢀdown mode, CKE must be held LOW; all other input  
signals except RESET are "Don't Care". CKE LOW must be maintained until tCKE,min is satisfied.  
VDDQ can be turned off during powerꢀdown. Prior to exiting powerꢀdown, VDDQ must be within its minimum/maximum  
operating range.  
No refresh operations are performed in powerꢀdown mode except SelfꢀRefresh powerꢀdown. The maximum duration  
in nonꢀSelfꢀRefresh powerꢀdown mode is only limited by the refresh requirements outlined in the Refresh command  
section.  
The powerꢀdown state is asynchronously exited when CKE is driven HIGH. CKE HIGH must be maintained until  
tCKE,min is satisfied. A valid, executable command can be applied with powerꢀdown exit latency tXP after CKE  
goes HIGH. Powerꢀdown exit latency is defined in the AC timing parameter table.  
Clock frequency change or Clock Stop is inhibited during tCMDCKE, tCKELCK, tCKCKEH, tXP, tMRWCKEL and  
tZQCKE periods.  
If powerꢀdown occurs when all banks are idle, this mode is referred to as idle powerꢀdown. if powerꢀdown occurs  
Version 1.0  
299  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
when there is a row active in any bank, this mode is referred to as active powerꢀdown. And If powerꢀdown occurs  
when Self Refresh is in progress, this mode is referred to as Self Refresh powerꢀdown in which the internal refresh is  
continuing in the same way as Self Refresh mode.  
When CA, CK and/or CS ODT is enabled via MR11 OP[6:4] and also via MR22 or CAꢀODT pad setting, the rank  
providing ODT will continue to terminate the command bus in all DRAM states including powerꢀdown.  
Basic PowerꢀDown Entry and Exit Timing  
Notes:  
1. Input clock frequency can be changed or the input clock can be stopped or floated during power-down, provided that upon exiting  
power-down, the clock is stable and within specified limits for a minimum of tCKCKEH of stable clock prior to power-down exit and the  
clock frequency is between the minimum and maximum specified frequency for the speed grade in use.  
Version 1.0  
300  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read and Read with Auto Precharge to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until the end of the burst operation.  
2. Minimum Delay time from Read Command or Read with Auto-Precharge Command to falling edge of CKE signal is as follows.  
Read Post-amble = 0.5nCK : MR1 OP[7]=[0] : (RL x tCK) + tDQSCK(Max) + ((BL/2) x tCK) + 1tCK  
Read Post-amble = 1.5nCK : MR1 OP[7]=[1] : (RL x tCK) + tDQSCK(Max) + ((BL/2) x tCK) + 2tCK  
Version 1.0  
301  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write and Mask Write to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until the end of the burst operation.  
2. Minimum Delay time from Write Command or Mask Write Command to falling edge of CKE signal is as follows.  
(WL x tCK) + tDQSS(Max) + tDQS2DQ(Max) + ((BL/2) x tCK) + tWR  
3. This timing is applied regardless of DQ ODT Disable/Enable setting: MR11[OP2:0].  
4. This timing diagram only applies to the Write and Mask Write Commands without Auto-Precharge.  
Version 1.0  
302  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Write with Auto Precharge and Mask Write with Auto Precharge to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until the end of the burst operation.  
2. Delay time from Write with Auto-Precharge Command or Mask Write with Auto-Precharge Command to falling edge of CKE signal is  
more than (WL x tCK) + tDQSS(Max) + tDQS2DQ(Max) + ((BL/2) x tCK) + (nWR x tCK) + (2 x tCK)  
3. Internal Precharge Command  
4. This timing is applied regardless of DQ ODT Disable/Enable setting: MR11[OP2:0].  
Version 1.0  
303  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Refresh entry to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until tCMDCKE is satisfied.  
Activate Command to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until tCMDCKE is satisfied.  
Version 1.0  
304  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Precharge Command to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until tCMDCKE is satisfied.  
Mode Register Read to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until the end of the burst operation.  
2. Minimum Delay time from Mode Register Read Command to falling edge of CKE signal is as follows:  
Read Post-amble = 0.5nCK : MR1 OP[7]=[0] : (RL x tCK) + tDQSCK(Max) + ((BL/2) x tCK) + 1tCK  
Read Post-amble = 1.5nCK : MR1 OP[7]=[1] : (RL x tCK) + tDQSCK(Max) + ((BL/2) x tCK) + 2tCK  
Version 1.0  
305  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Mode Register Write to PowerꢀDown Entry  
Notes:  
1. CKE must be held HIGH until tMRWCKEL is satisfied.  
2. This timing is the general definition for Power Down Entry after Mode Register Write Command.  
When a Mode Register Write Command changes a parameter or starts an operation that requires special timing  
longer than tMRWCKEL, that timing must be satisfied before CKE is driven low.  
Changing the Vref(DQ) value is one example, in this case the appropriate Vref_time-Short/Middle/Long must be satisfied.  
Multi purpose Command for Start ZQ Calibration to PowerꢀDown Entry  
Notes:  
1. ZQ Calibration continues if CKE goes low after tZQCKE is satisfied.  
Version 1.0  
306  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
PowerꢀDown AC Timing  
Min/  
Parameter  
Symbol  
Data Rate  
Unit Notes  
Max  
Power Down Timing  
CKE minimum pulse width  
tCKE  
Min  
Max(7.5ns, 4nCK)  
-
(HIGH and LOW pulse width)  
Delay from valid command to CKE input LOW  
Valid Clock Requirement after CKE Input low  
Valid CS Requirement before CKE Input Low  
Valid CS Requirement after CKE Input low  
Valid Clock Requirement before CKE Input High  
Exit power- down to next valid command delay  
Valid CS Requirement before CKE Input High  
Valid CS Requirement after CKE Input High  
Valid Clock and CS Requirement after CKE Input low  
after MRW Command  
tCMDCKE  
tCKELCK  
tCSCKE  
Min  
Min  
Min  
Min  
Min  
Min  
Min  
Min  
Max(1.75ns, 3nCK)  
Max(5ns, 5nCK)  
1.75  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
ns  
1
1
tCKELCS  
tCKCKEH  
tXP  
Max(5ns, 5nCK)  
Max(1.75ns, 3nCK)  
Max(7.5ns, 5nCK)  
1.75  
1
1
tCSCKEH  
tCKEHCS  
Max(7.5ns, 5nCK)  
tMRWCKEL  
tZQCKE  
Min  
Min  
Max(14ns, 10nCK)  
Max(1.75ns, 3nCK)  
ns  
ns  
1
1
Valid Clock and CS Requirement after CKE Input low  
after ZQ Calibration Start Command  
Notes:  
1. Delay time has to satisfy both analog time(ns) and clock count(nCK).  
For example, tCMDCKE will not expire until CK has toggled through at least 3 full cycles  
(3 *tCK) and 1.75ns has transpired.  
The case which 3nCK is applied to is shown below.  
tCMDCKE Timing  
Version 1.0  
307  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Input Clock Stop and Frequency Change  
LPDDR4 SDRAMs support input clock frequency change during CKE LOW under the following conditions:  
• tCK(abs)min is met for each clock cycle;  
• Refresh requirements apply during clock frequency change;  
• During clock frequency change, only REFab or REFpb commands may be executing;  
• Any Activate or Precharge commands have executed to completion prior to changing the frequency;  
• The related timing conditions (tRCD, tRP) have been met prior to changing the frequency;  
• The initial clock frequency shall be maintained for a minimum of 4 clock cycles after CKE goes LOW;  
• The clock satisfies tCH(abs) and tCL(abs) for a minimum of 2 clock cycles prior to CKE going HIGH  
After the input clock frequency is changed and CKE is held HIGH, additional MRW commands may be required to  
set the WR, RL etc. These settings may need to be adjusted to meet minimum timing requirements at the target  
clock frequency.  
LPDDR4 devices support clock stop during CKE LOW under the following conditions:  
• CK is held LOW and CK is held HIGH or both are floated during clock stop;  
• Refresh requirements apply during clock stop;  
• During clock stop, only REFab or REFpb commands may be executing;  
• Any Activate or Precharge commands have executed to completion prior to stopping the clock;  
• The related timing conditions (tRCD, tRP) have been met prior to stopping the clock;  
• The initial clock frequency shall be maintained for a minimum of 4 clock cycles after CKE goes LOW;  
• The clock satisfies tCH(abs) and tCL(abs) for a minimum of 2 clock cycles prior to CKE going HIGH  
LPDDR4 devices support input clock frequency change during CKE HIGH under the following conditions:  
• tCK(abs)min is met for each clock cycle;  
• Refresh requirements apply during clock frequency change;  
• Any Activate, Read, Write, Precharge, Mode Register Write, or Mode Register Read commands must have  
executed to completion, including any associated data bursts prior to changing the frequency;  
• The related timing conditions (tRCD, tWR, tWRA, tRP, tMRW, tMRR, etc.) have been met prior to changing  
the frequency;  
• CS shall be held LOW during clock frequency change;  
• During clock frequency change, only REFab or REFpb commands may be executing;  
• The LPDDR4 SDRAM is ready for normal operation after the clock satisfies tCH(abs) and tCL(abs) for a  
minimum of 2*tCK+tXP.  
After the input clock frequency is changed, additional MRW commands may be required to set the WR, RL etc.  
These settings may need to be adjusted to meet minimum timing requirements at the target clock frequency.  
Version 1.0  
308  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
LPDDR4 devices support clock stop during CKE HIGH under the following conditions:  
• CK is held LOW and CK is held HIGH during clock stop;  
• CS shall be held LOW during clock clock stop;  
• Refresh requirements apply during clock stop;  
• During clock stop, only REFab or REFpb commands may be executing;  
• Any Activate, Read, Write, MPC(WRFIFO,RDFIFO,RDDQCAL), Precharge, Mode Register Write or Mode  
Register Read commands must have executed to completion, including any associated data bursts and extra 4  
clock cycles must be provided prior to stopping the clock;  
• The related timing conditions (tRCD, tWR, tRP, tMRW, tMRR, tZQLAT, etc.) have been met prior to stopping  
the clock;  
• Read with auto preꢀcharge and write with auto preꢀcharge commands need extra 4 clock cycles in addition to  
the related timing constraints, nWR and nRTP, to complete the operations.  
• REFab, REFpb, SRE, SRX and MPC(Zqcal Start)commands are required to have 4 additional clocks prior to  
stopping the clock same as CKE=L case.  
• The LPDDR4 SDRAM is ready for normal operation after the clock is restarted and satisfies tCH(abs) and  
tCL(abs) for a minimum of 2*tCK+tXP.  
Version 1.0  
309  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Truth Tables  
Operation or timing that is not specified is illegal, and after such an event, in order to guarantee proper operation, the  
LPDDR4 device must be reset or power-cycled and then restarted through the specified initialization sequence  
before normal operation can continue.  
CKE signal has to be held High when the commands listed in the command truth table input.  
Version 1.0  
310  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Command Truth Table  
CK  
SDRAM Command  
CS CA0  
CA1  
CA2  
CA3  
CA4  
CA5  
Notes  
edge  
Deselect(DES)  
L
X
R1  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
R1  
R2  
1,2  
H
L
L
OP0  
L
L
OP1  
L
L
OP2  
L
L
OP3  
L
L
OP4  
H
OP6  
OP5  
AB  
V
Multi Purpose Command  
(MPC)  
1,9  
H
L
Precharge(PRE)  
(Per Bank, All Bank)  
Refresh(REF)  
1,2,3,4  
1,2,3,4  
1,2  
BA0  
L
BA1  
L
BA2  
L
V
V
H
L
H
L
AB  
V
(Per Bank, All Bank)  
BA0  
L
BA1  
L
BA2  
L
V
V
H
L
H
H
V
Self Refresh Entry(SRE)  
Write-1(WR-1)  
V
V
V
H
L
L
BA0  
L
L
BA1  
L
H
BA2  
H
L
V
L
L
C9  
H
BL  
AP  
V
1,2,3,6,7,9  
1,2  
H
L
Self Refresh Exit(SRX)  
Mask Write-1(MRW-1)  
RFU  
H
L
L
BA0  
L
L
BA1  
L
H
BA2  
H
H
V
H
L
C9  
H
L
AP  
V
1,2,3,5,6,9  
1,2  
H
L
H
L
L
BA0  
L
H
BA1  
H
L
BA2  
L
L
V
L
C9  
H
BL  
AP  
C8  
C7  
L
Read-1(RD-1)  
1,2,3,6,7,9  
1,8,9  
H
L
L
CAS-2  
(Write-2, Mask Write-2,Read-2, MRR-2, MPC)  
C2  
L
C3  
H
C4  
L
C5  
H
C6  
L
H
L
RFU  
1,2  
V
V
H
L
L
L
H
H
L
H
L
H
L
V
RFU  
1,2  
H
L
H
OP7  
Mode Register Write-1(MRW-1)  
Mode Register Write-2(MRW-2)  
Mode Register Read-1(MRR-1)  
RFU  
1,2,11  
1,2,11  
1,2,12  
1,2  
MA0 MA1  
MA2 MA3  
MA4 MA5  
H
L
L
OP0  
L
H
OP1  
H
H
OP2  
H
L
OP3  
H
H
OP4  
L
OP6  
OP5  
V
H
L
MA0 MA1  
L
MA2 MA3  
H
MA4 MA5  
H
H
L
H
H
V
V
H
L
H
BA0  
H
L
BA1  
H
R12  
BA2  
R6  
R13  
V
R14  
R10  
R8  
R15  
R11  
R9  
Activate-1(ACT-1)  
1,2,3,10  
1,10,13  
H
L
R7  
R3  
Activate-2(ACT-2)  
R0  
R1  
R2  
R4  
R5  
Version 1.0  
311  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
1. All LPDDR4 commands except for Deselect are 2 clock cycle long and defined by states of CS and CA[5:0] at the first rising edge of clock.  
Deselect command is 1 clock cycle long.  
2. "V" means "H" or "L" (a defined logic level). "X" means don’t care in which case CA[5:0] can be floated.  
3. Bank addresses BA[2:0] determine which bank is to be operated upon.  
4. AB "HIGH" during Precharge or Refresh command indicates that command must be applied to all banks and bank address is a don’t care.  
5. Mask Write-1 command supports only BL 16. For Mark Write-1 command, CA5 must be driven LOW on first rising clock cycle (R1).  
6. AP "HIGH" during Write-1, Mask Write-1 or Read-1 commands indicates that an Auto-Precharge will occur to the bank associated with  
the Write, Mask Write or Read command.  
7. If Burst Length on-the-fly is enabled, BL "HIGH" during Write-1 or Read-1 command indicates that Burst Length should be set on-the-Fly  
to BL=32. BL "LOW" during Write-1 or Read-1 command indicates that Burst Length should be set on-the-fly to BL=16. If Burst Length  
on-the-fly is disabled, then BL must be driven to defined logic level "H" or "L".  
8. For CAS-2 commands (Write-2 or Mask Write-2 or Read-2 or MRR-2 or MPC (Only Write FIFO, Read FIFO & Read DQ Calibration), C[1:0]  
are not transmitted on the CA[5:0] bus and are assumed to be zero. Note that for CAS-2 Write-2 or CAS-2 Mask Write-2 command, C[3:2]  
must be driven LOW.  
9. Write-1 or Mask Write-1 or Read-1 or Mode Register Read-1 or MPC (Only Write FIFO, Read FIFO & Read DQ Calibration) command must  
be immediately followed by CAS-2 command consecutively without any other command in between. Write-1 or Mask Write-1 or Read-1  
or Mode Register Read-1 or MPC (Only Write FIFO, Read FIFO & Read DQ Calibration) command must be issued first before issuing CAS-2  
command. MPC (Only Start & Stop DQS Oscillator, Start & Latch ZQ Calibration) commands do not require CAS-2 command; they require  
two additional DES or NOP commands consecutively before issuing any other commands..  
10. Activate-1 command must be immediately followed by Activate-2 command consecutively without any other command in between.  
Activate-1 command must be issued first before issuing Activate-2 command. Once Activate-1 command is issued, Activate-2 command  
must be issued before issuing another Activate-1 command.  
11. MRW-1 command must be immediately followed by MRW-2 command consecutively without any other command in between. MRW-1  
command must be issued first before issuing MRW-2 command.  
12. MRR-1 command must be immediately followed by CAS-2 command consecutively without any other command in between. MRR-1  
command must be issued first before issuing CAS-2 command.  
13. In case of the densities which not to use R17 and R18 as row address, R17 and R18 must both be driven High for every ACT-2 command to  
maintain backward compatibility.  
Version 1.0  
312  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
TRR Mode ꢀ Target Row Refresh  
A LPDDR4 SDRAM's row has a limited number of times a given row can be accessed within a refresh period  
(tREFW * 2) prior to requiring adjacent rows to be refreshed. The Maximum Activate Count (MAC) is the maximum  
number of activates that a single row can sustain within a refresh period before the adjacent rows need to be  
refreshed. The row receiving the excessive actives is the Target Row (TRn), the adjacent rows to be refreshed are  
the victim rows. When the MAC limit is reached on TRn, either the LPDRR4 SDRAM receive all (R * 2) Refresh  
Commands before another row activate is issued, or the LPDRR4 SDRAM should be placed into Targeted Row  
Refresh (TRR) mode. The TRR Mode will reꢀfresh the rows adjacent to the TRn that encountered tMAC limit.  
If LPDDR4 SDRAM supports Unlimited MAC value: MR24 [OP2:0=000] and MR24 [OP3=1], Target Row Refresh  
operation is not required. Even though LPDDR4 SDRAM allows to set MR24 [OP7=1]: TRR mode enable, in this  
case LPDDR4 SDRAM's behavior is vendor specific. For example, a certain LPDDR4 SDRAM may ignore MRW  
command for entering/exiting TRR mode or a certain SDRAM may support commands related TRR mode. See  
vendor device datasheets for details about TRR mode definition at supporting Unlimited MAC value case.  
There could be a maximum of two target rows to a victim row in a bank. The cumulative value of the acꢀtivates from  
the two target rows on a victim row in a bank should not exceed MAC value as well.  
MR24 fields required to support the new TRR settings. Setting MR24 [OP7=1] enables TRR Mode and setting MR24  
[OP7=0] disables TRR Mode. MR24 [OP6:OP4] defines which bank (BAn) the target row is located in (See MR24  
table for details).  
The TRR mode must be disabled during initialization as well as any other LPDRR4 SDRAM calibration modes. The  
TRR mode is entered from a DRAM Idle State, once TRR mode has been entered, no other Mode Register  
commands are allowed until TRR mode is completed, except setting MR24 [OP7=0] to interrupt and reissue the  
TRR mode is allowed.  
When enabled; TRR Mode is selfꢀclearing; the mode will be disabled automatically after the completion of defined  
TRR flow; after the 3rd BAn precharge has completed plus tMRD. Optionally the TRR mode can also be exited via  
another MRS command at the completion of TRR by setting MR24 [OP7=0]; if the TRR is exited via another MRS  
command, the value written to MR24 [OP6:OP4] are don’t cares.  
Version 1.0  
313  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
TRR Mode Operation  
1. The following steps must be performed when TRR mode is enabled. This mode requires all three ACT (ACT1,  
ACT2 and ACT3) and three corꢀresponding PRE commands (PRE1, PRE2 and PRE3) to complete TRR mode. A  
Precharge All (PREA) commands issued while LPDDR4 SDRAM is in TRR mode will also perform precharge to  
BAn and counts towards a PREn command.  
2. Prior to issuing the MRW command to enter TRR mode, the SDRAM should be in the idle state. A MRW  
command must be issued with MR24 [OP7=1] and MR24 [OP6:4] defining the bank in which the targeted row is  
located. All other MR24 bits should remain unchanged.  
3. No activity is to occur in the DRAM until tMRD has been satisfied. Once tMRD has been satisfied, the only  
commands to BAn allowed are ACT and PRE until the TRR mode has been completed.  
4. The first ACT to the BAn with the TRn address can now be applied, no other command is allowed at this point. All  
other banks must remain inactive from when the first BAn ACT command is issued until [(1.5 * tRAS) + tRP] is  
satisfied.  
5. After the first ACT to the BAn with the TRn address is issued, a PRE to BAn is to be issued (1.5 * tRAS) later; and  
then followed tRP later by the second ACT to the BAn with the TRn address. Once the 2nd activate to the BAn is  
issued, nonBAn banks are allowed to have activity.  
6. After the second ACT to the BAn with the TRn address is issued, a PRE to BAn is to be issued tRAS later and then  
followed tRP later by the third ACT to the BAn with the TRn address.  
7. After the third ACT to the BAn with the TRn address is issued, a PRE to BAn would be issued tRAS later; and once  
the third PRE has been issued, nonBAn banks are not allowed to have activity until TRR mode is exited. The TRR  
mode is completed once tRP plus tMRD is satisfied.  
8. TRR mode must be completed as specified to guarantee that adjacent rows are refreshed. Anyꢀtime the TRR  
mode is interrupted and not completed, the interrupted TRR Mode must be cleared and then subsequently  
performed again. To clear an interrupted TRR mode, an MR24 change is required with setting MR24 [OP7=0],  
MR24 [OP6:4] are don’t care, followed by three PRE to BAn, tRP time in between each PRE command. The  
complete TRR sequence (Steps 2ꢀ7) must be then reꢀissued and completed to guarantee that the adjacent rows  
are refreshed.  
9. Refresh command to the LPDRR4 SDRAM or entering SelfꢀRefresh mode is not allowed while the DRAM is in  
TRR mode.  
Version 1.0  
314  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
TRR Mode  
Notes:  
1. TRn is targeted row.  
2. Bank BAn represents the bank in which the targeted row is located.  
3. TRR mode self-clears after tMRD + tRP measured from 3rd BAn precharge PRE3 at clock edge Th4.  
4. TRR mode or any other activity can be re-engaged after tRP + tMRD from 3rd BAn precharge PRE3.  
PRE_ALL also counts if issued instead of PREn. TRR mode is cleared by DRAM after PRE3 to the BAn bank.  
5. Activate commands to BAn during TRR mode do not provide refreshing support, i.e. the Refresh counter is unaffected.  
6. The DRAM must restore the degraded row(s) caused by excessive activation of the targeted row (TRn) necessary to meet refresh  
requirements.  
7. A new TRR mode must wait tMRD+tRP time after the third precharge.  
8. BAn may not be used with any other command.  
9. ACT and PRE are the only allowed commands to BAn during TRR Mode.  
10. Refresh commands are not allowed during TRR mode.  
11. All DRAM timings are to be met by DRAM during TRR mode such as tFAW. Issuing of ACT1, ACT2 and ACT3 counts towards tFAW budget.  
Version 1.0  
315  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Post Package Repair (PPR)  
LPDDR4 supports Fail Row address repair as optional feature and it is readable through MR25 OP[7:0] PPR  
provides simple and easy repair method in the system and Fail Row address can be repaired by the electrical  
programming of Electricalꢀfuse scheme.  
With PPR, LPDDR4 can correct 1Row per Bank.  
Electricalꢀfuse cannot be switched back to unꢀfused states once it is programmed. The controller should prevent  
unintended the PPR mode entry and repair.  
Fail Row Address Repair  
The following is procedure of PPR.  
1. Before entering 'PPR' mode, All banks must be Precharged  
2. Enable PPR using MR4 bit "OP4=1" and wait tMRD  
3. Issue ACT command with Fail Row address  
4. Wait tPGM to allow DRAM repair target Row Address internally then issue PRE  
5. Wait tPGM_Exit after PRE which allow DRAM to recognize repaired Row address RAn  
6. Exit PPR with setting MR4 bit "OP4=0"  
7. LPDDR4 will accept any valid command after tPGMPST  
8. Repeat steps in 'Reset Initialization with Stable Power' section  
9. In More than one fail address repair case, Repeat Step 2 to 8  
Once PPR mode is exited, to confirm if target row is repaired correctly, host can verify by writing data into the target  
row and reading it back after PPR exit with MR4 [OP4=0] and tPGMPST.  
Version 1.0  
316  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
The following Timing diagram shows PPR operation.  
PPR Timing  
Notes:  
1. During tPGM, any other commands (including refresh) are not allowed on each die.  
2. With one PPR command, only one row can be repaired at one time per die.  
3. When PPR procedure is done, reset command is required before normal operation.  
4. During PPR, memory contents is not refreshed and may be lost.  
5. Assert Reset_n below 0.2 X VDD2. Reset_n needs to be maintained LOW for minimum tPW_RESET. CKE must be pulled LOW at least  
10ns before deassserting Reset.  
6. After RESET command, follow steps 4 to 10 in ‘Voltage Ramp and Device Initialization’ section.  
PPR Timing Parameters  
Parameter  
Symbol  
tPGM  
Min  
1000  
15  
Max  
Unit  
ms  
ns  
PPR Programming Time  
PPR Exit Time  
-
-
-
-
tPGM_Exit  
tPGMPST  
tCKPGM  
New Address Setting time  
PPR Programming Clock  
50  
us  
1.25  
ns  
Version 1.0  
12/2018  
317  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Absolute Maximum DC Ratings  
Stresses greater than those listed may cause permanent damage to the device.  
This is a stress rating only, and functional operation of the device at these or any other conditions above those  
indicated in the operational sections of this specification is not implied. Exposure to absolute maximum rating  
conditions for extended periods may affect reliability.  
Parameter  
Symbol  
Min  
-0.4  
-0.4  
-0.4  
-0.4  
-55  
Max  
2.1  
Unit Notes  
V
V
1
1
1
V
V
V
supply voltage relative to V  
V
V
DD1  
DD2  
DDQ  
SS  
DD1  
1.5  
supply voltage relative to V  
SS  
DD2  
VDDQ  
VIN, VOUT  
TSTG  
1.5  
V
supply voltage relative to V  
SSQ  
1.5  
V
Voltage on any ball except V  
Storage Temperature  
relative to V  
DD1  
SS  
125  
°C  
2
Notes:  
1. See “Power-Ramp” for relationships between power supplies.  
2. Storage Temperature is the case surface temperature on the center/top side of the LPDDR4 device. For the measurement  
conditions, please refer to JESD51-2.  
Version 1.0  
318  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AC & DC Operating Conditions  
Recommended DC Operating Conditions  
DRAM  
Symbol  
VDD1  
Min  
1.70  
1.06  
0.57  
Typ  
1.80  
1.10  
0.6  
Max  
1.95  
1.17  
0.65  
Unit  
V
Notes  
1,2  
Core 1 Power  
Core 2 Power/Input Buffer Power  
I/O Buffer Power  
VDD2  
V
1,2,3  
2,3,4,5  
VDDQ  
V
Notes:  
1. VDD1 uses significantly less current than VDD2.  
2. The volttage range is for DC voltage only. DC is defined as the voltage supplied at the DRAM and is inclusive of all noise up to  
20MHz at the DRAM package ball.  
3. The voltage noise tolerance from DC to 20MHz exceeding a pk-pk tolerance of 45mV at the DRAM ball is not included in the  
TdlVW.  
4. VDDQ(max) may be extended to 0.67V as an option in case the operating clock frequency is equal or less than 800Mhz.  
5. Pull up, pull down and ZQ calibration tolerance spec is valid only in normal VDDQ tolerance range (0.57 V - 0.65 V).  
Input Leakage Current  
Parameter/Condition  
Symbol  
Min  
Max  
Unit  
Notes  
Input Leakage current  
-4  
4
uA  
1,2  
I
L
Notes:  
1. For CK, CK, DQ, CKE, CS, CA, ODT_CA and RESET. Any Input 0V ≤ VIN ≤ VDD2 (All other pins not under test = 0V).  
2. CA ODT is disabled for CK, CK, CS and CA.  
Input/Output Leakage Current  
Parameter/Condition  
Symbol  
Min  
Max  
Unit  
Notes  
Input/Output Leakage current  
IOZ  
-5  
5
uA  
1,2  
Notes:  
1. For DQ, DQS, DQS and DMI. Any I/O 0V ≤ VOUT ≤ VDDQ.  
2. I/Os status are disabled: High Impedance and ODT off.  
Version 1.0  
319  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Operating Temperature Range  
Parameter/Condition  
Symbol  
Min  
Max  
Unit  
Standard  
-40  
85  
C°  
T
OPER  
Notes:  
1. Operating Temperature is the case surface temperature on the center-top side of the LPDDR4 device. For the measurement  
conditions, please refer to JESD51-2.  
2. Either the device case temperature rating or the temperature sensor (See “Temperature Sensor”) may be used to set an  
appropriate refresh rate, determine the need for AC timing de-rating and/or monitor the operating temperature. When using the  
temperature sensor, the actual device case temperature may be higher than the TOPER rating that applies for the Standard or  
Extended Temperature Ranges. For example, TCASE may be above 85°C when the temperature sensor indicates a temperature of  
less than 85 °C.  
Version 1.0  
320  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AC and DC Input/Output Measurement levels  
1.1V High speed LVCMOS (HS_LLVCMOS)  
Standard specifications  
All voltages are referenced to ground except where noted.  
DC electrical characteristics  
LPDDR4 Input Level for CKE  
This definition applies to CKE_A/B.  
Parameter  
Symbol  
VIH(AC)  
VIL(AC)  
VIH(DC)  
VIL(DC)  
Min  
Max  
V +0.2  
DD2  
Unit Notes  
Input high level (AC)  
Input low level (AC)  
Input high level (DC)  
Input low level (DC)  
V
V
V
V
1
1
0.75*V  
-0.2  
DD2  
0.25*V  
DD2  
0.65* V  
-0.2  
V
DD2  
+0.2  
DD2  
0.35*V  
DD2  
Notes:  
1. Refer LPDDR4 AC Over/Undershoot section  
Classꢀ1 LPDDR4 Input AC timing definition for CKE  
Notes:  
1. AC level is guaranteed transition point.  
2. DC level is hysteresis.  
LPDDR4 Input Level for Reset and ODT_CA  
This definition applies to Reset and ODT_CA  
Parameter  
Symbol  
VIH(AC)  
VIL(AC)  
Min  
Max  
V +0.2  
DD2  
Unit Notes  
Input high level (AC)  
Input low level (AC)  
V
V
1
1
0.80*V  
-0.2  
DD2  
0.20*V  
DD2  
Notes:  
1. Refer LPDDR4 AC Over/Undershoot section  
LPDDR4 Input AC timing definition for Reset and ODT_CA  
Version 1.0  
321  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
AC Over/Undershoot  
LPDDR4 AC Over/Undershoot  
Parameter  
Specification  
Maximum peak Amplitude allowed for overshoot area  
Maximum peak Amplitude allowed for undershoot area  
Maximum overshoot area above VDD/VDDQ  
0.35V  
0.35V  
0.8V-ns  
0.8V-ns  
Maximum undershoot area below VSS/VSSQ  
AC Overshoot and Undershoot Definition for Address and Control Pins  
Version 1.0  
322  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Differential Input Voltage  
Differential Input Voltage for CK  
CK Differential Input Voltage  
CK differential input voltage  
Data Rate  
3733  
Parameter  
Symbol  
Unit Notes  
Min  
Max  
CK differential input voltage  
Vindiff_CK  
360  
-
mV  
1
Notes:  
1. The peak voltage of Differential CK signals is calculated in a following equation.  
Vindiff_CK = (Max Peak Voltage) - (Min Peak Voltage)  
Max Peak Voltage = Max(f(t))  
Min Peak Voltage = Min(f(t))  
f(t) = VCK - VCK  
Version 1.0  
323  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Peak voltage calculation method  
The peak voltage of Differential Clock signals are calculated in a following equation.  
VIH.DIFF.Peak Voltage = Max(f(t))  
VIL.DIFF.Peak Voltage = Min(f(t))  
f(t) = VCK ꢀ VCK  
Definition of differential Clock Peak Voltage  
Notes:  
1. VREFCA is LPDDR4 SDRAM internal setting value by VREF Training.  
Version 1.0  
324  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
SingleꢀEnded Input Voltage for Clock  
The minimum input voltage need to satisfy both Vinse_CK, Vinse_CK_High/Low specification at input  
receiver.  
Clock SingleꢀEnded Input Voltage  
Notes:  
1. VREFCA is LPDDR4 SDRAM internal setting value by VREF Training.  
Clock SingleꢀEnded input voltage  
Data Rate  
Parameter  
Symbol  
3733  
Unit Notes  
Min  
Max  
Clock Single-Ended  
input voltage  
Vinse_CK  
180  
-
mV  
mV  
mV  
Clock Single-Ended input  
voltage High from VREFDQ  
Clock Single-Ended input  
voltage Low from VREFDQ  
Vinse_CK_High  
Vinse_CK_Low  
90  
90  
-
-
Version 1.0  
12/2018  
325  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Differential Input Slew Rate Definition for Clock  
Input slew rate for differential signals (CK, CK) are defined and measured as shown in the following figure and table.  
Differential Input Slew Rate Definition for CK, CK  
Notes:  
1. Differential signal rising edge from VILdiff_CK to VIHdiff_CK must be monotonic slope.  
2. Differential signal falling edge from VIHdiff_CK to VILdiff_CK must be monotonic slope.  
Differential Input Slew Rate Definition for CK, CK  
Measured  
Description  
Defined by  
From  
To  
Differential input slew rate for  
rising edge(CK - CK)  
Differential input slew rate for  
falling edge(CK - CK)  
VILdiff_CK VIHdiff_CK  
VIHdiff_CK VILdiff_CK  
|VILdiff_CK - VIHdiff_CK|/DeltaTRdiff  
|VILdiff_CK - VIHdiff_CK|/DeltaTFdiff  
Differential Input Level for CK, CK  
Data Rate  
3733  
Parameter  
Symbol  
Unit Notes  
Min  
145  
-
Max  
-
Differential Input High  
Differential Input Low  
VIHdiff_CK  
VILdiff_CK  
mV  
mV  
-145  
Differential Input Slew Rate for CK, CK  
Data Rate  
3733  
Parameter  
Symbol  
Unit Notes  
Min  
Max  
Differential Input  
Slew Rate for Clock  
SRIdiff_CK  
2
14  
V/ns  
Version 1.0  
12/2018  
326  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Differential Input Cross Point Voltage  
The cross point voltage of differential input signals (CK, CK) must meet the requirements in the following table.  
The differential input cross point voltage VIX is measured from the actual cross point of true and complement signals  
to the mid level that is V  
CA.  
REF  
Vix Definition (Clock)  
Notes:  
1. The base level of Vix_CK_FR/RF is VREFCA that is LPDDR4 SDRAM internal setting value by VREF Training.  
Corss point voltage for differential input signals (Clock)  
Data Rate  
Parameter  
Symbol  
3733  
Unit Notes  
Min  
Max  
Clock Differential input  
cross point voltage ratio  
Vix_CK_ratio  
-
25  
%
1,2  
Notes:  
1. Vix_CK_Ratio is defined by this equation: Vix_CK_Ratio = Vix_CK_FR/|Min(f(t))|  
2. Vix_CK_Ratio is defined by this equation: Vix_CK_Ratio = Vix_CK_RF/Max(f(t))  
Version 1.0  
327  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Differential Input Voltage for DQS  
The minimum input voltage need to satisfy both Vindiff_DQS and Vindiff_DQS /2 specification at input  
receiver and their measurement period is 1UI(tCK/2). Vindiff_DQS is the peak to peak voltage centered on  
0 volts differential and Vindiff_DQS /2 is max and min peak voltage from 0V.  
DQS Differential Input Voltage  
DQS differential input voltage  
Data Rate  
3733  
Parameter  
Symbol  
Unit Notes  
Min  
Max  
DQS differential input voltage Vindiff_DQS  
Notes:  
340  
-
mV  
1
1. The peak voltage of Differential DQS signals is calculated in a following equation.  
Vindiff_DQS = (Max Peak Voltage) - (Min Peak Voltage)  
Max Peak Voltage = Max(f(t))  
Min Peak Voltage = Min(f(t))  
f(t) = VDQS - VDQS  
Version 1.0  
328  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Peak voltage calculation method  
The peak voltage of Differential Clock signals are calculated in a following equation.  
VIH.DIFF.Peak Voltage = Max(f(t))  
VIL.DIFF.Peak Voltage = Min(f(t))  
f(t) = VDQS ꢀ VDQS  
Definition of differential DQS Peak Voltage  
Notes:  
1. VrefDQ is LPDDR4 SDRAM internal setting value by Vref Training.  
Version 1.0  
329  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
SingleꢀEnded Input Voltage for DQS  
The minimum input voltage need to satisfy both Vinse_DQS, Vinse_DQS_High/Low specification at input  
receiver.  
DQS SingleꢀEnded Input Voltage  
Notes:  
1. VrefDQ is LPDDR4 SDRAM internal setting value by Vref Training.  
DQS SingleꢀEnded Input voltage  
Data Rate  
Parameter  
Symbol  
3733  
Unit Notes  
Min  
Max  
DQS Single-Ended  
input voltage  
Vinse_DQS  
170  
-
mV  
mV  
mV  
DQS Single-Ended input  
voltage High from VREFDQ  
DQS Single-Ended input  
voltage Low from VREFDQ  
Vinse_DQS_High  
Vinse_DQS_Low  
85  
85  
-
-
Version 1.0  
330  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Differential Input Slew Rate Definition for DQS  
Input slew rate for differential signals (DQS, DQS) are defined and measured as shown in the following figure and  
table.  
Differential Input Slew Rate Definition for DQS, DQS  
Notes:  
1. Differential signal rising edge from VILdiff_DQS to VIHdiff_DQS must be monotonic slope.  
2. Differential signal falling edge from VIHdiff_DQS to VILdiff_DQS must be monotonic slope.  
Differential Input Slew Rate Definition for DQS, DQS  
Measured  
Description  
Defined by  
From  
To  
Differential input slew rate for  
rising edge(DQS - DQS)  
Differential input slew rate for  
falling edge(DQS - DQS)  
VILdiff_DQS VIHdiff_DQS |VILdiff_DQS - VIHdiff_DQS|/DeltaTRdiff  
VIHdiff_DQS VILdiff_DQS |VILdiff_DQS - VIHdiff_DQS|/DeltaTFdiff  
Differential Input Level for DQS, DQS  
Data Rate  
Parameter  
Symbol  
3733  
Unit Notes  
Min  
120  
-
Max  
-
Differential Input High  
Differential Input Low  
VIHdiff_DQS  
VILdiff_DQS  
mV  
mV  
-120  
Differential Input Slew Rate for DQS, DQS  
Data Rate  
3733  
Parameter  
Symbol  
Unit Notes  
Min  
Max  
Differential Input  
Slew Rate for Clock  
SRIdiff  
2
14  
V/ns  
Differential Input Cross Point Voltage  
The cross point voltage of differential input signals (DQS, DQS) must meet the requirements in following table.  
The differential input cross point voltage VIX is measured from the actual cross point of true and complement signals  
to the mid level that is V  
DQ.  
REF  
Version 1.0  
331  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Vix Definition (DQS)  
Notes:  
1. The base level of Vix_DQS_FR/RF is VrefDQ that is LPDDR4 SDRAM internal setting value by Vref Training.  
Differential Input Slew Rate for DQS, DQS  
Data Rate  
Parameter  
Symbol  
3733  
Unit Notes  
Min  
Max  
DQS Differential input  
cross point voltage ratio  
Vix_DQS_ratio  
-
20  
%
1,2  
Notes:  
1. Vix_DQS_Ratio is defined by this equation: Vix_DQS_Ratio = Vix_DQS_FR/|Min(f(t))|  
2. Vix_DQS_Ratio is defined by this equation: Vix_DQS_Ratio = Vix_DQS_RF/Max(f(t))  
Version 1.0  
332  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
LPDDR4 Input Level for ODT(ca) input  
Symbol  
Min  
Max  
+0.2  
DD2  
Unit  
Notes  
VIHODT  
VILODT  
ODT Input High Level  
ODT Input Low Level  
0.75*V  
-0.2  
V
V
V
DD2  
0.25*V  
DD2  
Version 1.0  
12/2018  
333  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Single Ended Output Slew Rate  
Single Ended Output Slew Rate Definition  
Output Slew Rate (singleꢀended) for 0.6V V  
DDQ  
Value  
Parameter  
Symbol  
Unit  
1
2
Min  
3.0  
Max  
SRQse  
-
9
V/ns  
-
Single-ended Output Slew Rate (VOH = V  
*0.5)  
DDQ  
Output slew-rate matching Ratio (Rise to Fall)  
Description:  
0.8  
1.2  
SR- Slew Rate  
Q- Query Output (like in DQ, which stands for Data-in, Query-Output)  
Se- Single-ended Signals  
Notes:  
1. Measured with output reference load.  
2. The ratio of pull-up to pull-down slew rate is specified for the same temperature and voltage, over the entire temperature and voltage  
range. For a given output, it represents the maximum difference between pull-up and pull-down drivers due to process variation.  
3. The output slew rate for falling and rising edges is defined and measured between VOL(AC)=0.2*VOH(DC) and VOH(AC)= 0.8*VOH(DC).  
4. Slew rates are measured under average SSO conditions, with 50% of DQ signals per data byte switching.  
Version 1.0  
334  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Differential Output Slew Rate  
Differential Output Slew Rate Definition  
Differential Output Slew Rate for 0.6V V  
DDQ  
Value  
Parameter  
Symbol  
Unit  
1
2
Min  
Max  
18  
SRQdiff  
6
V/ns  
Differential Output Slew Rate (VOH = V  
*0.5)  
DDQ  
Description:  
SR- Slew Rate  
Q- Query Output (like in DQ, which stands for Data-in, Query-Output)  
Se- Single-ended Signals  
Notes:  
1. Measured with output reference load.  
2. The output slew rate for falling and rising edges is defined and measured between VOL(AC)=0.2*VOH(DC) and VOH(AC)= 0.8*VOH(DC).  
3. Slew rates are measured under average SSO conditions, with 50% of DQ signals per data byte switching.  
Version 1.0  
335  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Overshoot and Undershoot for LVSTL  
AC Overshoot/Undershoot Specification  
Data Rate  
3733  
0.3  
Parameter  
Unit  
Maximum peak amplitude allowed for overshoot area.  
Maximum peak amplitude allowed for undershoot area.  
Maximum area above VDD.  
Max  
V
Max  
Max  
Max  
0.3  
V
0.1  
V-ns  
V-ns  
Maximum area below VSS.  
0.1  
Notes:  
1. VDD2 stands for VDD for CA[5:0], CK, CK, CS, CKE and ODT. VDD stands for VDDQ for DQ, DMI, DQS and DQS.  
2. VSS stands for VSS for CA[5:0], CK, CK, CS, CKE and ODT. VSS stands for VSSQ for DQ, DMI, DQS and DQS.  
3. Maximum peak amplitude values are referenced from actual VDD and VSS values.  
4. Maximum area values are referenced from maximum operating VDD and VSS values.  
Overshoot and Undershoot Definition  
Version 1.0  
336  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
LPDDR4 Driver Output Timing Reference load  
These 'Timing Reference Loads’ are not intended as a precise representation of any particular system environment  
or a depiction of the actual load presented by a production tester. System designers should use IBIS or other  
simulation tools to correlate the timing reference load to a system environment. Manufacturers correlate to their  
production test conditions, generally one or more coaxial transmission lines terminated at the tester electronics.  
Driver Output Reference Load for Timing and Slew Rate  
Notes:  
1. All output timing parameter values are reported with respect to this reference load.  
This reference load is also used to report slew rate.  
Version 1.0  
337  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
LVSTL (Low Voltage Swing Terminated Logic) IO System  
LVSTL I/O cell is comprised of pullꢀup, pullꢀdown dirver and a terminator. The basic cell is shown in figure below.  
LVSTL I/O Cell  
Version 1.0  
338  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
To ensure that the target impedance is achived the LVSTL I/O cell is designed to calibrated as below procedure.  
1. First calibrate the pullꢀdown device against a 240 Ohm resister to VDDQ via the ZQ pin.  
• Set Strength Control to minimum setting.  
• Increase drive strength until comparator detects data bit is less than VDDQ/2.  
• NMOS pullꢀdown device is calibrated to 240 Ohms.  
pullꢀdown calibration  
2. Then calibrate the pullꢀup device against the calibrated pullꢀdown device.  
• Set VOH target and NMOS controller ODT replica via MRS  
(VOH can be automatically controlled by ODT MRS).  
• Set Strength Control to minimum setting.  
• Increase drive strength until comparator detects data bit is grater than VOH target.  
• NMOS pullꢀup device is now calibrated to VOH target.  
pullꢀup calibration  
Version 1.0  
339  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Input / Output Capacitance  
Symbol  
Parameter  
Min  
0.5  
0.0  
0.5  
-0.1  
0.7  
0.0  
Max  
Unit  
pF  
Input capacitance :  
0.9  
0.09  
0.9  
C
CK  
CK, CK  
Input capacitance delta :  
CK, CK  
pF  
C
DCK  
Input capacitance:  
pF  
C
I
all other input-only pins  
Input capacitance delta:  
all other input-only pins  
Input/output capacitance :  
DQ, DQS, DQS, DMI  
0.1  
pF  
C
DI  
1.3  
pF  
C
IO  
Input/output capacitance delta :  
DQS, DQS  
0.1  
pF  
C
DDQS  
Input/output capacitance delta :  
DQ, DMI  
-0.1  
0.0  
0.1  
0.5  
pF  
pF  
C
DIO  
Input/output capacitance : ZQ  
C
ZQ  
Notes:  
1. This parameter applies to die devices only (does not include package capacitance).  
2. This parameter is not subject to production testing. It is verified. The capacitance is measured according to JEP147  
(procedure for measuring input capacitance using a vector network analyzer), with VDD1, VDD2, VDDQ, VSS, VSSQ applied  
and all other pins floating.  
3. Absolute value of CCK - CCK.  
4. CI applies to CS, CKE, and CA[5:0].  
5. CDI = CI – 0.5 × (CCK + CCK)  
6. DMI loading matches DQ and DQS.  
7. Absolute value of CDQS and CDQS.  
8. CDIO = CIO – 0.5 × (CDQS + CDQS) in byte-lane.  
Version 1.0  
340  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
IDD Specification Parameters and Test Conditions  
IDD Measurement Conditions  
The following definitions and conditions are used in the IDD measurement tables unless stated otherwise:  
• LOW: VIN ≤ VIL(DC)max  
• HIGH: VIN ≥ VIH(DC)min  
• STABLE: Inputs are stable at a HIGH or LOW level  
• SWITCHING: See Tables bellow  
Definition of Switching for CA Input Signal  
CK edge  
CKE  
R1  
HIGH  
LOW  
H
R2  
R3  
R4  
R5  
R6  
R7  
R8  
HIGH  
LOW  
H
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
CS  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
CA0  
L
H
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
H
L
CA1  
H
H
CA2  
H
H
L
H
L
H
CA3  
H
H
L
H
L
H
CA4  
H
H
L
H
L
H
CA5  
H
H
H
H
Notes:  
1. CS must always be driven HIGH.  
2. 50% of CA bus is changing between HIGH and LOW once per clock for the CA bus.  
3. The above pattern is used continuously during IDD measurement for IDD values that require switching on the CA bus.  
Version 1.0  
341  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
IDD Measurement Conditions (Continued)  
CA pattern for IDD4R for BL=16  
Clock Cycle  
CKE  
CS  
Command  
CA0  
CA1  
CA2  
CA3  
CA4  
CA5  
Number  
N
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
Read-1  
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
H
H
H
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
L
L
L
L
N+1  
N+2  
CAS-2  
H
L
N+3  
N+4  
DES  
DES  
L
L
N+5  
L
L
N+6  
DES  
L
L
N+7  
DES  
L
L
N+8  
Read-1  
H
H
H
H
L
L
N+9  
H
H
H
L
N+10  
N+11  
N+12  
N+13  
N+14  
N+15  
CAS-2  
DES  
DES  
DES  
DES  
L
L
L
L
L
L
Notes:  
1. BA[2:0]=010, CA[9:4]=000000 or 111111, Burst Order CA[3:2]=00 OR 11(Same as LRDDR3 IDD4R Spec)  
2. Difference from LPDDR3 Spec : CA pins are kept low with DES CMD to reduce ODT current.  
Version 1.0  
342  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CA pattern for IDD4W for BL=16  
Clock Cycle  
Number  
N
CKE  
CS  
Command  
CA0  
CA1  
CA2  
CA3  
CA4  
CA5  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
L
H
L
L
L
L
L
L
L
H
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
L
L
L
L
Write-1  
N+1  
N+2  
H
L
CAS-2  
N+3  
N+4  
DES  
DES  
DES  
DES  
L
L
N+5  
L
L
N+6  
L
L
N+7  
L
L
N+8  
L
L
Write-1  
CAS-2  
N+9  
H
H
L
H
H
H
L
N+10  
N+11  
N+12  
N+13  
N+14  
N+15  
Notes:  
DES  
DES  
DES  
DES  
L
L
L
L
L
L
L
1. BA[2:0]=010, CA[9:4]=000000 or 111111 (Same as LRDDR3 IDD4W Spec)  
2. Difference from LPDDR3 Spec :  
No burst ordering  
CA pins are kept low with DES CMD to reduce ODT current.  
Version 1.0  
343  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4W (DBI off) for BL=16  
DBI OFF Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
2
4
6
4
8
4
0
4
2
4
6
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
No. of 1’s  
Notes:  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
16  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
16  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
16  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
16  
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
16  
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
16  
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
16  
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
16  
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
4
2
4
0
4
8
4
2
4
6
4
8
4
0
4
1. Simplified pattern compared with last showing.  
Same data pattern was applied to DQ[4], DQ[5], DQ[6], DQ[7] for reducing complexity for IDD4W/R pattern programming.  
Version 1.0  
344  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4R (DBI off) for BL=16  
DBI OFF Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
2
4
6
4
8
4
0
4
2
4
6
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
No. of 1’s  
Notes:  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
16  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
16  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
16  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
16  
1
0
0
1
1
0
0
1
0
1
1
0
0
1
1
0
16  
1
0
0
1
1
0
0
1
0
1
1
0
0
1
1
0
16  
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
16  
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
16  
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
6
4
2
4
0
4
8
4
2
4
6
4
1. Same data pattern was applied to DQ[4], DQ[5], DQ[6], DQ[7] for reducing complexity for IDD4W/R pattern programming.  
Version 1.0  
345  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4W (DBI on) for BL=16  
DBI ON Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
4
0
4
2
4
3
4
1
4
0
4
2
4
3
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
No. of 1’s  
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
8
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
8
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
8
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
8
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
8
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
8
1
0
1
1
0
1
0
0
1
1
1
0
0
0
0
1
16  
1
0
1
1
0
1
0
0
1
1
1
0
0
0
0
1
16  
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
8
3
4
2
4
0
4
1
4
2
4
3
4
1
4
0
4
DBI enabled burst  
Version 1.0  
346  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4R (DBI on) for BL=16  
DBI ON Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
4
0
4
2
4
3
4
1
4
0
4
2
4
3
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
No. of 1’s  
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
8
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
8
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
8
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
8
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
8
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
16  
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
16  
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
8
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
4
0
4
3
4
2
4
0
4
1
4
2
4
3
4
Version 1.0  
347  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
IDD Measurement Conditions (Continued)  
CA pattern for IDD4R for BL=32  
Clock Cycle  
CKE  
CS  
Command  
CA0  
CA1  
CA2  
CA3  
CA4  
CA5  
Number  
N
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
Read-1  
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
H
H
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
L
L
L
L
L
L
L
L
L
L
L
L
N+1  
N+2  
CAS-2  
N+3  
N+4  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
Read-1  
L
N+5  
L
N+6  
L
N+7  
L
N+8  
L
N+9  
L
N+10  
N+11  
N+12  
N+13  
N+14  
N+15  
N+16  
N+17  
N+18  
N+19  
N+20  
N+21  
N+22  
N+23  
N+24  
N+25  
N+26  
N+27  
N+28  
N+29  
N+30  
N+31  
L
L
L
L
L
L
H
H
H
H
L
CAS-2  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
L
L
L
L
L
L
L
L
L
L
L
Notes:  
1. BA[2:0]=010, CA[9:4]=000000 or 111111, Burst Order CA[4:2]=000 OR 111  
Version 1.0  
348  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CA pattern for IDD4W for BL=32  
Clock Cycle  
Number  
CKE  
CS  
Command  
CA0  
CA1  
CA2  
CA3  
CA4  
CA5  
N
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
HIGH  
LOW  
HIGH  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
LOW  
Write-1  
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
H
H
L
L
L
L
L
L
L
L
L
L
L
L
N+1  
N+2  
CAS-2  
N+3  
N+4  
DES  
DES  
N+5  
N+6  
DES  
N+7  
DES  
N+8  
DES  
N+9  
DES  
N+10  
N+11  
N+12  
N+13  
N+14  
N+15  
N+16  
N+17  
N+18  
N+19  
N+20  
N+21  
N+22  
N+23  
N+24  
N+25  
N+26  
N+27  
N+28  
N+29  
N+30  
N+31  
DES  
DES  
DES  
DES  
DES  
DES  
Write-1  
CAS-2  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
DES  
Notes:  
1. BA[2:0]=010, CA[9:4]=000000 or 111111  
Version 1.0  
349  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4W (DBI off) for BL=32  
DBI OFF Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
2
4
6
4
8
4
0
4
2
4
6
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
4
2
4
0
4
8
4
2
4
6
4
8
4
0
4
Version 1.0  
350  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4W (DBI off) for BL=32 (Cont’d)  
DBI OFF Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL32  
BL33  
BL34  
BL35  
BL36  
BL37  
BL38  
BL39  
BL40  
BL41  
BL42  
BL43  
BL44  
BL45  
BL46  
BL47  
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
2
4
6
4
8
4
0
4
2
4
6
4
BL48  
BL49  
BL50  
BL51  
BL52  
BL53  
BL54  
BL55  
BL56  
BL57  
BL58  
BL59  
BL60  
BL61  
BL62  
BL63  
No. of 1’s  
Notes:  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
32  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
32  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
32  
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
32  
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
32  
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
32  
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
32  
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
32  
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
4
2
4
0
4
8
4
2
4
6
4
8
4
0
4
1. Simplified pattern compared with last showing.  
Same data pattern was applied to DQ[4], DQ[5], DQ[6], DQ[7] for reducing complexity for IDD4W/R pattern programming.  
Version 1.0  
12/2018  
351  
Nanya Technology Corp.  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4R (DBI off) for BL=32  
DBI OFF Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
2
4
6
4
8
4
0
4
2
4
6
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
1
0
0
1
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
1
0
0
1
0
1
1
0
0
1
1
0
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
6
4
2
4
0
4
8
4
2
4
6
4
Version 1.0  
352  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4R (DBI off) for BL=32 (Cont’d)  
DBI OFF Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL32  
BL33  
BL34  
BL35  
BL36  
BL37  
BL38  
BL39  
BL40  
BL41  
BL42  
BL43  
BL44  
BL45  
BL46  
BL47  
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
2
4
6
4
8
4
0
4
2
4
6
4
BL48  
BL49  
BL50  
BL51  
BL52  
BL53  
BL54  
BL55  
BL56  
BL57  
BL58  
BL59  
BL60  
BL61  
BL62  
BL63  
No. of 1’s  
Notes:  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
32  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
32  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
32  
1
1
0
0
1
1
0
0
0
0
1
1
0
0
1
1
32  
1
0
0
1
1
0
0
1
0
1
1
0
0
1
1
0
32  
1
0
0
1
1
0
0
1
0
1
1
0
0
1
1
0
32  
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
32  
1
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
32  
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
4
0
4
6
4
2
4
0
4
8
4
2
4
6
4
1. Same data pattern was applied to DQ[4], DQ[5], DQ[6], DQ[7] for reducing complexity for IDD4W/R pattern programming.  
Version 1.0  
353  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4W (DBI on) for BL=32  
DBI ON Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
4
0
4
2
4
3
4
1
4
0
4
2
4
3
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
0
0
0
1
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
3
4
2
4
0
4
1
4
2
4
3
4
1
4
0
4
Version 1.0  
354  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4W (DBI on) for BL=32 (Cont’d)  
DBI ON Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL32  
BL33  
BL34  
BL35  
BL36  
BL37  
BL38  
BL39  
BL40  
BL41  
BL42  
BL43  
BL44  
BL45  
BL46  
BL47  
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
4
0
4
2
4
3
4
1
4
0
4
2
4
3
4
BL48  
BL49  
BL50  
BL51  
BL52  
BL53  
BL54  
BL55  
BL56  
BL57  
BL58  
BL59  
BL60  
BL61  
BL62  
BL63  
No. of 1’s  
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
16  
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
16  
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
16  
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
16  
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
16  
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
16  
1
0
1
1
0
1
0
0
1
1
1
0
0
0
0
1
32  
1
0
1
1
0
1
0
0
1
1
1
0
0
0
0
1
32  
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
16  
3
4
2
4
0
4
1
4
2
4
3
4
1
4
0
4
DBI enabled burst  
Version 1.0  
355  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4R (DBI on) for BL=32  
DBI ON Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL0  
BL1  
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
4
0
4
2
4
3
4
1
4
0
4
2
4
3
4
BL2  
BL3  
BL4  
BL5  
BL6  
BL7  
BL8  
BL9  
BL10  
BL11  
BL12  
BL13  
BL14  
BL15  
BL16  
BL17  
BL18  
BL19  
BL20  
BL21  
BL22  
BL23  
BL24  
BL25  
BL26  
BL27  
BL28  
BL29  
BL30  
BL31  
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
1
4
0
4
3
4
2
4
0
4
1
4
2
4
3
4
Version 1.0  
356  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Data pattern for IDD4R (DBI on) for BL=32 (Cont’d)  
DBI ON Case  
DQ[7]  
DQ[6]  
DQ[5]  
DQ[4]  
DQ[3]  
DQ[2]  
DQ[1]  
DQ[0]  
DBI  
No. of 1’s  
BL32  
BL33  
BL34  
BL35  
BL36  
BL37  
BL38  
BL39  
BL40  
BL41  
BL42  
BL43  
BL44  
BL45  
BL46  
BL47  
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
4
0
4
2
4
3
4
1
4
0
4
2
4
3
4
BL48  
BL49  
BL50  
BL51  
BL52  
BL53  
BL54  
BL55  
BL56  
BL57  
BL58  
BL59  
BL60  
BL61  
BL62  
BL63  
No. of 1’s  
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
16  
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
16  
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
16  
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
16  
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
16  
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
16  
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
32  
0
0
0
1
1
0
1
1
0
1
0
0
1
1
1
0
32  
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
16  
1
4
0
4
3
4
2
4
0
4
1
4
2
4
3
4
Version 1.0  
357  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
IDD Specifications  
IDD values are for the entire operating voltage range, and all of them are for the entire standard range, with the  
exception of IDD6ET which is for the entire elevated temperature range.  
LPDDR4 IDD Specification Parameters and Operating Conditions  
Parameter/Condition  
Symbol  
Power Supply  
Notes  
Operating one bank active-precharge current:  
tCK = tCKmin; tRC = tRCmin;  
CKE is HIGH;   
IDD0  
IDD0  
VDD  
VDD  
1
1
CS is HIGH between valid commands;   
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
2
2
3
IDD0  
VDD  
Q
Q
Idle power-down standby current:   
tCK = tCKmin;   
IDD2P  
VDD  
1
2
1
2
CKE is LOW;   
CS is LOW;   
IDD2P  
VDD  
All banks are idle;   
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
3
3
3
3
IDD2P  
VDD  
Q
Q
Idle power-down standby current with clock stop:   
CK =LOW, CK =HIGH;   
CKE is LOW;   
IDD2PS  
IDD2PS  
VDD  
VDD  
1
1
CS is LOW;   
2
2
All banks are idle;   
CA bus inputs are STABLE;   
Data bus inputs are STABLE;  
ODT disabled  
IDD2PS  
VDD  
Q
Q
Idle non power-down standby current:   
tCK = tCKmin;   
IDD2N  
IDD2N  
VDD  
VDD  
1
1
CKE is HIGH;   
CS is LOW;   
2
2
All banks are idle;   
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
IDD2N  
VDD  
Q
Q
Idle non power-down standby current with clock stoped:  
CK =LOW, CK =HIGH;   
CKE is HIGH;   
IDD2NS  
IDD2NS  
VDD  
VDD  
1
2
1
2
CS is LOW;   
All banks are idle;   
CA bus inputs are STABLE;   
Data bus inputs are STABLE;  
ODT disabled  
IDD2NS  
VDD  
Q
Q
Version 1.0  
358  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Parameter/Condition  
Symbol  
Power Supply  
Notes  
Active power-down standby current:   
IDD3P  
VDD  
1
2
1
2
tCK = tCKmin;   
CKE is LOW;   
CS is LOW;   
IDD3P  
VDD  
One bank is active;   
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
3
IDD3P  
VDD  
Q
Q
Active power-down standby current with clock stop:   
CK=LOW, CK =HIGH;   
CKE is LOW;   
IDD3PS  
IDD3PS  
VDD  
VDD  
1
1
CS is LOW;   
2
2
One bank is active;   
CA bus inputs are STABLE;   
Data bus inputs are STABLE;  
ODT disabled  
4
4
4
5
IDD3PS  
VDD  
Q
Q
Active non power-down standby current:   
tCK = tCKmin;   
IDD3N  
IDD3N  
VDD  
VDD  
1
1
CKE is HIGH;   
CS is LOW;   
2
2
One bank is active;   
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
IDD3N  
VDD  
Q
Q
Active non power-down standby current with clock stoped:  
CK=LOW, CK =HIGH;   
IDD3NS  
IDD3NS  
VDD  
VDD  
1
2
1
2
CKE is HIGH;   
CS is LOW;   
One bank is active;   
CA bus inputs are STABLE;   
Data bus inputs are STABLE;  
ODT disabled  
IDD3NS  
VDD  
Q
Q
Operating burst read current:   
tCK = tCKmin;   
VDD1  
VDD2  
IDD4R  
1
CS is LOW between valid commands;   
One bank is active;   
IDD4R  
2
BL = 16 or 32; RL = RL(min);   
CA bus inputs are SWITCHING;   
50% data change each burst transfer  
ODT disabled  
IDD4R  
VDD  
Q
Q
Version 1.0  
359  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Parameter/Condition  
Symbol  
Power Supply  
Notes  
Operating burst write current:   
VDD1  
IDD4W  
1
2
tCK = tCKmin;   
CS is LOW between valid commands;   
One bank is active;   
IDD4W  
VDD  
2
BL = 16 or 32; WL = WL(min);   
CA bus inputs are SWITCHING;   
50% data change each burst transfer;  
ODT disabled  
4
IDD4W  
VDD  
Q
Q
All Bank Refresh Burst current:   
tCK = tCKmin;   
IDD5  
IDD5  
VDD  
VDD  
1
1
CKE is HIGH between valid commands;   
tRC = tRFCabmin;   
2
2
Burst refresh;  
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
4
4
4
IDD5  
VDD  
Q
Q
All Bank Refresh Average current:   
tCK = tCKmin;   
IDD5AB  
IDD5AB  
VDD  
VDD  
1
1
CKE is HIGH between valid commands;   
tRC = tREFI;   
2
2
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
IDD5AB  
VDD  
Q
Q
Per Bank Refresh Average current:   
tCK = tCKmin;   
IDD5PB  
IDD5PB  
VDD  
VDD  
1
2
1
2
CKE is HIGH between valid commands;   
tRC = tREFI/8;   
CA bus inputs are SWITCHING;   
Data bus inputs are STABLE;  
ODT disabled  
IDD5PB  
VDD  
Q
Q
Version 1.0  
360  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
IDD Specifications (Continued)  
Parameter/Condition  
Symbol  
Power Supply  
Notes  
Self refresh current (Standard Temperature Range):   
CK=LOW, CK=HIGH;  
6,7,8.10  
IDD6  
VDD  
1
2
1
2
CKE is LOW;   
6,7,8.10  
4,6,7,8.10  
7,8,11  
IDD6  
VDD  
CA bus inputs are STABLE;   
Data bus inputs are STABLE;  
Maximum 1x Self-Refresh Rate;  
ODT disabled  
IDD6  
VDD  
Q
Q
Self refresh current (Extended Temperature Range):   
CK=LOW, CK =HIGH;  
IDD6  
IDD6  
IDD6  
VDD  
VDD  
ET1  
ET2  
ETQ  
1
CKE is LOW;   
7,8,11  
CA bus inputs are STABLE;   
Data bus inputs are STABLE;  
Maximum 1x Self-Refresh Rate;  
ODT disabled  
2
4,7,8,11  
VDD  
Q
Notes:  
1. Published IDD values are the maximum of the distribution of the arithmetic mean and are measured at 85.  
2. ODT disabled: MR11[2:0] = 000B.  
3. IDD current specifications are tested after the device is properly initialized.  
4. Measured currents are the summation of VDDQ and VDD2.  
5. Guaranteed by design with output load of 5pf and RON = 40Ohm.  
6. The 1x self refresh rate is the rate at which the device is refreshed internally during self refresh, before going into the extended  
temperature range.  
7. This is the general definition that applies to full-array SELF REFRESH.  
8. Supplier datasheets may contain additional Self Refresh IDD values for temperature subranges within the Standard or elevated  
Temperature Range.  
9. For all IDD measurements, VIHCKE = 0.8 × VDD2, VILCKE = 0.2 × VDD2  
10. IDD6 85°C is guaranteed, IDD6 45°C is typical of the distribution of the arithmetic mean.  
11. IDD6ET is typical value, is sampled only, and is not tested  
12. Dual Channel devices are specified in dual channel operation (both channels operating together).  
Version 1.0  
361  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Reserved for IDD Specifications  
IDD Specifications and Measurement Conditions  
Version 1.0  
362  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Reserved for IDD Specifications  
IDD Specifications and Measurement Conditions  
IDD6 Partial Array Selfꢀrefresh current;  
Version 1.0  
363  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Electrical Characteristics and Recommended AC Timing  
Clock Specification  
The jitter specified is a random jitter meeting a Gaussian distribution. Input clocks violating the min/max values may result in  
malfunction of the LPDDR4 device.  
Definitions and Calculations  
Symbol  
Description  
Calculation  
Notes  
tCK(avg) is calculated as the average clock period across  
any consecutive 200 cycle window, where each clock  
period is calculated from rising edge to rising edge.  
Unit ‘tCK(avg)’ represents the actual clock average  
tCK(avg) of the input clock under operation. Unit ‘nCK’  
represents one clock cycle of the input clock, counting the  
actual clock edges.  
tCK(avg) and nCK  
tCK(avg) may change by up to +/ꢀ1% within a 100 clock  
cycle window, provided that all jitter and timingspecs are  
met.  
tCK(abs) is defined as the absolute clock period, as  
measured from one rising edge to the next consecutive  
rising edge.  
tCK(abs)  
tCK(abs) is not subject to production test.  
tCH(avg) is defined as the average high pulse width, as  
calculated across any consecutive 200 high pulses.  
tCH(avg)  
tCL(avg) is defined as the average low pulse width, as  
calculated across any consecutive 200 low pulses.  
tCL(avg)  
tCH(abs) is the absolute instantaneous clock high pulse  
width, as measured from one rising edge to the following  
falling edge.  
tCH(abs)  
tCH(abs) is not subject to production test.  
tCL(abs) is the absolute instantaneous clock low pulse  
width, as measured from one falling edge to the following  
rising edge.  
tCL(abs)  
tJIT(per)  
tCL(abs) is not subject to production test.  
tJIT(per) is the single period jitter defined as the largest  
deviation of any signal tCK from tCK(avg).  
tJIT(per) is not subject to production test.  
tJIT(per),act  
tJIT(per),act is the actual clock jitter for a given system.  
tJIT(per),allowed  
tJIT(per),allowed is the specified allowed clock period jitter.  
tJIT(cc) is defined as the absolute difference in clock period  
between two consecutive clock cycles.  
tJIT(cc)  
tJIT(cc) defines the cycle to cycle jitter.  
tJIT(cc) is not subject to production test.  
Version 1.0  
364  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Clock AC Timings  
(Data rate 3733 Specifications and conditions)  
Min/  
Max  
Parameter  
Clock Timing  
Symbol  
3733  
Unit  
0.535  
100  
ns  
Min  
Max  
Min  
Max  
Min  
Max  
Min  
Min  
Max  
Min  
Max  
Min  
Average Clock Period  
tCK(avg)1  
tCH(avg)  
ns  
0.46  
0.54  
0.46  
0.54  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
ps  
Average high pulse width  
Average low pulse width  
Absolute Clock Period  
tCL(avg)  
tCK(abs)  
tCH(abs)  
tCK(avg)min +  
tJIT(per)min  
0.43  
0.57  
0.43  
0.57  
-36  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
ps  
Absolute clock HIGH  
pulse width  
Absolute clock LOW  
pulse width  
tCL(abs)  
Clock period jitter  
tJIT(per)  
tJIT(cc)  
36  
-
ps  
ps  
ps  
Max  
Min  
Max  
Maximum Clock Jitter  
between consecutive  
cycles  
72  
NOTE1 Clock Frequency herewith is a reference base on JEDEC's. Precise tCK setting needs to follow where defined on speed compatible  
table in section “Operating frequency”, exceptional setting please confirm with NTC.  
Version 1.0  
365  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Temperature Derating for AC timing  
Data Rate  
3733  
Min/  
Max  
Parameter  
Symbol  
Unit  
DQS output access time from CK/CK (derated)  
RAS-to-CAS delay (derated)  
tDQSCK  
tRCD  
tRC  
Max  
Min  
Min  
Min  
Min  
Min  
3600  
ps  
ns  
ns  
ns  
ns  
ns  
tRCD + 1.875  
tRC + 3.75  
ACTIVATE-to- ACTIVATE command period (derated)  
Row active time (derated)  
tRAS  
tRP  
tRAS + 1.875  
tRP + 1.875  
tRRD + 1.875  
Row precharge time (derated)  
Active bank A to active bank B (derated)  
tRRD  
NOTE1 Timing derating applies for operation at 85°C to 105°C.  
Version 1.0  
366  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CA Rx voltage and timing  
The command and address (CA) including CS input receiver compliance mask for voltage and timing is shown in the  
figure below. All CA, CS signals apply the same compliance mask and operate in single data rate mode.  
The CA input receiver mask for voltage and timing is shown in the figure below is applied across all CA pins. The  
receiver mask (Rx Mask) defines the area that the input signal must not encroach in order for the DRAM input  
receiver to be expected to be able to successfully capture a valid input signal; it is not the valid dataꢀeye.  
CA Receiver (Rx) mask  
Across pin VREFCA voltage variation  
Vcent_CA(pin mid) is defined as the midpoint between the largest Vcent_CA voltage level and the smallest  
Vcent_CA voltage level across all CA and CS pins for a given DRAM component. This clarifies that any DRAM  
component level variation must be accounted for within the DRAM CA Rx mask. The component level VREF will be  
set by the system to account for Ron and ODT settings.  
Version 1.0  
367  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
CA Timings at the DRAM Pins  
All of the timing terms in figure below are measured from the CK/CK to the center(midpoint) of the TcIVW window  
taken at the VcIVW_total voltage levels centered around Vcent_CA(pin mid).  
CA TcIPW and SRIN_cIVW definition (for each input pulse)  
Notes:  
1. SRIN_cIVW=VcIVW_Total/(tr or tf), signal must be monotonic within tr and tf range.  
CA VIHL_AC definition (for each input pulse)  
Version 1.0  
368  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DRAM CMD/ADR, CS  
Speed Grade  
Min/  
Max  
Symbol  
Parameter  
Unit  
Note  
3733  
150  
VcIVW  
TcIVW  
Rx Mask voltage -p-p  
Rx timing window  
Max  
Max  
mV  
1,2,3  
0.3  
tck(avg)min 1,2,3  
CA AC input pulse amplitude  
pk-pk  
VIHL_AC  
TcIPW  
Min  
180  
mV  
4,7  
5
CA input pulse width  
Min  
Min  
Max  
0.6  
1
tck(avg)min  
SRIN_cIVW Input Slew Rate over VcIVW  
V/ns  
6
7
Notes:  
1. CA Rx mask voltage and timing parameters at the pin including voltage and temperature drift.  
2. Rx mask voltage VcIVW total(max) must be centered around Vcent_CA(pin mid).  
3. Defined over the CA internal VREF range. The Rx mask at the pin must be within the internal VREF CA range irrespective of the  
input signal common mode.  
4. CA only input pulse signal amplitude into the receiver must meet or exceed VIHL AC at any point over the total UI. No timing  
requirement above level. VIHL AC is the peak to peak voltage centered around Vcent_CA(pin mid) such that VIHL_AC/2 min  
must be met both above and below Vcent_CA.  
5. CA only minimum input pulse width defined at the Vcent_CA(pin mid).  
6. Input slew rate over VcIVW Mask centered at Vcent_CA(pin mid).  
7. VIHL_AC does not have to be met when no transitions are occurring.  
Version 1.0  
369  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DRAM Data Timing  
Read data timing definitions tQH and tDQSQ across on DQ signals per DQS group  
Read data timing tQW valid window defined per DQ signal  
Version 1.0  
370  
Nanya Technology Corp.  
All Rights Reserved. ©  
12/2018  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read output timings  
Speed Grade  
Min/  
Symbol  
Parameter  
Unit  
Notes  
Max  
1600/1867  
2133/2400  
Data Timing  
DQS, DQS to DQ Skew total, per group,  
per access (DBIDisabled)  
DQ output hold time total from DQS,  
DQS  
UI*  
UI*  
tDQSQ  
Max  
Min  
0.18  
0.18  
1
1
tQH  
min(tQSH, tQSL)  
min(tQSH, tQSL)  
(DBIꢀDisabled)  
DQ output window time total, per pin  
(DBIꢀDisabled)  
UI*  
UI*  
tQW_total  
tQW_dj  
Min  
0.75  
TBD  
0.73  
TBD  
1,4  
DQ output window time deterministic,  
per pin (DBIDisabled)  
Max  
1,4,3  
DQS, DQS to DQ Skew total,per group,  
UI*  
tDQSQ_DBI per access  
Max  
0.18  
0.18  
1
(DBIꢀEnabled)  
DQ output hold time total from DQS,  
DQS  
min(tQSH_DBI,  
tQSL_DBI)  
min(tQSH_DBI,  
tQSL_DBI)  
UI*  
UI*  
tQH_DBI  
Min  
1
(DBIꢀEnabled)  
DQ output window time total, per pin  
(DBIꢀEnabled)  
tQHW_total_DBI  
Max  
0.75  
0.73  
1,4  
Data Strobe Timing  
DQS, DQS differential output low time  
tCL(abs)  
ꢀ0.05  
tCL(abs)  
ꢀ0.05  
tQSL  
tQSH  
Min  
Min  
Min  
Min  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
4,5  
4,6  
5,7  
5,7  
(DBIꢀDisabled)  
DQS, DQS differential output high time  
(DBIꢀDisabled)  
tCH(abs)  
ꢀ0.05  
tCH(abs)  
ꢀ0.05  
DQS, DQS differential output low time  
(DBIꢀEnabled)  
tCL(abs)  
ꢀ0.045  
tCL(abs)  
ꢀ0.045  
tQSL_DBI  
tQSH_DBI  
DQS, DQS differential output high time  
(DBIꢀEnabled)  
tCH(abs)  
ꢀ0.045  
tCH(abs)  
ꢀ0.045  
*Unit UI = tCK(avg)min/2  
Notes:  
1. The deterministic component of the total timing. Measurement method tbd.  
2. This parameter will be characterized and guaranteed by design.  
3. This parameter is function of input clock jitter. These values assume the min tCH(abs) and tCL(abs). When the input clock jitter min  
tCH(abs) and tCL(abs) is 0.44 or greater of tck(avg) the min value of tQSL will be tCL(abs)ꢀ0.04 and tQSH will be tCH(abs) ꢀ0.04.  
4. tQSL describes the instantaneous differential output low pulse width on DQS ꢀ DQS, as measured from on falling edge to the next  
consecutive rising edge.  
5. tQSH describes the instantaneous differential output high pulse width on DQS ꢀ DQS, as measured from on falling edge to the next  
consecutive rising edge.  
6. This parameter is function of input clock jitter. These values assume the min tCH(abs) and tCL(abs). When the input clock jitter min  
tCH(abs) and tCL(abs) is 0.44 or greater of tck(avg) the min value of tQSL will be tCL(abs)ꢀ0.04 and tQSH will be tCH(abs) ꢀ0.04.  
Version 1.0  
371  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Read output timings (Continued)  
Speed Grade  
3733  
Min/  
Symbol  
Parameter  
Unit  
Notes  
Max  
Data Timing  
DQS, DQS to DQ Skew total, per group,  
per access (DBIDisabled)  
DQ output hold time total from DQS,  
DQS  
UI*  
UI*  
tDQSQ  
Max  
Min  
0.18  
1
1
tQH  
min(tQSH, tQSL)  
(DBIꢀDisabled)  
DQ output window time total, per pin  
(DBIꢀDisabled)  
UI*  
UI*  
tQW_total  
tQW_dj  
Min  
0.7  
1,4  
DQ output window time deterministic,  
per pin (DBIDisabled)  
Max  
TBD  
1,4,3  
DQS, DQS to DQ Skew total,per group,  
UI*  
tDQSQ_DBI per access  
Max  
0.18  
1
(DBIꢀEnabled)  
DQ output hold time total from DQS,  
DQS  
UI*  
UI*  
tQH_DBI  
Min  
min(tQSH_DBI, tQSL_DBI)  
0.7  
1
(DBIꢀEnabled)  
DQ output window time total, per pin  
(DBIꢀEnabled)  
tQHW_total_DBI  
Max  
1,4  
Data Strobe Timing  
DQS, DQS differential output low time  
tQSL  
tQSH  
Min  
Min  
Min  
Min  
tCL(abs)ꢀ0.05  
tCH(abs)ꢀ0.05  
tCL(abs)ꢀ0.045  
tCH(abs) ꢀ0.045  
tCK(avg)  
tCK(avg)  
tCK(avg)  
tCK(avg)  
4,5  
4,6  
5,7  
5,7  
(DBIꢀDisabled)  
DQS, DQS differential output high time  
(DBIꢀDisabled)  
DQS, DQS differential output low time  
(DBIꢀEnabled)  
tQSL_DBI  
tQSH_DBI  
DQS, DQS differential output high time  
(DBIꢀEnabled)  
*Unit UI = tCK(avg)min/2  
Notes:  
1. The deterministic component of the total timing. Measurement method tbd.  
2. This parameter will be characterized and guaranteed by design.  
3. This parameter is function of input clock jitter. These values assume the min tCH(abs) and tCL(abs). When the input clock jitter min  
tCH(abs) and tCL(abs) is 0.44 or greater of tck(avg) the min value of tQSL will be tCL(abs)ꢀ0.04 and tQSH will be tCH(abs) ꢀ0.04.  
4. tQSL describes the instantaneous differential output low pulse width on DQS ꢀ DQS, as measured from on falling edge to the next  
consecutive rising edge.  
5. tQSH describes the instantaneous differential output high pulse width on DQS ꢀ DQS, as measured from on falling edge to the next  
consecutive rising edge.  
6. This parameter is function of input clock jitter. These values assume the min tCH(abs) and tCL(abs). When the input clock jitter min  
tCH(abs) and tCL(abs) is 0.44 or greater of tck(avg) the min value of tQSL will be tCL(abs)ꢀ0.04 and tQSH will be tCH(abs) ꢀ0.04.  
Version 1.0  
372  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ Rx voltage and timing  
The DQ input receiver mask for voltage and timing is shown figure below is applied per pin. The "total" mask  
(VdIVW_total, TdiVW_total) defines the area the input signal must not encroach in order for the DQ input receiver to  
successfully capture an input signal with a BER of lower than TBD1. The mask is a receiver property and it is not the  
valid dataꢀeye.  
DQ Receiver (Rx) mask  
Across pin VREFDQ voltage variation  
Vcent_DQ(pin_mid) is defined as the midpoint between the largest Vcent_DQ voltage level and the smallest  
Vcent_DQ voltage level across all DQ pins for a given DRAM component. This clarifies that any DRAM component  
level variation must be accounted for within the DRAM Rx mask. The component level VREF will be set by the system  
to account for Ron and ODT settings.  
Version 1.0  
373  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ to DQS tDQS2DQ & tDQDQ Timings at the DRAM pins referenced from the internal latch  
Notes:  
1. tDQS2DQ is measured at the center(midpoint) of the TdiVW window.  
2. DQz represents the max tDQS2DQ in this example  
3. DQy represents the min tDQS2DQ in this example  
All of the timing terms in DQ to DQS are measured from the DQS/DQS to the center(midpoint) of the TdIVW window  
taken at the VdIVW_total voltage levels centered around Vcent_DQ(pin_mid). The timings at the pins are referenced  
with respect to all DQ signals center aligned to the DRAM internal latch. The data to data offset is defined as the  
difference between the min and max tDQS2DQ for a given component.  
Version 1.0  
374  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DQ TdIPW and SRIN_dIVW definition (for each input pulse)  
Notes:  
1. SRIN_dIVW=VdIVW_Total/(tr or tf), signal must be monotonic within tr and tf range.  
DQ VIHL_AC definition (for each input pulse)  
Version 1.0  
375  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DRAM DQs In Receive Mode  
Speed Grade  
Min/  
Symbol  
Parameter  
Unit  
Notes  
a
Max  
2133/2400  
1600/1867  
Max  
mV  
UI*  
VdIVW_total Rx Mask voltage ꢀpꢀp total  
140  
140  
1,2,3,4  
1,2,4  
Rx timing window total  
TdIVW_total  
Max  
Max  
0.22  
TBD  
0.22  
TBD  
(At VdIVW voltage levels)  
Rx timing window 1 bit toggle  
TdIVW_1bit  
1,2,4  
12  
UI*  
(At VdIVW voltage levels)  
mV  
UI*  
VIHL_AC  
DQ AC input pulse amplitude pkꢀpk  
Input pulse width (At Vcent_DQ)  
Min  
Min  
Min  
Max  
Max  
180  
0.45  
200  
800  
30  
180  
0.45  
200  
800  
30  
5,13  
6
TdIPW_DQ  
ps  
tDQS2DQ  
tDQDQ  
DQ to DQS offset  
DQ to DQ offset  
7
ps  
8
9
ps/°C  
tDQS2DQ_temp DQ to DQS offset temperature variation Max  
0.6  
33  
0.6  
33  
tDQS2DQ_Volt DQ to DQS offset voltage variation  
Max  
Min  
ps/50mV  
10  
1
1
SRIN_dIVW Input Slew Rate over VdIVW_total  
tDQS2DQ_rank DQ to DQS offset rank to rank  
V/ns  
11  
Max  
7
7
Mas  
200  
200  
ps  
14,15,16  
2rank  
variation  
A. The Rx voltage and absolute timing requirements apply for all DQ operating frequencies at or below 1600 for all speed bins.  
For example TdIVW_total(ps) = 137.5ps at or below 1600 operating frequencies.  
*UI=tCK(avg)min/2  
Notes:  
1. Data Rx mask voltage and timing parameters are applied per pin and includes the DRAM DQ to DQS voltage AC noise  
impact for frequencies >20 MHz and max voltage of 45mv pkꢀpk from DCꢀ20MHz at a fixed temperature on the  
package. The voltage supply noise must comply to the component MinꢀMax DC operating conditions.  
2. The design specification is a BER <TBD. The BER will be characterized and extrapolated if necessary using a dual  
dirac method.  
3. Rx mask voltage VdIVW total(max) must be centered around Vcent_DQ(pin_mid).  
4. Vcent_DQ must be within the adjustment range of the DQ internal Vref..  
5. DQ only input pulse amplitude into the receiver must meet or exceed VIHL AC at any point over the total UI. No timing  
requirement above level. VIHL AC is the peak to peak voltage centered around Vcent_DQ(pin_mid) such that  
VIHL_AC/2 min must be met both above and below Vcent_DQ.  
6. DQ only minimum input pulse width defined at the Vcent_DQ(pin_mid). .  
7. DQ to DQS offset is within byte from DRAM pin to DRAM internal latch. Includes all DRAM process, voltage and  
temperature variation.  
8. DQ to DQ offset defined within byte from DRAM pin to DRAM internal latch for a given component.  
9. TDQS2DQ max delay variation as a function of temperature.  
Version 1.0  
376  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
10. TDQS2DQ max delay variation as a function of the DC voltage variation for VDDQ and VDD2. It includes the VDDQ and  
VDD2 AC noise impact for frequencies > 20MHz and max voltage of 45mv pkꢀpk from DCꢀ20MHz at a fixed temperature  
on the package. For tester measurement VDDQ = VDD2 is assumed.  
11. Input slew rate over VdIVW Mask centered at Vcent_DQ(pin_mid).  
12. Rx mask defined for a one pin toggling with other DQ signals in a steady state.  
13. VIHL_AC does not have to be met when no transitions are occurring.  
14. The same voltage and temperature are applied to tDQS2DQ_rank2rank.  
15. tDQS2DQ_rank2rank parameter is applied to multiꢀranks per byte lane within a package consisting of the same  
design dies.  
16. tDQS2DQ_rabk2rank support was added to JESD209ꢀ4B, some older devices designed to support JESD209ꢀ4 and  
JESD209ꢀ4A may not support this parameter. Refer to vendor datasheet.  
Version 1.0  
377  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
DRAM DQs In Receive Mode (Continued)  
Speed Grade  
3733  
Min/  
Max  
Symbol  
Parameter  
Unit  
Notes  
Max  
Max  
mV  
UI*  
VdIVW_total Rx Mask voltage ꢀpꢀp total  
130  
1,2,3,4  
1,2,4  
Rx timing window total  
TdIVW_total  
0.25  
TBD  
(At VdIVW voltage levels)  
Rx timing window 1 bit toggle  
TdIVW_1bit  
1,2,4  
12  
UI*  
Max  
(At VdIVW voltage levels)  
mV  
UI*  
VIHL_AC  
DQ AC input pulse amplitude pkꢀpk  
Input pulse width (At Vcent_DQ)  
Min  
Min  
180  
5,13  
6
TdIPW_DQ  
0.45  
Min  
Max  
Max  
250  
700  
30  
ps  
tDQS2DQ  
tDQDQ  
DQ to DQS offset  
DQ to DQ offset  
7
ps  
8
9
ps/°C  
tDQS2DQ_temp DQ to DQS offset temperature variation Max  
0.4  
25  
1
tDQS2DQ_Volt DQ to DQS offset voltage variation  
Max  
Min  
ps/50mV  
10  
SRIN_dIVW Input Slew Rate over VdIVW_total  
tDQS2DQ_rank DQ to DQS offset rank to rank  
V/ns  
ps  
11  
Max  
7
Mas  
200  
14,15,16  
2rank  
variation  
*UI=tCK(avg)min/2  
Notes:  
1. Data Rx mask voltage and timing parameters are applied per pin and includes the DRAM DQ to DQS voltage AC noise  
impact for frequencies >20 MHz and max voltage of 45mv pkꢀpk from DCꢀ20MHz at a fixed temperature on the  
package. The voltage supply noise must comply to the component MinꢀMax DC operating conditions.  
2. The design specification is a BER <TBD. The BER will be characterized and extrapolated if necessary using a dual  
dirac method.  
3. Rx mask voltage VdIVW total(max) must be centered around Vcent_DQ(pin_mid).  
4. Vcent_DQ must be within the adjustment range of the DQ internal Vref..  
5. DQ only input pulse amplitude into the receiver must meet or exceed VIHL AC at any point over the total UI. No timing  
requirement above level. VIHL AC is the peak to peak voltage centered around Vcent_DQ(pin_mid) such that  
VIHL_AC/2 min must be met both above and below Vcent_DQ.  
6. DQ only minimum input pulse width defined at the Vcent_DQ(pin_mid). .  
7. DQ to DQS offset is within byte from DRAM pin to DRAM internal latch. Includes all DRAM process, voltage and  
temperature variation.  
8. DQ to DQ offset defined within byte from DRAM pin to DRAM internal latch for a given component.  
9. TDQS2DQ max delay variation as a function of temperature.  
10. TDQS2DQ max delay variation as a function of the DC voltage variation for VDDQ and VDD2. It includes the VDDQ and  
VDD2 AC noise impact for frequencies > 20MHz and max voltage of 45mv pkꢀpk from DCꢀ20MHz at a fixed temperature  
on the package. For tester measurement VDDQ = VDD2 is assumed.  
Version 1.0  
378  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
11. Input slew rate over VdIVW Mask centered at Vcent_DQ(pin_mid).  
12. Rx mask defined for a one pin toggling with other DQ signals in a steady state.  
13. VIHL_AC does not have to be met when no transitions are occurring.  
14. The same voltage and temperature are applied to tDQS2DQ_rank2rank.  
15. tDQS2DQ_rank2rank parameter is applied to multiꢀranks per byte lane within a package consisting of the same  
design dies.  
16. tDQS2DQ_rabk2rank support was added to JESD209ꢀ4B, some older devices designed to support JESD209ꢀ4 and  
JESD209ꢀ4A may not support this parameter. Refer to vendor datasheet.  
Version 1.0  
379  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
Preliminary  
Features, specification, functions and  
operations are not finalized  
NTC Proprietary  
Level: Property  
4Gb SLC NAND + 4Gb LPDDR4X  
NM4484NSPAXAE  
Revision History  
Rev  
Page  
Modified  
Description  
Released  
1.0  
Preliminary Release  
12/2018  
Version 1.0  
380  
Nanya Technology Corp.  
12/2018  
All Rights Reserved. ©  
http://www.nanya.com/  

相关型号:

NM4484NSPAXAE-3ENA

MCP Specification
NANYA

NM4484NSPAXAE_19

MCP Specification
NANYA

NM4492W

IC 2K X 9 STANDARD SRAM, 5 ns, CQFP64, CERAMIC, FP-64, Static RAM
TI

NM4492W10

x9 Synchronous SRAM
ETC

NM4492W5

x9 Synchronous SRAM
ETC

NM4492W7

x9 Synchronous SRAM
ETC

NM450M022012035050

Aluminum Electrolytic Capacitor, Polarized, Aluminum, 450V, 20% +Tol, 20% -Tol, 220uF
YAGEO

NM450M068022051095

Aluminum Electrolytic Capacitor, Polarized, Aluminum, 450V, 20% +Tol, 20% -Tol, 680uF
YAGEO

NM450M100022051100

Aluminum Electrolytic Capacitor, Polarized, Aluminum, 450V, 20% +Tol, 20% -Tol, 1000uF
YAGEO

NM450M150022051130

Aluminum Electrolytic Capacitor, Polarized, Aluminum, 450V, 20% +Tol, 20% -Tol, 1500uF
YAGEO

NM450M220028063115

Aluminum Electrolytic Capacitor, Polarized, Aluminum, 450V, 20% +Tol, 20% -Tol, 2200uF
YAGEO

NM450M330032076130

Aluminum Electrolytic Capacitor, Polarized, Aluminum, 450V, 20% +Tol, 20% -Tol, 3300uF
YAGEO