LMK05028 [TI]
低抖动双通道网络同步器时钟;型号: | LMK05028 |
厂家: | TEXAS INSTRUMENTS |
描述: | 低抖动双通道网络同步器时钟 时钟 |
文件: | 总95页 (文件大小:3242K) |
中文: | 中文翻译 | 下载: | 下载PDF数据表文档文件 |
Support &
Community
Product
Folder
Order
Now
Tools &
Software
Technical
Documents
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
具有 EEPROM 的 LMK05028 低抖动双通道网络同步器时钟
1 特性
•
•
3.3V 电源,提供 1.8V、2.5V 或 3.3V 输出
工业温度范围:-40°C 至 +85°C
1
•
两个独立 PLL 通道具有如下特性:
–
–
输出 ≥ 100MHz 时,抖动为 150fs RMS
2 应用
频率为 122.88MHz 时,在 100Hz 偏移频率处
的相位噪声为 –112dBc/Hz
•
SyncE (G.8262)、SONET/SDH(Stratum 3/3E、
G.813、GR-1244、GR-253)、IEEE 1588 PTP
从时钟,或光传输网络 (G.709)
–
–
–
无中断切换:50ps 相位瞬态(采用相位抑制)
具有快速锁定功能的可编程环路带宽
•
•
•
•
电信和企业线卡
使用低成本 TCXO/OCXO 实现符合标准的同步
和保持模式
无线基站 (BTS),无线回程
测试和测量、广播基础设施及医学超声
–
任何输入到任何输出频率转换
适用于 FPGA、DSP、ASIC 和 CPU 器件的抖动和
漂移衰减、精确频率转换和低抖动时钟生成
•
•
四个参考时钟输入
–
–
基于优先级的输入选择
在缺失参考时实现数字保持
3 说明
具有可编程驱动器的八个时钟输出
LMK05028 是一款高性能网络同步器时钟器件,提供
抖动消除、时钟生成、高级时钟监控和卓越的无中断切
换性能,可满足通信基础设施和工业应用的严格时序
要求。该器件具有低抖动和高 PSNR 性能,可降低高
速串行链路中的误码率 (BER)。
–
–
多达 6 个不同的输出频率
AC-LVDS、AC-CML、AC-LVPECL、HCSL 和
1.8V 或 2.5V LVCMOS 输出格式
•
•
加电后自定义时钟的 EEPROM/ROM(2)
灵活的配置选项
该器件具有两个 PLL 通道,最多可生成八个输出时钟
(抖动低至 150fs RMS)。每个 PLL 域可从任意四个
基准输入中进行选择,以实现输出同步。
–
–
输入和输出为 1Hz (1PPS) 至 750MHz
XO:10MHz 至 100MHz,TCXO:10MHz 至
54MHz
–
DCO 模式:步长 < 1ppt,可实现精确的频率和
相位控制(IEEE 1588 从运行)
器件信息(1)
器件型号
LMK05028
封装
VQFN (64)
封装尺寸(标称值)
–
–
–
零延迟,可实现确定性相位偏移
稳健的时钟监控和状态
I2C 或 SPI 接口
9.00mm x 9.00mm
(1) 如需了解所有可用封装,请参阅数据表末尾的可订购产品附
录。
(2) 请联系 TI 现场销售人员询问定制工厂预编程器件。
•
出色的电源噪声抑制 (PSNR) 性能
简化方框图
VDD
VDDO
TCXO/
OCXO
3.3 V 1.8 / 2.5 / 3.3 V
Power Conditioning
÷R
TCXO_IN
XO
4
÷OD
÷OD
÷OD
OUT7
OUT6
0:3
0:3
5
PLL1
IN0
IN1
IN2
IN3
APLL
0
1
÷
÷
REF
TCXO
DPLL
DPLL
OUT5
OUT4
VCO
0:3
DCO
Output
Muxes
Differential or
LVCMOS
Hitless
Switching
TCXO
XO
OUT3
OUT2
PLL2
0:3
÷OD
÷OD
÷OD
APLL
2
3
÷
÷
REF
DPLL
TCXO
DPLL
÷R
VCO
DCO
0:5
0:5
OUT1
EEPROM,
Registers
I2C/SPI
LMK05028
Low-Jitter Dual-Channel
Network Synchronizer Clock
OUT0
ROM
Differential or
LVCMOS
LOGIC I/O
STATUS
Device Control and Status
1
本文档旨在为方便起见,提供有关 TI 产品中文版本的信息,以确认产品的概要。 有关适用的官方英文版本的最新信息,请访问 www.ti.com,其内容始终优先。 TI 不保证翻译的准确
性和有效性。 在实际设计之前,请务必参考最新版本的英文版本。
English Data Sheet: SNAS724
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
目录
9.4 Device Functional Modes........................................ 55
9.5 Programming .......................................................... 64
9.6 Register Maps......................................................... 71
10 Application and Implementation........................ 72
10.1 Application Information.......................................... 72
10.2 Typical Application ............................................... 75
10.3 Do's and Don'ts..................................................... 81
11 Power Supply Recommendations ..................... 81
11.1 Power Supply Bypassing ...................................... 81
12 Layout................................................................... 82
12.1 Layout Guidelines ................................................. 82
12.2 Layout Example .................................................... 83
12.3 Thermal Reliability................................................. 83
13 器件和文档支持 ..................................................... 84
13.1 器件支持................................................................ 84
13.2 文档支持................................................................ 84
13.3 接收文档更新通知 ................................................. 84
13.4 社区资源................................................................ 84
13.5 商标....................................................................... 84
13.6 静电放电警告......................................................... 84
13.7 术语表 ................................................................... 84
14 机械、封装和可订购信息....................................... 85
1
2
3
4
5
6
特性.......................................................................... 1
应用.......................................................................... 1
说明.......................................................................... 1
修订历史记录 ........................................................... 2
说明 (续).............................................................. 3
Pin Configuration and Functions......................... 4
6.1 Device Start-Up Modes............................................. 7
Specifications......................................................... 8
7.1 Absolute Maximum Ratings ...................................... 8
7.2 ESD Ratings.............................................................. 8
7.3 Recommended Operating Conditions....................... 8
7.4 Thermal Information.................................................. 9
7.5 Electrical Characteristics........................................... 9
7.6 Timing Diagrams..................................................... 15
7.7 Typical Characteristics............................................ 17
Parameter Measurement Information ................ 18
8.1 Output Clock Test Configurations........................... 18
Detailed Description ............................................ 20
9.1 Overview ................................................................. 20
9.2 Functional Block Diagrams ..................................... 21
9.3 Feature Description................................................. 28
7
8
9
4 修订历史记录
注:之前版本的页码可能与当前版本有所不同。
Changes from Original (February 2018) to Revision A
Page
•
将器件状态从“预告信息”更改为“生产数据”.............................................................................................................................. 1
2
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
5 说明 (续)
每个 PLL 通道支持实现抖动和漂移衰减的可编程环路带宽,同时支持能够实现灵活频率配置的分频率转换。每个
PLL 通道支持的同步选项包括采用相位消除的无中断切换、数字保持、步长 <1ppt 可实现精确时钟控制(IEEE
1588 PTP 从运行)的 DCO 模式,以及用于实现确定性输入到输出相位偏移的零延迟模式。先进的基准输入监控
块可确保稳健的时钟故障检测并在发生基准缺失 (LOR) 时帮助将输出时钟干扰降至最低。
该器件可使用低频 TCXO/OCXO 实现自由运行型或保持型频率稳定性,从而在 LOR 期间确保同步符合标准;此
外,在保持型频率稳定性和漂移不重要时,该器件也可使用标准 XO。该器件可通过 I2C 或 SPI 接口实现完全编
程,在通电后支持通过内部 EEPROM 或 ROM 进行自定义频率配置。EEPROM 可厂家预编程并且可在系统内编
程。
AC-LVPECL 输出,fIN = 25MHz,fTCXO = 10MHz (OCXO),fXO = 48.0048MHz,fTCXO-TDC = 20MHz
图 1. 122.88MHz 输出相位噪声(3 环路)
版权 © 2018, Texas Instruments Incorporated
3
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
6 Pin Configuration and Functions
RGC Package
64-Pin VQFN
Top View
IN0_P
IN0_N
1
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
CAP_APLL1
2
LF1
VDD_IN0
VDD_IN3
IN3_P
3
PDN
4
GPIO0/SYNCN
XO_N
5
IN3_N
6
XO_P
CAP_DIG
VDD_DIG
VDD_IN2
IN2_P
7
VDD_XO
GPIO4/FDEC1
GPIO3/FINC1
LF2
8
GND
9
10
11
12
13
14
15
16
IN2_N
CAP_APLL2
VDD_APLL2
SCL/SCK
SDA/SDI
OUT3_P
OUT3_N
GPIO5/FINC2
GPIO6/FDEC2
IN1_P
IN1_N
VDD_IN1
Not to scale
Pin Functions
PIN
TYPE(1)
DESCRIPTION
NAME
POWER
NO.
Ground / Thermal Pad.
GND
PAD
G
The exposed pad must be connected to PCB ground for proper electrical and thermal performance.
A 7x7 via pattern is recommended to connect the IC ground pad to the PCB ground layers.
VDD_IN0
VDD_IN1
VDD_IN2
VDD_IN3
3
16
9
P
P
P
P
Core Supply (3.3 V) for Reference Inputs 0 to 3.
Place a nearby 0.1-µF bypass capacitor on each pin.
4
(1) G = Ground, P = Power, I = Input, O = Output, I/O = Input or Output, A = Analog.
4
Copyright © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Pin Functions (continued)
PIN
NAME
VDD_XO
TYPE(1)
DESCRIPTION
NO.
42
19
49
37
8
P
P
P
P
P
P
P
P
P
P
P
Core Supply (3.3 V) for XO and TCXO Inputs.
Place a nearby 0.1-µF bypass capacitor on each pin.
VDD_TCXO
VDD_APLL1
VDD_APLL2
VDD_DIG
VDDO_0
VDDO_1
VDDO_23
VDDO_45
VDDO_6
VDDO_7
CORE BLOCKS
LF1
Core Supply (3.3 V) for PLL1, PLL2, and Digital Blocks.
Place a nearby 0.1-µF bypass capacitor on each pin.
21
25
30
50
59
63
Output Supply (1.8, 2.5, or 3.3 V) for Clock Outputs 0 to 7.
Place a nearby 0.1-µF bypass capacitor on each pin.
47
39
48
38
7
A
A
A
A
A
External Loop Filter Capacitor for APLL1 and APLL2.
Place a nearby 0.1-µF capacitor on each pin.
LF2
CAP_APLL1
CAP_APLL2
CAP_DIG
INPUT BLOCKS
IN0_P
External Bypass Capacitors for APLL1, APLL2, and Digital Blocks.
Place a nearby 10-µF bypass capacitor on each pin.
1
2
I
I
I
I
I
I
I
I
I
IN0_N
DPLL Reference Clock Inputs 0 to 3.
IN1_P
14
15
10
11
5
Each input pair can accept a differential or single-ended clock signal for synchronizing the DPLLs.
Each pair has a programmable input type with internal termination to support AC- or DC-coupled
clocks. A single-ended LVCMOS clock can be applied to the P input with the N input pulled down to
ground. An unused input pair can be left floating. LVCMOS input mode is recommended for input
frequencies from 5 MHz down to 1 Hz (1 PPS or pulse-per-second).
IN1_N
IN2_P
IN2_N
IN3_P
IN3_N
6
XO_P
43
XO Input.
This input pair can accept a differential or single-ended clock signal from a low-jitter local oscillator
to lock the APLLs. This input has a programmable input type with internal termination to support
AC- or DC-coupled clocks. A single-ended LVCMOS clock (up to 2.5 V) can be applied to the P
input with the N input pulled down to ground.
XO_N
44
18
I
TCXO Input.
This input can accept an AC-coupled sinewave, clipped-sinewave, or single-ended clock signal from
a stable oscillator (TCXO/OCXO) to lock the TCXO-DPLL if used by a DPLL configuration. The
input swing must be less than 1.3 Vpp before AC-coupling to the input pin, which has weak internal
biasing of 0.6 V and no internal termination. Leave pin floating if unused.
TCXO_IN
I
OUTPUT BLOCKS
OUT0_P
22
23
27
26
31
32
34
33
O
O
O
O
O
O
O
O
OUT0_N
Clock Outputs 0 to 3 Bank.
OUT1_P
Each programmable output driver pair can support AC-LVDS, AC-CML, AC-LVPECL, HCSL, or
1.8/2.5-V LVCMOS clocks (one or two per pair). Unused differential outputs should be terminated if
active or left floating if disabled through registers.
The OUT[0:3] bank requires at least one clock from the PLL2 domain if enabled. This bank is
preferred for PLL2 clocks to minimize output crosstalk.
OUT1_N
OUT2_P
OUT2_N
OUT3_P
OUT3_N
Copyright © 2018, Texas Instruments Incorporated
5
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Pin Functions (continued)
PIN
TYPE(1)
DESCRIPTION
NAME
OUT4_P
NO.
51
52
54
53
57
58
62
61
O
O
O
O
O
O
O
OUT4_N
OUT5_P
OUT5_N
OUT6_P
OUT6_N
OUT7_P
OUT7_N
Clock Outputs 4 to 7 Bank.
Each programmable output driver pair can support AC-LVDS, AC-CML, AC-LVPECL, HCSL, or
1.8/2.5-V LVCMOS clocks (one or two per pair). Unused differential outputs should be terminated if
active or left floating if disabled through registers.
The OUT[4:7] bank requires at least one clock from the PLL1 domain. This bank is preferred for
PLL1 clocks to minimize output crosstalk.
O
(2)(3)
LOGIC CONTROL / STATUS
Device Start-Up Mode Select (3-level, 1.8-V compatible).
This input selects the device start-up mode that determines the memory page used to initialize the
registers, serial interface, and logic pin functions. The input level is sampled only at device power-
on reset (POR).
HW_SW_CTRL
PDN
64
46
35
I
I
See Table 1 for start-up mode descriptions and logic pin functions.
Device Power-Down (active low).
When PDN is pulled low, the device is in hard reset and all blocks including the serial interface are
powered down. When PDN is pulled high, the device is started according to device mode selected
by HW_SW_CTRL and begins normal operation with all internal circuits reset to their initial state.
I2C Serial Data I/O (SDA) or SPI Serial Data Input (SDI). See Table 1.
The default 7-bit I2C address is 11000xxb, where the MSB bits (11000b) are initialized from on-chip
EEPROM and the LSB bits (xxb) are determined by the logic input pins. When HW_SW_CTRL is 0,
the LSBs are determined by the GPIO[2:1] input levels during POR. When HW_SW_CTRL is 1, the
LSBs are fixed to 00b.
SDA/SDI
I/O
SCL/SCK
36
45
24
60
40
41
12
13
56
I
I
I2C Serial Clock Input (SCL) or SPI Serial Clock Input (SCK). See Table 1.
GPIO0/SYNCN
GPIO1/SCS
GPIO2/SDO
GPIO3/FINC1
GPIO4/FDEC1
GPIO5/FINC2
GPIO6/FDEC2
STATUS1
I
I/O
I
Multifunction Inputs or Outputs.
See Table 1.
I
I/O
I/O
I/O
Status Outputs [1:0].
Each output has programmable status signal selection, driver type (3.3-V LVCMOS or open-drain),
and status polarity. Open-drain requires an external pullup resistor. Leave pin floating if unused.
STATUS0
55
I/O
INSEL0_1
INSEL0_0
INSEL1_1
INSEL1_0
17
20
29
28
I
I
I
I
Manual Reference Input Selection for DPLL1.
INSEL0_[1:0] = 00b (IN0), 01b (IN1), 10b (IN2), or 11b (IN3). Leave pin floating if unused.
Manual Reference Input Selection for DPLL2.
INSEL1_[1:0] = 00b (IN0), 01b (IN1), 10b (IN2), or 11b (IN3). Leave pin floating if unused.
(2) Internal resistors: PDN pin has 200-kΩ pullup to VDD. Each HW_SW_CTRL, GPIO, and STATUS pin has a 150-kΩ bias to VIM
(approximately 0.8 V) when PDN = 0 or 400-kΩ pulldown when PDN = 1. Each INSEL pin has an 85-kΩ pullup to 1.8 V when PDN = 0
or 400-kΩ pulldown when PDN = 1.
(3) Unless otherwise noted: Logic inputs are 2-level, 1.8-V compatible inputs. Logic outputs are 3.3-V LVCMOS levels.
6
Copyright © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
6.1 Device Start-Up Modes
The HW_SW_CTRL input pin selects the device start-up mode that determines the memory page used to
initialize the registers, serial interface, and logic pin functions at power-on reset. The initial register settings
determine the device's frequency configuration stored in the internal EEPROM (NVM) or ROM. After start-up, the
device registers can be accessed through the selected serial interface for device status monitoring or
programming, and the logic pins will function as defined by the mode configuration.
Table 1. Device Start-Up Modes
HW_SW_CTRL
INPUT LEVEL(1)
START-UP
MODE
MODE DESCRIPTION
Registers are initialized from EEPROM, and I2C interface is enabled. Logic pins:
•
•
•
•
•
SDA, SCL: I2C Data, I2C Clock (open-drain). Pins require external pullups > 1 kΩ.
GPIO0: Output SYNC Input (active low). Tie pin high externally if not used.
GPIO[2:1](1): I2C Address LSB Select Input (00, 01, 10, 11b)
EEPROM + I2C
(Soft pin mode)
0
GPIO[4:3](2): DPLL1 DCO Frequency Decrement and Increment Inputs (active high)
GPIO[6:5](2)(3): DPLL2 DCO Frequency Decrement and Increment Inputs (active high), or
Status Outputs
Registers are initialized from EEPROM, and SPI interface is enabled. Logic pins:
•
•
•
•
SDA, SCL: SPI Data Input (SDI), SPI Clock (SCK)
GPIO1: SPI Chip Select (SCS)
GPIO2: SPI Data Output (SDO)
Float
EEPROM + SPI
(Soft pin mode)
(VIM
)
GPIO[6:3, 0]: Same as for HW_SW_CTRL = 0
Registers are initialized from the ROM page selected by GPIO pins, and I2C interface is enabled.
Logic pins:
•
•
•
•
SDA, SCL: I2C Data, I2C Clock (open-drain). Pins require external pullups > 1 kΩ.
GPIO[3:0](1): ROM Page Select Input (0000 to 1111b)
GPIO[6:5](3): Status Outputs
ROM + I2C
(Hard pin mode)
1
GPIO4: Not used during POR
After POR, GPIO[6:3] can function the same as for HW_SW_CTRL = 0 if enabled by registers.
(1) The input levels on these pins are sampled only during POR.
(2) These GPIO pins are only functional when the DCO feature and FINC/FDEC pin controls are enabled by registers.
(3) As status outputs, the GPIO[6:5] pins have the same configuration options as the STATUS[1:0] pins.
NOTE
To ensure proper start-up into EEPROM + SPI Mode, the HW_SW_CTRL, STATUS0, and
STATUS1 pins must all be floating or biased to VIM (0.8-V typical) before the PDN pin is
pulled high. These three pins momentarily operate as 3-level inputs and get sampled at
the low-to-high transition of PDN to determine the device start-up mode during POR. If
any of these pins are connected to a host device (MCU or FPGA), TI recommends using
external biasing resistors on each pin (10-kΩ pullup to 3.3 V with 3.3-kΩ pulldown to GND)
to set the inputs to VIM during POR. After power-up, the STATUS pins can operate as
LVCMOS outputs to overdrive the external resistor bias for normal status operation.
Copyright © 2018, Texas Instruments Incorporated
7
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
7 Specifications
7.1 Absolute Maximum Ratings
Over operating free-air temperature range (unless otherwise noted)(1)
MIN
-0.3
-0.3
-0.3
-0.3
-0.3
MAX
3.6
UNIT
V
VDD(2)
VDDO(3)
VIN
Core supply voltages
Output supply voltages
3.6
V
Input voltage range for clock and logic inputs
Output voltage range for logic outputs
Output voltage range for clock outputs
Junction temperature
VDD+0.3
VDD+0.3
VDDO+0.3
150
V
VOUT_LOGIC
VOUT
V
V
Tj
°C
°C
Tstg
Storage temperture range
-65
150
(1) Stresses beyond those listed under Absolute Maximum Ratings may cause permanent damage to the device. These are stress ratings
only, which do not imply functional operation of the device at these or any other conditions beyond those indicated under Recommended
Operating Conditions. Exposure to absolute-maximum-rated conditions for extended periods may affect device reliability.
(2) VDD refers to all core supply pins or voltages. All VDD core supplies should be powered-on before the PDN is pulled high to trigger the
internal power-on reset (POR).
(3) VDDO refers to all output supply pins or voltages. VDDO_x refers to the output supply for a specific output channel, where x denotes
the channel index.
7.2 ESD Ratings
VALUE
UNIT
Human body model (HBM), per ANSI/ESDA/JEDEC JS-001, all pins(1)
±2000
Electrostatic
discharge
V(ESD)
V
Charged device model (CDM), per JEDEC specification JESD22-C101, all
pins(2)
±750
(1) JEDEC document JEP155 states that 500-V HBM allows safe manufacturing with a standard ESD control process.
(2) JEDEC document JEP157 states that 250-V CDM allows safe manufacturing with a standard ESD control process.
7.3 Recommended Operating Conditions
Over operating free-air temperature range (unless otherwise noted)
MIN
3.135
1.71
1.71
2.375
0
NOM
MAX
3.465
3.465
1.89
UNIT
VDD(1)
VDDO_x(2)
VDDO_x(2)
VDDO_x(2)
VIN
Core supply voltages
3.3
1.8, 2.5, 3.3
1.8
V
V
Output supply voltage for AC-LVDS/CML/LVPECL or HCSL driver
Output supply voltage for 1.8-V LVCMOS driver(3)
Output supply voltage for 2.5-V LVCMOS driver(3)
Input voltage range for clock and logic inputs
Junction temperature
Power supply ramp time(4)
EEPROM program cycles(5)
V
2.5
2.625
3.465
135
V
V
TJ
°C
ms
cycles
-
tVDD
0.01
100
nEEcyc
100
SROUT
Output slew rate mode(6)
Fast
(1) VDD refers to all core supply pins or voltages. All VDD core supplies should be powered-on before internal power-on reset (POR).
(2) VDDO refers to all output supply pins or voltages. VDDO_x refers to the output supply for a specific output channel, where x denotes
the channel index.
(3) The LVCMOS driver supports full rail-to-rail swing when VDDO_x is 1.8 V or 2.5 V ±5%. When VDDO_x is 3.3 V, the LVCMOS driver
will not fully swing to the positive rail due to the dropout voltage of the output channel's internal LDO regulator.
(4) Time for VDD to ramp monotonically above 2.7 V for proper internal power-on reset. For slower or non-monotonic VDD ramp, hold PDN
low until after VDD voltages are valid.
(5) nEEcyc specifies the maximum EEPROM program cycles allowed for customer programming. The initial count of factory-programmed
cycles is non-zero due to production tests, but factory-programmed cycles are excluded from the nEEcyc limit. The total number of
EEPROM program cycles can be read from the 8-bit NVM count status register (NVMCNT), which automatically increments by 1 on
each successful programming cycle. TI does not ensure EEPROM endurance if the nEEcyc limit is exceeded by the customer.
(6) Clock output characteristics are specified for all driver types with the output slew rate mode programmed to Fast. Some specifications
may not be met with the output slew rate programmed to Nominal.
8
Copyright © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
7.4 Thermal Information
LMK05028
THERMAL METRIC(1) (2) (3)
RGC (VQFN)
UNIT
64 PINS
20.5
6.3
RθJA
Junction-to-ambient thermal resistance
Junction-to-case (top) thermal resistance
Junction-to-board thermal resistance
°C/W
°C/W
°C/W
°C/W
°C/W
°C/W
RθJC(top)
RθJB
RθJC(bot)
ψJT
4.9
Junction-to-case (bottom) thermal resistance
Junction-to-top characterization parameter
Junction-to-board characterization parameter
0.3
0.1
ψJB
4.8
(1) For more information about traditional and new thermal metrics, see the Semiconductor and IC Package Thermal Metrics application
report, SPRA953.
(2) The thermal information is based on a 10-layer 200 mm x 250 mm board with 49 thermal vias (7 x 7 pattern, 0.3 mm holes).
(3) ΨJB can allow the system designer to measure the board temperature (TPCB) with a fine-gauge thermocouple and back-calculate the
device junction temperature, TJ = TPCB + (ΨJB x Power). Measurement of ΨJB is defined by JESD51-6.
7.5 Electrical Characteristics
Over Recommended Operating Conditions (unless otherwise noted)
PARAMETER
Power Supply Characteristics
IDD_IN0,
TEST CONDITIONS
MIN
TYP
MAX
UNIT
Core Supply Current
IDD_IN1,
IDD_IN3
3.5
10
mA
(VDD_INx)
Core Supply Current
(VDD_IN2)
IDD_IN2
IDD_XO
6
14
33
mA
mA
Core Supply Current
(VDD_XO)
25
Configuration A(1)
1
6
4
9
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
Core Supply Current
(VDD_TCXO)
IDD_TCXO
IDD_PLL1
IDD_PLL2
IDD_DIG
Configuration B(2)
Configuration A(1)
Configuration B(2)
Configuration A(1)
Configuration B(2)
Configuration A(1)
Configuration B(2)
AC-LVDS
160
185
138
160
34
188
217
160
187
59
Core Supply Current
(VDD_PLL1)
Core Supply Current
(VDD_PLL2)
Core Supply Current
(VDD_DIG)
42
70
22
28
Output Supply Current(3)
(VDDO_x = 3.3 V ± 5%)
AC-CML
24
32
IDDO_x
AC-LVPECL
27
34
HCSL
33
42
AC-LVDS (x2)
AC-CML (x2)
AC-LVPECL (x2)
HCSL (x2)
32
40
Output Supply Current(4)
(VDDO_x = 3.3 V ± 5%)
37
45
IDDO_x
IDDPDN
41
51
55
67
Total Supply Current
(all VDD and VDDO pins, 3.3 V)
Device powered-down (PDN pin held
low)
40
(1) Configuration A (All blocks on except TCXO_IN and both TCXO-DPLLs): fIN[0:3] = 25 MHz, fXO = 48.0048 MHz, TCXO_IN disabled. Both
DPLL[1:2] in 2-loop mode, fVCO1 = 5 GHz, fVCO2 = 5.5296 GHz, PLL1_P1 = 8, PLL2_P1 = 9.
(2) Configuration B (All blocks on): fIN[0:3] = 25 MHz, fXO = 48.0048 MHz, fTCXO = 10 MHz. Both DPLL[1:2] in 3-loop mode, fVCO1 = 5 GHz,
fVCO2 = 5.5296 GHz, PLL1_P1 = 8, PLL2_P1 = 9.
(3) IDDO_x includes supply current for output divider and one output driver with fOUT = 156.25 MHz or 122.88 MHz.
(4) IDDO_x includes supply current for output divider and two output drivers with fOUT = 156.25 MHz or 122.88 MHz.
Copyright © 2018, Texas Instruments Incorporated
9
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Electrical Characteristics (continued)
Over Recommended Operating Conditions (unless otherwise noted)
PARAMETER
TEST CONDITIONS
MIN
TYP
MAX
UNIT
Reference Input Characteristics (INx)
Differential input(6)
LVCMOS input
5
1E-6
1
750
250
fIN
Input frequency range(5)
MHz
VIN-SE
VIDpp
dV/dt
IIN
Single-ended input voltage swing LVCMOS input, DC-coupled to INx_P
V
V
Differential input voltage swing,
Differential input
0.4
0.2
2
peak-peak (|VP – VN|)(7)
Input slew rate(5)
V/ns
µA
pF
50-Ω and 100-Ω internal terminations
Input leakage
disabled
-350
350
CIN
Input capacitance
Single-ended, each pin
2
XO Input Characteristics (XO)
fCLK
Input frequency range(5)
10
1
100
2.6
MHz
V
VIN-SE
Single-ended input voltage swing LVCMOS input, DC-coupled to XO_P
Differential input voltage swing,
Differential input
VIDpp
0.4
2
V
peak-peak (|VP – VN|)(7)
dV/dt
IDC
Input slew rate(5)
0.2
40
V/ns
%
Input duty cycle
60
50-Ω and 100-Ω internal terminations
IIN
Input leakage
disabled
-350
350
µA
pF
CIN
Input capacitance
Single-ended, each pin
1
TCXO/OCXO Input Characteristics (TCXO_IN)
fTCXO
VIN
Input frequency(5)
Input voltage swing
Input bias voltage
Input slew rate(5)
Input duty cycle
10
54
MHz
V
AC-coupled
0.8
1.3
VBIAS
dV/dt
IDC
Weak internal bias
0.6
10
V
0.2
40
V/ns
%
60
CIN
Input capacitance
pF
APLL/VCO Charateristics
fVCO1 VCO1 Frequency range
fVCO2 VCO2 Frequency range
(5) Parameter is specified by characterization and is not tested in production.
4.8
5.5
5.4
6.2
GHz
GHz
(6) For a differential input clock below 5 MHz, TI recommends to disable the differential input amplitude monitor and enable at least one
other monitor (frequency, window detectors) to validate the input clock. Otherwise, consider using an LVCMOS clock for an input below
5 MHz.
(7) Minimum limit applies for the minimum setting of the differential input amplitude monitor.
10
Copyright © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Electrical Characteristics (continued)
Over Recommended Operating Conditions (unless otherwise noted)
PARAMETER
TEST CONDITIONS
MIN
TYP
MAX
200
0.4
UNIT
1.8-V LVCMOS Output Characteristics (OUTx)
fOUT
VOH
VOL
IOH
Output frequency(5)
Output high voltage
Output low voltage
Output high current
Output low current
Output rise/fall time(5)
1E-6
1.2
MHz
V
IOH = 1 mA
IOL = 1 mA
V
-23
24
mA
mA
ps
IOL
tR/tF
20% to 80%
250
Same post divider, output divide
values, and output type
100
1.5
ps
ns
tSK
Output-to-output skew(5)
Same post divider, output divide
values, LVCMOS-to-DIFF
Output phase noise floor
(fOFFSET > 10 MHz)
Output duty cycle(5) (8)
PNFLOOR
66.66 MHz
-155
50
dBc/Hz
ODC
ROUT
45
55
%
Output impedance
Ω
2.5-V LVCMOS Output Characteristics (OUTx)
fOUT
VOH
VOL
IOH
Output frequency(5)
Output high voltage
Output low voltage
Output high current
Output low current
Output rise/fall time(5)
1E-6
1.9
200
MHz
V
IOH = 1 mA
IOL = 1 mA
0.525
V
-48
55
mA
mA
ps
IOL
tR/tF
20% to 80%
250
Same post divider, output divide
values, and output type
100
1.5
ps
ns
tSK
Output-to-output skew(5)
Same post divider, output divide
values, LVCMOS-to-DIFF
Output phase noise floor
(fOFFSET > 10 MHz)
Output duty cycle(5) (8)
PNFLOOR
66.66 MHz
-155
50
dBc/Hz
ODC
ROUT
45
55
%
Output impedance
Ω
AC-LVDS Output Characteristics (OUTx)
fOUT
VOD
Output frequency(5) (9)
750
450
MHz
mV
Output voltage swing (VOH - VOL
)
fOUT > 25 MHz
250
100
400
Differential output voltage swing,
peak-to-peak
VODpp
VOS
tSK
2×VOD
V
mV
ps
ps
ps
Output common mode
Output-to-output skew(5)
430
100
350
150
Same post divider, output divide
values, and output type
20% to 80%, < 300 MHz
225
75
tR/tF
Output rise/fall time(5)
± 100 mV around center point, ≥ 300
MHz
Output phase noise floor
(fOFFSET > 10 MHz)
Output duty cycle(5) (8)
PNFLOOR
ODC
156.25 MHz
-160
dBc/Hz
%
45
55
(8) Parameter is specified for PLL outputs divided from either VCO domain.
(9) An output frequency over the fOUT max spec is possible, but the output swing may be less than the VOD min spec.
Copyright © 2018, Texas Instruments Incorporated
11
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Electrical Characteristics (continued)
Over Recommended Operating Conditions (unless otherwise noted)
PARAMETER
TEST CONDITIONS
MIN
TYP
MAX
UNIT
AC-CML Output Characteristics (OUTx)
fOUT
VOD
Output frequency(5) (9)
750
800
MHz
mV
Output voltage swing (VOH - VOL
)
400
150
600
Differential output voltage swing,
peak-to-peak
VODpp
VOS
tSK
2×VOD
V
mV
ps
ps
ps
Output common mode
Output-to-output skew(5)
550
100
300
125
Same post divider, output divide
values, and output type
20% to 80%, < 300 MHz
150
50
tR/tF
Output rise/fall time(5)
± 100 mV around center point, ≥ 300
MHz
PNFLOOR
ODC
Output duty cycle(5) (8)
Output duty cycle(5)
156.25 MHz
-160
dBc/Hz
%
45
55
AC-LVPECL Output Characteristics (OUTx)
fOUT
VOD
Output frequency(5) (9)
750
MHz
mV
Output voltage swing (VOH - VOL
)
500
0.3
850
1000
Differential output voltage swing,
peak-to-peak
VODpp
VOS
tSK
2×VOD
V
V
Output common mode
Output-to-output skew(5)
0.7
100
300
100
Same post divider, output divide
values, and output type
ps
ps
ps
20% to 80%, < 300 MHz
150
25
tR/tF
Output rise/fall time(5)
± 100 mV around center point, ≥ 300
MHz
Output phase noise floor
(fOFFSET > 10 MHz)
Output duty cycle(5) (8)
PNFLOOR
ODC
156.25 MHz
-162
dBc/Hz
%
45
55
HCSL Output Characteristics (OUTx)
fOUT
VOH
VOL
Output frequency(5) (9)
Output high voltage
Output low voltage
400
880
150
MHz
mV
600
-150
mV
Same post divider, output divide
values, and output type
tSK
Output-to-output skew(5)
Output slew rate(5)
100
6
ps
Measured from -150 mV to +150 mV
on the differential waveform
dV/dt
2.5
45
V/ns
Output phase noise floor (fOFFSET
> 10 MHz)
Output duty cycle(5) (8)
PNFLOOR
ODC
100 MHz
100 MHz
-158
dBc/Hz
%
55
3-Level Logic Input Characteristics (HW_SW_CTRL, STATUS[1:0])
VIH
Input high voltage
1.4
0.7
V
V
Input floating with internal bias and
PDN pulled low
VIM
Input mid voltage
0.9
VIL
IIH
Input low voltage
Input high current
Input low current
Input capacitance
0.4
40
40
V
VIH = VDD
VIL = GND
-40
-40
µA
µA
pF
IIL
CIN
2
12
Copyright © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Electrical Characteristics (continued)
Over Recommended Operating Conditions (unless otherwise noted)
PARAMETER
TEST CONDITIONS
MIN
TYP
MAX
UNIT
2-Level Logic Input Characteristics (PDN, GPIO[6:0], SDI, SCK, SCS, INSELx_[1:0])
VIH
VIL
IIH
Input high voltage
Input low voltage
Input high current
Input low current
Input capacitance
1.2
V
0.6
40
40
V
VIH = VDD
VIL = GND
-40
-40
µA
µA
pF
IIL
CIN
2
Logic Output Characteristics (STATUS[1:0], GPIO[6:5], SDO)
VOH
VOL
Output high voltage
Output low voltage
IOH = 1 mA
IOL = 1 mA
1.2
V
V
0.6
20
20% to 80%, LVCMOS mode, 1 kΩ to
GND
tR/tF
Output rise/fall time
500
ps
SPI Timing Requirements (SDI, SCK, SCS, SDO)
fSCK
t1
SPI clock rate
MHz
ns
SCS to SCK setup time
SDI to SCK setup time
SDI to SCK hold time
SCK high time
10
10
10
25
25
10
20
10
t2
ns
t3
ns
t4
ns
t5
SCK low time
ns
t6
SCK to SDO valid read-back data
SCS pulse width
ns
t7
ns
t8
SDI to SCK hold time
ns
I2C Interface Characteristics (SDA, SCL)
VIH
VIL
IIH
Input high voltage
Input low voltage
Input leakage
1.2
-15
V
V
0.5
15
µA
pF
V
CIN
VOL
Input capacitance
Output low voltage
2
IOL = 3 mA
0.3
100
400
Standard
kHz
kHz
µs
µs
µs
µs
ns
µs
ns
ns
ns
µs
fSCL
I2C clock rate
Fast mode
tSU(START)
tH(START)
tW(SCLH)
tW(SCLL)
tSU(SDA)
tH(SDA)
START condition setup time
START condition hold time
SCL pulse width high
SCL pulse width low
SDA setup time
SCL high before SDA low
SCL low after SDA low
0.6
0.6
0.6
1.3
100
0
SDA hold time
SDA valid after SCL low
0.9
300
300
300
tR(IN)
SDA/SCL input rise time
SDA/SCL input fall time
SDA output fall time
tF(IN)
tF(OUT)
CBUS ≤ 400 pF
tSU(STOP)
STOP condition setup time
0.6
1.3
Bus free time between STOP and
START
tBUS
µs
Copyright © 2018, Texas Instruments Incorporated
13
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Electrical Characteristics (continued)
Over Recommended Operating Conditions (unless otherwise noted)
PARAMETER
Other Characteristics
TEST CONDITIONS
MIN
TYP
MAX
UNIT
tPHO
Input-to-output phase offset
Zero delay mode
2
-70
-55
-70
-45
ns
VDDO_x = 2.5 V or 3.3 V, AC-DIFF or
HCSL output
Spur induced by power supply
noise (VN = 50 mVpp)(10) (11)
PSNR
VDDO_x = 2.5 V, LVCMOS output
dBc
dBc
VDDO_x = 1.8 V, AC-DIFF or HCSL
output
Spur induced by power supply
noise (VN = 25 mVpp)(10) (11)
PSNR
SPUR
VDDO_x = 1.8 V, LVCMOS output
fOUTx = 156.25 MHz, fOUTy = 155.52
MHz, AC-DIFF or HCSL (same output
type for both channels)
Spur level due to output-to-output
crosstalk (adjacent channels)(11)
-75
PLL Clock Output Performance Characteristics
RMS phase jitter
(12 kHz to 20 MHz)
156.25 MHz AC-DIFF or HCSL output,
fXO = 48.0048 MHz
RJ
150
250 fs RMS
dBc/Hz
PNTDC
Output close-in phase noise
(fOFFSET = 100 Hz)
122.88 MHz AC-DIFF or HCSL, 3-loop
mode, fXO = 48.0048 MHz, fTCXO = 10
MHz, fTCXO-TDC = 20 MHz, BWREF = 5
Hz, BWTCXO = 400 Hz
-112
0.01 to
4000
BW
JPK
DPLL bandwidth range(12)
Programmed bandwidth setting
Hz
dB
DPLL closed-loop jitter
peaking(13)
fIN = 25 MHz, fOUT = 10 MHz, DPLL
BW = 0.1 Hz or 10 Hz
0.1
Jitter modulation = 10 Hz,
25.78125 Gbps
JTOL
Jitter tolerance
6455
UI p-p
Valid for a single switchover event
between two clock inputs at the same
frequency
Phase transient during hitless
switch
tHITLESS
± 50
ps
Valid for a single switchover event
between two clock inputs at the same
frequency
Frequency transient during hitless
switch
fHITLESS
± 10
20
ppb
ms
From rising edge of PDN to free-
running output clocks
tSTARTUP
Initial PLL clock start-up time(14)
(10) PSNR is the single-sideband spur level (in dBc) measured when sinusoidal noise with ampitude VN and frequency between 100 kHz
and 1 MHz is injected onto VDD and VDDO_x pins.
(11) DJSPUR (ps pk-pk) = [2 × 10(dBc/20) / (π × fOUT) × 1E6], where dBc is the PSNR or SPUR level (in dBc) and fOUT is the output frequency
(in MHz).
(12) Actual loop bandwidth may be lower. Applies to REF-DPLL and TCXO-DPLL. The valid loop bandwidth range may be constrained by
the DPLL loop mode and REF-TDC and/or TCXO-TDC frequencies used in a given configuration.
(13) The TICS Pro software configures the closed-loop jitter peaking for 0.1 dB or less based on the programmed DPLL bandwidth setting.
(14) Assumes XO input clock is stable in frequency and amplitude before rising edge of PDN, PLLs start-up using parallel calibration mode,
VCO wait timers set to 0.4 ms, PLL wait timers set to 3 ms, and outputs auto-mute during APLL lock only (DPLL auto-mute options
disabled).
14
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
7.6 Timing Diagrams
t1
t4
t5
SCK
t2
–
SDI Write/Read
SDO Read
W/R
A14
D0/A0
DON‘T CARE
A13...D1/A1
t6
DON‘T CARE
D7
D1
D0
t7
SCS
t8
图 2. SPI Timing Parameters
ACK
STOP
STOP
START
tW(SCLL)
tf(SM)
tW(SCLH)
tr(SM)
VIH(SM)
VIL(SM)
SCL
th(START)
tSU(SDATA)
tr(SM)
th(SDATA)
tSU(START)
tBUS
tSU(STOP)
tf(SM)
VIH(SM)
VIL(SM)
SDA
图 3. I2C Timing Diagram
OUTx_P
OUTx_N
VOD
80%
0 V
VODpp = 2 × |VOD
|
20%
tR
tF
图 4. Differential Output Voltage and Rise/Fall Time
版权 © 2018, Texas Instruments Incorporated
15
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Timing Diagrams (接下页)
80%
VOUT,SE
OUT_REFx/2
20%
tR
tF
图 5. Single-Ended Output Voltage and Rise/Fall Time
INx_P
INx_P
Single Ended
Differential
INx_N
tPHO,DIFF
OUTx_P
Differential, PLL
OUTx_N
OUTx_P
OUTx_N
tSK,DIFF,INT
Differential, PLL
tSK,SE-DIFF,INT
Single Ended, PLL
OUTx_P/N
tPHO, SE
tSK,SE,INT
OUTx_P/N
Single Ended, PLL
图 6. Differential and Single-Ended Output Skew and Phase Offset
16
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
7.7 Typical Characteristics
Unless otherwise noted: VDD = 3.3 V, VDDO = 1.8 V, TA = 25°C, BWAPLL = 500 kHz, AC-LVPECL output.
The PLL output clock phase noise at different frequency offsets are determined by different noise contributors, such as
external clock input sources (REF IN, OCXO, XO) and internal noise sources (PLL, VCO), as well as the configured PLL loop
bandwidths (BWREF-DPLL, BWTCXO-DPLL, BWAPLL). The phase noise profile shown for each external clock source (fSOURCE) was
normalized to the PLL output frequency (fOUT) by adding 20×LOG10(fOUT / fSOURCE) to the measured source's phase noise.
-20
-40
-20
-40
19.44 MHz REF IN
12.8 MHz OCXO
48 MHz XO
25 MHz REF IN
10 MHz OCXO
48 MHz XO
156.25 MHz OUT
122.88 MHz OUT
-60
-60
-80
-80
-100
-120
-140
-160
-180
-100
-120
-140
-160
-180
1
10
100
1000 10000 100000 1E+6 1E+7
Frequency Offset (Hz)
1
10
100
1000 10000 100000 1E+6 1E+7
Frequency Offset (Hz)
D005
D003
Output Phase Jitter = 175-fs RMS (12 kHz to 20 MHz)
Output Phase Jitter = 160-fs RMS (12 kHz to 20 MHz)
fIN = 25 MHz, fTCXO = 10 MHz (OCXO), fXO = 48.0048 MHz,
fIN = 25 MHz, fTCXO = 10 MHz (OCXO) , fXO = 48.0048 MHz,
fTCXO-TDC = 20 MHz, BWREF-DPLL = 4 Hz, BWTCXO-DPLL = 200 Hz
图 7. 156.25-MHz Output Phase Noise (3-Loop)
With Phase Noise of External Inputs
fTCXO-TDC = 20 MHz, BWREF-DPLL = 5 Hz, BWTCXO-DPLL = 400 Hz
图 8. 122.88-MHz Output Phase Noise (3-Loop)
With Phase Noise of External Inputs
-20
-20
19.44 MHz REF IN
12.8 MHz OCXO
12.8 MHz OCXO
48 MHz XO
155.52 MHz OUT
48 MHz XO
156.25 MHz OUT
-40
-60
-40
-60
-80
-80
-100
-120
-140
-160
-180
-100
-120
-140
-160
-180
1
10
100
1000 10000 100000 1E+6 1E+7
Frequency Offset (Hz)
1
10
100
1000 10000 100000 1E+6 1E+7
Frequency Offset (Hz)
D004
D002
Output Phase Jitter = 170-fs RMS (12 kHz to 20 MHz)
Output Phase Jitter = 170-fs RMS (12 kHz to 20 MHz)
fTCXO = 12.8 MHz (OCXO), fXO = 48.0048 MHz,
fTCXO-TDC = 25.6 MHz, BWTCXO-DPLL = 200 Hz
fIN = 19.44 MHz, fTCXO = 12.8 MHz (OCXO), fXO = 48.0048 MHz,
fTCXO-TDC = 25.6 MHz, BWREF-DPLL = 1 Hz, BWTCXO-DPLL = 200 Hz
图 9. 155.52-MHz Output Phase Noise (3-Loop)
图 10. 156.25-MHz Output Phase Noise (2-Loop TCXO)
With Phase Noise of External Inputs
With Phase Noise of External Inputs
版权 © 2018, Texas Instruments Incorporated
17
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
8 Parameter Measurement Information
8.1 Output Clock Test Configurations
High-impedance
probe
LVCMOS
LMK05028
Oscilloscope
2 pF
图 11. LVCMOS Output Test Configuration
Phase Noise/
Spectrum
Analyzer
LVCMOS
LMK05028
Copyright © 2018, Texas Instruments Incorporated
图 12. LVCMOS Output Phase Noise Test Configuration
Oscilloscope
(50-ꢀ inputs)
AC-LVPECL, AC-LVDS, AC-CML
LMK05028
图 13. AC-LVPECL, AC-LVDS, AC-CML Output AC Test Configuration
Phase Noise/
Spectrum Analyzer
LMK05028
Balun
AC-LVPECL, AC-LVDS, AC-CML
图 14. AC-LVPECL, AC-LVDS, AC-CML Output Phase Noise Test Configuration
0 ꢀ
Oscilloscope
(50-ꢀ inputs)
LMK05028
HCSL
0 ꢀ
图 15. HCSL Output Test Configuration
18
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Output Clock Test Configurations (接下页)
Opt œ 33 ꢀ
HCSL
Phase Noise/
LMK05028
Spectrum
Analyzer
Balun
Opt œ 33 ꢀ
HCSL
50 ꢀ
50 ꢀ
Copyright © 2018, Texas Instruments Incorporated
图 16. HCSL Output Phase Noise Test Configuration
Sine wave
Modulator
Power Supply
Phase Noise/
Spectrum
Analyzer
Signal Generator
LMK05028
Device Output
Balun
Reference
Input
Copyright © 2018, Texas Instruments Incorporated
Single-sideband spur level measured in dBc with a known noise amplitude and frequency injected onto the device
power supply.
图 17. Power Supply Noise Rejection (PSNR) Test Configuration
版权 © 2018, Texas Instruments Incorporated
19
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9 Detailed Description
9.1 Overview
The LMK05028 is a high-performance network synchronizer clock device that provides jitter cleaning, clock
generation, advanced clock monitoring, and superior hitless switching performance to meet the stringent timing
requirements of communications infrastructure and industrial applications.
The LMK05028 features four reference inputs, two independent PLL channels, and eight output clocks with RMS
phase jitter of 150-fs typical. The flexible PLL channels provide programmable loop bandwidths for input jitter and
wander attenuation and fractional-N PLL frequency synthesis to generate any output frequency from any input
frequency. Each PLL channel has three phase-locked loops comprised of two digital PLLs (DPLLs) and one
analog PLL (APLL) with a low-noise integrated VCO. Each channel supports three-loop or two-loop mode
configurations to optimize clock performance and solution cost for different use cases.
The reference input muxes support automatic input selection or manual input selection through software or pin
control. The reference switchover event will be hitless with superior phase transient performance (50-ps typical).
The reference clock input monitoring block monitors the clock inputs and will perform a switchover or holdover
when a loss of reference (LOR) is detected. A LOR condition can be detected upon any violation of the threshold
limits set for the input monitors, which include amplitude, frequency, missing pulse, runt pulse, and 1-PPS (pulse-
per-second) detectors. The threshold limits for each input detector can be set and enabled independently per
clock input. The tuning word history monitor feature allows the initial output frequency accuracy upon entry into
holdover to be determined by the historical average frequency when locked, minimizing the frequency and phase
disturbance during a LOR condition.
The device has eight outputs with programmable drivers, allowing up to eight differential clocks, eight LVCMOS
pairs (two outputs per pair), or a combination of both. The output clocks can be selected from either PLL/VCO
domain by the output muxes. A 1-PPS output can be supported on outputs 0 and 7. The output dividers have a
SYNC feature to allow multiple outputs to be phase-aligned. If needed, zero delay can be enabled to achieve a
deterministic phase offset between any specified PLL output clock and its selected input clock.
To support IEEE 1588 PTP slave clock or other clock steering applications, each PLL channel also supports
DCO mode with <1-ppt (part per trillion) frequency resolution for precise frequency and phase adjustment
through external software or pin control.
The device is fully programmable through I2C or SPI and features custom start-up frequency configuration with
the internal EEPROM, which is custom factory pre-programmable and in-system programmable. Internal LDO
regulators provide excellent PSNR to reduce the cost and complexity of the power delivery network. The clock
input and PLL monitoring status can be observed through the status pins and interrupt registers for full diagnostic
capability.
9.1.1 ITU-T G.8262 (SyncE) Standards Compliance
The LMK05028 meets the applicable requirements of the ITU-T G.8262 (SyncE) standard. See the Application
Report, ITU-T G.8262 Compliance Test Result for the LMK05028 (SNAA315).
20
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.2 Functional Block Diagrams
TCXO/OCXO determines output frequency stability.
(e.g. Df/f0 ≤ ±4.6 ppm for SyncE)
XO determines output jitter
in 12 kHz to 20 MHz band.
VDD (x9)
3.3 V
VDDO (x6)
1.8 / 2.5 / 3.3 V
XO_N
XO_P
TCXO_IN
DPLL1 Zero Delay
Clock
Bus
Feedback
TCXO/XO
Monitors
Power Conditioning
TCXO
(all blocks)
4
×2
XO
OUT[4:7]
Clocks
0
1
2
3
System
÷M
÷ODB
11-b
÷OD
20-b
Ref Inputs
OUT7
OUT6
INSEL0_[1:0]
IN0
PLL Channels
PLL1
0
1
2
3
÷OD
20-b
VCO1:
4.8 to 5.4 GHz
REF A
÷R
IN1
IN2
IN3
Hitless
Switching
and
Priority
Selection
0
1
APLL
÷P1
÷P2
REF
DPLL
TCXO
DPLL
OUT5
OUT4
0
1
2
3
÷OD
20-b
DCO1
FDEV
VCO1 FB
VCO2 FB
VCO2 FB (IN5)
OUT3
OUT2
0
1
2
3
÷OD
20-b
PLL2
REF B
÷R
Hitless
Switching
and
Priority
Selection
2
3
÷P1
÷P2
VCO2:
APLL
REF
DPLL
TCXO
DPLL
0
1
2
3
4
5
5.5 to 6.2 GHz
÷OD
20-b
OUT1
OUT0
VCO1 FB
(IN4)
Input
Monitors
DCO2
FDEV
PLL
Monitors
INSEL1_[1:0]
0
1
2
3
4
5
Digital
÷ODB
11-b
÷OD
20-b
SDA
SCL
GPIO1/SCS
GPIO2/SDO
Ref Bypass Mux
EEPROM
ROM
I2C/
SPI
Registers
LMK05028
TCXO
REF A
REF B
0
1
2
5
Low-Jitter Dual-Channel
Network Synchronizer Clock
HW_SW_CTRL
PDN
GPIO0/SYNCN
STATUS[0:1]
Outputs
OUT[0:3]
Device Control and Status
DPLL2 Zero Delay
Feedback
SYNC
LF2
CAP
(3x)
LF1
图 18. Top-Level Device Block Diagram
版权 © 2018, Texas Instruments Incorporated
21
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Functional Block Diagrams (接下页)
9.2.1 PLL Architecture Overview
图 19 shows the 3-loop architecture implemented the same for both PLL channels with exception of the VCO
frequency range. Each channel has three phase-locked loops with two digital PLLs (REF-DPLL and TCXO-
DPLL) and one analog PLL (APLL) with integrated VCO. The REF-DPLL and TCXO-DPLL are each comprised
of a time-to-digital converter (TDC), digital loop filter (DLF), feedback prescaler (PR), and 40-bit fractional
feedback (FB) divider with sigma-delta-modulator (SDM or "MASH"). The APLL is comprised of a reference
frequency doubler (×2), phase-frequency detector (PFD), loop filter (LF), fractional feedback (N) divider with
SDM, and VCO.
XO
VCOx FB
×2
÷M
TCXO
VCO1 FB
VCO2 FB
Cascaded PLL
option (3)
5-bit
PLLx Channel
Loopback
Dividers
VCO1: 4.8 to 5.4 GHz
VCO2: 5.5 to 6.2 GHz
REF Mux
TCXO Mux
0
1 2
0
1
2
3
4
5
IN0
IN1
IN2
IN3
Post
Dividers
REF-DPLL
TCXO-DPLL
APLL
fREF-TDC
fTCXO-TDC
fPD
VCO
fVCO
fVCO/P1
÷R
×2
÷P1
TDC
DLF
TDC
DLF
PFD
LF
16-bit
(x6)
/4 to /9,
/11, /13
/2 to /17
÷PR
/2 to /17
÷PR
fVCO/P2
÷FB
40-bit Frac-N SDM
÷FB
40-bit Frac-N SDM
÷N
40-bit Frac-N SDM
VCO1 FB
VCO2 FB
÷P2
To
Output
Muxes
DPLL feedback clock
SDM control (1)
DCO option (2)
1
0
DCO
FDEV
FINC/FDEC
38-bit
DCO loop select
(1) Fractional divider SDM control node depends on the selected PLL mode configuration.
(2) DCO frequency adjustments can be software or pin controlled.
(3) PLL cascading options: a) PLL1 VCO1 Loopback to PLL2, b) PLL2 VCO2 Loopback to PLL1.
图 19. PLL Architecture (One Channel)
表 2 summarizes the PLL mode configuration options available in each channel. These modes support a wide
range of use cases depending to the clock functionality and performance required in the application. Most
applications will use 2-loop REF-DPLL or 3-loop mode for network synchronization clock features such as
programmable loop bandwidth for jitter and wander attenuation, hitless switching, precise digital holdover, DCO
frequency steering, and/or zero delay mode.
表 2. PLL Mode Configuration Options
FRACTIONAL SDM ENABLED (SDM CONTROL FROM)
FREE-RUN /
HOLDOVER
CLOCK
PLL MODE
DCO MODE OPTION
REF-DPLL
TCXO-DPLL
APLL
1-Loop (APLL only)
2-Loop REF-DPLL
2-Loop TCXO-DPLL
3-Loop
–
–
Y (Free-run from XO)
Y (REF-DPLL)
–
XO
XO
Y (DCO option)
–
–
REF-DPLL SDM
TCXO-DPLL SDM
REF-DPLL SDM
Y (DCO option)
Y (REF-DPLL)
Y (TCXO-DPLL)
Y (TCXO-DPLL)
TCXO
TCXO
Y (DCO option)
22
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
The following sections describe the basic principle of operation for 2-loop and 3-loop modes. See PLL Operating
Modes for more details on the PLL modes of operation including holdover.
9.2.2 3-Loop Mode
In 3-loop mode, the TCXO/OCXO source determines the free-run and holdover frequency stability and accuracy,
and the XO source determines the output phase noise and jitter performance over the 12-kHz to 20-MHz
integration band. 3-loop mode allows the use of a cost-effective, low-frequency TCXO/OCXO (such as 10 or 12.8
MHz) to support standards-compliant frequency stability and low loop bandwidth (≤10 Hz) required in
synchronization applications like SyncE and SONET/SDH.
The principle of operation for 3-loop mode is as follows. After power-on reset and initialization, the APLL locks
the VCO to the external XO input clock and operates in free-run mode. Once the external TCXO/OCXO input
clock is detected, the TCXO-DPLL begins lock acquisition. The TCXO TDC compares the phase of the
TCXO/OCXO clock and the TCXO FB divider clock (from the VCO) and generates a digital correction word
corresponding to the phase error. The correction word is filtered by the TCXO DLF, and its output controls the
APLL N divider SDM to pull the VCO frequency until it is locked to the TCXO/OCXO clock. After a valid reference
input is selected, the REF-DPLL enters lock acquisition mode. The REF TDC compares the phase of the
selected input clock and the REF FB divider clock (from the VCO) and generates a digital correction word. The
correction word is filtered by the REF DLF, and its output controls the TCXO FB divider SDM which translates to
a frequency offset to the TCXO TDC. This frequency correction propagates through the TCXO-DPLL which then
controls the APLL N divider SDM to pull the VCO frequency until it is locked to the selected reference input
clock.
If DCO mode is enabled on the REF-DPLL, a frequency deviation step value (FDEV) can be programmed and
used to adjust (increment or decrement) the REF FB divider SDM, where the frequency adjustment effectively
propagates through the 3 nested loops to the VCO output.
To ensure proper loop stability in 3-loop mode, the REF-DPLL has the lowest loop bandwidth (BWREF-DPLL ≤ 80
Hz, typical), the TCXO-DPLL has a higher loop bandwidth (BWREF-DPLL × 50 ≤ BWTCXO-DPLL ≤ 4 kHz), and the
APLL has the highest bandwidth (BWAPLL is approximately 500 kHz typical).
When operating in 3-loop mode and all reference inputs to the REF-DPLL are lost, the PLL channel will enter
holdover mode and operate similar to 2-loop TCXO-DPLL mode.
XO
VCOx FB
×2
÷M
TCXO
VCO1 FB
VCO2 FB
5 bit
PLLx Channel
Loopback
Dividers
VCO1: 4.8 to 5.4 GHz
VCO2: 5.5 to 6.2 GHz
REF Mux
TCXO Mux
0
1 2
0
1
2
3
4
5
IN0
IN1
IN2
IN3
Post
Dividers
REF-DPLL
TCXO-DPLL
APLL
fREF-TDC
fTCXO-TDC
fPD
VCO
fVCO
fVCO/P1
÷R
×2
÷P1
TDC
DLF
TDC
DLF
PFD
LF
16-bit
(x6)
/4 to /9,
/11, /13
/2 to /17
÷PR
/2 to /17
÷PR
fVCO/P2
÷FB
40-bit Frac-N SDM
÷FB
40-bit Frac-N SDM
÷N
40-bit Frac-N SDM
VCO1 FB
VCO2 FB
÷P2
To
Output
Muxes
DPLL feedback clock
DCO option
1
0
DCO
FDEV
FINC/FDEC
38-bit
图 20. 3-Loop Mode with DCO Option
版权 © 2018, Texas Instruments Incorporated
23
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.2.2.1 PLL Output Clock Phase Noise Analysis in 3-Loop Mode
The following plot shows an example PLL clock output phase noise profile in 3-loop mode. The PLL output clock
phase noise at different frequency offsets are determined by different noise contributors, such as external clock
input sources (REF IN, OCXO, XO) and internal noise sources (PLL, VCO), as well as the configured PLL loop
bandwidths (BWREF-DPLL, BWTCXO-DPLL, BWAPLL). The phase noise profile shown for each external clock source
(fSOURCE) was normalized to the PLL output frequency (fOUT) by adding 20×LOG10(fOUT / fSOURCE) to the measured
source's phase noise. The PLL output phase noise can be analyzed as follows:
1. Below BWREF-DPLL, the REF input noise contributes to output.
2. Above BWREF-DPLL, the REF input noise is attenuated by REF-DPLL bandwidth with up to 60-dB/decade roll-
off.
3. Below BWTCXO-DPLL, the OCXO noise and TDC noise of the TCXO-DPLL determine the output noise here.
The close-in phase noise is –112 dBc/Hz at 100-Hz offset.
4. Above BWTCXO-DPLL, the OCXO noise is attenuated. Below BWAPLL, the APLL noise dominates as the XO
noise contribution is much lower.
5. Above BWAPLL, the VCO noise dominates and the XO noise is attenuated.
6. AC-LVPECL output noise floor is -163 dBc/Hz at 10-MHz offset.
1
BWREF-DPLL
5 Hz
BWTCXO-DPLL
400 Hz
2
3
4
BWAPLL
500 kHz
5
6
Output Phase Jitter = 160-fs RMS (12 kHz to 20 MHz)
AC-LVPECL output, fIN = 25 MHz, fTCXO = 10 MHz (OCXO), fXO = 48.0048 MHz, fTCXO-TDC = 20 MHz
图 21. 122.88-MHz Output Phase Noise (3-Loop) With Phase Noise of External Inputs
24
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.2.3 2-Loop REF-DPLL Mode
For applications that do not require the higher frequency stability in holdover mode and added cost of a
TCXO/OCXO, 2-loop mode allows the XO input to determine the free-run and holdover frequency accuracy and
also supports higher loop bandwidth.
The principle of operation for 2-loop REF-DPLL mode is similar to 3-loop mode except the TCXO-DPLL stage is
bypassed. After power-on reset and initialization, the APLL locks the VCO to the external XO input clock and
operates in free-run mode. After a valid reference input is selected, the REF-DPLL enters lock acquisition mode.
The REF TDC compares the phase of the selected input clock and the REF FB divider clock (from the VCO) and
generates a digital correction word. The correction word is filtered by the REF DLF, and its output controls the
APLL N divider SDM to pull the VCO frequency until it is locked to the selected reference input clock.
If DCO mode is enabled on the REF-DPLL, a frequency deviation step value (FDEV) can be programmed and
used to adjust (increment or decrement) the REF FB divider SDM, where the frequency adjustment effectively
propagates through the two nested loops to the VCO output.
In 2-loop mode, the REF-DPLL loop bandwidth (BWREF-DPLL) must be less than fINx/50 and less than the
maximum bandwidth of 4 kHz.
When operating in 2-loop mode and all reference inputs to the REF-DPLL are lost, the PLL channel will enter
holdover mode and operate similar to 1-loop APLL only mode.
VCOx FB
XO
PLLx Channel
Loopback
Dividers
VCO1: 4.8 to 5.4 GHz
VCO2: 5.5 to 6.2 GHz
REF Mux
0
1
2
3
4
5
IN0
IN1
IN2
IN3
Post
Dividers
REF-DPLL
APLL
fREF-TDC
fPD
VCO
fVCO/P1
÷R
×2
÷P1
TDC
DLF
PFD
LF
16-bit
(x6)
fVCO
/4 to /9,
/11, /13
/2 to /17
÷PR
fVCO/P2
÷FB
40-bit Frac-N SDM
÷N
VCO1 FB
VCO2 FB
40-bit Frac-N SDM
÷P2
To
Output
Muxes
DPLL feedback clock
DCO option
1
0
DCO
FDEV
FINC/FDEC
38-bit
图 22. 2-Loop REF-DPLL Mode with DCO Option
版权 © 2018, Texas Instruments Incorporated
25
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.2.4 2-Loop TCXO-DPLL Mode
The principle of operation for 2-loop TCXO-DPLL mode is similar to 3-loop mode except the REF-DPLL stage is
bypassed. After power-on reset and initialization, the APLL locks the VCO to the external XO input clock and
operates in free-run mode. Once the external TCXO/OCXO input clock is detected, the TCXO-DPLL begins lock
acquisition. The TCXO TDC compares the phase of the TCXO/OCXO clock and the TCXO FB divider clock (from
the VCO) and generates a digital correction word corresponding to the phase error. The correction word is
filtered by the TCXO DLF, and its output controls the APLL N divider SDM to pull the VCO frequency until it is
locked to the TCXO/OCXO clock.
If DCO mode is enabled on the TCXO-DPLL, a frequency deviation step value (FDEV) can be programmed and
used to adjust (increment or decrement) the TCXO FB divider SDM, where the frequency adjustment effectively
propagates through the two nested loops to the VCO output.
In 2-loop mode, the TCXO-DPLL loop bandwidth (BWTCXO-DPLL) must less fTCXO/50 and less than the maximum
bandwidth of 4 kHz.
When operating in 2-loop mode and the TCXO input is lost, the PLL channel will operate similar to 1-loop APLL
only mode.
VCOx FB
XO
×2
÷M
TCXO
VCO1 FB
VCO2 FB
5-bit
Loopback
Dividers
VCO1: 4.8 to 5.4 GHz
VCO2: 5.5 to 6.2 GHz
TCXO Mux
0 1 2
Post
Dividers
TCXO-DPLL
APLL
fTCXO-TDC
fPD
VCO
fVCO/P1
×2
÷P1
TDC
DLF
PFD
LF
fVCO
/4 to /9,
/11, /13
/2 to /17
÷PR
fVCO/P2
÷FB
40-bit Frac-N SDM
÷N
40-bit Frac-N SDM
÷P2
To
Output
Muxes
DPLL feedback clock
PLLx Channel
DCO option
1
DCO
FDEV
FINC/FDEC
0
38-bit
图 23. 2-Loop TCXO-DPLL Mode With DCO Option
26
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.2.5 PLL Configurations for Common Applications
表 3. Example PLL Configurations Based on Common Application Design Parameters
KEY DESIGN GOALS
TYPICAL LOOP CONFIGURATION
WANDER
ATTENUATION
OR TCXO/OCXO
HOLDOVER
MARKET SEGMENT /
APPLICATION
BEST CLOSE-IN
PHASE NOISE
(100-Hz OFFSET)
HITLESS
SWITCHING
REF-DPLL
PLL MODE
BANDWIDTH
SyncE EEC Opt. 1,
SDH G.813 Opt. 1
YES
YES
–
–
YES
YES
3-Loop
3-Loop
1 to 10 Hz
0.1 Hz
SyncE EEC Opt. 2,
SONET GR-253
SyncE/IEEE 1588
PTP Slave
3-Loop with
DCO Mode Enabled
See above for SyncE
EEC Opt. 1 or 2
YES
–
–
–
YES
YES
YES
1-PPS Input
3-Loop
0.02 Hz
3-Loop with
DCO Mode Option
Wireless/BTS
YES
YES
1 to 20 Hz
Test Instrumentation
(for example, 10-MHz
Ref. In)
Optional
YES
YES
3-Loop
1 to 40 Hz
Medical Imaging
OTN/OTU
–
–
YES
–
Optional
–
2-Loop TCXO
2-Loop REF
100 to 400 Hz
100 to 300 Hz
Broadcast
(Genlock)
Optional
Optional
–
–
YES (TCXO)
–
2-Loop REF
2-Loop REF
1 to 10 Hz
Other Jitter Cleaning
10 to 100 Hz
版权 © 2018, Texas Instruments Incorporated
27
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.3 Feature Description
The following sections describe the features and functional blocks of the LMK05028.
9.3.1 Oscillator Input (XO_P/N)
The XO input is the reference clock for the fractional-N APLLs. The combination of XO and APLL determines the
jitter and phase noise performance of the output clocks. For optimal performance, the XO frequency should be at
least 48 MHz and have a non-integer frequency relationship with the VCO frequencies so the APLLs operate in
fractional mode. When the TCXO input is not used by either PLL channel, the XO input determines the output
frequency accuracy and stability in free-run or holdover modes.
The XO input buffer has programmable input on-chip termination and AC-coupled input biasing configurations as
shown in 图 24.
The buffered XO path also drives the input monitoring blocks as well as output muxes, allowing buffered copies
of the XO input on OUT0 and/or OUT1.
LMK05028
28 pF
XO_P
100 kꢀ
S1
S2
VAC-DIFF
(weak bias)
Differential or
Single-Ended*
S3
50 ꢀ
100 ꢀ
XO path
S2
100 kꢀ
28 pF
XO_ N
*Supports 2.5-V
single-ended swing
S1
50 ꢀ
图 24. XO Input Buffer
表 4 lists the typical XO input buffer configurations for common clock interface types.
表 4. XO Input Buffer Modes
INTERNAL SWITCH SETTINGS
XO_DIFF_TYPE
INPUT TYPES
INTERNAL TERM. (S1, S2)(1)
INTERNAL BIAS (S3)(2)
LVDS, CML, LVPECL, LVCMOS
(DC-coupled)
0h
1h
3h
4h
OFF
OFF
OFF
LVDS, CML, LVPECL
(AC-coupled)
ON (1.3 V)
ON (1.3 V)
OFF
LVDS, CML, LVPECL
(AC-coupled, internal 100-Ω)
100 Ω
50 Ω
HCSL
(DC-coupled, internal 50-Ω)
(1) S1, S2: OFF = External termination is assumed.
(2) S3: OFF = External input bias or DC coupling is assumed.
28
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.2 TCXO/OCXO Input (TCXO_IN)
The TCXO input is the reference clock to the TCXO-DPLL loop in each PLL channel. When the PLL channel
uses the TCXO-DPLL, the TCXO input source determines the close-in phase noise and wander performance
(MTIE/TDEV) when the DPLL is locked, as well as the frequency accuracy and stability in free-run and holdover
modes. A TCXO input with high phase noise floor should have minimal or no impact on the output jitter
performance, provided the TCXO loop bandwidth is designed low enough to attenuate its noise contribution. This
input can be driven from a low-frequency TCXO, OCXO, or external traceable clock that conforms to the
frequency accuracy and holdover stability requirements of the application. TCXO and OCXO frequencies of 10 to
12.8 MHz are widely available and cost-effective options.
The TCXO input can accept an AC-coupled single-ended clock (sine, clipped-sine, or square wave) and has an
internal weak bias of about 0.6 V. The input voltage swing should be less than 1.3 Vpp and terminated before
AC-coupling to the pin. If unused, the TCXO input buffer can be powered down by register bit and the pin can be
left floating.
The buffered TCXO path also drives the input monitoring blocks as well as the TCXO/Ref bypass mux to the
output muxes, allowing a buffered copy of the TCXO input on OUT0 and/or OUT1.
LMK05028
11 kꢀ
1.2-V
TCXO_IN
LVCMOS
~0.6 V
Internal
(weak bias)
TCXO path
图 25. TCXO Input Buffer
版权 © 2018, Texas Instruments Incorporated
29
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.3.3 Reference Inputs (INx_P/N)
The reference inputs (IN0 to IN3) can accept differential or single-ended clocks to synchronize any of the PLL
channels. Each input has programmable input type, termination, and AC-coupled input biasing configurations as
shown in 图 26. Each input buffer drives the reference input mux of both DPLL blocks. The DPLL input mux can
select from any of the reference inputs. Any DPLL can switch between inputs with different frequencies provided
they can be divided-down to a common frequency by DPLL R dividers (DPLLy_REFx_RDIV). The reference
input paths also drive the various detector blocks for reference input monitoring and validation. The selected
reference of each DPLL input (before the R divider) can be routed through the TCXO/Ref bypass mux and output
muxes, allowing a buffered copy of either DPLL reference input on OUT0 and/or OUT1.
LMK05028
To LVCMOS input
slew rate detector
3.6 kꢀ
S4
28 pF
INx_P
100 kꢀ
S1
S2
Differential or
Single-Ended*
VAC-DIFF
(weak bias)
S3
50 ꢀ
100 ꢀ
REF path
S2
100 kꢀ
28 pF
INx_ N
*Supports 3.3-V
S-E input swing
S1
50 ꢀ
图 26. Reference Input Buffer
表 5 lists the reference input buffer configurations for common clock interface types.
表 5. Reference Input Buffer Modes
INTERNAL SWITCH SETTINGS
LVCMOS SLEW
RATE DETECT
(S4)(3)
REFx_TYPE
INPUT TYPES
INTERNAL TERM.
(S1, S2)(1)
INTERNAL BIAS
(S3)(2)
LVDS, CML, LVPECL
(DC-coupled)
0h
1h
3h
OFF
OFF
OFF
OFF
OFF
OFF
LVDS, CML, LVPECL
(AC-coupled)
ON (1.3 V)
ON (1.3 V)
LVDS, CML, LVPECL
(AC-coupled, internal 100-Ω)
100 Ω
HCSL
4h
8h
50 Ω
OFF
OFF
OFF
ON
(DC-coupled, internal 50-Ω)
LVCMOS
OFF
(1) S1, S2: OFF = External termination is assumed.
(2) S3: OFF = External input bias or DC coupling is assumed.
(3) S4: OFF = Differential input amplitude detector is used for all input types except LVCMOS.
30
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.4 Clock Input Interfacing and Termination
图 27 through 图 34 show the recommended input interfacing and termination circuits. Unused clock inputs can
be left floating or pulled down.
VDD
Rs
R1
LVCMOS
Driver
XO_P
XO_N
50 W
LMK05028
(ROUT
)
Rs = 50 œ ROUT
R2
VDD
3.3 V
2.5 V
1.8 V
R1 (ꢀ) R2 (ꢀ)
125
0
375
open
open
0
图 27. Single-Ended LVCMOS to XO Input (XO_P)
VDD
TCXO
_IN
Rs
Clock Driver
(ROUT
50 W
LMK05028
)
Swing
<1.3 Vpp
Rt
Driver
Rs (ꢀ)
100
68
Rt (ꢀ)
50
50
3.3-V LVCMOS
2.5-V LVCMOS
1.8-V LVCMOS
33
50
(Clipped) Sinewave
open
0
图 28. Single-Ended LVCMOS or Sinewave to TCXO Input (TCXO_IN)
Rs
LVCMOS
3.3V LVCMOS
LMK05028
Driver
Copyright © 2018, Texas Instruments Incorporated
图 29. Single-Ended LVCMOS (1.8, 2.5, 3.3 V) to Reference (INx_P)
版权 © 2018, Texas Instruments Incorporated
31
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
LVPECL Driver
LVPECL
LMK05028
50 ꢀ
50 ꢀ
VDD_IN - 2
Copyright © 2018, Texas Instruments Incorporated
图 30. DC-Coupled LVPECL to Reference (INx) or XO Inputs
LMK05028
LVDS Driver
100 ꢀ
LVDS
Copyright © 2018, Texas Instruments Incorporated
图 31. DC-Coupled LVDS to Reference (INx) or XO Inputs
CML
Driver
LMK05028
CML
Copyright © 2018, Texas Instruments Incorporated
图 32. DC-Coupled CML (Source Terminated) to Reference (INx) or XO Inputs
50 ꢀ
HCSL
Driver
LMK05028
HCSL
50 ꢀ
Copyright © 2018, Texas Instruments Incorporated
图 33. HCSL (Load Terminated) to Reference (INx) or XO Inputs
32
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Driver
LVDS
RB (ꢀ)
open
open
150
LMK05028
Differential
Driver
100 ꢀ
CML*
3.3-V LVPECL
2.5-V LVPECL
HCSL
82
Internal input biasing
RB
RB
50
*CML driver has 50-ꢀ pull-up
图 34. AC-Coupled Differential to Reference (INx) or XO Inputs
9.3.5 Reference Input Mux Selection
For each REF-DPLL block, the reference input mux selection can be done automatically using an internal state
machine with a configurable input priority scheme, or manually through software register control or hardware pin
control. The input mux can select from IN0 to IN3. Additionally, DPLL1 can select IN5 as an internal loopback
clock divided-down from PLL2's VCO (VCO2 FB clock), and DPLL2 can select IN4 as an internal loopback clock
from PLL1's VCO (VCO1 FB clock).
The priority for all inputs can be assigned for each DPLL through registers. The priority ranges from 0 to 6, where
0 means Ignored (never select) and 1 to 6 are highest (1st) to lowest (6th) priority. When two or more inputs are
configured with the same priority setting, the reference input with the lowest index (INx) will be given higher
priority.
The currently selected reference input for each DPLL can be read through the status pin or register.
9.3.5.1 Automatic Input Selection
There are two automatic input selection modes that can be set by a register: Auto Revertive and Auto Non-
Revertive.
•
Auto Revertive: In this mode, the DPLL automatically selects the valid input with the highest configured
priority. If a clock with higher priority becomes valid, the DPLL will automatically switch over to that clock
immediately.
•
Auto Non-Revertive: In this mode, the DPLL automatically selects the highest priority input that is valid. If a
higher priority input because valid, the DPLL will not switch-over until the currently selected input becomes
invalid.
9.3.5.2 Manual Input Selection
There are two manual input selection modes that can be set by a register: Manual with Auto-Fallback and
Manual with Auto-Holdover. In either manual mode, the input selection can be done through register control (表
6) or hardware pin control (表 7).
•
Manual with Auto-Fallback: In this mode, the manually selected reference is the active reference until it
becomes invalid. If the reference becomes invalid, the DPLL will automatically fallback to the highest priority
input that is valid or qualified. If no prioritized inputs are valid, the DPLL will enter holdover mode (if tuning
word history is valid) or free-run mode. The DPLL will exit holdover mode when the selected input becomes
valid.
•
Manual with Auto-Holdover: In this mode, the manually selected reference is the active reference until it
becomes invalid. If the reference becomes invalid, the DPLL will automatically enter holdover mode (if tuning
word history is valid) or free-run mode. The DPLL will exit holdover mode when the selected input becomes
valid.
表 6. Manual Input Selection by Register Bits
DPLLx_REF_MAN_REG_SEL[2:0]
DPLLx_REF_MAN_SEL BIT
SELECTED INPUT
BITS
000b
001b
010b
011b
0
0
0
0
IN0
IN1
IN2
IN3
版权 © 2018, Texas Instruments Incorporated
33
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
表 6. Manual Input Selection by Register Bits (接下页)
DPLLx_REF_MAN_REG_SEL[2:0]
BITS
DPLLx_REF_MAN_SEL BIT
SELECTED INPUT
100b
101b
0
0
VCO1 Loopback to DPLL2
VCO2 Loopback to DPLL1
表 7. Manual Input Selection by Hardware Pins
INSELx_[1:0] PINS
DPLLx_REF_MAN_SEL BIT
SELECTED INPUT
00b
01b
10b
11b
1
1
1
1
IN0
IN1
IN2
IN3
The reference input selection flowchart is shown in 图 35.
See Device POR and
PLL Initialization and
DPLL Modes Flowcharts
REF-DPLL
Locked
Yes: With
Auto-Holdover
Yes: Auto
Revertive
No
Input Select Mode
= Manual?
Input Select Mode
= Auto?
Yes: With
Auto-Fallback
LOR on
Yes: Auto
Non-Revertive
No
Selected Input, or
Higher Priority Input
Valid?
No
Loss of Ref (LOR)
on Selected Input?
Yes
No
Loss of Ref (LOR) on
Selected Input?
Yes
Yes
Holdover Mode
Holdover Mode
No
No
Higher Priority
Input Valid?
Manually Selected
Input Valid?
Yes: Auto-Switch
according to Priority
settings
Yes: Switch to
Selected Input
Lock Acquisition
(Fastlock, Hitless Switch)
图 35. Reference Input Selection Flowchart
9.3.6 Hitless Switching
Each REF-DPLL supports hitless switching through a proprietary phase cancellation scheme, which can be
enabled per DPLL. When hitless switching is enabled, it will prevent a phase transient (phase hit) from
propagating to the outputs when the two switched inputs have a fixed phase offset and are frequency-locked.
The inputs are frequency-locked when they have same exact frequency (0-ppm offset), or have frequencies that
are integer-related and can each be divided to a common frequency by integers. When hitless switching is
34
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
disabled, a phase hit equal to the phase offset between the two inputs will be propagated to the output at a rate
determined by the REF-DPLL fastlock bandwidth. The hitless switching specifications (tHITLESS and fHITLESS) are
valid for reference inputs with no wander. In the case where two inputs are switched but are not frequency-
locked, the output smoothly transitions to the new frequency with reduced transient. Hitless switching is not
supported for 1-PPS inputs.
9.3.7 Gapped Clock Support on Reference Inputs
Each DPLL supports locking to an input clock that has missing periods and is referred to as a gapped clock.
Gapping a clock severely increases its jitter, so the device provides the high input jitter tolerance and low loop
bandwidth necessary to generate a low-jitter periodic output clock. The resulting output will be a periodic non-
gapped clock with an average frequency of the input with its missing cycles. The gapped clock width cannot be
longer be than the reference clock period after the R divider (RINx / fINx). The reference input monitors should be
configured to avoid any flags due to the worst-case clock gapping scenario to achieve and maintain lock.
Reference switchover between two gapped clock inputs may violate the hitless switching specification if the
switch occurs during a gap in either input clock.
9.3.8 Input Clock and PLL Monitoring, Status, and Interrupts
The following section describes the input clock and PLL monitoring, status, and interrupt features.
XO
TCXO
DIFF: Min. Swing
Status Bits
LOS_TCXO
EN
EN
2
2
LOS
Amplitude
Frequency
LOS_FDET_TCXO
LOS_XO
LOS_FDET
LOS_FDET_XO
TCXO/XO Monitors (x2)
Ref Inputs (x4)
4
REF
Muxes
(x2)
Clock Status
PLL Channels
INx
÷R
(x2)
LOR_AMP1
LOR_FREQ1
LOR_MISSCLK1
REFSWITCH1
DPLL1
Selected
Input
Ref Input Monitors (x4)
EN
EN
EN
EN
EN
DIFF: Min. Swing
LVCMOS: Slew rate
Amplitude
Frequency
EN
Valid / Invalid ppm
Late detect window
Early detect window
Jitter threshold
4
LOR
Validation Timer
Starts when LOR‰0
REF[0:3]
Valid
Missing pulse
Runt pulse
LOR_AMP2
DPLL2
Selected
Input
LOR_FREQ2
LOR_MISSCLK2
REFSWITCH2
Valid time
Phase*
5
Detector Status (1 = fault)
REF[0:3]
Status
*Enable for 1-PPS in
LMK05028
图 36. Clock Monitors for References, XO, and TCXO Inputs
9.3.8.1 XO Input Monitoring
The XO input has amplitude and frequency monitors to help qualify the input before it can be used to lock the
APLLs.
The XO amplitude detector clears its LOS (loss-of-signal) flag when the differential input voltage swing (peak-to-
peak) is greater than the minimum threshold selected by the registers (400, 600, or 800 mVpp nominal). The
same threshold applies also for a single-ended LVCMOS input with the non-driven input pin pulled to ground. If
the input clock does not meet the amplitude threshold, the amplitude detector will set the LOS flag and disqualify
the input.
版权 © 2018, Texas Instruments Incorporated
35
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
The XO frequency detector clears its LOS_FDET flag when the input frequency is detected within the range of
about 10 MHz to 90 MHz. Above 90 MHz, the frequency detector should be bypassed for proper operation. The
XO frequency monitor uses a RC-based detector and cannot precisely detect if the XO input clock has sufficient
frequency stability to ensure successful VCO calibration during the PLL start-up when the external XO clock has
a slow or delayed start-up behavior. See Slow or Delayed XO Start-Up for more information.
The XO monitors can be bypassed through registers, so the input will always be considered valid by the PLL
control state machines. The XO's LOS status flags can be observed through the status pins and the status bits.
The XO LOS signal from the status pin is the logic-OR combination of both its amplitude and frequency monitor
flags.
9.3.8.2 TCXO Input Monitoring
The TCXO input has amplitude and frequency monitors to help qualify the input before it can be used to lock the
TCXO-DPLLs.
The TCXO amplitude detector determines if the input meets the minimum input slew rate threshold. The input
slew rate detector clears its LOS flag when the slew rate is faster than 0.2 V/ns on the clock edge selected by
the registers (rising edge, falling edge, or both edges). If the input clock does not meet the slew rate threshold on
the selected clock edge(s), the amplitude monitor will set the LOS flag and disqualify the input.
The TCXO frequency detector clears its LOS_FDET flag when the input frequency is detected within the range of
about 10 MHz to 90 MHz. Above 90 MHz, the frequency detector should be bypassed for proper operation.
The TCXO monitors can be bypassed through registers, so the input will always considered valid by the PLL
control state machines. The TCXO's LOS status flags can be observed through the status pins and the status
bits. The TCXO LOS signal from the status pin is the logic-OR combination of both its amplitude and frequency
monitor flags.
9.3.8.3 Reference Input Monitoring
Each DPLL reference clock input is independently monitored for input validation (qualification) before it is
available for input selection by either DPLL. The reference monitoring blocks include amplitude, frequency,
missing pulse, and runt pulse monitors. For a 1-PPS input, the phase valid monitor is supported and the
frequency, missing pulse, and runt pulse monitors are not supported. A validation timer sets the minimum time
for all enabled reference monitors to be clear of flags before an input is qualified.
The enablement and valid threshold for all reference monitors and validation timers are programmable per input.
The reference monitors and validation timers are optional to enable, but critical to achieve optimal transient
performance during holdover or switchover events and also to avoid selection of an unreliable or intermittent
clock input. If a given detector is not enabled, it will not set a flag and will be ignored. The status flag of any
enabled detector can be observed through the status pins for any reference input (selected or not selected). The
status flags of the enabled detectors can also be read through the status bits for the selected input of each
DPLL.
9.3.8.3.1 Reference Validation Timer
The validation timer sets the amount of time for each reference to be clear of flags from all enabled input
monitors before it is qualified and valid for selection. The validation timer and enable settings are programmable.
9.3.8.3.2 Amplitude Monitor
The reference amplitude detector determines if the input meets the amplitude-related threshold depending on the
input buffer configuration. For differential input mode, the amplitude detector clears its LOR_AMP flag when the
differential input voltage swing (peak-to-peak) is greater than the minimum threshold selected by the registers
(400, 500, or 600 mVpp nominal). For LVCMOS input mode, the input slew rate detector clears its LOR_AMP
flag when its slew rate is faster than 0.2 V/ns on the clock edge selected by the registers (rising edge, falling
edge, or both edges). If either the differential or LVCMOS input clock does not meet the specified thresholds, the
amplitude detector will set the LOR_AMP flag and disqualify the input.
Below about 5 MHz, the differential input detector may signal a false flag; in this case, the amplitude detector
should be disabled and at least one other input monitor (frequency, window detector) should be enabled to
validate the input clock. The LVCMOS input detector can be used for low-frequency clocks down to 1 Hz or 1
PPS.
36
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.8.3.3 Missing Pulse Monitor (Late Detect)
The missing pulse monitor uses a window detector to validate input clock pulses that arrive within the nominal
clock period plus a programmable late window threshold (TLATE). When an input pulse arrives before TLATE, the
pulse is considered valid and the missing pulse flag will be cleared. When an input pulse does not arrive before
TLATE (due to a missing or late pulse), the flag will be set immediately to disqualify the input.
Typically, TLATE should be set higher than the input's longest clock period (including cycle-to-cycle jitter), or
higher than the gap width for a gapped clock. The missing pulse monitor can act as a coarse frequency detector
with faster detection than the ppm frequency detector. The missing pulse monitor is supported for input
frequencies between 2 kHz and fVCO/48 and should be disabled when outside this range.
The missing pulse and runt pulse monitors operate from the same window detector block for each reference
input. The status flags for both these monitors are combined by logic-OR gate and can be observed through
status pin. The window detector flag for the selected DPLL input can also be observed through the corresponding
MISSCLK status bit.
9.3.8.3.4 Runt Pulse Monitor (Early Detect)
The runt pulse monitor uses a window detector to validate input clock pulses that arrive within the nominal clock
period minus a programmable early window threshold (TEARLY). When an input pulse arrives after TEARLY, the
pulse is considered valid and the runt pulse flag will be cleared. When an early or runt input pulse arrives before
TEARLY, the monitor will set the flag immediately to disqualify the input.
Typically, TEARLY should be set lower than the input's shortest clock period (including cycle-to-cycle jitter). The
early pulse monitor can act as a coarse frequency detector with faster detection than the ppm frequency
detector. The early pulse monitor is supported for input frequencies between 2 kHz and fVCO/48 and should be
disabled when outside this range.
Ideal Reference Period
Ideal Edge
Ideal Reference Input
(rising-edge triggered)
Early Pulse (Input disqualified
at this input rising edge)
Example A: Input with
Early (Runt) Pulse
Late Pulse (Input disqualified at falling
edge of previous valid window)
Example B: Input with
Missing (Late) Pulse
Gapped Clock (To avoid disqualifying input at the
missing clock cycle, set TLATE window > Gap width)
Example C: Input with
Missing (Gapped) Clock
Gap width
Valid
Invalid
Valid Windows
Valid Window size can be relaxed by reducing the Early Window
size and/or increasing the Late Window size.
Early Window
(TEARLY
)
Late Window
Window Step Size = 8 / fVCO
(TLATE
Minimum Valid Window
is ±3*(8 / fVCO
)
)
图 37. Early and Late Window Detector Examples
版权 © 2018, Texas Instruments Incorporated
37
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.3.8.3.5 Frequency Monitoring
The precision frequency detector measures the frequency offset (in ppm) for all input clocks relative to a 0-ppm
reference clock, which can be selected from either the XO or TCXO input. The valid and invalid ppm frequency
thresholds are configurable through the registers. The monitor will clear its LOR_FREQ flag when the relative
input frequency error is less than the valid ppm threshold. Otherwise, it will set the LOR_FREQ flag when the
relative input frequency error is greater than the invalid ppm threshold. The ppm delta between the valid and
invalid thresholds provides hysteresis to prevent the LOR_FREQ flag from toggling when the input frequency
offset is crossing these thresholds.
A frequency measurement averaging factor is also used in computing the frequency detector register settings. A
higher averaging factor increases the measurement delay to set or clear the flag, which allow more time for the
input frequency to settle, and can also provide better measurement resolution for an input with high drift or
wander. Note that higher averaging reduces the maximum frequency ppm thresholds that can be configured.
9.3.8.3.6 Phase Valid Monitor for 1-PPS Inputs
The phase valid monitor is designed specifically for 1-PPS input validation because the frequency and window
detectors do not support this mode. The phase valid monitor uses a window detector to validate 1-PPS input
pulses that arrive within the nominal clock period (TIN) plus a programmable jitter threshold (TJIT). When the input
pulse arrives within the counter window (TV), the pulse is considered valid and the phase valid flag will be
cleared. When the input pulse does not arrive before TV (due to a missing or late pulse), the flag will be set
immediately to disqualify the input. TJIT should be set higher than the worst-case input cycle-to-cycle jitter.
Ideal Edge
(TIN < TV
Counter resets at
valid edge (TIN‘ < TV)
Counter time-out (TIN‘‘ > TV).
Input is disqualified here
)
Late Pulse
(Large peak jitter)
Ideal Input Period
TIN
TIN‘
TIN‘‘
TIN‘ > TIN
TIN‘‘ >> TIN
Example:
1-PPS Input
TJIT
Valid Counter (TV)
TV = TIN + TJIT
TV
TV
图 38. 1-PPS Input Window Detector Example
38
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.8.4 PLL Lock Detectors
The loss-of-lock (LOL) status is available for each APLL and DPLL. The APLLs are monitored for loss-of-
frequency lock only. The REF-DPLLs are monitored for both loss-of-frequency lock (LOFL) and loss-of-phase
lock (LOPL). The DPLL lock threshold and loss-of-lock threshold are programmable for both LOPF and LOFL
detectors.
Each DPLL frequency lock detector will clear its LOFL flag when the DPLL's frequency error relative the selected
reference input is less than the lock ppm threshold. Otherwise, it will set the LOFL flag when the DPLL's
frequency error is greater than the unlock ppm threshold. The ppm delta between the lock and unlock thresholds
provides hysteresis to prevent the LOFL flag from toggling when the DPLL frequency error is crossing these
thresholds.
A measurement averaging factor is also used in computing the frequency lock detector register settings. A higher
averaging factor increases the measurement delay to set or clear the LOFL flag. Higher averaging may be useful
when locking to an input with high wander or when the DPLL is configured with a narrow loop bandwidth. Note
that higher averaging reduces the maximum frequency ppm thresholds that can be configured.
Each DPLL phase lock detector will clear its LOPL flag when the DPLL's phase error is less than the phase lock
threshold. Otherwise, it will set the LOPL flag when greater than the phase unlock threshold.
The APLL and DPLL lock detector flags can be observed through the status pins and the status bits.
PLL Channel (x2)
2
2
Status Bits
DPLL Frequency Lock
Detector
APLL Lock
Detector
LOL
Lock
Unlock
LOL_PLL1
LOFL_DPLL1
LOPL_DPLL1
HIST1
LOFL
Thresh
(ppm)
Thresh
(ppm)
PLL1
HLDOVR1
APLL
REF
DPLL
TCXO
DPLL
fREF-TDC
PLL Status
fVCO/P1
LOL_PLL2
LOFL_DPLL2
LOPL_DPLL2
HIST2
Free-run
Tuning Word
DPLL Phase Lock
Detector
2
LOPL
PLL2
Lock
Unlock
Tuning Word History
Count Delay
2
2
HLDOVR2
HIST Update
Holdover
Active
Thresh
(ns)
Thresh
(ns)
EN Average Ignore Hold
time time
图 39. DPLL and APLL Lock Detectors
版权 © 2018, Texas Instruments Incorporated
39
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.3.8.5 Tuning Word History
Each REF-DPLL domain has a tuning word history monitor block that determines the initial output frequency
accuracy upon entry into holdover. The tuning word can be updated from one of three sources depending on the
DPLL operating mode:
a. Locked Mode: From the output of the digital loop filter when locked
b. Holdover Mode: From the final output of the history monitor
c. Free Run Mode: From the free-run tuning word register (user defined)
When the history monitor is enabled and the DPLL is locked, it effectively averages the reference input frequency
by accumulating history from the digital loop filter output during a programmable averaging time (TAVG). Once the
input becomes invalid, the final tuning word value is stored to determine the initial holdover frequency accuracy.
Generally, a longer TAVG time will produce a more accurate initial holdover frequency. The stability of the 0-ppm
reference clock (XO or TCXO input) determines the long-term stability and accuracy of the holdover output
frequency.
There is also a separate programmable delay timer (TIGN) that can be set to ignore the history data that is
corrupted just prior to entry into holdover. The history data could be corrupted if a tuning word update occurs
while the input clock is failing and before it is detected by the input monitors. Both TAVG and TIGN times are
programmable through the HISTCNT and HISTDLY register bits, respectively, and are related to the REF-TDC
rate.
The tuning word history is initial cleared after a device hard reset or soft reset. The history monitor begins to
accumulate history once the DPLL locks to a new reference. The previous history will be cleared when a
switchover to a new reference occurs assuming the history persistence bit (HIST_HOLD) is not set. The history
can be manually cleared by asserting the history soft reset bit (HIST_SW_RST). If the history persistence bit is
set, the history monitor will not clear the previous history value during reference switchover, holdover exit, or
history soft reset. Whenever the tuning word is cleared, the history monitor waits for the first TAVG timer to expire
before storing the first tuning word value.
Initial start of history
Ref Lost
Ref Valid
Ref Valid
when LOFL‰ 0 only
LOR‰ 1
LOR‰ 0
LOR‰ 0
History
Reset
No History
History Data Accumulating
History Valid
History Data Accumulating
TAVG(0)
TAVG(1)
TIGN
TAVG(2..n)
History Delay*
Delay to ignore
history updates
prior to LOR.
Initial holdover
frequency determined
by averaged history.
History held*
or cleared on
holdover exit.
*Programmable
History Count*
Timer to average history data to
compute initial holdover frequency accuracy.
settings
Time
Free Run
Lock Acq.
Locked
Holdover
Lock Acq.
Locked
LOFL = 1, LOPL = 1
LOFL‰ 0, then LOPL‰ 0
LOFL = 0, LOPL‰ 1
LOFL = 0, LOPL = 1 LOFL = 0, LOPL‰ 0
图 40. Tuning Word History Windows
If the TAVG period is set very long (minutes or hours) to obtain a more precise historical average frequency, it is
possible for a switchover or holdover event to occur before the first tuning word is stored and available for use.
To overcome this, there is an intermediate history update option (HIST_INTMD). If the history is reset, then the
intermediate average can be updated at intervals of TAVG/2K , where K = HIST_INTMD to 0, during the first TAVG
period only. If HIST_INTMD = 0, there is no intermediate update and the first average is stored after the first
TAVG period. However, if HIST_INTMD = 4, then four intermediate averages are taken at TAVG/16, TAVG/8, TAVG/4,
and TAVG/2, as well as at TAVG. After the first TAVG period, all subsequent history updates occur at the TAVG
period.
When no tuning word history exists, the free-run tuning word value (TUNING_FREE_RUN) determines the initial
holdover output frequency accuracy.
40
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.8.6 Status Outputs
STATUS[1:0] and GPIO[6:5] pins can be configured to output various status signals and interrupt flag for device
diagnostic and debug purposes. The status signal, output driver type, and output polarity settings are
programmable. The status output signals available at these pins for each device block monitored are listed in 公
式 10. When the status signal is asserted, the status output will be active high (assuming the status polarity is not
set to active low).
表 8. Status Pin Signals Available per Device Block
DEVICE BLOCK MONITORED
STATUS SIGNAL (ACTIVE HIGH)
XO Input Loss of Signal (LOS)
TCXO Input Loss of Signal (LOS)
APLLx Lock Detected (LOL)
PLLx VCO Calibration Active
APLLx N Divider, div-by-2
XO
TCXO
APLL1, APLL2
EEPROM
EEPROM Active
All Inputs and PLLs
Interrupt (INTR)
REFx Monitor Divider Output, div-by-2
REFx Amplitude Monitor Fault
REFx Frequency Monitor Fault
REFx Missing or Early Pulse Monitor Fault
REFx Validation Timer Active
REFx Phase Validation Monitor Fault
DPLLx R Divider, div-by-2
REF0 to REF3 (IN0 to IN3)
DPLLx REF N Divider, div-by-2
DPLL TCXO M Divider, div-by-2
DPLLx TCXO N Divider, div-by-2
DPLLx REFn Selected
DPLL1, DPLL2
DPLLx Holdover Active
DPLLx Reference Switchover Event
DPLLx Tuning History Update
DPLLx Loss of Lock (LOFL)
9.3.8.7 Interrupt
Any of the four status pins can be configured as a device interrupt output pin. The interrupt configuration is set
through registers. When the interrupt is enabled, the interrupt flag can be triggered from any combination of
interrupt status indicators, including LOS for the XO, TCXO, and DPLL-selected inputs, LOL for each DPLL and
APLL, and holdover and switchover events for each DPLL. Any status indicator can be masked so it will not
trigger the interrupt pin. Any unmasked status indicator can have its polarity inverted before it is combined at the
interrupt AND/OR gate and output to the status pin.
When the interrupt output is enabled and an interrupt flag is asserted by one or more fault conditions, the host
device can read the sticky status registers to identify which flags were set, resolve any fault conditions in the
system, and clear the flag by writing 0 to clear the sticky bits that were set.
版权 © 2018, Texas Instruments Incorporated
41
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
INTR
Polarity
INTR
Mask
INTR
Flag*
INT_AND_OR
Status Bits
F
F
F
F
LOS_FDET_TCXO
LOS_FDET_XO
LOS_TCXO
LOS_XO
2
2
2
2
2
Status Pins (x4)
LOL_PLL[1:2]
F
F
F
F
F
INTR
Enable
LOFL_DPLL[1:2]
INTR
AND/OR
Gate
Polarity
Type
0xA
LOPL_DPLL[1:2]
HIST[1:2]
Status
Select
STATUS0
HLDOVR[1:2]
Other
status
signals
STATUS1
GPIO5
2
F
F
F
F
LOR_AMP[1:2]
LOR_FREQ[1:2]
2
2
2
GPIO6
LOR_MISSCLK[1:2]
REFSWITCH[1:2]
Live Status Registers
0x00D to 0x00F
Sticky Status Registers
0x016 to 0x018
*Write 0 to clear INTR flag bits
LMK05028
图 41. Status and Interrupt
9.3.9 PLL Channels
图 42 shows the 3-loop architecture implemented the same for both PLL channels with exception of the VCO
frequency range. Each PLL channel can be configured independently in the different PLL modes described in
PLL Architecture Overview.
XO
VCOx FB
×2
÷M
TCXO
VCO1 FB
VCO2 FB
5-bit
PLLx Channel
Loopback
Dividers
VCO1: 4.8 to 5.4 GHz
VCO2: 5.5 to 6.2 GHz
REF Mux
TCXO Mux
0
1 2
0
1
2
3
4
5
IN0
IN1
IN2
IN3
Post
Dividers
REF-DPLL
TCXO-DPLL
APLL
fREF-TDC
fTCXO-TDC
fPD
VCO
fVCO
fVCO/P1
÷R
×2
÷P1
TDC
DLF
TDC
DLF
PFD
LF
16-bit
(x6)
/4 to /9,
/11, /13
/2 to /17
÷PR
/2 to /17
÷PR
fVCO/P2
÷FB
40-bit Frac-N SDM
÷FB
40-bit Frac-N SDM
÷N
40-bit Frac-N SDM
VCO1 FB
VCO2 FB
÷P2
To
Output
Muxes
SDM control
DPLL feedback clock
图 42. PLL Architecture (One Channel)
42
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.9.1 PLL Frequency Relationships
The following equations provide the PLL frequency relationships required to achieve closed-loop operation
according to the selected PLL mode. The TICS Pro programming software can be used to generate valid divider
settings based on the desired frequency plan configuration and PLL mode. The equations are applicable to both
PLL channels.
•
•
•
•
For 1-loop APLL mode, the condition in 公式 1 must be met.
For 3-loop mode, the conditions in 公式 1, 公式 2, 公式 3, and 公式 4 must be met.
For 2-loop mode (REF-DPLL), the conditions in 公式 1, 公式 3, and 公式 4 must be met.
For 2-loop mode (TCXO-DPLL), the conditions in 公式 1 and 公式 2 must be met.
公式 1 relates to the APLL:
fVCO = fXO × DXO × (INTAPLL + NUMAPLL/ DENAPLL
)
where
•
•
•
•
•
•
fVCO: VCO frequency
fXO: XO input frequency
DXO: APLL XO doubler (1 = disabled, 2 = enabled)
INTAPLL: APLL N divider integer value (9 bits, 1 to 29-1)
NUMAPLL: APLL N divider numerator value (40 bits, 0 to 240-1)
DENAPLL: APLL N divider denominator value (fixed, 240)
(1)
公式 2 relates to the TCXO-DPLL:
fVCO = (fTCXO × DTCXO / MTCXO) × P1PLL × PRTCXO × (INTTCXO + NUMTCXO/ DENTCXO
)
where
•
•
•
•
•
•
•
•
fTCXO: TCXO/OCXO input frequency
DTCXO: TCXO input doubler (1 = disabled, 2 = enabled)
MTCXO: TCXO input divide value (5 bits, 1 to 32)
P1PLL: PLL primary post-divider value (4 to 9, 11, 13)
PRTCXO: TCXO-DPLL FB prescaler divide value (2 to 17)
INTTCXO: TCXO-DPLL FB divider integer value (30 bits, 1 to 230-1)
NUMTCXO: TCXO-DPLL FB divider numerator value (40 bits, 0 to 240-1)
DENTCXO: TCXO-DPLL FB divider denominator value (fixed, 240)
(2)
公式 3 relates to the REF-DPLL:
fVCO = (fINx / RINx) × P1PLL × PRREF × (INTREF + NUMREF/ DENREF
)
where
•
•
•
•
•
•
fINx: Reference input frequency (x = 0 to 3) or VCO loopback frequency (x = 4 or 5)
RINx: Reference input divide value (16 bits, 1 to 216-1) (x = 0 to 5)
PRREF: REF-DPLL FB prescaler divide value (2 to 17)
INTREF: REF-DPLL FB divider integer value (30 bits, 1 to 230-1)
NUMREF: REF-DPLL FB divider numerator value (40 bits, 0 to 240-1)
DENREF: REF-DPLL FB divider denominator value (40 bits, 1 to 240)
(3)
公式 4 relates to any reference inputs assigned to a DPLL reference mux to achieve a constant REF-TDC rate
required for proper input switchover.
fREF-TDC = fIN0/RIN0 =fIN1/RIN1 = fIN2/RIN2 = fIN3/RIN3
(4)
版权 © 2018, Texas Instruments Incorporated
43
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
公式 5, 公式 6, 公式 7, 公式 8, and 公式 9 relate to the output frequency according to the output channel mux
selection (CHxMUX).
fCHxMUX = fVCOy / PnPLLy when PLLy post-divider is selected
fCHxMUX = fXO when XO is selected (OUT0 or OUT1)
fCHxMUX = fTCXO/REF when TCXO or REF is selected (OUT0 or OUT1)
fOUTx = fCHxMUX / ODOUTx (OUT1 to OUT6)
(5)
(6)
(7)
(8)
fOUTx = fCHxMUX / (DIVAOUTx × DIVBOUTx) (OUT0 or OUT7 only)
where
•
•
•
•
•
•
fCHxMUX: Output channel mux frequency (from PLL post-divider, XO, or TCXO/Ref Bypass mux)
fTCXO/REF: TCXO, DPLL1 Ref, or DPLL2 Ref input frequency (selected by TCXO/Ref Bypass mux)
fOUTx: Output clock frequency (x = 0 to 7)
PnPLLy: PLLy P1 (primary) or P2 (secondary) post-divider value (4 to 9, 11, 13)
ODOUTx: Output divide value (20 bits, 1 to 220-1)
ODBOUTx: Output MSB divide value for OUT0 or OUT7 (11 bits, 1 to 211-1)
(9)
9.3.9.2 Analog PLL (APLL)
The APLL has a 40-bit fractional-N divider to support high-resolution frequency synthesis, wide output frequency
range, very low phase noise and jitter, and the ability to tune its VCO frequency through sigma-delta modulator
(SDM) control.
The APLL XO doubler doubles the XO input frequency into the phase frequency detector (PFD) input. The APLL
multiplies the PFD frequency by the total N divider value to generate the VCO clock. The desired VCO output to
PFD input frequency ratio is the total value of N (INT + NUM/DEN) applied to the SDM to tune the VCO
frequency.
In free-run mode, the APLL uses a low-jitter XO input as a initial reference clock to lock the internal voltage
controlled oscillator (VCO). The PFD compares the fractional-N divided clock with the XO doubler frequency and
generates a control signal. The control signal is filtered by the APLL loop filter to generate the VCO’s control
voltage that sets its output frequency. The SDM modulates the N divider ratio to get the desired fractional ratio
between the PFD input and the VCO output.
In 2-loop or 3-loop mode, the APLL's SDM is controlled by one of the DPLL loops to pull the VCO frequency into
lock with the DPLL reference input.
9.3.9.3 APLL XO Doubler
The APLL has a XO doubler that can be enabled to double the PFD frequency up to 200 MHz. The doubler adds
minimal noise and is useful for raising the PFD frequency for better phase noise and jitter and also to avoid
spurs. When the PFD frequency is increased, the flat portion of the APLL phase noise improves.
9.3.9.4 APLL Phase Frequency Detector (PFD) and Charge Pump
The APLL PFD frequency can operate from 10 MHz to 200 MHz, but the APLL performance is optimized for
frequencies of 96 MHz or higher. The PLL has programmable charge pump settings of 1.6, 3.2, 4.8, or 6.4 mA.
9.3.9.5 APLL Loop Filter
The APLL supports programmable loop bandwidth from 100 kHz and 1 MHz. The loop filter components can be
programmed to optimize the APLL bandwidth depending on the XO frequency and phase noise without changing
any external components. The LF1 and LF2 pins each require an external C2 capacitor to ground, typically 0.1-
µF. 图 43 shows the APLL loop filter structure between the PFD/charge pump output and VCO control input.
44
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
VCO
Programmable
Loop Filter
R3
R4
PFD /
Charge Pump
C1
R2
C3
C4
LMK05028
LF1, LF2
C2
图 43. Loop Filter Structure of Each APLL
9.3.9.6 APLL Voltage Controlled Oscillator (VCO)
Each PLL contains fully-integrated LC-based oscillators with very low phase noise. The VCO takes the voltage
from the loop filter and converts this into a frequency. The tuning range of VCO1 is 4.8 to 5.4 GHz, and the
tuning range of VCO2 is 5.5 to 6.2 GHz. The two VCO frequency ranges are spaced apart to cover a wide range
of frequency configurations and to help minimize cross-coupling between the two PLL domains.
9.3.9.6.1 VCO Calibration
Each PLL's VCO must be calibrated to ensure that the PLL can achieve lock and deliver optimal phase noise
performance. Fundamentally, the VCO calibration establishes an optimal operating point within the tuning range
of the VCO. The VCO calibration is executed automatically during initial PLL start-up after device power-on, hard
reset, or soft reset once the XO input is detected by its input monitor. To ensure proper VCO calibration, it is
critical for the XO clock to be stable in amplitude and frequency prior to the start of VCO calibration; otherwise,
the VCO calibration can fail and prevent start-up of the PLL and its output clocks. Prior to VCO calibration and
APLL lock, the output drivers are typically held in the mute state (configurable per output) to prevent spurious
output clocks.
To trigger VCO calibration for one PLL channel without affecting the other channel, this can be achieved through
host programming by either entering/exiting PLL power-down (PLLx_PDN register bit) or by asserting a PLL soft-
reset (SWRxPLL register bit).
9.3.9.7 APLL VCO Post-Dividers (P1, P2)
The APLL has a primary (P1) and secondary (P2) VCO post-divider for flexible clock frequency planning. All
post-divider supports divide by 4 to 9, 11, or 13. The post-divider clocks for both PLLs are distributed to all output
channel muxes for selection. The primary (P1) post-divider output is also fed back to the FB divider paths of the
REF-DPLL and TCXO-DPLL to close the loops.
After the P1 divider and DPLL fractional FB divider values have been determined for closed-loop operation, the
P1 divider value should not be modified dynamically because it would affect the FB divider clock frequency to the
TDC of the DPLL. If the P1 divider must be changed, it is necessary to re-compute the DPLL FB divider values.
Also, changing any PLL post-divider value requires a PLL soft-reset (or device soft-reset) to reset the divider for
proper operation.
9.3.9.8 APLL Fractional N Divider (N) With SDM
The APLL fractional N divider includes a 9-b integer portion (INT), 40-b numerator portion (NUM), a fixed 40-b
denominator portion (DEN), and sigma-delta modulator. The INT and NUM are programmable, while the
denominator is fixed to 240 for highest frequency resolution (step size) on the output. The total N divider value is:
N = INT + NUM / DEN.
版权 © 2018, Texas Instruments Incorporated
45
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.3.9.9 REF-DPLL Reference Divider (R)
The reference clock input paths to each REF-DPLL features a 16-b reference divider (R) for each clock input
(IN0 to IN3). The output of each R divider sets the frequencies to the reference input mux and the TDC rate of
the REF-DPLL. There are also two additional R dividers for the internal VCO loopback clocks (IN4 and IN5) that
could be used in cascaded PLL configurations. IN4 refers to the VCO1 loopback clock to DPLL2 reference input,
and IN5 refers to the VCO2 loopback clock to DPLL1 reference input. To support hitless switching between
inputs with different frequencies, the R divider can be used to divide the clocks to a single common frequency to
the REF-DPLL TDC input.
9.3.9.10 TCXO/OCXO Input Doubler and M Divider
The TCXO/OCXO input features a frequency doubler followed by a 5-b M divider. The M divider output is sent to
the TCXO mux of both TCXO-DPLLs.
9.3.9.11 TCXO Mux
Each PLL channel has a TCXO mux to select the TCXO-DPLL input from either the TCXO M divider clock, or the
VCO loopback clock from the opposite PLL channel when PLL cascading is used. When the TCXO M divider is
selected, the M Divider frequency sets the TCXO-TDC rate. When the VCO loopback clock is selected, the VCO
loopback divider frequency from the opposite PLL sets the TCXO-TDC rate.
9.3.9.12 REF-DPLL and TCXO-DPLL Time-to-Digital Converter (TDC)
The TDCs for the REF-DPLL and TCXO-DPLL operate up 30 MHz. The TDC resolution is fine enough to achieve
in-band phase noise of –112 dBc/Hz at 100-Hz offset for a 122.88-MHz output.
When the REF mux selects a reference input clock, the REF-DPLL TDC rate is:
fREF-TDC = fINx / RINx
(10)
(11)
(12)
When the REF mux selects the VCO loopback clock, the REF-DPLL TDC rate is:
fREF-TDC = fVCOa / (48 × DIVDPLLa_CLK_FB) / RINx
When the TCXO mux selects the M divider clock, the TCXO-DPLL TDC rate is:
fTCXO-TDC = fTCXO × DTCXO / MTCXO
When the TCXO mux selects the VCO loopback clock, the TCXO-DPLL TDC rate is:
fTCXO-TDC = fVCOa / (48 × DIVDPLLa_CLK_FB
)
where
•
•
fVCOa: VCO frequency fed back from the first PLL channel (PLLa) in a cascaded configuration
DIVDPLLy_CLK_FB: VCO loopback divide value (3 to 32) from PLLa
(13)
9.3.9.13 REF-DPLL and TCXO-DPLL Loop Filter
Each REF-DPLL and TCXO-DPLL supports programmable loop bandwidth from 10 mHz to 4 kHz and can
achieve jitter peaking below 0.1 dB (typical). The low-pass jitter transfer characteristic of each DPLL attenuates
its reference input noise with up to 60-dB/decade roll-off above the loop bandwidth.
In 3-loop mode, the REF-DPLL loop filter output modulates the TCXO-DPLL's SDM, and the TCXO-DPLL loop
filter output correspondingly modulates the APLL's SDM to steer the APLL VCO into lock with the selected REF-
DPLL input.
In 2-loop REF-DPLL mode, the TCXO-DPLL is not used and the REF-DPLL loop filter output controls the APLL's
SDM to steer the VCO frequency into lock with the selected REF-DPLL input.
In 2-loop TCXO-DPLL mode, the REF-DPLL is not used and the TCXO-DPLL loop filter output controls the
APLL's SDM to steer the VCO frequency into lock with the TCXO input.
46
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.9.14 REF-DPLL and TCXO-DPLL Feedback Dividers
The feedback path of each REF-DPLL and TCXO-DPLL has a feedback prescaler (PR) followed by a fractional
FB divider. The prescaler divides the PLL primary post-divider (P1) clock by a programmable value from 2 to 17,
which then clocks the FB divider. The FB divider of each REF-DPLL and TCXO-DPLL includes a 30-b integer
portion (INT), 40-b numerator portion (NUM), and 40-b denominator portion (DEN). The total FB divider value is:
FB = INT + NUM / DEN. All DPLL feedback dividers are programmable, except for the DENTCXO (fixed, 240). The
FB divider clock must match the TDC rate determined by the TDC input path of the respective DPLL.
The REF-DPLL TDC rate is:
fREF-TDC = fVCO / (P1PLL × PRREF × FBREF
)
(14)
(15)
The TCXO-DPLL TDC rate is:
fTCXO-TDC = fVCO / (P1PLL × PRTCXO × FBTCXO
)
9.3.10 Output Clock Distribution
The output clock distribution blocks shown in 图 44 includes six output muxes, a TCXO/Ref Bypass mux, six
output dividers, and eight programmable output drivers. The output dividers support output synchronization
(SYNC) to allow phase synchronization between two or more output channels. Also, each output bank (OUT[0:3]
and OUT[4:7]) has separate a zero-delay feedback path to support the zero-delay mode option available on each
DPLL channel.
版权 © 2018, Texas Instruments Incorporated
47
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Clock Bus
0
1
2
÷ODB
11-b
÷OD
20-b
OUT7
OUT6
3
4
XO
0
1
2
÷OD
20-b
PLL1
OUT[4:7] Bank
preferred for
PLL1 clocks
0
1
3
÷P1
÷P2
/4 to /9,
/11, /13
VCO1
4.8 to 5.4
GHz
Output Channel Configuration
OUT5
OUT4
0
1
2
3
Power-
down
Output Type,
Slew Rate
÷OD
20-b
Mux
DPLL feedback
2
÷OD
Auto SYNC
at POR only
OUT3
OUT2
0
1
2
3
2
÷OD
20-b
PLL2
SYNC EN
(1)
Auto Mute,
Mute Level
2
3
÷P1
/4 to /9,
/11, /13
VCO2
SYNC
0
1
2
3
4
5
5.5 to 6.2
GHz
OUT[0:3] Bank
preferred for
PLL2 clocks
÷P2
÷OD
20-b
OUT1
OUT0
DPLL feedback
Ref Bypass Mux
0
1
2
3
4
5
TCXO
REF B
REF A
0
1
2
5
÷ODB
11-b
÷OD
20-b
SYNC_SW
SYNC
LMK05028
GPIO0/SYNCN
(active-low pin)
(1) SYNC Enable logic: SYNC EN = ((CH[z]_SYNCEN) && (PLL[x]_[y]_CH[zz]_SYNC_BANK)) where [x]: PLL = 1 or 2,
[y]: Post-Div = PRI or SEC (÷P1 or P2), [z]: CH = 0, 1, 23, 45, 6, or 7, and [zz]: Bank = 03 or 47.
图 44. Output Clock Distribution
9.3.11 Output Channel Muxes
Each of the six output channels has as output mux. Each output mux for OUT2 through OUT7 can individually
select between the PLL1 and PLL2 post-divider clocks. Each output mux for OUT0 and OUT1 can individually
select between the PLL1 and PLL2 post-divider clocks, the XO clock, or one of the clocks from the TCXO/Ref
Bypass Mux.
9.3.11.1 TCXO/Ref Bypass Mux
The TCXO/Ref bypass mux can select between the TCXO clock, the selected DPLL1 input (REF A), or the
selected DPLL2 input (REF B). The bypass clocks are primarily intended for diagnostic purposes and not
optimized for lowest phase noise or jitter.
9.3.12 Output Dividers
Each of the six output channels has an output divider after the output mux. OUT2 and OUT3 share an output
divider, as do OUT4 and OUT5. OUT0, OUT1, OUT6, and OUT7 have their own output dividers. The output
divider is used to generate the final output clock frequency from the source selected by the output mux.
48
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
OUT1 to OUT6 channels have 20-bit dividers that can support output frequencies from 2 kHz to 750 MHz (or up
to the maximum fOUT frequency for the configured output driver type). It is possible to configure the PLL post-
divider and output divider to achieve higher clock frequencies, but the driver's output swing may fall out of
specification depending on the output type).
OUT0 and OUT7 channels each have cascaded 11-bit (MSB) and 20-bit (LSB) output dividers to support output
frequencies from 1 Hz (1 PPS) to 750 MHz. In this case, the total output divide value is the product of the MSB
and LSB output divider values.
Each output divider is powered from the same VDDO_x supply used for the clock output drivers. The output
divider can be powered down if not used to reduce power. Each output divider is automatically powered down
when its output driver is powered down, or when both output drivers are powered down for OUT[2:3] or
OUT[4:5].
9.3.13 Clock Outputs (OUTx_P/N)
Each clock output can be individually configured as a differential driver (AC-LVDS/CML/LVPECL), HCSL driver,
or LVCMOS driver (1.8 V or 2.5 V). Otherwise, it can be powered down if not used. OUT2 and OUT3 share an
output supply, as do OUT4 and OUT5. OUT0, OUT1, OUT6, and OUT7 have their own output supplies. Each
output supply can be separately powered by 1.8 V, 2.5 V, or 3.3 V for a differential or HCSL output, or 1.8 V or
2.5 V for an LVCMOS output. Each output channel has its own internal LDO regulator to provide excellent power
supply noise rejection (PSNR) and minimize supply-noise induced jitter and spurs. The output clock
specifications (for example, output swing, phase noise, jitter, and so forth) for differential and HCSL drivers are
not sensitive to the VDDO_x voltage because these driver modes are powered through the channel's internal
LDO regulator. When an output channel is left unpowered, the channel does not generate any clocks and will not
interfere with other output channels that are powered-on.
表 9. Output Driver Modes
OUT_x_TYPE
00h
OUTPUT TYPE
Disabled
10h
AC-LVDS
14h
AC-CML
18h
AC-LVPECL
2Ch
2Dh
30h
HCSL (External 50-Ω to GND)
HCSL (Internal 50-Ω to GND)
LVCMOS (HiZ / HiZ)
LVCMOS (HiZ / –)
LVCMOS (HiZ / +)
LVCMOS (Low / Low)
LVCMOS (– / HiZ)
LVCMOS (– / –)
32h
33h
35h
38h
3Ah
3Bh
LVCMOS (– / +)
3Ch
3Eh
LVCMOS (+ / HiZ)
LVCMOS (+ / –)
3Fh
LVCMOS (+ / +)
版权 © 2018, Texas Instruments Incorporated
49
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.3.13.1 AC-Differential Output (AC-DIFF)
The differential output driver uses a switched-current mode type shown in 图 45. A programmable tail current of
4, 6, or 8 mA (nominal) is used to achieve VOD swing compatible with AC-coupled LVDS, CML, or LVPECL
receivers, respectively, across a 100-Ω differential termination. The differential output driver is ground-referenced
(similar to a HCSL driver), meaning the differential output has a low common-mode voltage (VOS).
The differential driver is internally biased and does not need any external pullup or pulldown resistors, unlike
conventional CML or LVPECL drivers. The differential output should be interfaced through external AC-coupling
to a differential receiver with proper input termination and biasing.
VDDO_x
LDO
I1 = 4 mA
From
output
channel
Output tail current (I1 + I2) can be programmed
to 4, 6, or 8 mA for LVDS-, CML-, and LVPECL-
compatible swing across AC-coupled
P
N
P
N
clk_p
50-W single-ended or 100-W differential load.
clk_n
OUTx_P
I2 = 0, 2, or 4 mA
P
N
P
N
OUTx_N
图 45. AC-LVDS/CML/LVPECL Output Driver Structure
9.3.13.2 HCSL Output
The HCSL output driver is an open-drain type, which should be DC-coupled to an HCSL receiver. The HCSL
output has programmable, internal 50-Ω to ground (on P and N) for source termination, which can be enabled
when the receiver does not provide input (load) termination. If the internal termination is disabled, external 50-Ω
termination to ground (on P and N) is required at either the driver side (source terminated) or receiver side (load
terminated).
9.3.13.3 LVCMOS Output (1.8 V, 2.5 V)
The LVCMOS driver has two outputs per pair. Each output on P and N can be configured for normal polarity,
inverted polarity, or disabled as HiZ or static low level. The LVCMOS output high level (VOH) is determined by the
VDDO_x voltage of 1.8 V or 2.5 V for rail-to-rail LVCMOS output voltage swing. If a VDDO_x voltage of 3.3 V is
applied, the VOH level not will not swing to the VDDO_x rail due to the dropout voltage of the channel's internal
LDO regulator.
Because an LVCMOS output clock is a high-swing and unbalanced signal, it can be a strong aggressor and
couple noise onto other jitter-sensitive differential output clocks. If an LVCMOS clock is required from an output
pair, configure the pair with both outputs enabled but with opposite polarity (+/– or –/+) and leave the unused
output floating with no trace connected.
50
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.13.4 Output Auto-Mute During LOL or LOS
Each output driver can automatically mute or squelch its clock when the selected output mux clock source is
invalid, as configured by its CH_x_MUTE bit. If the selected clock source is derived from a PLL post-divider
output, the source can be invalid based on the LOL status of each PLL by configuring the APLL and DPLL mute
control bits (MUTE_APLLx_LOCK, MUTE_DPLLx_LOCK, MUTE_DPLLx_TCXO). If the selected source is a
bypass clock (XO or TCXO), the source is invalid when a LOS is detected on the input. The mute level can be
configured per output channel by its CHx_MUTE_LVL bits, where the mute level depends on the configured
output driver type (Differential/HCSL or LVCMOS). The mute level for a differential or HCSL driver can be set to
output common mode, differential high, or differential low levels. The mute level for a LVCMOS driver pair can be
set to output low level for each of its outputs (P and N) independently. When auto-mute is disabled or bypassed
(CH_x_MUTE = CHx_MUTE_LVL = 0), the output clock can have incorrect frequency or be unstable before and
during the VCO calibration if derived from a PLL. For this reason, the mute bypass mode should only be used for
diagnostic or debug purposes.
9.3.14 Glitchless Output Clock Start-Up
When output auto-mute is enabled, any output derived from a PLL will start up in synchronous fashion without
clock glitches when PLL lock is achieved after any the following events: device power-on, exiting hard reset
(PDN pin), exiting soft reset (RESET_SW bit), or exiting PLL reset (PLLx_PDN bit). The output clock will also
start up without glitches after any of the following events: VDDO_x is ramped (even when delayed after the
device initialization), exiting channel soft reset (CHxPWDN bit), or deassertion of output SYNC (assuming
SYNC_MUTE bit is set).
9.3.15 Clock Output Interfacing and Termination
图 46 to 图 50 show the recommended output interfacing and termination circuits. Unused clock outputs can be
left floating and powered down by programming.
LVCMOS
2.5 V LVCMOS
LMK05028
Receiver
Copyright © 2018, Texas Instruments Incorporated
Same applies for 1.8-V LVCMOS output to 1.8-V LVCMOS receiver.
图 46. 2.5-V LVCMOS Output to 2.5-V LVCMOS Receiver
LVDS
Receiver
LMK05028
AC-LVDS
100 ꢀ
Copyright © 2018, Texas Instruments Incorporated
图 47. AC-LVDS Output to LVDS Receiver With Internal Termination/Biasing
版权 © 2018, Texas Instruments Incorporated
51
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
50 ꢀ
50 ꢀ
CML
Receiver
LMK05028
AC-CML
Copyright © 2018, Texas Instruments Incorporated
图 48. AC-CML Output to CML Receiver With Internal Termination/Biasing
AC-LVPECL
LVPECL Receiver
LMK05028
50 ꢀ
50 ꢀ
VDD_IN œ 1.3 V
图 49. AC-LVPECL Output to LVPECL Receiver With External Termination/Biasing
33 ꢀ (optional)
HCSL
Receiver
LMK05028
HCSL
33 ꢀ (optional)
50 ꢀ
50 ꢀ
If HCSL Internal Termination (50-Ω to GND) is enabled, replace 33-Ω with 0-Ω and remove 50-Ω external
resistors.
图 50. HCSL Output to HCSL Receiver With External Source Termination
9.3.16 Output Synchronization (SYNC)
Output SYNC can be used to align two or more output clocks to be phase-aligned at a common rising edge by
allowing the output dividers to exit reset on the same PLL post-divider clock cycle. Any output dividers selecting
the same PLL post-divider can be synchronized together as a SYNC group by triggering a SYNC event through
the hardware pin or software bit.
52
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
The following requirements must be met establish a SYNC group for two or more output channels:
•
•
Output dividers have their respective sync enabled (CHx_SYNCEN bit = 1).
Output dividers have their output mux selecting the same PLL primary post-divider (for example, PLL1 P1 or
PRI, PLL1 P2 or SEC).
•
The PLL post-divider (PRI and/or SEC) must have the applicable sync bank bit(s) enabled for the output
divider bank(s). Examples:
–
PLL1_PRI_CH47_SYNC_BANK should be set when output dividers in OUT[4:7] bank will be synced to
PLL1 P1 (PRI).
–
PLL1_SEC_CH03_SYNC_BANK should be set when output dividers in OUT[0:3] bank will be synced to
PLL1 P1 (SEC).
A SYNC event can be asserted by the hardware GPIO0/SYNCN pin (active low) or the SYNC_SW register bit
(active high). When SYNC is asserted, the SYNC-enabled dividers held are reset and clock outputs are muted.
The divider reset and output muting is done synchronously, allowing the outputs to finish their final clock cycle (to
avoid a short clock period) before the actual SYNC event. When SYNC is deasserted, the outputs will start with
their initial clock phases synchronized or aligned. SYNC can also be used to mute any SYNC-enabled outputs to
prevent output clocks from being distributed to down-stream devices until they are configured and ready to
accept the incoming clock. The SYNC signal is internally qualified or sampled by the internal digital system clock
that runs at 10 MHz nominal. The negative pulse applied to the SYNCN input pin should be greater than 200 ns
to be captured by the internal digital system clock. SYNC deassertion can take two cycles of the digital clock
before the outputs are released.
Output channels with their sync disabled (CHx_SYNCEN bit = 0) will not be affected by a SYNC event and will
continue normal output operation as configured. Also, VCO and PLL post-divider clocks do not stop running
during the SYNC so they can continue to source any output channels that do not require synchronization. Output
dividers with divide-by-1 (divider bypass mode) are not gated during the SYNC event. Also, SYNC should be
disabled and is not supported when the output mux is selecting the XO, TCXO, or DPLL reference clocks.
表 10. Output Synchronization
GPIO0 PIN
SYNC_SW BIT
OUTPUT DIVIDER AND DRIVER STATE
Output driver(s) muted and output divider(s) reset
0
0→1
1
1
1→0
0
Outputs in a SYNC group are unmuted with their initial clock phases aligned
Normal output driver/divider operation as configured
图 51 shows an example of the SYNC timing example for a SYNC group. The SYNC group is comprised of
OUT0, OUT6, and OUT7 dividers, which are sourced by the PLL1 P1 (primary) post-divider. Notice that the
output divider reset and output mute is applied synchronously by waiting until the last output clock in the group
goes low (OUT7).
VCO
PLL Post Divider
PLL1_P1 = 4
(Primary Post-div)
Output Channel Dividers
SYNCed outputs muted
Phase Aligned
OUT0_DIV = 4
OUT6_DIV = 5
Up to 2 Digital clock cycles
OUT7_DIV = 6
Internal SYNC
(Qualified by
Digital clock)
Internal SYNC starts and waits for last output in the
—SYNC group“ to go low before Step 2.
Dividers SYNCed and drivers muted. Each
driver‘s mute level is programmable.
1
2
3
SYNCed dividers released.
OUT[0,6,7] constitute a —SYNC group“ by selecting the same PLL1_P1 post-divider with its SYNC Banks enabled (PLL1_PRI_CHx_SYNC_BANK = 1) and by having their respective channel SYNC enabled (CHx_SYNC_EN = 1).
(1) The VCO clock and PLL post-divider clock do not stop running during the SYNC.
图 51. Output SYNC Group Timing Example
版权 © 2018, Texas Instruments Incorporated
53
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.3.17 Zero-Delay Mode (ZDM) Configuration
Zero-delay mode can be enabled to achieve zero phase delay between the selected reference input clock and
the output clocks of a DPLL. As shown in 图 52 and 图 53, DPLL1 supports zero-delay for OUT4/5, OUT6, and
OUT7, while DPLL2 supports zero-delay for OUT0, OUT1, and OUT2/3. Any output that requires the zero-delay
feature should be derived from the PLL's P1 (primary) post-divider and have zero-delay enabled
(DPLLx_ZDM_EN bit = 1). Then, one of the outputs per DPLL can be selected as the primary zero-delay output
(by O_CHx_y_ZERODLY_EN bit). Other outputs that need zero-delay can be synchronized with the primary
zero-delay output by comprising a SYNC group (see Output Synchronization (SYNC)). ZDM and DCO mode
should not be enabled at the same time within a PLL channel.
When the DPLL is acquiring lock to the reference input, the initial phase lock is governed by the DPLL fastlock
bandwidth. Once phase lock is detected, the final output phase alignment with the input reference is governed by
the normal DPLL loop bandwidth. The same phase lock and alignment process also occurs when exiting
holdover or after a switchover event.
PLL1
OUT[4:7] Channels (x3)
INx
÷R
To TDC for
phase offset
cancellation
fVCO1/P1
÷OD
OUT
0
ZDM
Feedback
DPLL1 ZDM
Channel
DPLL1_
ZDM_EN
OUT4/5
OUT6
3
2
1
0
OUT7
ZDM off
O_CH4_7_ZERODLY_EN
图 52. DPLL1 ZDM Configuration for OUT4 to OUT7
PLL2
OUT[0:3] Channels (x3)
INx
÷R
To TDC for
phase offset
cancellation
fVCO2/P1
÷OD
OUT
2
ZDM
Feedback
DPLL2 ZDM
Channel
DPLL2_
ZDM_EN
OUT2/3
OUT1
3
2
1
0
OUT0
ZDM off
O_CH0_3_ZERODLY_EN
图 53. DPLL2 ZDM Configuration for OUT0 to OUT3
54
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
9.3.18 PLL Cascading With Internal VCO Loopback
PLL cascading can be used when the second PLL (either PLL1 or PLL2) must be precisely locked to the
frequency of the first PLL. The internal VCO loopback configuration options are implemented identically on both
PLL channels, allowing PLL2 to be cascaded after PLL1 or vice versa. The internal VCO loopback clock from the
first PLL can drive the REF-DPLL input path or the TCXO-DPLL input path if the TCXO loopback enable control
is set. The second PLL can have reference validation enabled to qualify the VCO loopback clock from the first
PLL stage to ensure that the second PLL stage has a stable and valid clock input from the first PLL stage before
it acquires lock. The VCO loopback clock can be validated based on when the first PLL stage achieves
frequency lock and/or phase lock on the REF-DPLL or frequency lock on the TCXO-DPLL. Once the VCO
loopback clock is validated based on the enabled criteria, then the second PLL stage can begin to acquire lock.
The VCO loopback dividers, loopback mux, and loopback reference validation options are programmable.
In the example shown in 图 54, PLL2 is cascaded and locked to PLL1's internal VCO1 loopback clock through
the two loopback dividers (fixed and programmable) and TCXO loopback muxes. PLL2 operates with a wide loop
bandwidth to precisely track the DCO frequency adjustments applied to PLL1. This effectively applies DCO
adjustments to both clock domains simultaneously, which would not be possible if both loops were operating in
parallel (not cascaded) with separate DCO controls.
TCXO
XO
PLL1 (TCXO DCO)
0
fVCO1
APLL
REF
DPLL
TCXO
DPLL
/3 to /32
÷CLK
FB
÷48
FB Div EN
VCO1
FB clk
DCO Control
(from FPGA)
DCO
TCXO Loopback Enable
PLL2 (2-loop TCXO)
0
1
TCXO mux
fTCXO-TDC2
XO
1
REF
mux
APLL
fVCO2
REF
DPLL
TCXO
DPLL
fTCXO-TDC2 = fVCO1 / 48 / DPLL1_CLK_FB_DIV
图 54. PLL Cascading Example With DCO Frequency Steering
9.4 Device Functional Modes
9.4.1 Device Start-Up Modes
The LMK05028 can start up in one of three device modes depending on the 3-level input level sampled on the
HW_SW_CTRL pin during power-on reset (POR):
•
•
•
HW_SW_CTRL = 0: EEPROM + I2C Mode (Soft pin mode)
HW_SW_CTRL = VIM / Float: EEPROM + SPI Mode (Soft pin mode)
HW_SW_CTRL = 1: ROM + I2C Mode (Hard pin mode)
版权 © 2018, Texas Instruments Incorporated
55
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Device Functional Modes (接下页)
The device start-up mode determines:
•
The memory bank (EEPROM or ROM) used to initialize the register settings that sets the frequency
configuration.
•
•
The serial interface (I2C or SPI) used for register access.
The logic pin functionality for device control and status.
After start-up, the I2C or SPI interface is enabled for register access to monitor the device status and control (or
reconfigure) the device if needed. The register map configurations are the same for I2C and SPI.
Table 1 summarizes the device start-up mode and corresponding logic pin functionality.
图 55 shows the device power-on reset configuration sequence.
Power-On Reset
(POR)
Device POR
Configuration Sequence
0
PDN?
Outputs muted
(Hard reset)
1
HW_SW_CTRL = 0
GPIO[2:1] = 00b to 11b
(Select I2C Addr.)
HW_SW_CTRL = 1
GPIO[3:0] = 0000b to 1111b
(Select ROM Page)
Start-up Mode?
(sample pin states)
HW_SW_CTRL = Float
STATUS[1:0] = Float
EEPROM + I2C
EEPROM + SPI
ROM + I2C
(Soft Pin) Mode
(Soft Pin) Mode
(Hard Pin) Mode
Registers initialized from EEPROM/ROM (hard reset only).
I2C/SPI, Control, Status pins activated.
Device Block
Configuration
Soft reset
RESET_SW: 0‰1
All blocks reset to initial states.
Register and EEPROM programming available.
Normal Operation
See PLL Initialization
Flowchart
图 55. Device POR Configuration Sequence
9.4.1.1 EEPROM Mode
In EEPROM mode, the device's frequency configuration is loaded to the registers from the non-volatile
EEPROM. A single user-defined register page can be programmed to the EEPROM to generate a custom
frequency configuration on start-up. The EEPROM image can be pre-programmed at factory test or programmed
in-system through the serial interface. The EEPROM supports up to 100 programming cycles to facilitate clock
reconfiguration for system-level prototyping, debug, and optimization.
The EEPROM image can store a single register page or frequency configuration. A factory pre-programmed
device with custom EEPROM image would be assigned by TI with a unique orderable part number (OPN).
56
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Device Functional Modes (接下页)
TI suggests using EEPROM mode when any of the following is required:
•
•
A single custom start-up frequency configuration is needed from a single OPN.
A host device is available to program the registers and EEPROM (if desired) with a new configuration after
power-up through I2C or SPI.
•
SPI protocol is required for register access because SPI is not supported in ROM mode.
9.4.1.2 ROM Mode
In ROM mode, the device's frequency configuration is loaded to the registers from one of 16 register pages in
ROM selected by the GPIO[3:0] control pins. All register pages in the ROM image are factory-set in hardware
(mask ROM) and not software programmable. Only the I2C interface is available after start-up in ROM mode.
A benefit of ROM over EEPROM is a custom ROM image can support up to 16 different pin-selectable frequency
configurations from a single OPN. A factory preset device with custom ROM image would be assigned by TI with
a unique OPN.
9.4.2 PLL Operating Modes
Both PLL channels have identical functionality and modes of operation, but each are configured and operate
independently. The operating mode for both channels can be different at any time. The following sections
describe the PLL modes of operation shown in 图 56.
版权 © 2018, Texas Instruments Incorporated
57
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Device Functional Modes (接下页)
See Device POR
and PLL Initialization
Flowchart
No valid input
available
Free-run Mode(1)
Initial frequency accuracy
determined by free-run
tuning word register.
(1) Free-run/Holdover Mode frequency
stability determined by TCXO or XO.
DCO mode available on TCXO-DPLL
in these modes.
No
Valid Input
Available for
Selection? (2)
(2) See Input Selection Flowchart.
Yes
Lock Acquisition
(Fastlock, Hitless Switch)
Phase-locked to
selected input
Yes
Valid Input
Available for
Selection? (2)
No
REF-DPLL Locked
DCO Mode available.
No
Loss of Ref (LOR) on
Selected Input? (2)
Holdover Mode(1)
Initial holdover frequency
accuracy determined by
averaged history data.
Yes
No
Yes
Is Tuning Word
History Valid?
图 56. PLL Operating Mode
58
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Device Functional Modes (接下页)
9.4.2.1 Free-Run Mode
After device POR configuration and initialization, the APLL will automatically lock to the XO clock once it is
detected by its input monitor. The output clock frequency accuracy and stability in free-run mode are equal to
that of the XO input. If the TCXO input is used, the TCXO-DPLL will lock to the TCXO/OCXO clock once it is
detected by its input monitor, and the output clock frequency accuracy and stability in free-run mode are equal to
that of the TCXO/OCXO input. The reference inputs remain invalid (unqualified) during free-run mode.
9.4.2.2 Lock Acquisition
The PLL channel constantly monitors its assigned reference inputs for a valid input clock. When at least one
valid input clock is detected, the PLL will exit free-run mode or holdover mode and initiate lock acquisition
through the REF-DPLL. The device supports the Fastlock feature where the REF-DPLL temporarily engages a
wider loop bandwidth to reduce the lock time. Once lock acquisition is done, the loop bandwidth is set to its
normal configured loop bandwidth setting (BWREF-DPLL).
9.4.2.3 Locked Mode
Once locked, the PLL output clocks will be frequency and phase locked to its selected DPLL input clock. While
the DPLL is locked, the output clocks will not be affected by frequency drift on the XO or TCXO inputs. The REF-
DPLL has a programmable frequency lock detector and phase lock detectors to indicate loss of frequency lock
(LOFL) and loss of phase lock (LOPL) status flags, which can be observed through the status pins or status bits.
Once frequency lock is detected (LOFL → 0), the tuning word history monitor (if enabled) will begin to
accumulate history data that is used to determine the initial output frequency accuracy upon entry into holdover
mode.
9.4.2.4 Holdover Mode
When a loss of reference (LOR) condition is detected and no valid input is available, the PLL will enter holdover
mode. If the tuning word history is valid, the initial output frequency accuracy upon entry into holdover will be
pulled to the computed average frequency accuracy just prior to the loss of reference. If no history exists, the
holdover frequency accuracy will be determined by the free-run tuning word register (user programmable). The
initial holdover frequency accuracy depends on the DPLL loop bandwidth and the elapsed time used for historical
averaging. In general, the longer the historical average time, the more accurate the initial holdover frequency
assuming the 0-ppm reference clock is drift-free. The stability of the 0-ppm reference clock (either XO or TCXO
input) determines the long-term stability and accuracy of the holdover output frequency. Upon entry into
holdover, the LOPL flag will be asserted (LOPL → 1); however, the LOFL flag will not be asserted as long as the
holdover frequency accuracy does not drift beyond of the programmed loss-of-frequency-lock threshold. When a
valid input becomes available for selection, the PLL will exit holdover mode and automatically phase lock with the
new input clock without any output glitches.
版权 © 2018, Texas Instruments Incorporated
59
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Device Functional Modes (接下页)
9.4.3 PLL Start-Up Sequence
图 57 shows the general sequence for PLL start-up after device configuration. This sequence is also applicable
after a device soft-reset or individual PLL soft-reset. To ensure proper VCO calibration, it is critical for the
external XO clock to be stable in amplitude and frequency prior to the start of VCO calibration; otherwise, the
VCO calibration can fail and prevent start-up of the PLL and its output clocks.
Device Configured / Reset.
See Device POR
Configuration Flowchart
XO Detected
VCO Calibration
PLL Initialization Sequence
CAL Done
Outputs locked to XO frequency.
Outputs with DPLL auto-mute disabled are un-muted.
Outputs are auto-SYNCed if enabled.
APLL Locked
(Free-run from XO)
TCXO Detected
Outputs locked to TCXO/OCXO frequency.
Skip step if 2-loop REF-DPLL mode is
configured (TCXO-DPLL is not used).
2-loop TCXO-
DPLL mode
TCXO-DPLL
DCO Mode control
available
TCXO-DPLL Locked
(Free-run from
TCXO)
Input Monitoring (fastest to slowest detector):
1. Missing and/or Early clock detector
REF-DPLL modes
2. Amplitude or Slew rate detector
3. Frequency (ppm) detector
4. 1-PPS phase valid detector (skip #1 and 3)
5. After enabled detectors are valid, validation timer
starts and must finish for input to be qualified.
Ref. Input
Validation
Valid Input Selected
REF-DPLL
Lock Acquisition
Fastlock bandwidth temporarily asserted
during lock acquisition, if enabled.
REF-DPLL
Outputs locked to selected input clock frequency.
Outputs with DPLL auto-mute enabled are un-muted.
The configured DPLL loop bandwidth is asserted.
DPLL frequency- and phase-lock detectors are monitored.
If Zero-Delay Mode (ZDM) is enabled, output will have
deterministic input-to-output phase relationship.
DCO Mode control
available (ZDM
must be disabled)
REF-DPLL
Locked
See DPLL Modes and
Input Selection
Flowcharts
图 57. PLL Start-Up Sequence
60
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Device Functional Modes (接下页)
9.4.4 Digitally-Controlled Oscillator (DCO) Mode
To support IEEE 1588 slave clock and other clock steering applications, each PLL channel supports DCO mode
to allow precise output clock frequency adjustment of less than 1 ppt/step. DCO mode can be enabled on either
REF-DPLL or TCXO-DPLL loop when operating in locked mode.
The DCO frequency step size can be programmed through the frequency deviation or FDEV register
(DPLLy_FDEV bits). The FDEV step value is an offset added to or subtracted from the current numerator value
of the DPLL's fractional FB divider SDM that determines the DCO frequency offset at the VCO output.
The DCO frequency increment (FINC) or frequency decrement (FDEC) updates can be controlled through
software control or pin control. DCO updates through software control are always available through I2C or SPI by
writing to the DPLLy_FDEV_REG_UPDATE register bit. Writing a 0 will increment the DCO frequency by the
programmed step size, while writing a 1 will decrement it.
The pin control paths to each DCO block must be enabled through registers. Once enabled, a positive pulse on
the GPIO3/FINC1 or GPIO4/FDEC1 pin will apply a corresponding DCO update to DPLL1. Similarly, a positive
pulse on the GPIO5/FINC2 or GPIO6/FDEC2 pin will apply a corresponding DCO update to DPLL2. The
minimum positive pulse width applied to the FINC or FDEC pins should be greater than 100 ns to be captured by
the internal sampling clock. The DCO update rate should limited to less than 1 MHz when using pin control.
LMK05028
fTCXO-TDC
PLL1
DPLL numerator is incremented or decremented by the
DCO FDEV step word on the rising-edge of FINC or FDEC.
APLL
DPLLy_FDEV = (Reqd_ppb_step / 109) × DENDPLL × INTDPLL
REF
DPLL
TCXO
DPLL
fREF-TDC
fVCO1/P
FINC/FDEC Pin Control
FINC
DPLL1_IGNORE
_GPIO_PIN
1
0
FDEV
Step
FDEC
GPIO3_FDEV_EN
GPIO4_FDEV_EN
DPLL1_FDEV_EN
38-bit
DPLL1_FDEV
GPIO3/FINC1
GPIO4/FDEC1
DPLL1_DCO_SEL
_REF_TCXOb
DPLL2_IGNORE
_GPIO_PIN
fTCXO-TDC
GPIO5_FDEV_EN
GPIO6_FDEV_EN
DPLL2_FDEV_EN
PLL2
GPIO5/FINC2
GPIO6/FDEC2
APLL
REF
DPLL
TCXO
DPLL
fVCO2/P
fREF-TDC
FINC/FDEC Register Control
FINC
DPLL1_FDEV_REG_UPDATE
1
0
FDEV
Step
FDEC
0x24B[0]
I2C/SPI
38-bit
0x251[0]
DPLL2_DCO_SEL
_REF_TCXOb
DPLL2_FDEV
Write:
0 = FINC
1 = FDEC
DPLL2_FDEV_REG_UPDATE
图 58. DCO Mode Control Options
版权 © 2018, Texas Instruments Incorporated
61
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Device Functional Modes (接下页)
9.4.4.1 DCO Frequency Step Size
公式 16 shows the formula to compute the DPLLy_FDEV register value required to meet the specified DCO
frequency step size in ppb (part-per-billion) when DCO mode is enabled for the REF-DPLL (when
DPLLy_DCO_SEL_REF_TCXOB = 1).
DPLLy_FDEV = ( Reqd_ppb / 109 ) × DENREF / ( fINx / RINx ) × fVCOy / ( P1PLLy × PRREF
)
where
•
•
•
•
•
•
•
•
•
y: PLL index (1 or 2)
DPLLy_FDEV: Frequency deviation value (0 to 238–1)
Reqd_ppb: Required DCO frequency step size (in ppb)
DENREF: REF-DPLLy feedback divider denominator value (1 to 240)
fINx: Reference input frequency (x = 0, 1, 2, 3)
RINx: Reference input divide value (1 to 216–1) (x = 0, 1, 2, 3)
fVCOy: VCOy frequency
P1PLLy = PLLy primary post-divide value (4 to 9, 11, 13)
PRREF: REF-DPLLy feedback prescaler divide value (2 to 17)
(16)
公式 17 shows the formula to compute the DPLLy_FDEV register value required to meet the specified DCO
frequency step size (in ppb) when DCO mode is enabled for the TCXO-DPLL (when
DPLLy_DCO_SEL_REF_TCXOB = 0).
DPLLy_FDEV = ( Reqd_ppb / 109 ) × DENTCXO / ( fTCXO × D / M ) × fVCOy / ( P1PLLy × PRTCXO
)
where
•
•
•
•
•
DENTCXO: TCXO-DPLLy feedback divider denominator value (fixed, 240)
fTCXO: TCXO/OCXO input frequency
DTCXO: TCXO/OCXO input doubler (1 = disabled, 2 = enabled)
MTCXO: TCXO/OCXO input divide value (1 to 32)
PRTCXO: TCXO-DPLLy feedback prescaler divide value (2 to 17)
(17)
9.4.4.2 DCO Direct-Write Mode
An alternate method to update the DCO frequency is to take the current numerator value of the DPLL's fractional
feedback divider, compute the adjusted numerator value by adding or subtracting the DPLLy_FDEV step value
computed above, and to write the adjusted numerator value through I2C or SPI.
62
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Device Functional Modes (接下页)
9.4.5 Zero-Delay Mode (ZDM)
Each PLL channel supports the zero-delay mode option to achieve a known and deterministic phase relationship
between the reference and output clock. ZDM is supported for 2-loop and 3-loop modes through the REF-DPLL.
Once PLL is locked with ZDM enabled, the PLL will have minimal phase delay (phase offset) between its
reference input and the output clocks. The input-to-output phase offset (tPHO) will be repeatable after exiting
holdover, after a switchover event, and after device start-up. Note that ZDM and DCO mode should not be
enabled at the same time within a PLL channel.
As shown in 图 59, PLL1 supports zero-delay for OUT4/5, OUT6, and OUT7. PLL2 supports zero-delay for
OUT0, OUT1, and OUT2/3 through as similar ZDM configuration. See Zero-Delay Mode (ZDM) Configuration.
PLL1
OUT[4:7] Channels (x3)
INx
÷R
To TDC for
phase offset
cancellation
fVCO1/P1
÷OD
OUT
0
ZDM
Feedback
DPLL1 ZDM
Channel
DPLL1_
ZDM_EN
OUT4/5
OUT6
3
2
1
0
OUT7
ZDM off
O_CH4_7_ZERODLY_EN
图 59. DPLL1 ZDM Configuration for OUT4 to OUT7
9.4.6 Cascaded PLL Operation
Each PLL channel has a VCO loopback clock (VCOx FB) routed internally to the input paths of the opposite
DPLLs to support PLL cascading, where the VCO loopback clock from the first PLL stage (either PLL1 or PLL2)
is used as an input reference to the second PLL stage. PLL cascading allows the second PLL must be precisely
locked to the frequency of the first PLL. The internal VCO loopback configuration options are implemented
identically on both PLL channels, allowing PLL2 to be cascaded after PLL1 or vice versa. The loopback
configuration options include programmable VCO clock dividers, clock muxes to loop-back to either REF-DPLL or
TCXO-DPLL inputs, and loopback clock validation to control the PLL lock sequence for the second PLL stage.
The internal VCO loopback option eliminates the need for external clock loopback, which would otherwise require
the designer to dedicate an output buffer, an input buffer, and external routing to support cascaded PLL
operation. See PLL Cascading With Internal VCO Loopback.
版权 © 2018, Texas Instruments Incorporated
63
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
9.5 Programming
9.5.1 Interface and Control
A system host device (MCU or FPGA) can use either I2C or SPI to access the register, SRAM, and EEPROM
maps. The register and EEPROM map configurations are the same for I2C and SPI. The device can be
initialized, controlled, and monitored through register access during normal operation (not hard reset by PDN =
0). Some device features can also be controlled and monitored through the external logic control and status pins.
In the absence of a host, the LMK05028 can self-start from its on-chip EEPROM or ROM page depending on the
state of HW_SW_CTRL pin. The EEPROM or ROM page is used to initialize the registers upon device POR. The
EEPROM configuration can be custom programmed through the register interface by either I2C or SPI. The ROM
configurations are fixed in hardware and cannot be modified.
图 60 shows the device control pin, register, and memory interfaces. The arrows refer to the control interface
directions between the different blocks.
The register map has 800 data bytes. Some registers (such as status, internal test/diagnostic bit fields) do not
need to be written or accessed during device initialization.
The SRAM/EEPROM has one register page with 509 data bytes. The SRAM/EEPROM map has fewer bytes
because not all bit fields are mapped from the register space. To program the EEPROM, it is necessary to write
the register contents to SRAM (internal register commit or direct write), then Program EEPROM with the register
contents from SRAM. The EEPROM cannot be written directly from the registers.
The ROM has sixteen register pages, and each page has 509 data bytes (same as EEPROM). The ROM
contents are fixed in hardware and cannot be modified.
Select ROM
start-up mode
(Page 0 to 15)
Mask ROM
(16 Pages)
Addr: 0x000 to 0x1FCF
Data: 8143 bytes
(509 bytes / Page)
- Initialize Registers on POR (ROM pin mode)
- Read ROM
STATUS0
STATUS1
HW_SW_CTRL
PDN
Memory I/F
GPIO6
Control/
Status Pins
Device
Control
and
GPIO5
Registers
Device Blocks
(Inputs, PLLs,
Outputs, etc.)
Block-level I/F
Serial I/F
GPIO4
Addr: 0x000 to 0x31F
Data: 800 bytes
Status
GPIO3
GPIO0
GPIO2/SDO
GPIO1/SCS
SCL/SCK
SDA/SDI
I2C/SPI
Pins
Memory I/F
Memory I/F
- Initialize Registers on POR (EEPROM mode)
- Read EEPROM
- Write SRAM (Commit registers)
- Read SRAM
- EEPROM status (NVM busy, CRC error)
Program
EEPROM
SRAM
NVM EEPROM
Addr: 0x000 to 0x1FC
Data: 509 bytes
Addr: 0x000 to 0x1FC
Data: 509 bytes
Select EEPROM
start-up mode
图 60. Device Control, Register, and Memory Interfaces
64
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Programming (接下页)
9.5.2 I2C Serial Interface
When started in I2C mode (HW_SW_CTRL = 0 or 1), the LMK05028 operates as an I2C slave and supports bus
rates of 100 kHz (standard mode) and 400 kHz (fast mode). Slower bus rates can work as long as the other I2C
specifications are met.
In EEPROM mode, the LMK05028 can support up to four different I2C addresses depending on the GPIO[2:1]
pins. The 7-bit I2C address is 11000xxb, where the two LSBs are determined by the GPIO[2:1] input levels
sampled at device POR and the five MSBs (11000b) are initialized from EEPROM. In ROM mode, the two LSBs
are fixed to 00b and the five MSB (11000b) are initialized from ROM.
Write Transfer
1
7
1
1
S
Slave Address
Wr
A
8
8
1
1
Register Address High
A
Register Address Low
A
8
1
1
Data Byte
A
P
Read Transfer
1
7
1
1
S
Slave Address
Wr
A
8
8
1
1
Register Address High
A
Register Address Low
A
7
Slave Address
8
1
1
1
A
1
Sr
Rd
1
Data Byte
A
P
Legend
S
Sr
Start condition sent by master device
Write bit = 0 sent by master device
Acknowledge sent by master device
Stop condition sent by master device
|
|
|
Repeated start condition sent by master device
Read bit = 1 sent by master device
Wr Rd
A
P
N
A
Acknowledge sent by slave device
Not-acknowledge sent by master device
Data sent by master
|
|
Not-acknowledge sent by slave device
Data sent by slave
N
Data
Data
图 61. I2C Byte Write and Read Transfers
版权 © 2018, Texas Instruments Incorporated
65
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Programming (接下页)
9.5.2.1 I2C Block Register Transfers
The device supports I2C block write and block read register transfers as shown in 图 62
Block Write Transfer
1
7
1
1
S
Slave Address
Wr
A
8
8
1
1
Register Address High
A
Register Address Low
A
8
1
8
1
1
Data Byte
A
Data Byte
A
P
Block Read Transfer
1
7
1
1
S
Slave Address
Wr
A
8
8
1
1
Register Address High
A
Register Address Low
A
7
Slave Address
8
1
1
1
A
1
Sr
Rd
8
1
1
Data Byte
A
Data Byte
A
P
图 62. I2C Block Register Transfers
9.5.3 SPI Serial Interface
When started in SPI mode (HW_SW_CTRL = float), the device uses a 4-wire SPI interface with SDI, SCK, SDO,
and SCS signals. The host device must present data to the device MSB first. A message includes a transfer
direction bit (W/R), a 15-bit address field (A14 to A0), and a 8-bit data field (D7 to D0) as shown in 图 63. The
W/R bit is 0 for a SPI write and 1 for a SPI read.
MSB
1
LSB
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Order of Transmission
Bit Definition
A
10
A
9
A
8
A
7
A
6
A
5
A
4
A
3
A
2
A
1
A
0
D
7
D
6
D
5
D
4
D
3
D
2
D
1
D
0
A
A
A
A
First Out
14 13 12 11
Message Field Definition
Register Address (15 bits)
Data Payload (8 bits)
图 63. SPI Message Format
A message frame is initiated by asserting SCS low. The frame ends when SCS is deasserted high. The first bit
transferred is the W/R bit. The next 15 bits are the register address, and the remaining 8 bits are data. On write
transfers, data is committed in bytes as the final data bit (D0) is clocked in on the rising edge of SCK. If the write
access is not an even multiple of 8 clocks, the trailing data bits are not committed. On read transfers, data bits
are clocked out from the SDO pin on the falling edges of SCK.
66
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Programming (接下页)
9.5.3.1 SPI Block Register Transfer
The device supports a SPI block write and block read transfers. A SPI block transfer is exactly (2 + N) bytes
long, where N is the number of data bytes to write or read. The host device (SPI master) is only required to
specify the lowest address of the sequence of addresses to be accessed. The device will automatically increment
the internal register address pointer if the SCS pin remains low after the host finishes the initial 24-bit
transmission sequence. Each transfer of 8 bits (a data payload width) results in the device automatically
incrementing the address pointer (provided the SCS pin remains active low for all sequences).
9.5.4 Register Map Generation
The TICS Pro software tool for EVM programming has a step-by-step design flow to enter the user's clock design
parameters, calculate the frequency plan, and generate the device register settings for the desired configuration.
The register settings can be exported (in hex format) to enable host programming of the LMK05028 on start-up.
The TICS Pro setup file can also be provided to TI for device configuration review, optimization, and to enable
factory pre-programmed sample devices.
9.5.5 General Register Programming Sequence
For applications that use a system host device to program the initial LMK05028 configuration after start-up, this
general procedure can be followed from the register map data generated and exported from TICS Pro:
1. Apply power to the device to start in I2C or SPI mode. The PDN pin should be pulled high or driven high.
2. Write 9Ah to R12 to disable device auto-start and mute the outputs during programming.
3. Write 1Ah to R12 to exit soft-reset (this does not reset register values).
4. Write the register settings from lower to higher addresses (R0 to R647) while applying the following register
mask (Do not modify mask bits = 1):
–
–
–
Mask R12 = A5h
Mask R167 = FFh
Mask R174 = FFh
5. Write 1b to R12[7] to assert soft-reset. This does not reset register values.
6. Write 1b to R12[7]) to begin normal device operation (starting with VCO calibration and PLL start-up).
7. See EEPROM Programming Using Register Commit (Method #1) to store the active configuration to the
EEPROM to enable self-startup on the next power cycle.
9.5.6 EEPROM Programming Flow
Before the EEPROM can be programmed, it is necessary to program the desired configuration to the SRAM
through the memory control registers. The register data can be written to SRAM by committing the active register
configuration internally (Method #1), or by using direct writes to SRAM (Method #2). Method #1 requires the
active registers be first programmed to the desired configuration, but does not require knowledge of the
SRAM/EEPROM map. Method #2 bypasses any writes to the active registers configuration, allowing the device
to continue normal operation without disruption while the SRAM/EEPROM are programmed. The EEPROM
programming flow for the two methods are different and described as follows.
9.5.6.1 EEPROM Programming Using Register Commit (Method #1)
This sequence can be followed to program EEPROM from the active registers. This requires the register data in
the register map format:
1. Program the desired configuration to the active registers (see General Register Programming Sequence).
2. Write SRAM Using Register Commit.
3. Program EEPROM.
9.5.6.1.1 Write SRAM Using Register Commit
The SRAM array is volatile shadow memory mapped to a subset of the active configuration registers that
determine the device's frequency configuration at start-up.
版权 © 2018, Texas Instruments Incorporated
67
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Programming (接下页)
After the active registers have been programmed, they can be internally committed to the SRAM through the
following sequence:
1. Write 0h to R175 (REGCOMMIT_PG byte).
2. Write 40h to R167 (REGCOMMIT bit, self-clearing). This commits the current register data to the SRAM
internally.
9.5.6.1.2 Program EEPROM
The EEPROM array is non-volatile memory mapped directly from the SRAM array.
After the register settings have been written to SRAM (by either Method #1 or #2), the EEPROM can be
programmed through the following sequence:
1. Write EAh to R174 (NVMUNLK). This unlocks the EEPROM to allow programming.
2. Write 03h to R167 (NVM_ERASE_PROG bits). This programs the EEPROM from the entire SRAM contents.
The total erase/program cycle takes about 230 ms.
–
NOTE: Steps 1 and 2 must be atomic writes without any other register transactions in-between.
3. (optional) Read or poll R167[2] (NVMBUSY bit). When this bit cleared, the EEPROM programming is done.
4. (optional) Write 00h to R174. This locks the EEPROM to protect against inadvertent programming.
On the next power-up or hard reset, the device can self-start in EEPROM mode from the newly programmed
configuration. Also, the NVMCNT register value will be incremented by 1 after power-up or hard reset to reflect
total number of EEPROM programming cycles completed.
9.5.6.2 EEPROM Programming Using Direct SRAM Writes (Method #2)
This sequence can be followed to program EEPROM by writing SRAM directly to avoid disruption to the current
device configuration. This requires the register data in the SRAM/EEPROM map format.
1. Write SRAM Using Direct Writes.
2. Program EEPROM.
9.5.6.2.1 Write SRAM Using Direct Writes
This SRAM write method can be used if it is required to store a different device configuration to EEPROM without
disrupting the current operational state of the device. This method requires the SRAM/EEPROM map data to be
already generated, which can be done by the TICS Pro software. The SRAM can be directly written without
modifying the active configuration registers through the following sequence:
1. Write the most significant 5 bits of the SRAM address to R169 (MEMADR byte 1) and write the least
significant 8 bits of the SRAM address to R170 (MEMADR byte 0).
2. Write the SRAM data byte to R172 (RAMDAT byte) for the SRAM address specified in the previous step in
the same register transaction.
–
–
–
–
Any additional write transfers in same transaction will cause the SRAM address pointer to be auto-
incremented and a subsequent write will take place at the next SRAM address.
Byte or Block write transfers to R172 can be used to write the entire SRAM map sequentially from Byte 0
to 508 (509 bytes total).
Alternatively, it is valid to write R169 and R170 before each write to R172 to set the memory address
explicitly (that is, bypass the memory pointer auto-increment).
Access to the SRAM will terminate at the end of current write transaction. Note that reading the RAMDAT
register will also auto-increment the memory address pointer.
9.5.7 Read SRAM
The contents of the SRAM can be read back, one word at a time, starting with that of the requested address
through the following sequence. This sequence can be used to verify the contents written to SRAM before an
EEPROM program cycle.
1. Write the most significant 5 bits of the SRAM address to R169 (MEMADR byte 1) and write the least
significant 8 bits of the SRAM address to R170 (MEMADR byte 0).
68
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Programming (接下页)
2. The SRAM data located at the address specified in the step above can be obtained by reading R172
(RAMDAT byte) in the same register transaction.
–
–
–
Any additional read transfers that is part of the same transaction will cause the SRAM address to be auto-
incremented and a subsequent read will take place at the next SRAM address.
Byte or Block read transfers from R172 can be used to read the entire SRAM map sequentially from Byte
0 to 508 (509 bytes total).
Access to SRAM will terminate at the end of current register transaction.
9.5.8 Read EEPROM
The contents of the EEPROM can be read back, one word at a time, starting with that of the requested address
through the following sequence. This sequence can be used to verify the contents written to EEPROM after an
EEPROM program cycle.
1. Write the most significant 5 bit of the EEPROM address in R169 (MEMADR byte 1) and write the least
significant 8 bits of the EEPROM address in R170 (MEMADR byte 0).
2. The EEPROM data located at the address specified in the step above can be obtained by reading R171
(NVMDAT byte) in the same register transaction.
–
–
–
Any additional access that is part of the same transaction will cause the EEPROM address to be
incremented and a read will take place of the next EEPROM address.
Byte or Block read transfers from R171 can be used to read the entire EEPROM map sequentially from
Byte 0 to 508 (509 bytes total).
Access to EEPROM will terminate at the end of current register transaction.
版权 © 2018, Texas Instruments Incorporated
69
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Programming (接下页)
9.5.9 EEPROM Start-Up Mode Default Configuration
The device is factory pre-programmed with the following EEPROM default configuration.
表 11. LMK05028 Default Configuration
SYSTEM CLOCKS
FREQUENCY (Hz)
INPUT TYPE
0-PPM REF CLK
XO
48,000,000
-
DC-DIFF(ext. term)
XO
TCXO
CLOCK INPUTS
FREQUENCY (Hz)
156,250,000
156,250,000
27,000,000
DPLL REF MUX
DPLL 1,2
INPUT TYPE
IN0
IN1
IN2
IN3
AC-DIFF(ext. term)
AC-DIFF(ext. term)
AC-DIFF(ext. term)
AC-DIFF(ext. term)
DPLL 1,2
DPLL 1,2
10,000,000
DPLL 1,2
MANUAL REGISTER
SELECTION
DPLL1 INPUT SELECT MODE
Auto Revertive
DPLL1 INPUT
AUTO PRIORITY
IN0
IN0
IN1
IN2
IN3
1st
2nd
3rd
4th
MANUAL REGISTER
SELECTION
DPLL2 INPUT SELECT MODE
Auto Revertive
DPLL2 INPUT
AUTO PRIORITY
IN0
IN0
IN1
IN2
IN3
1st
2nd
3rd
4th
CLOCK OUTPUTS
OUT0
FREQUENCY (Hz)
322,265,625.0
122,880,000.0
122,880,000.0
122,880,000.0
122,880,000.0
122,880,000.0
122,880,000.0
322,265,625.0
OUTPUT MUX
PLL 1
OUTPUT TYPE
AC-LVDS
OUT1
PLL 2
AC-LVDS
OUT2
PLL 2
AC-LVDS
OUT3
PLL 2
AC-LVPECL
AC-LVPECL
AC-LVDS
OUT4
PLL 2
OUT5
PLL 2
OUT6
PLL 2
AC-LVDS
OUT7
PLL 1
AC-LVDS
PLL CONFIGURATION
PLL MODE
REF-DPLL BW (Hz)
TCXO-DPLL BW (Hz)
PLL1
PLL2
2-loop (REF-DPLL)
2-loop (REF-DPLL)
100
100
-
-
REF INPUT MONITORS (1)
VALIDATION TIMER (s)
FREQ DET VALID (ppm)
FREQ DET INVALID (ppm)
IN0
IN1
IN2
IN3
0.1
0.1
0.1
0.1
-
-
-
-
-
-
-
-
1-PPS JITTER THRESHOLD
REF INPUT MONITORS (2)
LATE DETECT WINDOW (μs)
EARLY DETECT WINDOW (μs)
(μs)
IN0
IN1
IN2
IN3
-
-
-
-
-
-
-
-
-
-
-
-
70
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Programming (接下页)
表 11. LMK05028 Default Configuration (接下页)
DPLL LOCK DETECT
DPLL1
FREQ LOCK (ppm)
FREQ UNLOCK (ppm)
1
1
3
3
DPLL2
DCO MODE
DPLL1
DCO MODE SELECT
DCO Disabled
DCO STEP SIZE (ppb)
-
-
DPLL2
DCO Disabled
ZERO DELAY MODE
DPLL1
ZDM FEEDBACK CLOCK
Disabled
DPLL2
Disabled
9.6 Register Maps
See the LMK05028 Programming Guide (SNAU233) for the register map with register descriptions.
版权 © 2018, Texas Instruments Incorporated
71
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
10 Application and Implementation
注
Information in the following applications sections is not part of the TI component
specification, and TI does not warrant its accuracy or completeness. TI’s customers are
responsible for determining suitability of components for their purposes. Customers should
validate and test their design implementation to confirm system functionality.
10.1 Application Information
10.1.1 Device Start-Up Sequence
The device start-up sequence is shown in 图 64. If an output channel's VDDO_x is delayed after the device POR,
the output channel is held in reset and its output is muted. Once VDDO_x is ramped above its threshold of about
1.5 V, the output channel is held in reset until its programmable timeout counter expires before the output driver
is unmuted and clock starts up without any glitches.
72
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
Application Information (接下页)
Power-On Reset
(POR)
Device POR
Configuration Sequence
0
PDN?
Outputs muted
(Hard reset)
1
HW_SW_CTRL = 1
GPIO[3:0] = 0000b to 1111b
(Select ROM Page)
HW_SW_CTRL = 0
GPIO[2:1] = 00b to 11b
(Select I2C Addr.)
Start-up Mode?
(sample pin states)
HW_SW_CTRL = Float
STATUS[1:0] = Float
EEPROM + I2C
EEPROM + SPI
ROM + I2C
(Soft Pin) Mode
(Soft Pin) Mode
(Hard Pin) Mode
Registers initialized from EEPROM/ROM (hard reset only).
I2C/SPI, Control, Status pins activated.
Device Block
Configuration
Soft reset
RESET_SW: 0‰1
All blocks reset to initial states.
Register and EEPROM programming available.
XO Detected
VCO Calibration
PLL Initialization Sequence
CAL Done
Outputs locked to XO frequency.
Outputs with DPLL auto-mute disabled are un-muted.
Outputs are auto-SYNCed if enabled.
APLL Locked
(Free-run from XO)
TCXO Detected
Outputs locked to TCXO/OCXO frequency.
Skip step if 2-loop REF-DPLL mode is
configured (TCXO-DPLL is not used).
2-loop TCXO-
DPLL mode
TCXO-DPLL
DCO Mode control
available
TCXO-DPLL Locked
(Free-run from
TCXO)
Input Monitoring (fastest to slowest detector):
1. Missing and/or Early clock detector
REF-DPLL modes
2. Amplitude or Slew rate detector
3. Frequency (ppm) detector
4. 1-PPS phase valid detector (skip #1 and 3)
5. After enabled detectors are valid, validation timer
starts and must finish for input to be qualified.
Ref. Input
Validation
Valid Input Selected
REF-DPLL
Lock Acquisition
Fastlock bandwidth temporarily asserted
during lock acquisition, if enabled.
REF-DPLL
DCO Mode control
available (ZDM must be
disabled)
Outputs locked to selected input clock frequency.
Outputs with DPLL auto-mute enabled are un-muted.
The configured DPLL loop bandwidth is asserted.
DPLL frequency- and phase-lock detectors are monitored.
If Zero-Delay Mode (ZDM) is enabled, output will have
deterministic input-to-output phase relationship.
REF-DPLL
Locked
See DPLL Modes and
Input Selection
Flowcharts
图 64. Device Start-Up Sequence
版权 © 2018, Texas Instruments Incorporated
73
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
Application Information (接下页)
10.1.2 Power Down (PDN) Pin
The PDN pin (active low) can be used for device power-down and used to initialize the POR sequence. When
PDN is pulled low, the entire device is powered down and the serial interface is disabled. When PDN is pulled
high, the device POR sequence is triggered to begin the device start-up sequence and normal operation as
depicted in 图 64. If the PDN pin is toggled to issue a momentary hard reset, the negative pulse applied to the
PDN pin should be greater than 200 ns to be captured by the internal digital system clock.
表 12. PDN Control
PDN PIN STATE
DEVICE OPERATION
Device is disabled
Normal operation
0
1
10.1.3 Power Rail Sequencing, Power Supply Ramp Rate, and Mixing Supply Domains
10.1.3.1 Mixing Supplies
The LMK05028 incorporates flexible power supply architecture. While all VDD core supplies should be powered
by the same 3.3-V rail, the individual output supplies can be powered from separate 1.8-V, 2.5-V or 3.3-V rails.
This can allow all output supplies at 1.8 V to minimize power consumption. It can also allow mixed output driver
levels simultaneously, for example, a 2.5-V LVCMOS clock from a 2.5-V rail and other differential clocks from a
1.8-V rail.
10.1.3.2 Power-On Reset (POR) Circuit
The LMK05028 integrates a built-in power-on reset (POR) circuit that holds the device in reset until all of the
following conditions have been met:
•
•
All VDD core supplies have ramped above 2.72 V
PDN pin has ramped above 1.2 V (VIH)
10.1.3.3 Powering Up From a Single-Supply Rail
As long as all VDD core supplies are driven by the same 3.3-V supply rail that ramp in a monotonic manner from
0 V to 3.135 V, irrespective of the ramp time, then there is no requirement to add a capacitor on the PDN pin to
externally delay the device power-up sequence. As shown in 图 65, the PDN pin can be left floating or otherwise
driven by a system host device to meet the clock sequencing requirements in the system.
VDD_PLLx, VDD_INx,
VDD_DIG, VDD_XO,
VDD_TCXO, PDN
3.135 V
Decision Point 2:
VDD_PLLx/VDD_INx/
VDD_DIG / VDD_XO /
VDD_TCXO ≥ 2.72 V
VDD_DIG
200 kꢀ
Decision Point 1:
PDN ≥ 1.2 V
PDN
0 V
图 65. Recommendation for Power Up From a Single-Supply Rail
74
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
10.1.3.4 Power Up From Split-Supply Rails
If some VDD core supplies are driven from different supply rails, TI recommends starting the PLL calibration after
all of the core supplies have settled at 3.135 V. This can be realized by delaying the PDN low-to-high transition.
The PDN input incorporates a 200-kΩ resistor to VDD_DIG and as shown in 图 66, a capacitor from the PDN pin
to GND can be used to form an R-C time constant with the internal pullup resistor. This R-C time constant can be
designed to delay the low-to-high transition of PDN until all the core supplies have settled at 3.135 V.
Alternatively, the PDN pin can be driven high by a system host or power management device to delay the device
power-up sequence until all VDD supplies have ramped.
3.135 V
VDD_INx, VDD_TCXO, VDD_XO
VDD_PLLx, VDD_DIG
PDN
VDD_DIG
Decision Point 1b:
VDD_PLLx,
VDD_DIG ≥ 2.72 V
Decision Point 1a:
VDD_INx, VDD_TCXO,
VDD_XO ≥ 2.72 V
200 kꢀ
PDN
Decision Point 2:
PDN ≥ 1.2 V
CPDN
Delay
0 V
图 66. Recommendation for Power Up From Split-Supply Rails
10.1.3.5 Non-Monotonic or Slow Power-Up Supply Ramp
In case the VDD core supplies ramp with a non-monotonic manner or with a slow ramp time from 0 V to 3.135 V
of over 100 ms, TI recommends delaying the VCO calibration until after all of the core supplies have settled at or
above 3.135 V. This can be realized by delaying the PDN low-to-high transition using one of the methods
described in 图 66.
10.1.4 Slow or Delayed XO Start-Up
Because the external XO clock input is used as the reference input for the VCO calibration, the XO input
amplitude and frequency must be stable before the start of VCO calibration to ensure successful PLL lock and
output start-up. If the XO clock is not stable prior to VCO calibration, the VCO calibration can fail and prevent
PLL lock and output clock start-up.
If the XO clock has a slow start-up time or glitches on power-up (due to a slow or non-monotonic power supply
ramp, for example),TI recommends to delay the start of VCO calibration until after the XO is stable. This could be
achieved by delaying the PDN low-to-high transition until after the XO clock has stabilized using one of the
methods described in 图 66. It is also possible to issue a device soft-reset after the XO clock has stabilized to
manually trigger the VCO calibration and PLL start-up sequence.
10.2 Typical Application
图 67 shows a reference schematic to help implement the LMK05028 and its peripheral circuitry. Power filtering
examples are given for the core supply pins and independent output supply pins. Single-ended LVCMOS, AC-
coupled differential, and HCSL clock interfacing examples are shown for the clock input and output pins. An
external LVCMOS oscillator drives an AC-coupled voltage divider network as an example to interface the 3.3-V
LVCMOS output to meet the input voltage swing specified for the XO or TCXO inputs. The required external
capacitors are placed close to the LMK05028 and shown with the recommended values. External pullup and
pulldown resistor options at the logic I/O pins set the default input states. The I2C or SPI pins and other logic I/O
pins can be connected to a host device (not shown) to program and control the LMK05028 and monitor its status.
This example assumes the device will start up from EEPROM mode with an I2C interface (HW_SW_CTRL = 0).
版权 © 2018, Texas Instruments Incorporated
75
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
图 67. LMK05028 Reference Schematic Example
76
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
10.2.1 Design Requirements
In a typical application, the following design requirements or parameters should be considered to implement the
overall clock solution:
1. Device initial configuration: Host programmed (MCU or FPGA) or factory pre-programmed.
2. Device start-up mode and serial interface. Typically, this will be EEPROM + I2C or SPI mode.
3. XO frequency, signal type, and phase noise or jitter
4. TCXO frequency and stability if any of the following is required:
–
–
–
Standard-compliant frequency stability (such as SyncE, SONET/SDH, IEEE 1588)
Lowest possible close-in phase noise at offsets ≤ 100 Hz
Narrow DPLL bandwidth ≤ 10 Hz
5. For each PLL domain, determine the following:
–
–
–
–
Input clocks: frequency, buffer mode, priority, and input selection mode
Output clocks: frequency, buffer mode
DPLL loop mode, loop bandwidth, and market segment
DCO mode or Zero delay
6. Input clock and PLL monitoring options
7. Status outputs and interrupt flag
8. Power supply rails
10.2.2 Detailed Design Procedure
In a typical application, TI recommends the following steps:
1. The LMK05028 GUI in the TICS Pro programming software has a step-by-step design flow to enter the
design parameters, calculate the frequency plan for each PLL domain, and generate the register settings for
the desired configuration. The register settings can be exported (in hex format) to enable host programming
or factory pre-programming.
–
If using a generic (non-custom) device, a host device can program the register settings through the serial
interface after power-up and issue a soft-reset (by RESET_SW bit) to start the device. The host can also
store the settings to the EEPROM to allow self-startup with these register settings on subsequent power-
on reset cycles.
–
Alternatively, a LMK05028 setup file for TICS Pro (.tcs) can be sent to TI to request custom factory pre-
programmed devices.
2. Tie the HW_SW_CTRL pin to ground to select EEPROM+I2C mode, or bias the pin to VIM through the weak
internal resistors or external resistors to select EEPROM+SPI mode. Determine the logic I/O pin assignments
for control and status functions. See Device Start-Up Modes.
–
Connect I2C/SPI and logic I/O pins (1.8-V compatible levels) to the host device pins with the proper I/O
direction and voltage levels.
3. Select a XO frequency by following Oscillator Input (XO_P/N).
–
–
Choose a XO with target phase jitter performance < 300 fs RMS (12 kHz to 20 MHz).
For a 3.3-V LVCMOS driver, follow the OSC clock interface example in 图 67. Power the OSC from a low-
noise LDO regulator or optimize its supply filtering to avoid supply-induced jitter on the XO clock.
–
TICS Pro: Configure the XO input buffer mode to match the XO driver interface requirements. See 表 4.
4. If a TCXO/OCXO is needed, select the frequency by following TCXO/OCXO Input (TCXO_IN).
–
Choose a TCXO/OCXO that meets the frequency stability and accuracy requirements required for the
output clocks during free-run or holdover.
–
–
For a 3.3-V LVCMOS driver, follow the OSC clock interface example in 图 67.
A (clipped) sinewave TCXO/OCXO with less than 1.3-Vpp swing can be simply AC-coupled to the input
pin.
–
TICS Pro: The TCXO/OCXO input buffer is enabled when either PLL channel uses the TCXO-DPLL.
版权 © 2018, Texas Instruments Incorporated
77
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
5. For each PLL domain, wire the clock I/O in the schematic and use TICS Pro to configure the device settings
as follows:
–
Reference inputs: Follow the LVCMOS or differential clock input interface examples in 图 67 or Clock
Input Interfacing and Termination.
–
TICS Pro: Configure the reference input buffer modes to match the reference clock driver interface
requirements. See 表 5.
–
LVCMOS clock input should be used for input frequencies below 5 MHz when amplitude monitoring is
enabled.
–
–
TICS Pro: Configure the DPLL input selection modes and input priorities. See Reference Input Mux
Selection.
TICS Pro: Output clock assignment guidelines to minimize crosstalk and spurs.
–
–
–
–
OUT[4:7] bank requires at least one clock from the PLL1 domain. OUT[4:7] bank is preferred for PLL1
clocks.
OUT[0:3] bank requires at least one clock from the PLL2 domain (if PLL2 is enabled). OUT[0:3] bank
is preferred for PLL2 clocks.
Group identical output frequencies (or harmonic frequencies) on adjacent channels, and use the
output pairs with a single divider (OUT2/3 or OUT4/5) when possible to minimize power.
Separate clock outputs when the difference of the two frequencies, |fOUTx – fOUTy|, falls within the jitter
integration bandwidth (12 kHz to 20 MHz, for example). Any outputs that are potential aggressors
should be separated by at least four static pins (power pin, logic pin, or disabled output pins) to
minimize potential coupling. If possible, separate these clocks by the placing them on opposite output
banks, which are on opposite sides of the chip for best isolation.
–
–
Avoid or isolate any LVCMOS output (strong aggressor) from other jitter-sensitive differential output
clocks. If a LVCMOS output is required, use dual complementary LVCMOS mode (+/- or -/+) with the
unused LVCMOS output left floating with no trace. Furthermore, the output slew rate could also be
slowed to Normal mode (CHx_SLEW_RATE bit) to reduce the coupling strength of an LVCMOS
output.
If not all outputs pairs are used in the application, consider connecting OUT0 and/or OUT1 to a pair of
RF coaxial test structures for testing purposes (such as SMA, SMP ports). OUT0 and OUT1 are
capable of selecting a buffered copy of the XO clock or the TCXO/Ref Bypass clock as well as any
PLL post-divider clock.
–
–
TICS Pro: Configure output divider and drivers.
–
–
Configure the output driver modes to match the receiver clock input interface requirements. See 表 9.
Configure any output SYNC groups that need their output phases synchronized. See Output
Synchronization (SYNC).
–
Configure the output auto-mute modes, output mute levels, and APLL and DPLL mute options. See
Output Auto-Mute During LOL or LOS.
Clock output Interfacing: Follow the single-ended or differential clock output interface examples in 图 67
or Clock Output Interfacing and Termination.
–
–
Differential outputs should be AC-coupled and terminated and biased at the receiver inputs.
HCSL outputs should have 50-Ω termination to GND (at source or load side) unless the internal
source termination is enabled by programming.
–
LVCMOS outputs have internal source termination to drive 50-Ω traces directly. LVCMOS VOH level is
determined by VDDO voltage (1.8 V and 2.5 V).
–
–
TICS Pro: Configure the PLL loop mode. See PLL Configurations for Common Applications.
–
–
–
3-Loop mode: Supports standards-compliant synchronization using a low-cost holdover TCXO/OCXO,
very low loop bandwidths (≤10 Hz), or both. 3-loop mode also supports 1-PPS input synchronization.
2-Loop REF-DPLL mode: Supports higher loop bandwidth (>10 Hz) and relaxed holdover stability
without a TCXO/OCXO.
2-Loop TCXO-DPLL mode: Locks to a TCXO/OCXO input and is typically used with DCO mode
enabled for external clock steering (such as IEEE 1588 PTP).
TICS Pro: Configure the REF-DPLL loop bandwidth.
–
Below the loop bandwidth, the reference noise is added to the REF-TDC noise floor (and the XO
noise in 2-loop mode). Above the loop bandwidth, the reference noise will be attenuated with roll-off
78
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
up to 60 dB/decade. The optimal bandwidth depends on the relative phase noise between the
reference input and the XO (2-loop mode) or the TCXO (3-loop mode).
–
–
3-Loop mode: The bandwidth is typically 10 mHz to 10 Hz to attenuate wander, or determined by the
applicable standard specification.
2-Loop mode: The bandwidth is typically 10 Hz or higher. Target a bandwidth below 200 Hz if the PLL
VCO frequency is an integer multiple of the reference input frequency.
–
–
TICS Pro: Configure the TCXO-DPLL loop bandwidth.
–
The optimal bandwidth depends on the relative phase noise between the TCXO/OCXO and the XO.
Below the loop bandwidth, the TCXO/OCXO noise is added to the TCXO-TDC noise floor. Above the
loop bandwidth, the OCXO/TCXO noise will be attenuated.
–
–
3-Loop mode: The bandwidth should be at least 50x higher than the REF-DPLL bandwidth for loop
stability.
2-Loop TCXO-DPLL mode: Target a bandwidth below 300 Hz if the PLL VCO frequency is an integer
multiple of the TCXO/OCXO frequency.
TICS Pro: Configure the Market Segment parameter to optimize the DPLL for the desired use case.
–
–
–
SyncE/SONET: REF-TDC rate is targeted for approximately 400 kHz. Hitless switching is enabled.
This supports SyncE/SONET and other use cases using a narrow loop bandwidth (≤10 Hz) in 3-loop
mode with a TCXO/OCXO to set the frequency stability and wander performance.
Wireless/BTS: REF-TDC rate is maximized for lowest in-band TDC noise contribution. Hitless
switching is enabled. Supports wireless and other use cases where close-in phase noise is critical.
This is used to achieve –112 dBc/Hz at 100-Hz offset for a 122.88-MHz output.
OTN/JitterAtten: REF-TDC rate is targeted for approximately 1 MHz. Hitless switching is disabled.
Supports OTN/OTU and traditional jitter cleaning use cases with wider bandwidths (>10 Hz) in 2-loop
mode and relaxed holdover frequency accuracy (no TCXO/OCXO).
–
–
TICS Pro: If clock steering is needed (such as for IEEE 1588 PTP), configure DCO mode for the REF or
TCXO loop and enter the frequency step size (in ppb). The FDEV step register will be computed
according to DCO Frequency Step Size. To allow DCO frequency updates using the external control pins,
enable the FINC/FDEC functionality on the needed GPIO pins by setting the appropriate register bits
(GPIO[3:6]_FDEV_EN).
TICS Pro: If deterministic input-to-output clock phase is needed, configure Zero-Delay mode and select
the primary output channel that must be phase-aligned with the input. For DPLL1, any output from the
OUT[4:7] bank can be selected for zero-delay feedback. For DPLL2, any output from the OUT[0:3] bank
can be selected for zero-delay feedback. See Zero-Delay Mode (ZDM) Configuration.
6. TICS Pro: Configure the reference input monitoring options for each reference input. Disable the monitor
when not required or when the input operates beyond the monitor's supported frequency range. See
Reference Input Monitoring.
–
Amplitude monitor: Set the LVCMOS detected slew rate edge or the differential input amplitude threshold
to monitor input signal quality. Disable the monitor for a differential input below 5 MHz or else use
LVCMOS input clock.
–
–
Frequency monitor: Set the valid and invalid thresholds (in ppm).
Missing pulse monitor: Set the late window threshold (TLATE) to allow for the longest expected input clock
period, including worst-case cycle-to-cycle jitter. For a gapped clock input, set TLATE based on the number
of allowable missing clock pulses.
–
–
–
Runt pulse monitor: Set the early window threshold (TEARLY) to allow for the shortest expected input clock
period, including worst-case cycle-to-cycle jitter.
1-PPS Phase validation monitor: Set the phase validation jitter threshold, including worst-case input
cycle-to-cycle jitter.
Validation timer: Set the amount of time the reference input must be qualified by all enabled input
monitors before the input is valid for selection.
7. TICS Pro: Configure the DPLL lock detect and tuning word history monitoring options for each channel. See
PLL Lock Detectors and Tuning Word History.
–
DPLL tuning word history: Set the history count/averaging time (TAVG), history delay/ignore time (TIGN),
and intermediate averaging option.
–
DPLL frequency lock and phase lock detectors: Set the lock and unlock thresholds for each detector.
版权 © 2018, Texas Instruments Incorporated
79
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
8. TICS Pro: Configure each status output pin and interrupt flag as needed. See Status Outputs and Interrupt.
–
Select the desired status signal selection, status polarity, and driver mode (3.3-V LVCMOS or open-
drain). Open-drain requires an external pullup resistor.
–
If the Interrupt is enabled and selected as a status output, configure the mask bit for any interrupt source
to be ignored, interrupt flag polarity, and the combinational gate logic (AND/OR) as desired.
10.2.3 Application Curves
AC-LVPECL output, fIN = 25 MHz, fTCXO = 10 MHz (OCXO), fXO = 48.0048 MHz, fTCXO-TDC = 20 MHz
图 68. 122.88-MHz Output Phase Noise (3-Loop), Wireless Use Case
AC-LVPECL output, fIN = 19.44 MHz, fTCXO = 12.8 MHz (OCXO), fXO = 48.0048 MHz, fTCXO-TDC = 25.6 MHz,
BWREF-DPLL = 1 Hz, BWTCXO-DPLL = 200 Hz
图 69. 155.52-MHz Output Phase Noise (3-Loop), Telecom Use Case
80
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
10.3 Do's and Don'ts
•
•
•
•
•
Power all VDD pins with proper supply decoupling and bypassing connect like shown in 图 67.
Power down unused blocks through registers to minimize power consumption.
Leave unused clock outputs floating and powered down through register control.
Leave unused clock inputs floating.
For EEPROM+SPI Mode: Leave HW_SW_CTRL and STATUS[1:0] pins floating during POR to ensure proper
start-up. These pins has internal biasing to VIM internally.
–
If HW_SW_CTRL or either STATUS pin is connected to a host device (MCU or FPGA), the external
device must be configured with high-impedance input (no pullup or pulldown resistors) to avoid conflict
with the internal bias to VIM. If needed, external biasing resistors (10-kΩ pullup to 3.3 V and 3.3-kΩ
pulldown) can be connected on each STATUS pin to bias the inputs to VIM during POR.
•
•
Use a low-noise, high-PSRR LDO regulator to power the external oscillators used to drive the XO and TCXO
inputs. Typically, oscillator jitter performance is typically impacted by switching noise on its power supply.
Include a dedicated serial port to the I2C or SPI pins of the LMK05028.
–
This allows off-board programming for device bring-up, debug, and diagnostics using TI's USB hardware
interface and software GUI tools.
11 Power Supply Recommendations
11.1 Power Supply Bypassing
图 70 shows two general placements of power supply bypass capacitors on either the back side or the
component side of the PCB. If the capacitors are mounted on the back side, 0402 components can be employed.
For component side mounting, use 0201 body size capacitors to facilitate signal routing. A combination of
component side and back side placement can be used. Keep the connections between the bypass capacitors
and the power supply on the device as short as possible. Ground the other side of the capacitor using a low-
impedance connection to the ground plane.
Back Side
Component Side
(Does not indicate actual location of the LMK05028 supply pins)
图 70. Generalized Placement of Power Supply Bypass Capacitors
版权 © 2018, Texas Instruments Incorporated
81
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
12 Layout
12.1 Layout Guidelines
•
Isolate input, XO, TCXO/OCXO, and output clocks from adjacent clocks with different frequencies and other
nearby dynamic signals.
•
Consider the XO and TCXO/OCXO placement and layout in terms of supply/ground noise and thermal
gradients from adjacent circuitry (for example, power supplies, FPGA, ASIC) as well as system/board-level
vibration and shock. These factors can affect the frequency stability/accuracy and transient performance of
the oscillators.
•
•
Avoid impedance discontinuities on controlled-impedance 50-Ω single-ended (or 100-Ω differential) traces for
clock and dynamic logic signals.
Place bypass capacitors close to the VDD and VDDO pins on the same side as the LMK05028, or directly
below the IC pins on the back side of the PCB. Larger decoupling capacitor values can be placed further
away.
•
•
•
Place external capacitors close to the CAP_x and LFx pins.
If possible, use multiple vias to connect wide supply traces to the respective power islands or planes.
Use at least 7x7 through-hole via pattern to connect the IC ground/thermal pad to the PCB ground planes.
82
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
12.2 Layout Example
图 71. General PCB Ground Layout for Thermal Reliability (8+ Layers Recommended)
12.3 Thermal Reliability
The LMK05028 is a high-performance device. To ensure good electrical and thermal performance, it is
recommended to design a thermally-enhanced interface between the IC ground/thermal pad and the PCB ground
using at least 7x7 through-hole via pattern connected to multiple PCB ground layers like shown in 图 71.
版权 © 2018, Texas Instruments Incorporated
83
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
13 器件和文档支持
13.1 器件支持
13.1.1 时钟架构
时钟架构是一种在线工具,用于选择 LMK05028 部件。该工具同样支持其他 TI 时钟器件的部件选择、环路滤波器
设计和相位噪声仿真。如需时钟架构,请访问 www.ti.com.cn/clockarchitect。
13.1.2 TICS Pro
TICS Pro 是用于 EVM 编程的离线软件工具,也可以用生成寄存器映射,为特定应用的器件配置编程。如需 TICS
Pro,请访问 www.ti.com.cn/tool/cn/TICSPRO-SW。
13.2 文档支持
13.2.1 相关文档
请参阅如下相关文档:
•
•
•
《LMK05028 的 ITU-T G.8262 合规性测试结果》(SNAA315)
《LMK05028 编程指南》(SNAU233)
《LMK05028EVM 用户指南》(SNAU223)
13.3 接收文档更新通知
如需接收文档更新通知,请导航至 TI.com.cn 上的器件产品文件夹。单击右上角的通知我 进行注册,即可每周接收
产品信息更改摘要。有关更改的详细信息,请查看任何已修订文档中包含的修订历史记录。
13.4 社区资源
下列链接提供到 TI 社区资源的连接。链接的内容由各个分销商“按照原样”提供。这些内容并不构成 TI 技术规范,
并且不一定反映 TI 的观点;请参阅 TI 的 《使用条款》。
TI E2E™ 在线社区 TI 的工程师对工程师 (E2E) 社区。此社区的创建目的在于促进工程师之间的协作。在
e2e.ti.com 中,您可以咨询问题、分享知识、拓展思路并与同行工程师一道帮助解决问题。
设计支持
TI 参考设计支持 可帮助您快速查找有帮助的 E2E 论坛、设计支持工具以及技术支持的联系信息。
13.5 商标
E2E is a trademark of Texas Instruments.
All other trademarks are the property of their respective owners.
13.6 静电放电警告
ESD 可能会损坏该集成电路。德州仪器 (TI) 建议通过适当的预防措施处理所有集成电路。如果不遵守正确的处理措施和安装程序 , 可
能会损坏集成电路。
ESD 的损坏小至导致微小的性能降级 , 大至整个器件故障。 精密的集成电路可能更容易受到损坏 , 这是因为非常细微的参数更改都可
能会导致器件与其发布的规格不相符。
13.7 术语表
SLYZ022 — TI 术语表。
这份术语表列出并解释术语、缩写和定义。
84
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
14 机械、封装和可订购信息
以下页面包含机械、封装和可订购信息。这些信息是指定器件的最新可用数据。数据如有变更,恕不另行通知,且
不会对此文档进行修订。如需获取此数据表的浏览器版本,请查阅左侧的导航栏。
版权 © 2018, Texas Instruments Incorporated
85
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
PACKAGE OUTLINE
RGC0064J
VQFN - 1 mm max height
S
C
A
L
E
1
.
5
0
0
PLASTIC QUAD FLATPACK - NO LEAD
A
9.15
8.85
B
PIN 1 INDEX AREA
9.15
8.85
1.0
0.8
C
SEATING PLANE
0.08 C
0.05
0.00
2X 7.5
SYMM
EXPOSED
THERMAL PAD
17
(0.2) TYP
32
33
16
SYMM
2X 7.5
65
7.7 0.1
1
48
0.30
64X
60X 0.5
64
0.18
49
0.1
C A B
PIN 1 ID
0.5
0.3
64X
0.05
4219012/A 05/2018
NOTES:
1. All linear dimensions are in millimeters. Any dimensions in parenthesis are for reference only. Dimensioning and tolerancing
per ASME Y14.5M.
2. This drawing is subject to change without notice.
3. The package thermal pad must be soldered to the printed circuit board for thermal and mechanical performance.
www.ti.com
86
版权 © 2018, Texas Instruments Incorporated
LMK05028
www.ti.com.cn
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
EXAMPLE BOARD LAYOUT
RGC0064J
VQFN - 1 mm max height
PLASTIC QUAD FLATPACK - NO LEAD
(
7.7)
SEE SOLDER MASK
DETAIL
49
SYMM
64X (0.6)
64
64X (0.24)
1
48
60X (0.5)
2X (3.6)
(
0.2) TYP
VIA
4X (1.19)
65
SYMM
(8.8)
(R0.05) TYP
16
33
32
17
4X (1.19)
2X (3.6)
(8.8)
LAND PATTERN EXAMPLE
EXPOSED METAL SHOWN
SCALE: 10X
0.07 MIN
ALL AROUND
0.07 MAX
ALL AROUND
METAL UNDER
SOLDER MASK
METAL EDGE
EXPOSED METAL
SOLDER MASK
OPENING
EXPOSED
METAL
SOLDER MASK
OPENING
NON SOLDER MASK
DEFINED
(PREFERRED)
SOLDER MASK DEFINED
SOLDER MASK DETAILS
4219012/A 05/2018
NOTES: (continued)
4. This package is designed to be soldered to a thermal pad on the board. For more information, see Texas Instruments literature
number SLUA271 (www.ti.com/lit/slua271).
5. Vias are optional depending on application, refer to device data sheet. If any vias are implemented, refer to their locations shown
on this view. It is recommended that vias under paste be filled, plugged or tented.
www.ti.com
版权 © 2018, Texas Instruments Incorporated
87
LMK05028
ZHCSHN9A –FEBRUARY 2018–REVISED APRIL 2018
www.ti.com.cn
EXAMPLE STENCIL DESIGN
RGC0064J
VQFN - 1 mm max height
PLASTIC QUAD FLATPACK - NO LEAD
5X
(1.19)
(0.595) TYP
64X (0.6)
64
49
64X (0.24)
1
48
60X (0.5)
(R0.05) TYP
5X (1.19)
65
(0.595) TYP
SYMM
(8.8)
33
16
17
32
SYMM
(8.8)
36X ( 0.99)
SOLDER PASTE EXAMPLE
BASED ON 0.125 MM THICK STENCIL
SCALE: 10X
EXPOSED PAD 65
60% PRINTED SOLDER COVERAGE BY AREA UNDER PACKAGE
4219012/A 05/2018
NOTES: (continued)
6. Laser cutting apertures with trapezoidal walls and rounded corners may offer better paste release. IPC-7525 may have alternate
design recommendations.
www.ti.com
88
版权 © 2018, Texas Instruments Incorporated
重要声明和免责声明
TI 均以“原样”提供技术性及可靠性数据(包括数据表)、设计资源(包括参考设计)、应用或其他设计建议、网络工具、安全信息和其他资
源,不保证其中不含任何瑕疵,且不做任何明示或暗示的担保,包括但不限于对适销性、适合某特定用途或不侵犯任何第三方知识产权的暗示
担保。
所述资源可供专业开发人员应用TI 产品进行设计使用。您将对以下行为独自承担全部责任:(1) 针对您的应用选择合适的TI 产品;(2) 设计、
验证并测试您的应用;(3) 确保您的应用满足相应标准以及任何其他安全、安保或其他要求。所述资源如有变更,恕不另行通知。TI 对您使用
所述资源的授权仅限于开发资源所涉及TI 产品的相关应用。除此之外不得复制或展示所述资源,也不提供其它TI或任何第三方的知识产权授权
许可。如因使用所述资源而产生任何索赔、赔偿、成本、损失及债务等,TI对此概不负责,并且您须赔偿由此对TI 及其代表造成的损害。
TI 所提供产品均受TI 的销售条款 (http://www.ti.com.cn/zh-cn/legal/termsofsale.html) 以及ti.com.cn上或随附TI产品提供的其他可适用条款的约
束。TI提供所述资源并不扩展或以其他方式更改TI 针对TI 产品所发布的可适用的担保范围或担保免责声明。IMPORTANT NOTICE
邮寄地址:上海市浦东新区世纪大道 1568 号中建大厦 32 楼,邮政编码:200122
Copyright © 2019 德州仪器半导体技术(上海)有限公司
PACKAGE OPTION ADDENDUM
www.ti.com
10-Dec-2020
PACKAGING INFORMATION
Orderable Device
Status Package Type Package Pins Package
Eco Plan
Lead finish/
Ball material
MSL Peak Temp
Op Temp (°C)
Device Marking
Samples
Drawing
Qty
(1)
(2)
(3)
(4/5)
(6)
LMK05028RGCR
LMK05028RGCT
ACTIVE
ACTIVE
VQFN
VQFN
RGC
RGC
64
64
2500 RoHS & Green
250 RoHS & Green
NIPDAU
Level-3-260C-168 HR
Level-3-260C-168 HR
-40 to 85
-40 to 85
K5028
K5028
NIPDAU
(1) The marketing status values are defined as follows:
ACTIVE: Product device recommended for new designs.
LIFEBUY: TI has announced that the device will be discontinued, and a lifetime-buy period is in effect.
NRND: Not recommended for new designs. Device is in production to support existing customers, but TI does not recommend using this part in a new design.
PREVIEW: Device has been announced but is not in production. Samples may or may not be available.
OBSOLETE: TI has discontinued the production of the device.
(2) RoHS: TI defines "RoHS" to mean semiconductor products that are compliant with the current EU RoHS requirements for all 10 RoHS substances, including the requirement that RoHS substance
do not exceed 0.1% by weight in homogeneous materials. Where designed to be soldered at high temperatures, "RoHS" products are suitable for use in specified lead-free processes. TI may
reference these types of products as "Pb-Free".
RoHS Exempt: TI defines "RoHS Exempt" to mean products that contain lead but are compliant with EU RoHS pursuant to a specific EU RoHS exemption.
Green: TI defines "Green" to mean the content of Chlorine (Cl) and Bromine (Br) based flame retardants meet JS709B low halogen requirements of <=1000ppm threshold. Antimony trioxide based
flame retardants must also meet the <=1000ppm threshold requirement.
(3) MSL, Peak Temp. - The Moisture Sensitivity Level rating according to the JEDEC industry standard classifications, and peak solder temperature.
(4) There may be additional marking, which relates to the logo, the lot trace code information, or the environmental category on the device.
(5) Multiple Device Markings will be inside parentheses. Only one Device Marking contained in parentheses and separated by a "~" will appear on a device. If a line is indented then it is a continuation
of the previous line and the two combined represent the entire Device Marking for that device.
(6)
Lead finish/Ball material - Orderable Devices may have multiple material finish options. Finish options are separated by a vertical ruled line. Lead finish/Ball material values may wrap to two
lines if the finish value exceeds the maximum column width.
Important Information and Disclaimer:The information provided on this page represents TI's knowledge and belief as of the date that it is provided. TI bases its knowledge and belief on information
provided by third parties, and makes no representation or warranty as to the accuracy of such information. Efforts are underway to better integrate information from third parties. TI has taken and
continues to take reasonable steps to provide representative and accurate information but may not have conducted destructive testing or chemical analysis on incoming materials and chemicals.
TI and TI suppliers consider certain information to be proprietary, and thus CAS numbers and other limited information may not be available for release.
In no event shall TI's liability arising out of such information exceed the total purchase price of the TI part(s) at issue in this document sold by TI to Customer on an annual basis.
Addendum-Page 1
PACKAGE OPTION ADDENDUM
www.ti.com
10-Dec-2020
Addendum-Page 2
PACKAGE MATERIALS INFORMATION
www.ti.com
28-Aug-2018
TAPE AND REEL INFORMATION
*All dimensions are nominal
Device
Package Package Pins
Type Drawing
SPQ
Reel
Reel
A0
B0
K0
P1
W
Pin1
Diameter Width (mm) (mm) (mm) (mm) (mm) Quadrant
(mm) W1 (mm)
LMK05028RGCR
LMK05028RGCT
VQFN
VQFN
RGC
RGC
64
64
2500
250
330.0
180.0
16.4
16.4
9.3
9.3
9.3
9.3
1.1
1.1
12.0
12.0
16.0
16.0
Q2
Q2
Pack Materials-Page 1
PACKAGE MATERIALS INFORMATION
www.ti.com
28-Aug-2018
*All dimensions are nominal
Device
Package Type Package Drawing Pins
SPQ
Length (mm) Width (mm) Height (mm)
LMK05028RGCR
LMK05028RGCT
VQFN
VQFN
RGC
RGC
64
64
2500
250
367.0
210.0
367.0
185.0
38.0
35.0
Pack Materials-Page 2
GENERIC PACKAGE VIEW
RGC 64
9 x 9, 0.5 mm pitch
VQFN - 1 mm max height
PLASTIC QUAD FLATPACK - NO LEAD
Images above are just a representation of the package family, actual package may vary.
Refer to the product data sheet for package details.
4224597/A
www.ti.com
重要声明和免责声明
TI 均以“原样”提供技术性及可靠性数据(包括数据表)、设计资源(包括参考设计)、应用或其他设计建议、网络工具、安全信息和其他资
源,不保证其中不含任何瑕疵,且不做任何明示或暗示的担保,包括但不限于对适销性、适合某特定用途或不侵犯任何第三方知识产权的暗示
担保。
所述资源可供专业开发人员应用TI 产品进行设计使用。您将对以下行为独自承担全部责任:(1) 针对您的应用选择合适的TI 产品;(2) 设计、
验证并测试您的应用;(3) 确保您的应用满足相应标准以及任何其他安全、安保或其他要求。所述资源如有变更,恕不另行通知。TI 对您使用
所述资源的授权仅限于开发资源所涉及TI 产品的相关应用。除此之外不得复制或展示所述资源,也不提供其它TI或任何第三方的知识产权授权
许可。如因使用所述资源而产生任何索赔、赔偿、成本、损失及债务等,TI对此概不负责,并且您须赔偿由此对TI 及其代表造成的损害。
TI 所提供产品均受TI 的销售条款 (http://www.ti.com.cn/zh-cn/legal/termsofsale.html) 以及ti.com.cn上或随附TI产品提供的其他可适用条款的约
束。TI提供所述资源并不扩展或以其他方式更改TI 针对TI 产品所发布的可适用的担保范围或担保免责声明。IMPORTANT NOTICE
邮寄地址:上海市浦东新区世纪大道 1568 号中建大厦 32 楼,邮政编码:200122
Copyright © 2020 德州仪器半导体技术(上海)有限公司
相关型号:
SI9130DB
5- and 3.3-V Step-Down Synchronous ConvertersWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9135LG-T1
SMBus Multi-Output Power-Supply ControllerWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9135LG-T1-E3
SMBus Multi-Output Power-Supply ControllerWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9135_11
SMBus Multi-Output Power-Supply ControllerWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9136_11
Multi-Output Power-Supply ControllerWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9130CG-T1-E3
Pin-Programmable Dual Controller - Portable PCsWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9130LG-T1-E3
Pin-Programmable Dual Controller - Portable PCsWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9130_11
Pin-Programmable Dual Controller - Portable PCsWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9137
Multi-Output, Sequence Selectable Power-Supply Controller for Mobile ApplicationsWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9137DB
Multi-Output, Sequence Selectable Power-Supply Controller for Mobile ApplicationsWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9137LG
Multi-Output, Sequence Selectable Power-Supply Controller for Mobile ApplicationsWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
SI9122E
500-kHz Half-Bridge DC/DC Controller with Integrated Secondary Synchronous Rectification DriversWarning: Undefined variable $rtag in /www/wwwroot/website_ic37/www.icpdf.com/pdf/pdf/index.php on line 217
-
VISHAY
©2020 ICPDF网 联系我们和版权申明