LTC6812-1 [ADI]
Current, Voltage, and Charge Monitor for High Voltage Battery Packs;型号: | LTC6812-1 |
厂家: | ADI |
描述: | Current, Voltage, and Charge Monitor for High Voltage Battery Packs 电池 |
文件: | 总80页 (文件大小:3353K) |
中文: | 中文翻译 | 下载: | 下载PDF数据表文档文件 |
LTC2949
Current, Voltage, and Charge Monitor for
High Voltage Battery Packs
FEATURES
DESCRIPTION
The LTC®2949 is a high precision current, voltage, tem-
perature,chargeandenergymeterforelectricalandhybrid
vehicles and other isolated current sense applications. It
infers charge and energy flowing in and out of the battery
pack by monitoring simultaneously the voltage drop over
up to two sense resistors and the battery pack voltage.
n
Measures Battery Stack Voltage, Current and Power
Indicates Accumulated Battery Charge and Energy
20-Bit Current Measurement with <1μV Offset
Built-In Isolated isoSPI™ or SPI Interface
LTC68xx/ADBMS68xx Compatible, Supports
Synchronous Measurements with Cell Monitors
Up to 12 Buffered Voltage Measurement Inputs
Up to 5 GPIOs, Configurable to Drive Ground, Supply
or Toggling at 400kHz
n
n
n
n
n
n
Low offset ΔΣ ADCs ensure accurate measurement of
voltage and current with insignificant power loss. Con-
tinuous integration of current and power ensures lossless
tracking of charge and energy delivered or received by
the battery pack.
n
n
n
n
n
n
n
n
n
n
High or Low Side Current Sense
0.3% Current and Voltage Accuracy
1% Energy and Charge Accuracy
True Average ADCs
The built-in serial interface can be configured to support
isolated isoSPI communication to the host or as SPI
interface.
2
I C EEPROM Interface to Store Board Calibration Factors
Threshold Registers for all Measured Quantities
Engineered for ISO26262 Compliant Systems
Open Wire Detection on Input Pins
The LTC2949 features 12 internally buffered high imped-
ance inputs (V1 to V12) for measuring voltages from
external sensors or resistor dividers allowing to measure
temperatures, HV-Link voltages, chassis isolation and
supervise contactor states. LTC2949 has up to five pro-
grammable digital outputs which can be set to ground,
supply or toggling at 400kHz.
Available in 48-Lead LQFP Package
AEC-Q100 Qualified for Automotive Applications
APPLICATIONS
n
Electric and Hybrid Vehicles
n
Isolated Current Sensing
Programmable threshold and tracking registers reduce
digital traffic to the host.
n
Backup Battery Systems
High Power Portable Equipment
n
All registered trademarks and trademarks are the property of their respective owners.
TYPICAL APPLICATION
Electric Vehicle Battery Meter
ꢂ
ꢄ
ꢀꢁ
ꢀAꢁ
ꢀꢁꢁꢂ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃAꢄꢀꢂꢅ
RꢆꢁꢀꢁꢄAꢅꢇꢆ
ꢀꢁꢂꢃꢄꢅ ꢆ
ꢀꢁꢀꢂ
Aꢀꢁꢁ
ꢀꢁꢂꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄꢅ
ꢀꢁꢂꢃꢄꢅꢄ
ꢀꢁꢂ
ꢀꢁꢂꢃ ꢄꢅꢀꢆAꢇꢈ
ꢀꢁꢀꢂ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁꢂꢃꢄꢅ A
ꢀꢁꢂꢃ
ꢀꢁꢂ
ꢀꢀꢁ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄ
ꢀꢁ
ꢀꢁꢂꢂ
ꢃꢄꢅꢆꢇꢄRꢈ
ꢀRꢁꢂ
ꢀAꢁꢂ
ꢀꢁꢁꢂ
ꢀꢀꢁ
ꢀꢁꢂꢃR
ꢃꢄꢃRꢅꢆ
ꢀꢁAꢂꢃ
ꢄꢅꢆꢁAꢇꢈ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂꢃꢄꢅ ꢆ
ꢀꢁ
ꢀꢁꢂꢃ ꢄꢅꢀꢆAꢇꢈ
ꢀAꢁꢂ
ꢀꢁ
ꢀAꢁꢂꢃR
ꢀꢁꢂ
ꢀꢁꢂꢃꢄꢅꢆ
ꢀꢁꢂ
ꢀꢁRRꢂꢃꢄ
ꢀꢅARꢆꢂ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ
ꢀꢁꢂꢃꢄꢅ
ꢀꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄꢅ ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢁꢂ
ꢀꢁꢂꢃꢄꢅ A
ꢀꢁꢂ
ꢀꢁꢂꢃAꢄꢅꢆ ꢇꢂꢈꢈꢉꢊꢀꢇAꢄꢀꢂꢊ
ꢄ
50µΩ
ꢀꢁ
ꢀꢁꢂꢃ
ꢂ
ꢀꢁ
ꢀAꢁ
ꢀꢁꢂꢁ ꢃAꢄꢅ
Rev A
1
Document Feedback
For more information www.analog.com
LTC2949
TABLE OF CONTENTS
Features............................................................................................................................ 1
Applications ....................................................................................................................... 1
Typical Application ............................................................................................................... 1
Description......................................................................................................................... 1
Absolute Maximum Ratings..................................................................................................... 3
Order Information................................................................................................................. 3
Pin Configuration ................................................................................................................. 3
Electrical Characteristics........................................................................................................ 4
Typical Performance Characteristics .........................................................................................10
Pin Functions.....................................................................................................................11
Pin Functions.....................................................................................................................12
Block Diagram....................................................................................................................13
Operation..........................................................................................................................14
Overview............................................................................................................................................................... 14
Modes of Operation .............................................................................................................................................. 15
Data Acquisition Channels .................................................................................................................................... 16
Power Measurement............................................................................................................................................. 19
Charge, Energy and Time...................................................................................................................................... 20
Overcurrent Comparators ..................................................................................................................................... 21
SERIAL INTERFACES ............................................................................................................22
Serial Interfaces Overview .................................................................................................................................... 22
4-Wire Serial Peripheral Interface (SPI) Physical Layer........................................................................................ 22
2-Wire Isolated Interface (isoSPI) Physical Layer................................................................................................. 23
Data Link Layer..................................................................................................................................................... 28
Network Layer....................................................................................................................................................... 28
REGISTER MAP...................................................................................................................41
REGISTER Description ..........................................................................................................42
MEMORY MAP AND Paging Mechanism............................................................................................................... 42
Register Map PAGE0............................................................................................................................................. 44
Register Map PAGE1 ............................................................................................................................................ 62
Application Information.........................................................................................................68
Temperature Measurement .................................................................................................................................. 68
Sense Resistor Temperature Compensation ......................................................................................................... 69
Current and Voltage Input Filtering ....................................................................................................................... 74
Powering the LTC2949 ......................................................................................................................................... 75
Package Description ............................................................................................................78
Revision History .................................................................................................................79
Typical Application ..............................................................................................................80
Related Parts.....................................................................................................................80
Rev A
2
For more information www.analog.com
LTC2949
ABSOLUTE MAXIMUM RATINGS
(Notes 1, 2)
Supply Pins
PIN CONFIGURATION
ꢔꢚꢒ ꢊꢖꢘꢝ
AVCC to AGND........................................ –0.3V to 14.5V
DVCC to DGND........................................ –0.3V to 14.5V
AVCC to DVCC........................................... –0.1V to 0.1V
DGND to AGND......................................... –0.1V to 0.1V
Analog Pins
ꢊꢀ
ꢊꢁ
ꢀ
ꢁ
ꢂ
ꢃ
ꢄ
ꢅ
ꢆ
ꢇ
ꢈ
ꢂꢅ ꢋꢌꢍ
ꢂꢄ ꢊRꢘꢗ
ꢊꢂ
ꢂꢃ ꢍꢏꢙꢚ
ꢊꢃ
ꢊꢄ
ꢋꢌꢍ
ꢊꢅ
ꢊꢆ
ꢊꢇ
ꢂꢂ ꢍꢏꢙꢖ
I1P, I1M, I2P, I2M ......................... –0.3V to V
+ 0.3V
AVCC
ꢂꢁ ꢋꢌꢍ
I1P to I1M, I2P to I2M .............................................. 1V
ꢃꢈ
ꢓꢌꢋ
ꢂꢀ ꢋꢌꢍ
ꢂꢉ ꢍꢎꢐ ꢛꢖꢕꢜ
ꢁꢈ ꢎꢍꢙ ꢛꢖꢒꢜ
ꢁꢇ ꢎꢋꢚ ꢛꢖꢐꢖAꢎꢜ
ꢁꢆ ꢎꢋꢖ ꢛꢖꢍꢕꢒꢜ
ꢁꢅ ꢖꢚꢊꢍꢍ
ꢁꢄ ꢐꢑꢒꢁ
VBATP, VBATM ............................. –0.3V to V
V1-V12......................................... –0.3V to V
+ 0.3V
+ 0.3V
AVCC
AVCC
ꢊꢈ ꢀꢉ
ꢊꢀꢉ ꢀꢀ
ꢊꢀꢀ ꢀꢁ
CLKO, DNC.........................................................(Note 3)
Digital Input/Output Pins
IOVCC, CLKI, CSB(IM), SCK(IP)................... –0.3V to 5V
SDI (ICMP), SDO (IBIAS)............................. –0.3V to 5V
SDA, SCL ................................................ –0.3V to 2.75V
Current In/Out of Pins
IP, IM.................................................................... 30mA
SDO (IBIAS) .............................................. –1mA to 9mA
V8-V12 ................................................................... 2mA
VREF (Note 4) ....................................................... 2mA
BYP1 (Note 4)......................................... –10mA to 0mA
BYP2 (Note 4)......................................... –10mA to 0mA
Operating Ambient Temperature Range
ꢏꢞꢘ ꢒAꢍꢙAꢓꢘ
ꢃꢇꢟꢏꢘAꢋ ꢛꢆꢠꢠ × ꢆꢠꢠꢜ ꢒꢏAꢎꢔꢖꢍ ꢏꢡꢗꢒ
ꢋꢌꢍꢢ ꢋꢚ ꢌꢚꢔ ꢍꢚꢌꢌꢘꢍꢔ
ꢤ ꢀꢄꢉꢥꢍꢦ θ ꢤ ꢁꢉꢧꢃꢅꢥꢍꢨꢝꢦ θ ꢤ ꢂꢧꢅꢇꢥꢍꢨꢝ
ꢣA ꢣꢍ
ꢘꢞꢒꢚꢎꢘꢋ ꢒAꢋ ꢛꢒꢖꢌ ꢃꢈꢜꢦ ꢍꢚꢌꢌꢘꢍꢔ ꢔꢚ Aꢓꢌꢋ
ꢔ
ꢣꢕAꢞ
LTC2949I .................................................–40°C to 85°C
LTC2949H.............................................. –40°C to 125°C
Storage Temperature Range…............... –65°C to 150°C
ORDER INFORMATION
AUTOMOTIVE PRODUCTS**
TRAY (250PC)
TAPE AND REEL (2000PC) PART MARKING*
LTC2949ILXE#3ZZTRPBF LTC2949LXE
LTC2949HLXE#3ZZTRPBF LTC2949LXE
PACKAGE DESCRIPTION
48-LEAD PLASTIC eLQFP
48-LEAD PLASTIC eLQFP
MSL RATING TEMPERATURE RANGE
LTC2949ILXE#3ZZPBF
LTC2949HLXE#3ZZPBF
3
3
–40°C to 85°C
–40°C to 125°C
Contact the factory for parts specified with wider operating temperature ranges. *The temperature grade is identified by a label on the shipping container.
Tape and reel specifications. Some packages are available in 500 unit reels through designated sales channels with #TRMPBF suffix.
**Versions of this part are available with controlled manufacturing to support the quality and reliability requirements of automotive applications. These
models are designated with a #3ZZ suffix. Only the automotive grade products shown are available for use in automotive applications. Contact your local
Analog Devices account representative for specific product ordering information and to obtain the specific Automotive Reliability reports for these models.
Rev A
3
For more information www.analog.com
LTC2949
ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating
temperature range, otherwise specifications are at TA = 25°C.
SYMBOL
PARAMETER
CONDITIONS
MIN
TYP
MAX
UNITS
Power Supply
l
l
l
l
V
V
V
Analog Supply Voltage
4.5
4.5
14
14
V
V
AVCC
DVCC
UVLO
Digital Supply Voltage
Supply Undervoltage Lockout Threshold
Average Supply Current into AVCC and DVCC
V
, V
Falling
4.5
20
V
AVCC DVCC
I
CC
Core State: STANDBY or MEASURE
Core State: SLEEP
16
9
mA
µA
µA
mA
mA
mA
mA
V
15
l
l
l
l
l
l
l
l
l
l
l
l
Core State: SLEEP
150
5.8
7.8
3
Additional DVCC Supply Current if isoSPI in
READY/ACTIVE States
R
+ R = 2k
READY
ACTIVE
READY
ACTIVE
4.8
6.1
2.1
2.5
2.5
B1
B1
B2
Note: ACTIVE State Current Assumes t = 1μs,
(Note 5)
CLK
R
+ R = 20k
B2
3.5
2.75
0
V
V
BYP1 Regulated Output Voltage
BYP1 Load Current
2.25
–10
–15
BYP1
mA
mV
V
BYP1 Load Regulation Error
BYP1 Undervoltage Lockout Threshold
BYP2 Regulation Output Voltage
BYP2 Load Current
I
I
= –10mA
= –10mA
0
LOAD
LOAD
2.25
3.6
0
3
3.25
170
V
BYP2
–10
–60
mA
mV
°C
BYP2 Load Regulation Error
Thermal Shutdown Temperature
Current Sense ADC
Resolution (No Missing Codes)
l
l
l
Slow Mode Filtered (Note 7)
Slow Mode (Note 7)
20
18
15
Bit
Bit
Bit
mV
mV
V
Fast Mode (Note 7)
Full-Scale Differential Input Voltage
Differential Input Voltage Range
Pin Voltage of I1P, I1M, I2P, I2M
Current Sense Quantization Step
V
V
-V , V -V
124
I1P I1M I2P I2M
l
l
VDIF
-V , V -V
I1P I1M I2P I2M
110
I
–0.11
V
+0.11
AVCC
Slow Mode Filtered
Slow Mode
237.5
950
nV
nV
µV
nA
Fast Mode
7.60371
l
CFPx
Input Leakage Current at CF1P, CF1M, I1P,
I1M, CF2P, CF2M, I2P, I2M
Core State = SLEEP/STANDBY
60
Differential Input Current from CF1P to CF1M,
CF2P to CF2M
Core State: MEASURE; Pin Voltages: 0V ≤
VDIF /100kΩ
µA
I
V
, V
, V , V , V
, V
,
CF1P CF1M I1P I1M CF2P CF2M
V
, V ≤ V
I2P I2M
AVCC
Noise
Slow Mode Filtered
Slow Mode
160
320
3
nV
nV
µV
RMS
RMS
Fast Mode
RMS
Gain Error
|VDIF | ≤ 110mV
0.15
0.3
1
%
I
l
l
l
l
%
µV
µV
%
Offset Voltage
IADCx, IxP, IxM = 0V
Slow Mode
Fast Mode
0
0
V
|
= V
= 5V
2
AVCC
DVCC
|
Total Unadjusted Error
VDIF ≥ 25mV
0.3
I
Rev A
4
For more information www.analog.com
LTC2949
ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating
temperature range, otherwise specifications are at TA = 25°C.
SYMBOL
PARAMETER
CONDITIONS
MIN
TYP
MAX
0.8211
4.8
UNITS
dB
l
Input Voltage Common Mode Rejection at DC
Input Sampling Frequency
Conversion Time
100
10.48
400
MHz
ms
Slow Mode Filtered
Slow Mode
100
ms
l
Fast Mode
0.782
ms
Voltage Measurement by Power ADC
l
l
Resolution (No Missing Codes)
Slow Mode (Note 7)
Fast Mode (Note 7)
18
15
Bit
Bit
V
VFS
Full-Scale Differential Input Voltage
Differential Input Voltage Range
Pin Voltage of VBATP, VBATM
V
V
V
V
– V
– V
6.14
V
VBATP
VBATP
VBATM
VBATM
l
l
l
VDIF
V
V
≥ 5V
< 5V
–0.1
–0.1
V
+0.1
AVCC
V
AVCC
AVCC
V -1.5
AVCC
V
LSB
Differential Input Voltage Quantization Step
Slow Mode
46.875
µV
µV
nA
MΩ
MΩ
%
V
Fast Mode
375.183
l
Input Leakage Current
Core State: SLEEP/STANDBY
Core State: MEASURE; Pin Voltages 0V
60
Differential Input Resistance
50
20
≤ V
, V
≤ V
BATP BATM AVCC
l
l
l
l
l
Gain Error
0.3
3
Offset
V
BATP
= V
= 0V
0
LSB
V
BATM
Voltage Total Unadjusted Error
Input Voltage Common Mode Rejection at DC
Noise
1V ≤ |VDIF | ≤ 4.8V
0.4
%
V
80
dB
Slow Mode (Note 7)
Fast Mode (Note 7)
3
µV
µV
RMS
RMS
30
Input Sampling Rate
Conversion Time
5.24
100
0.782
MHz
ms
Slow Mode
Fast Mode
l
0.8211
ms
Power Measurement by Power ADC
Resolution (No Missing Codes)
l
l
Slow Mode (Note 7)
Fast Mode (Note 7)
18
11
Bit
Bit
2
FS
Full-Scale Power
FS = VFS • VFS /R
/VDIV
v
0.76504
5.8368
1
[V /Ω]
P
P
V
I
ISENSE
17
2
LSB
Power Quantization Step
Power Offset
LSB = FS /2
µ[V /Ω]
P
P
P
POS
TUE
VDIF = 0
LSB
P
1
l
Power Total Unadjusted Error
1V ≤ |VDIF | ≤ 4.8V, 25mV ≤ |VDIF | ≤
0.9
%
P
V
I
110mV
RMS Noise
Slow Mode; VBATP – VBATM = 4.8V (Note 7)
Slow Mode; VBATP – VBATM = 0V (Note 7)
0.3
0.03
5.24
5.24
100
LSB
P
LSB
P
Input Sampling Frequency
Power Modulation Frequency
Conversion Time
MHz
MHz
ms
Slow Mode
Fast Mode
l
0.782
0.8211
ms
Rev A
5
For more information www.analog.com
LTC2949
ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating
temperature range, otherwise specifications are at TA = 25°C.
SYMBOL
Energy Measurement
TUE Energy Total Unadjusted Error
PARAMETER
CONDITIONS
MIN
TYP
MAX
UNITS
l
l
1V ≤ |VDIF | ≤ 4.8V, 25mV ≤ |VDIF | ≤
0.9
%
E
V
I
110mV, Ideal External Clock or 4MHz
Crystal
1V ≤ |VDIF | ≤ 4.8V, 25mV ≤ |VDIF | ≤
1.9
%
V
I
110mV, Internal Clock
Charge Measurement
TUE Charge Total Unadjusted Error
l
l
1V ≤ |VDIF | ≤ 4.8V, 25mV ≤ |VDIF | ≤
0.4
1.4
%
%
C
V
I
110mV, Ideal External Clock or 4MHz
Crystal
1V ≤ |VDIF | ≤ 4.8V, 25mV ≤ |VDIF | ≤
V
I
110mV, Internal Clock
Voltage Measurement by AUXILIARY ADC
l
Resolution (No Missing Codes)
(Note 7)
15
Bit
V
VFS
Full-Scale Differential Input Voltage
Differential Input Voltage Range
V
V
V
V
– V
– V
, V
– V
– V
6.14
AUX
VBATP
VBATP
VBATM MUXP
MUXN
l
l
l
VDIF
, V
VBATM MUXP
4.8
V
AUX
MUXN
Pin Voltage of VBATP, VBATM, V1 – V12, CF1P,
CF1M, CF2P, CF2M
≥ 5V
< 5V
–0.1
–0.1
V
V
+0.1
AVCC
V
AVCC
AVCC
–1.5
AVCC
V
LSB
Differential Voltage Quantization Step
Slow Mode
Fast Mode
375
375.183
1
µV
µV
nA
MΩ
%
AUX
l
l
l
l
l
l
Input Leakage Current
Differential Input Resistance
Gain Error
60
40
80
13
|VDIF | ≤ 4.8V
0.3
1
AUX
Offset
V
BATP
= V
= 0V
0
LSB
V
BATM
Total Unadjusted Error
Input Voltage Common Mode Rejection at DC
Sampling Rate
1V ≤ |VDIF | ≤ 4.8V
0.4
%
dB
V
5.24
MHz
ms
l
l
Conversion Time
0.782
0.8211
On-Die Temperature Measurement by AUXILIARY ADC
Resolution (No Missing Codes)
(Note 7)
Bit
K
Full-Scale Temperature
819.2
0.2
3
ΔT
Temperature Quantization Step
Total Unadjusted Error
Conversion Time
K
LSB
K
13.1
20
ms
K/W
Self-Heating
Supply Voltage Measurement by AUXILIARY ADC
Resolution (No Missing Codes)
Full-Scale Differential Input Voltage
A/DVCC Measurement Quantization Step
Total Unadjusted Error
l
l
l
(Note 7)
14
Bit
V
18.43
2.2583
2
mV
%
5
Conversion Time
6.55
ms
AUX MUX
Signal Range
–0.1
V
+0.1
AVCC
V
Rev A
6
For more information www.analog.com
LTC2949
ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating
temperature range, otherwise specifications are at TA = 25°C.
SYMBOL
PARAMETER
CONDITIONS
MIN
TYP
–250
250
MAX
UNITS
µA
l
l
Pull-Up Current Source
Pull-Down Current Source
Pin Voltage < V
− 3.0V
–150
AVCC
Pin Voltage > 2.5V
200
µA
Reference Voltages
VREF
Reference Pin Voltage
3
V
%
l
VREF Error
1
VREF Temperature Coefficient
VREF Long Term Drift
7
ppm/K
ppm/√kHr
mV
80
l
l
VREF Load Regulation Error
Internal Redundant Reference Voltage
VREF2 Error
–0.5mA ≤ I
≤ 0.5mA
–5
0
LOAD
VREF2
2.39
V
0.85
%
VREF2 Temperature Coefficient
VREF2 Long Term Drift
10
80
ppm/K
ppm/√kHr
Overcurrent Comparator
Pin Voltages I1P, I1M I2P, I2M
l
l
l
l
l
l
–0.11
V
+0.11
V
mV
mV
mV
µs
AVCC
Total Unadjusted Error
|Vthr| ≤ 103mV
|Vthr| > 103mV
|Vthr| = 310mV
5
10
20
Programmable Deglitch Time Delay
T
T
20, 80, 320µs
T
T
−10
T
T
+37
degl
degl
degl
degl
degl
degl
1280µs
−26
+56
µs
Digital Input CLKI
l
l
l
l
Logic Input Threshold
Input Current DC Current
Input Capacitance
0.4
2
V
μA
1
(Note 7)
10
25
pF
External Clock Frequency
General Purpose Outputs GPIOx
0.1
MHz
l
l
Low Level Output Voltage at GPIOx
I
I
= 0.5mA
0.4
V
V
GPIOx
GPIOx
High Level Output Voltage at GPIOx
= –0.25mA
V
DVCC
–0.5
l
GPIOx Toggling Frequency
370
400
430
kHz
SPI Interface DC Specification IOVCC, CSB, SCK, SDI, SDO
l
l
l
V
IOVCC
SPI Mode IOVCC Operating Voltage
Pin Voltages CSB, SCK, SDI, SDO
Logic Input Threshold (CSB, SCK, SDI)
1.8
4.5
V
V
V
V
IOVCC
0.3 •
IOVCC
0.7 •
V
IOVCC
V
l
l
l
DC Input Current (CSB, SCK, SDI)
Input Capacitance (CSB, SCK, SDI)
Low Level Output Voltage at SDO
1
µA
pF
V
(Note 7)
≥ 3.3V, I
10
V
= 3mA, 1.8V ≤ V
IOVCC
0.4
IOVCC
SDO
≤ 3.3V, I
= 1mA
SDO
SPI Timing Requirements (See Figure 7)
l
l
l
t
t
t
SCK Period
(Note 6)
1
μs
ns
ns
CLK
1
SDI Setup Time Before SCK Rising Edge
SDI Hold Time After SCK Rising Edge
25
25
2
Rev A
7
For more information www.analog.com
LTC2949
ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating
temperature range, otherwise specifications are at TA = 25°C.
SYMBOL
PARAMETER
CONDITIONS
MIN
200
200
0.65
0.8
1
TYP
MAX
UNITS
ns
l
l
l
l
l
l
l
t
t
t
t
t
t
SCK Low
t
t
= t + t ≥ 1μs
3 4
3
4
5
6
7
8
CLK
CLK
SCK High
= t + t ≥ 1μs
ns
3
4
CSB Rising Edge to CSB Falling Edge
SCK Rising Edge to CSB Rising Edge
CSB Falling Edge to SCK Rising Edge
SCK Falling Edge to SDO Valid
μs
(Note 6)
(Note 6)
μs
μs
(Note 9), VIOVCC ≥ 3.3V
(Note 9), VIOVCC < 3.3V
60
ns
150
ns
isoSPI DC Specifications (See Figure 10)
l
l
Voltage at IOVCC to Select isoSPI
0.5
2.1
V
V
IBIAS
Voltage on IBIAS Pin
READY/ACTIVE State
IDLE
1.9
2
0
V
V
l
l
l
l
l
l
l
l
I
Isolated Interface Bias Current
Isolated Interface Current Gain
R
= 2kΩ to 20kΩ
BIAS
0.1
18
17
1
22
24.5
1.4
1.5
1
mA
B
A
V ≤ 1V, I = 1mA
A
20
20
mA/mA
mA/mA
V
IB
B
I = 0.1mA
B
V
V
Transmitter Pulse Amplitude
V = |V – V
A
|
A
IP
IM
Threshold-Setting Voltage on ICMP Pin
Input Leakage Current on ICMP Pin
Leakage Current on IP and IM Pins
Receiver Comparator Threshold Voltage Gain
V
V
= A
• V
ICMP
0.2
0.4
V
ICMP
TCMP
ICMP
TCMP
= 0V to V
µA
BYP2
IDLE State, V or V = 0V to V
1
µA
IP
IM
BYP2
A
V
V
V
= V
/2 to V
– 0.2V,
0.5
0.6
V/V
TCMP
CM
ICMP
BYP2
BYP2
= 0.2V to 1.5V
Receiver Common Mode Bias
IP, IM Not Driving
V
V
–
V
CM
BYP2
/3 –
ICMP
167mV
l
Receiver Input Resistance
Single-Ended to IP, IM
27
35
43
kΩ
isoSPI IDLE/WAKE-UP Specifications (See Figure 3)
l
l
l
l
V
Differential Wake-Up Voltage
Dwell Time at V Before Wake Detection
t
= 240ns
DWELL
200
240
mV
ns
WAKE
DWELL
READY
IDLE
t
t
t
V
= 200mV
WAKE
WAKE
Start-Up Time After Wake Detection
Idle Timeout Duration
10
8
µs
4.3
6.4
ms
isoSPI Pulse Timing Specifications (See Figures 10,11)
l
l
l
l
l
l
l
t
t
t
t
t
t
t
Chip-Select Signal Filter
Chip-Select Valid Pulse Window
Data Half-Pulse Width
Data Signal Filter
Receiver
70
220
40
90
270
50
115
330
60
ns
ns
ns
ns
ns
ns
ns
FILT(CS)
WNDW(CS)
1/2PW(D)
FILT(D)
Receiver
Transmitter
Receiver
10
25
35
Data Pulse Inversion Delay
Data Valid Pulses Window
Data Return Delay
Transmitter
Receiver
40
55
69
INV(D)
70
90
110
625
WNDW(D)
485
RTN
2
I C Interface DC Specification (SCL, SDA)
Logic Input Threshold (SDA)
l
l
l
l
0.9
1.6
1
V
μA
pF
V
DC Input Current (SDA)
Input Capacitance (SDA)
(Note 7)
10
0.4
Low Level Output Voltage at SDA, SCL
I = 0.5mA
Rev A
8
For more information www.analog.com
LTC2949
ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating
temperature range, otherwise specifications are at TA = 25°C.
SYMBOL
PARAMETER
CONDITIONS
MIN
TYP
MAX
UNITS
2
I C Interface Timing Specification (SCL, SDA)
l
l
l
l
l
l
f
t
t
t
t
t
Maximum SCL Clock Frequency
SCL Low Period
8
10
kHz
µs
µs
µs
µs
µs
SCL(MAX)
SCLLO
80
80
30
30
30
SDA Low Period
SDALO
Bus Free Time Between STOP/START
Minimum Repeated START Setup Time
BUF(MIN)
SU,STA(MIN)
HD,STA(MIN)
Minimum Hold Time (Repeated) START
Condition
l
l
l
l
l
t
t
t
t
t
Minimum Setup Time for STOP Condition
Minimum Data Setup Time Input
Minimum Data Hold Time Input
Minimum Data Hold Time Output
Data Output Fall Time
30
30
µs
µs
ns
µs
ns
SU,STO(MIN)
SU,DAT(MIN)
HD,DAT(MIN)
HD,DATO
0
30
(Notes 7, 8)
20 + 0.1
OF
• C
B
Digital Core Timings (See Figure 3)
l
t
Core Boot-Up Time from SLEEP or POWER-OFF AVCC/DVCC Pins at Minimum Operating
100
ms
BOOT
to STANDBY
Voltage
l
l
l
l
l
t
t
t
t
t
Core STANDBY Cycle Time
(Note 10)
17
20
110
130
40
ms
ms
ms
ms
s
IDLE_CORE
Core MEASURE Cycle Time
(Note 11)
90
100
CONT
Memory Lock Request to Acknowledge Time
Memory Lock Request to Acknowledge Time
Core Status MEASURE
Core Status STANDBY
MLCK,M
MLCK,S
ACKN
Time from Core Entering STANDBY to Return to No Write of 0x0 to Reg. WKUPACK, No
SLEEP, When Wake-Up is not Confirmed
0.6
1.5
Write of 0x8 to Reg. OPCTRL
Time Base
TUE
TUE Time Base
Internal Clock
0.5
1
%
%
TB
l
Note 1: Stresses beyond those listed under Absolute Maximum Ratings
may cause permanent damage to the device. Exposure to any Absolute
Maximum Rating condition for extended periods may affect the device
reliability and lifetime.
Note 6: These timing specifications are dependent on the delay through
the cable, and include allowances for 50ns of delay each direction. 50ns
corresponds to 10m of CAT-5 cable (which has a velocity of propagation of
66% the speed of light). Use of longer cables would require derating these
specs by the amount of additional delay.
Note 2: Positive currents flow into pins, negative currents flow out of pins.
Minimum and maximum values refer to absolute values.
Note 7: Guaranteed by design and characterization, not subject to
production test.
Note 3: Do not apply a voltage or current source to these pins. They must
be unconnected, connected to capacitive loads or connected to a crystal
according to their pin description. Otherwise permanent damage may
occur.
Note 4: Do not apply a voltage source to these pins. Overloading these
pins might disrupt operation.
Note 8: C = capacitance of one bus line in pf (10pF < C < 400pF)
B
B
Note 9: These specifications do not include rise time of SDO due to pull up
resistance and load capacitance on SDO pin.
Note 10: Cycle time at which STATUS/FAULTS and V registers are
REF
updated.
Note 5: Active supply current (I ) is dependent on the amount of time
CC
Note 11: Cycle time at which STATUS/ALERT/FAULTS registers and all
slow channel measurement results are updated after the first update. The
first update after enabling any measurement is typically 50ms delayed.
that the output drivers are active on IP and IM. During those times I will
CC
increase by the 20 • I drive current. For the maximum data rate 1MHz, the
B
drivers are active approximately 5% of the time.
Rev A
9
For more information www.analog.com
LTC2949
TYPICAL PERFORMANCE CHARACTERISTICS
Current Measurement Gain Error
vs Temperature
Current Measurement Offset
vs Temperature
Current Measurement Offset
Distribution
ꢀꢁꢂꢀ
ꢀꢁꢂꢃ
ꢀ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀ
ꢀꢁ Rꢂꢃꢄꢅꢆꢇꢈ ꢉꢊꢅꢋeꢃ ꢌeꢆꢅ ꢍꢅꢂꢎꢄꢋ
ꢀꢁꢂꢀ
ꢀ
ꢀꢁꢀꢂ
ꢀꢁꢀꢀ
ꢀ
ꢀ
ꢀꢁꢂꢁꢃ
ꢀꢁꢂꢃꢁ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢁ
ꢀꢁ
ꢀꢁ
ꢀꢁ
ꢀ
ꢀ
ꢀ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀ
ꢀꢁꢂ
ꢀ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢁꢂꢃꢄ ꢅꢀꢆꢄAꢇꢃ ꢈꢉꢅꢊ
ꢀꢁꢂꢁ ꢃꢄꢅ
ꢀꢁꢂꢁ ꢃꢄꢀ
ꢀꢁꢂꢁ ꢃꢄꢅ
Current Measurement Noise Filter
Response
Power as Voltage Gain Error vs
Temperature
AUXADC Gain Error vs
Temperature
ꢀꢁꢂ
ꢀꢁꢂ
ꢀ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢀ
ꢀꢁꢀ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂ
ꢀ
ꢀꢁ ꢀꢁꢁ ꢀꢁ ꢀꢁꢂ ꢀꢁꢁꢂ ꢀꢁ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀRꢁꢂꢃꢁꢄꢅꢆ ꢇꢈꢉꢊ
ꢀꢁꢂꢃ ꢄꢂꢅe ꢆꢇꢁꢈeꢉeꢅ
ꢀꢁꢂꢃ ꢄꢂꢅe
ꢀꢁꢂꢃ ꢄꢅꢆe
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢁ ꢃꢄꢅ
ꢀꢁꢂꢁ ꢃꢄꢂ
ꢀꢁꢂꢁ ꢃꢄꢅ
Auxiliary ADC Measurement
Offset vs Temperature
Time Base Internal Clock TUE vs
Temperature
ADC Conversion Time Error Slow/
Fast vs Temperature
ꢀ
ꢀ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀ
ꢀ
ꢀ
ꢀꢁꢂ
ꢀ
ꢀ
ꢀꢁꢀ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁ
ꢀꢁ
ꢀꢁ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢁ ꢃꢄꢁ
ꢀꢁꢂꢁ ꢃꢄꢅ
ꢀꢁꢂꢁ ꢃꢄꢅ
Rev A
10
For more information www.analog.com
LTC2949
TYPICAL PERFORMANCE CHARACTERISTICS
VREF vs Temperature
VREF Load Regulation
VREF2 vs Temperature
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢂꢀ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢂꢂ
ꢀꢁꢂꢂꢂ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢃ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢄ
ꢀ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄ
ꢀ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀ
Rꢀꢁ
ꢀꢁAꢂ ꢃꢄRRꢅꢆꢇ ꢈꢉAꢊ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢁ ꢃꢄꢄ
ꢀꢁꢂꢁ ꢃꢄꢅ
ꢀꢁꢂꢁ ꢃꢄꢀ
AVCC/DVCC Supply Current
Standby/Measure vs Temperature
AVCC/DVCC Supply Current vs
Temperature
ꢀꢁ
ꢀꢁ
ꢀꢁ
ꢀꢁ
ꢀꢁ
ꢀ
ꢀꢁeeꢂ ꢃꢄꢅeꢆ ꢇ
ꢀ ꢁ
ꢀ ꢁꢂꢃ
ꢀꢁꢂꢂ
ꢀꢁꢂꢃꢄꢅꢆꢇꢈeꢂꢉꢊꢋeꢌ ꢍ
ꢀꢁ
ꢀꢁꢂꢃ
Aꢀꢁꢁ
Aꢀꢁꢁ ꢀꢁꢂꢂ
ꢀꢁꢁ
ꢀꢁ
ꢀ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂ ꢀꢁꢂ
ꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ ꢀꢁꢁ ꢀꢁꢂ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ ꢅꢆꢇꢈ
ꢀꢁꢂꢁ ꢃꢄꢂ
ꢀꢁꢂꢁ ꢃꢄꢅ
PIN FUNCTIONS
AVCC (Pin 19): Analog Supply Voltage. Bypass this pin to
AGND with a 0.1μF (or greater) capacitor. AVCC operating
range is 4.5V to 14V.
BYP2 (Pin 25): Internal 3.25V Supply Voltage. Bypass
BYP2 to DGND with a 1μF capacitor. Can supply external
circuitry(exampleSPIisolatorADuM141EorADuM4154)
with up to 10mA. Overloading might disrupt LTC2949
functionality.
AGND (Pin 18): Analog Ground. Bypass this pin to AVCC
with a 0.1μF (or greater) capacitor.
CF1P, CF1M (Pins 44, 43): Filter Capacitor Inputs for the
first current channel. Connect a 1µF capacitor between
CF1P and CF1M for filtering differential noise and fast cur-
rent variations. Connect 0.1µF capacitors between AGND
and the filter pins for damping high frequency common
mode variations.
BYP1 (Pin 16): Internal Supply Voltage. Bypass BYP1 to
DGND with a 1μF capacitor. BYP1 is regulated to 2.5V. Can
supply external circuitry (example EEPROM) with up to
10mA. Overloading might disrupt LTC2949 functionality.
Rev A
11
For more information www.analog.com
LTC2949
PIN FUNCTIONS
CF2P, CF2M (Pins 39, 40): Filter Capacitor Inputs for the
second current channel. Connect a 1µF capacitor between
CFI2P and CFI2M for filtering differential noise and fast
currentvariations.Connect0.1µFcapacitorsbetweenAGND
and the filter pins for damping high frequency common
mode variations.
preventsLTC2949togoautomaticallyintoSLEEPstateand
to execute HW memory BIST. Connect a 4.7k-10k pull-up
resistor from SDA to BYP1 to ensure correct operation of
auto-sleep and memory BIST.
SDI/ICMP (Pin 27): Serial Data Input in SPI mode or Iso-
lated Interface Comparator Voltage Threshold in isoSPI
mode. Tie ICMP to the resistor divider between IBIAS and
DGND to set the voltage threshold of the isoSPI receiver
comparators. The comparator thresholds are set to 1/2
the voltage on the ICMP pin.
CLKI (Pin 33): Clock Input. Connect to ground if internal
clock is used. For improved measurement accuracy, con-
necta4MHzcrystalbetweenCLKIandCLKOandmatching
capacitors to ground, or drive with an external clock. See
the Timebase Control section.
SDO/I
(Pin 28): Open Drain Serial Data Output in SPI
BIAS
CLKO (Pin 34): Clock Output. Connect a 4MHz crystal
between CLKO and CLKI if used; leave pin unconnected
otherwise.
mode or Isolated Interface Current Bias in isoSPI mode.
In SPI mode tie with a pullup resistor to IOVCC. In isoSPI
mode tie IBIAS to DGND through a resistor divider to set
theinterfaceoutputcurrentlevel.WhentheisoSPIinterface
is enabled, the IBIAS pin voltage is regulated to 2V. The
IP/IM output current drive is set to 20 times the current
IB, sourced from the IBIAS pin.
CSB/IM (Pin 30): Active Low Chip Select in SPI mode or
Isolated Interface Negative Input/Output in isoSPI mode.
DGND (Pin 17): Digital Ground. Connect to AGND.
DNC (Pins 6, 21, 22, 23, 24, 31, 32, 36, 37, 46): Do
V1, V2, V3, V4, V5, V6, V7 (Pins 1, 2, 3, 4, 5, 7, 8):
Voltage Measurement Inputs. Pins are internally buffered
before being applied to the AUXADC for ensuring high
input impedance (50MΩ) and low leakage. Can be left
floating if unused.
not connect.
DVCC (Pin 20): Supply Voltage. Bypass this pin to DGND
with a 1μF capacitor. Operating range is 4.5V to 14V.
I1P, I1M (Pins 45, 42): Differential Input of I1ADC and
overcurrent comparator 1. Tie to AGND if unused.
V8-V12/GPIO1-GPIO5 (Pins 9, 10, 11, 12, 13): General
PurposeVoltageIn–andDigitalOutputs.Pinsareinternally
buffered before being applied to the AUXADC for ensuring
high input impedance (50MΩ) and low leakage (<10nA).
Each pin can be switched to DVCC, switched to DGND or
to toggle at 400kHz (typ.) between DVCC and DGND. Pins
are tri-state in sleep mode. Can be left floating if unused.
I2P, I2M (Pins 38, 41): Differential Input of I2ADC and
overcurrent comparator 2. Tie to AGND if unused.
IOVCC (Pin 26): Serial Interface Configuration and Sup-
ply Pin. Tie pin to DGND for isoSPI communication. Tie
pin to a voltage ≥1.8V and ≤ 4.5V and bypass with 1µF
to DGND for standard SPI communication. In SPI mode
IOVCC supplies the digital input and output circuits of the
serial interface.
VBATP VBATM (Pins 48, 47): Battery Voltage Measure-
,
ment. The differential voltage between VBATP and VBATM
is internally buffered for ensuring high input impedance
(50MΩ) and low leakage (<10nA).
SCK/IP(Pin29):SerialClockInputinSPImodeorIsolated
Interface Positive Input/Output in isoSPI mode.
VREF (Pin 35): Reference Voltage Output. VREF provides
a buffered 3V reference voltage for temperature measure-
mentswithNTCs. Currentloadislimitedto0.5mA. Bypass
this pin to AGND with a 1μF capacitor.
2
SCL (Pin 14): I C Master Clock Open Drain Output. Con-
nect to clock input of EEPROM.
2
SDA (Pin 15): I C Data Input And Open Drain Output. Con-
EXPOSED PAD (Pin 49): Connect to AGND.
nect to data line of EEPROM. SDA driven low at power up
Rev A
12
For more information www.analog.com
LTC2949
BLOCK DIAGRAM
ꢀꢁꢂꢃꢄꢅꢆꢇ ꢈꢄꢃꢅꢁR
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢁꢂ
Aꢀꢁ
ꢀꢁAꢁꢂꢀ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃRꢁꢄ
ꢀꢁRRꢂꢃꢄ ꢅ ꢄꢆRꢂꢇꢆꢈꢉꢊꢇ
ꢀꢁRRꢂꢃꢄ ꢅ
ꢀꢁꢂ
ꢀꢁ ꢂꢃ ꢅ ꢀꢆꢆꢇꢈ
ꢀꢁꢂꢃ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁ
ꢀꢁ ꢂꢃ ꢅ ꢆꢇꢈꢉꢊꢋ
ꢀꢁꢂ
ꢀAꢁꢂ
ꢀꢁ
ꢀꢁRRꢂꢃꢄ ꢅ ꢄRAꢀꢆꢇꢃꢈ
ꢀꢁꢂꢃR ꢄ ꢅꢆRꢃꢇꢆꢁꢈꢉꢇ
ꢀꢁꢂꢃR ꢄ
ꢀꢁꢂꢃ
ꢀ ꢁꢂ ꢃ
ꢀꢁ ꢂꢃ ꢅ ꢀꢆꢆꢇꢈ
ꢀꢁꢂꢃ
ꢀꢁ
ꢀꢁꢂ
ꢀ
ꢀꢁ ꢂꢃ ꢅ ꢆꢇꢈꢉꢊꢋ
ꢀꢁAꢂꢃ
ꢀAꢁꢂ
ꢀꢁꢂꢃR ꢄ ꢅRAꢆꢇꢈꢉꢊ
ꢀꢁꢂꢃR ꢄ ꢅꢆRꢃꢇꢆꢁꢈꢉꢇ
ꢀꢁꢂꢃR ꢄ
ꢀꢁ ꢂꢃ ꢅ ꢀꢆꢆꢇꢈ
ꢀꢁAꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁ
ꢀ
ꢀꢁ ꢂꢃ ꢅ ꢆꢇꢈꢉꢊꢋ
ꢀꢁꢂ
ꢀAꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃR ꢄ ꢅRAꢆꢇꢈꢉꢊ
ꢀꢁRRꢂꢃꢄ ꢅ ꢄꢆRꢂꢇꢆꢈꢉꢊꢇ
ꢀꢁRRꢂꢃꢄ ꢅ
ꢀ ꢁꢂ ꢃ
ꢀꢁꢂ
ꢀꢁ ꢂꢃ ꢅ ꢀꢆꢆꢇꢈ
ꢀꢁꢂꢃ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁ
ꢀꢁ ꢂꢃ ꢅ ꢆꢇꢈꢉꢊꢋ
ꢀꢁꢂ
ꢀAꢁꢂ
ꢀꢁRRꢂꢃꢄ ꢅ ꢄRAꢀꢆꢇꢃꢈ
ꢀꢁRRꢂꢃꢄ ꢅꢆꢇꢄꢈRꢉ
ꢀꢁARꢂꢃ ꢄ ꢅ ꢆ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄꢅꢆꢇ ꢈꢄꢃꢅꢁR
ꢀꢁꢀRꢂꢃ ꢄ ꢅ ꢆ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢁꢂ
ꢀꢁARꢂꢃ ꢄ ꢃꢅꢃRꢂꢆ ꢇꢁRꢃꢈꢁꢉꢊꢋꢈ
ꢀꢁARꢂꢃ ꢄ ꢃꢅꢃRꢂꢆ ꢇRAꢀꢈꢉꢅꢂ
ꢀAꢁꢁꢂRꢃ ꢄꢅꢆꢁAꢇꢂ ꢁꢈRꢂꢉꢈꢅꢆꢊꢉ
ꢀAꢁꢁꢂRꢃ ꢄꢅꢆꢁAꢇꢂ
Aꢀꢁꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁ ꢂꢃ ꢅ ꢆꢇꢈꢉꢊꢋ
ꢀꢁ ꢂꢃ ꢅ ꢆꢇꢈꢉꢊ
ꢀꢁ ꢂꢃ ꢅ ꢀꢁꢆꢇ
Aꢀꢁꢂꢃꢀꢁ
ꢀꢁꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀAꢁꢁꢂRꢃ ꢄꢅꢆꢁAꢇꢂ ꢁRAꢈꢉꢊꢋꢇ
ꢀꢁꢂꢃꢁRAꢀꢄRꢁ
ꢀ
Aꢀꢁ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀAꢁꢂ ꢀꢃꢀꢄ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃꢄ
ꢀRꢁꢂꢃꢃ
ꢀRꢁꢂꢃꢄ
ꢀRꢁꢂꢃ
ꢀꢁꢂꢀꢃ
Aꢀꢁꢂ
ꢀꢀꢀ ꢁꢂꢂ RꢂꢃꢄꢁꢅꢂR ꢆA ꢀꢀꢀ
Aꢀꢁꢂꢃꢃ
ꢀꢁꢂꢀꢃR
ꢀꢁꢂꢃꢄꢅ
ꢀꢁꢂꢃꢄꢃꢄꢃ
ꢅꢀAꢆꢇꢂꢁꢈꢀꢉꢁ
Aꢀꢁꢂ
ꢀRꢁꢂꢃ
ꢃꢄꢅꢀ
ꢀRꢁꢂꢃ
ꢃꢄꢅꢆꢀ
ꢀꢁꢂꢃ
ꢀꢁꢂꢀꢃR
ꢀꢁꢂꢃꢄꢅ
ꢀꢁꢂꢀꢃ
ꢀ
ꢀꢁꢂꢃꢄꢅꢄAꢀ
ꢀꢁꢂ
ꢀ
ꢀꢁꢂꢃꢂꢄꢅꢆ
ꢀꢁꢂꢃ ꢀꢁꢂꢃ
ꢀꢁꢂ
ꢀꢁꢂꢀꢃꢄ ꢅ
ꢀꢁꢂꢃꢄꢅꢀꢁꢂꢃꢆ
ꢀꢁꢂꢃꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ ꢀꢁꢂꢂ
Aꢀꢁꢁ
ꢀRꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁA
ꢀꢁꢂ
* Measured value not accessible by user. Only used for internal diagnostics.
** VREF measurement value is only accessible by user from the AUX slow channel.
See also section 'Unused Input Pins V1-V12' for recommendation to allow VREF measurement from AUX fast channel.
See also 'Table 57. MUX Settings' for more details on AUX MUX configuration.
Rev A
13
For more information www.analog.com
LTC2949
OPERATION
OVERVIEW
measured in slow mode, and the LTC2949 will set the
corresponding bit in the Alert Register if a threshold is
exceeded.Programmableheartbeatfunctionsonuptotwo
GPIOs allow to signal any enabled alert over an isolation
barrier, independent of the serial interface. Those pins
toggle at 400kHz and stop toggling in case of an alert.
The LTC2949 is a high precision current, voltage, charge
and energy meter for electrical and hybrid vehicles or
other applications requiring isolated data acquisition.
Operating from supply voltages from 4.5 to 14V, it infers
charge and energy flowing in and out of the battery pack
by monitoring simultaneously the current through up to
two sense resistors and the battery pack voltage. Five rail-
to-rail low-offset ΔΣ ADCs ensure accurate measurement
of currents, voltage and power with insignificant power
loss. The LTC2949 uses instantaneous multiplication of
voltage and current at a high sampling rate to infer ac-
curately power even in presence of fast load variations.
Additionally to the inputs for measuring currents and
battery voltage, the LTC2949 features 12 analog input
pins (V1 to V12) for measuring external voltages. Using
its built-in multiplexer, the LTC2949 performs differential
high input impedance rail-to-rail voltage measurements
between any pair of input pins. Pins V8 to V12 can be
configured as high voltage digital outputs, swinging from
ground to digital supply voltage (DVCC) for controlling
externalcomponentssuchashighvoltagetransistors.One
automatic measurement cycle of current, voltage, power,
temperature, supply voltage and two programmable mul-
tiplexer settings takes 100ms in slow mode. The LTC2949
repeatedlyperformssuchmeasurementsandrecalculates
energy, charge, timeandupdatestheminimum/maximum
tracking and threshold registers, resulting in continuous
integration of current and power with lossless tracking
of charge and energy delivered or received by the battery
pack. An on-chip oscillator provides a 1% precise time
baseforcalculatingtotalcharge,energyandtime.Ifhigher
accuracy is required, a 4MHz crystal connected between
pin CLKI and CLKO or an external clock can be used.
The LTC2949 features a programmable analog overcur-
rent comparator for each current channel for applications
which require fast detection of overcurrent conditions. A
programmable deglitch filter allows to discard overcur-
rent conditions shorter than a predefined time duration*.
A 3V reference voltage output (VREF) is provided for
connecting external NTCs or voltage dividers allowing to
measuresignalsbelowground.Inslowmode,theLTC2949
provides means for linearizing temperature readings of up
to two external NTCs by solving Steinhart-Hart equations
with programmable coefficients. The LTC2949 can be
configuredtoautomaticallycompensateuser-programmed
temperature coefficients of low-cost shunt resistors by
using linearized NTC temperature readings.
The LTC2949 features programmable gain correction
factors to compensate for tolerances of external shunt
2
resistors and resistor dividers. A master I C interface and
dedicated commands allow to read from and write to an
externalEEPROMwhichcanbeusedforstoringcalibration
factors and the entire register content of the LTC2949 to
guarantee data retention without supply. Storing correc-
tion factors in an EEPROM enables a modular approach
to factory-calibration of application boards.
Athermalshutdowncircuittripsatdietemperaturesabove
150°C and resets the IC to default state, only the thermal
shutdown bit itself is not reset.
Measuredquantitiesarestoredininternalregistersacces-
sibleviatheonboardSPIorLTC-proprietaryisoSPIinterface
which allows fully isolated operation of the LTC2949. The
LTC2949wasdevelopedforcompatibilitywithLinear’sMul-
ticell Battery Monitors (LTC68XX). Various bus structures
ineitherSPIorisoSPI,multi-dropand/ordaisychainingare
possible. The LTC2949 supports a limited set of Linear’s
Battery Cell Monitor compatible commands for triggering
synchronous ADC conversions and reading back data.
For time critical applications, a fast mode is available
which reduces conversion times to 782µs. Data acquired
during fast operation is stored in four FIFOs which contain
each up to 1000 fast ADC readings from 4 synchronously
measured parameters. Reading data from the FIFO yields
simultaneouslyconvertedconversionresults,enablingbat-
teryimpedancetracking,currentprofilingormonitoringof
other fast events, such as the pre-charging voltage before
closing contactors. Thresholds can be set for parameters
*An overcurrent condition is signaled by a dedicated heartbeat pin for fastest response times.
Rev A
14
For more information www.analog.com
LTC2949
OPERATION
CORE LTC2949
Digital data acquired in isolated operation is transferred
over external capacitors or transformers across an isola-
tion barrier. Bridging potential differences of several kV is
achieved by choosing appropriate external components.
ꢀꢁꢂꢂꢃ
ꢅꢂꢐAꢄꢑꢒꢂ ꢂꢆꢐꢂ
ꢓꢅ ꢔꢀꢇ
ꢍꢎ
ꢀꢁꢂꢂꢃ ꢋ ꢌ
ꢍꢎ
ꢅꢓ ꢕAꢖꢂꢊꢃ
ꢔꢓꢅꢗꢑRꢉAꢄꢑꢓꢅ
ꢏ
ꢀꢁꢂꢂꢃ
All those features enable a wide variety of applications
beyond current and charge measurement, like measur-
ing isolation resistance, controlling pre-charge switches,
signalingalarmconditions,monitoringstateofcontactors,
etc. The LTC2949 offers various diagnosis functions to
support functional safety critical systems, see the Safety
Manual for more information.
ꢏ
ꢇꢓꢓꢄ
ꢍꢎ
ꢏ
Aꢔꢖꢅ
ꢀꢁꢂꢂꢃ ꢋ ꢌ
ꢍꢎ
ꢀꢁꢂꢂꢃ
ꢏ
ꢃꢓꢕꢂR ꢊꢃ
ꢍꢎꢇꢓꢓꢄꢏ
ꢀꢄAꢅꢆꢇꢈ
ꢉꢂAꢀꢊRꢂ
ꢀꢀꢜꢓꢄ ꢋ ꢌ
ꢓR
ꢔꢓꢅꢄ ꢋ ꢌ
ꢔꢓꢅ ꢝ ꢆꢓꢅꢂ
Aꢅꢆ
ꢔꢓꢅꢄ ꢋ ꢛ
ꢘꢙꢚꢙ ꢗꢛꢌ
MODES OF OPERATION
Core State Description
Figure 1. LTC2949 Operation State Diagram
t
willpreventtheLTC2949fromenteringSLEEPstate.
SLEEP
When all power supply voltages have risen above their
UVLO thresholds, the LTC2949 boots up, sets all registers
to their default state and enters after 1s its default SLEEP
state with a current consumption of 9µA (typ), prevent-
ing rapid discharge after insertion when being supplied
by a battery.
In SLEEP, internal analog supplies and BYP1 are switched
off, causing the UVLOA and UVLOD bits to be set when
the LTC2949 resumes from SLEEP. An internal always-on
regulatedvoltagesuppliesthememoryandguaranteesdata
retention during SLEEP. If AVCC or DVCC drop below the
UVLO threshold, the UVLO bit of the internal always-on
supply UVLOSTBY is set and a power-on-reset occurs,
resetting all registers to their default value. In STANDBY
state, all internal circuitry is active but no measurements
except the reference voltage (VREF) are being made.
From STANDBY, the LTC2949 can be instructed to go into
MEASURE state by setting the single shot (SSHOT) or
continuous (CONT) bit in the Operations Control Register
OPCTRL.
In SLEEP state, all GPIOs are tri-state and the LTC2949
monitorstheserialinterfaceandinitiatesthebootsequence
on a falling edge of CSB in SPI mode. In isoSPI mode the
isoSPIinterfacemustfirstbewokenupbyawake-uppulse
before a pulse generating a negative edge on the internal
CSB can be sent. During the boot sequence, the host can
polltheSLEEPbitintheOperationControlRegistertocheck
thattheLTC2949isawakeandinSTANDBYmode(seealso
Figure 20 about Wake-Up and Boot procedure). LTC2949
isoSPI State Description
enters STANDBY state maximum 100ms (t
) after the
BOOT
first falling edge of CSB. In STANDBY state all reference
voltages are powered up and a clock is provided to digital
circuits.TheLTC2949automaticallyreturnstoSLEEPstate
if no wake up confirmation command is received within 1
If the IOVCC pin is tied to a supply voltage ≥1.8V, the
LTC2949 operates in normal SPI mode and IOVCC sup-
plies the receiving circuit and output driving circuit for all
SPI signals.
second (t
) after entering STANDBY state. Wake up is
ACKN
Tying IOVCC to DGND enables the isoSPI port. The isoSPI
port has three different states: IDLE, READY and ACTIVE.
In IDLE state, the isoSPI port is powered down. Only
confirmed by writing 0x00 to register WKUPACK.
The LTC2949 enters SLEEP state ~200ms (t
) after
SLEEP
receivingasleepcommand.AnegativeedgeonCSBduring
Rev A
15
For more information www.analog.com
LTC2949
OPERATION
differential activity on IP-IM generates a wake-up signal
ꢈꢉꢊꢋꢌꢀ ꢌꢊꢍꢎ
ꢀꢁꢂꢃ
and the isoSPI port will enter READY state after t
READY
(10µs) and be ready to send or receive data. The current
consumption increases by several mA in READY. When
communication takes place the isoSPI port is in ACTIVE
state and supply current rises further depending on clock
frequency. In order to save power the isoSPI port enters
IDLE state when there has been no differential activity on
ꢔAꢕꢃꢑꢌ ꢋꢀꢖꢗAꢂ
ꢒꢎ
ꢀꢁꢂꢃ ꢆꢀꢏꢃꢐꢑꢆ
ꢒꢎ
ꢓ
RꢃAꢁꢄ
ꢓ
ꢀꢁꢂꢃ
RꢃAꢁꢄ
Aꢅꢆꢀꢇꢃ
ꢗꢐ Aꢅꢆꢀꢇꢀꢆꢄ ꢐꢗ
ꢈꢉꢊꢋꢌꢀ ꢌꢐRꢆ
IP-IM for more than t
(6.4ms typ.). Communication
IDLE
ꢆRAꢗꢋꢏꢀꢆꢝRꢃꢅꢃꢀꢇꢃ
to the LTC2949 core is only possible if the isoSPI port is
not in IDLE state. This means that even when the LTC2949
coreisinSTANDBYorMEASUREstateandtheisoSPIport
is in IDLE state, communication can only take place 10µs
ꢘꢙꢚꢙ ꢛꢜꢘ
(t ) after differential activity on IP-IM.
READY
Figure 2. isoSPI State Diagram
Figure 2 displays the sequence of states which the isoSPI
interface and the LTC2949 core go through from waking
up the interface until effectuating ADC measurements.
See also Figure 20 for recommended wake-up sequence
implementations.
CH1 and CH2 can be individually set to an 18-bit high
precision mode (slow mode, default) or a 15-bit fast
mode. Activating fast mode reduces conversion times
from 100ms to 782µs on the selected channel. The power
ADCs can be individually configured as voltage ADCs by
disabling the power multiplication after the input buffers
by setting the corresponding Power as Voltage (PasV) bit
in ADC Configuration Register (ADCCONF).
DATA ACQUISITION CHANNELS
The LT2949 has two current ADCs (I1ADC, I2ADC), two
power ADCs (P1ADC, P2ADC) and one Auxiliary ADC
(AUXADC). I1ADC and P1ADC are grouped together and
form data acquisition channel 1 (CH1), I2ADC and P2ADC
form channel 2 (CH2) and the AUXADC together with
auxiliary multiplexer (AUXMUX), die-temperature sensor
and supply voltage sensor form channel AUX (CHAUX).
Slow, High Precision Mode
By default, LTC2949’s acquisition channels are in slow
high precision mode where conversions of CH1 or CH2
take 100ms and yield 18-bit conversion results of current
and power or current and voltage, if PasV set. During
these 100ms the auxiliary channel (CHAUX) measures
ꢓꢁꢀꢈꢇRꢎ ꢏAꢐꢃꢍꢋꢊꢠ
ꢏRꢇꢂꢃ ꢚꢡꢚ ꢂꢁ ꢏꢐꢋꢊAꢓꢐ
ꢌꢂARꢂ ꢎꢃAꢌꢋRꢃꢎꢃꢀꢂꢠ
ꢏRꢇꢂꢃ ꢚꢡꢚꢢ ꢂꢁ ꢁꢊꢓꢂRꢉ
ꢓꢌꢔꢙꢚ
ꢓꢌꢔꢙꢛ
ꢆAꢂA ꢉAꢖꢃR
ꢇꢊꢍꢇꢎ
ꢓꢌꢔꢙꢚ
ꢓꢌꢔꢙꢛ
ꢞ
ꢆꢏꢃꢉꢉ
ꢞ
ꢆꢏꢃꢉꢉ
ꢗ
ꢏAꢐꢃ
ꢞ
ꢞ
ꢞ
ꢞ
ꢇꢆꢉꢃ
RꢃAꢆꢖ
ꢇꢆꢉꢃ
RꢃAꢆꢖ
ꢇꢆꢉꢃ
RꢃAꢆꢖ
ꢇꢆꢉꢃ
RꢃAꢆꢖ
Aꢓꢂꢇꢗꢃ
RꢃAꢆꢖ
ꢇꢆꢉꢃ
ꢟꢑꢒꢌꢊꢇ ꢌꢂAꢂꢃ
ꢓꢁRꢃ ꢌꢂAꢂꢃ
ꢞ
ꢔꢁꢁꢂ
ꢞ
ꢇꢆꢉꢃꢧꢓꢁRꢃ
ꢌꢉꢃꢃꢊ
ꢌꢂAꢀꢆꢔꢖ
ꢎꢃAꢌꢋRꢃ
ꢣꢤꢥꢤ ꢈꢚꢦ
ꢀꢁꢂꢃꢄ ꢂꢅAꢂ ꢂꢅꢃ ꢆꢇꢈꢈꢃRꢃꢀꢂꢇAꢉ ꢊꢋꢉꢌꢃ ꢁꢀ ꢇꢊꢍꢇꢎ ꢂꢁ ꢏAꢐꢃꢍꢋꢊ ꢂꢅꢃ ꢇꢑꢒꢌꢊꢇ ꢇꢀꢂꢃRꢈAꢓꢃ ꢓAꢀ ꢔꢃ RꢃAꢉꢇꢕꢃꢆ ꢔꢖ A ꢓꢌꢔ ꢆRꢇꢗꢃ ꢉꢁꢏ ꢊꢋꢉꢌꢃ ꢘꢓꢌꢔ ꢙ ꢚ ꢈꢁꢉꢉꢁꢏ ꢔꢖ ꢓꢌꢔ ꢙ ꢛꢜꢝ
Figure 3. Timing of IsoSPI and Core States
Rev A
16
For more information www.analog.com
LTC2949
OPERATION
sequentially six different quantities using its Round Robin
(RR) mode starting with VBATP – VBATM (BAT) then die-
temperature (TEMP), AVCC supply voltage (VCC), two
AUXMUX inputs (SLOT1 and SLOT2), selectable by the
Multiplexer Setting Registers and finally the Reference
Voltage(VREF). Furthermore, amovingaverageofthelast
four measurements of IADC1 and IADC2 in slow mode is
provided, yielding a 20-bit result. In slow, high precision
mode a single conversion or continuous conversions can
be triggered. The continuous slow mode (CONT) is the
most typical operation and also the prerequisite to make
fast conversions.
In continuous measurement mode, the host can poll for
the end of a measurement cycle by periodically checking
the corresponding time registers (TB1, TB2, TB3 or TB4)
for incrementation.
Accessing High Precision Results
At the end of each measurement cycle of 100ms, all result
register for the measured quantities are updated; in con-
tinuous mode, accumulated quantities are updated also.
Furthermore, the four preceding measurements of each
current channel are stored in Current History Registers
and their average in Current Average Registers, see the
Register Map section for more details. The completion
of register updates of the measurement results can be
detected by reading one of the time registers (TB1-TB4)
and looking for a changed value.
Single Shot Measurement Mode (SSHOT)
When bit SSHOT in the Operation Control Register is set,
the LTC2949 takes measurements of CH1 and CH2 as
well as the six auxiliary channel measurements described
above, updates the corresponding minimum, maximum
and threshold registers, resets bit SSHOT, sets the bit
UPDATE in Status Register and returns to the STANDBY
state.Notimemeasurementsaremadeandthechargeand
energy registers are not updated and therefore not com-
pared against minimum/maximum thresholds. The host
can poll the UPDATE bit in the STATUS Register to detect
thecompletionofthemeasurementcycle.Ameasurement
Fast Mode
The LTC2949 provides a fast mode with a reduced conver-
sion time of 782µs and a resolution of 15-bit. Fast mode
allows to start measurements at a precise point in time
andtherebyperformmeasurementssynchronizedwiththe
voltage measurements of LTC’s Battery Stack Monitors,
for example to deduce cell impedance of individual cells.
starts within 20ms (t
) after setting bit SSHOT.
IDLE_CORE
Fast Mode Configurations
The LTC2949 allows to set data acquisition channel 2 in
fast mode while data acquisition channel 1 stays in slow
high precision mode or to set both data acquisition chan-
nels 1 and 2 in fast mode by setting the corresponding bit
FACH1 or FACH2 in the Fast Control Register (FACTRL).
The auxiliary channel can be set to fast mode indepen-
dently of CH1 and CH2, converting only a single quantity
(selected by the Fast MUX Control Registers) instead of
Round-Robin (RR).
Continuous Measurement Mode (CONT)
When the bit CONT in the Operation Control is set, the
LTC2949repeatedlytakesmeasurementsofCH1andCH2as
wellasthesixauxiliarychannelmeasurements,recalculates
energy, charge, time and updates the minimum/maximum
tracking and status registers every 100ms. The start of
continuous high precision measurements can take up to
20ms (t
) after setting bit CONT. The current and
IDLE_CORE
power ADCs run continuously in this mode, ensuring that
nochargeorenergyismissed. IfapowerADCisconfigured
to measure voltage (PasV), the energy of the correspond-
ing channel is not accumulated. The LTC2949 remains in
continuous mode until bit CONT of the Operation Control
Register is reset by the user. If the SSHOT bit is set while
in continuous mode, the LTC2949 completes the current
measurement cycle and then enters single shot mode,
clearing the CONT bit in the Operation Control Register.
To enable short startup delays, the LTC2949 must be in
continuousmode(CONT=1)beforetriggeringfastconver-
sions. FastmeasurementsaretriggeredeitherbyanADCV
command (fast single shot) or by setting bit FACONV in
the Fast Control Register (FACTRL). The ADCV command
triggers a single fast conversion of the selected channels
right after the correct PEC is received, synchronous to
Rev A
17
For more information www.analog.com
LTC2949
OPERATION
LTC’s Battery Stack Monitors. If fast measurements are
triggered by setting FACONV=1, LTC2949 executes a
sequence of fast conversions until FACONV or CONT is
reset. Samples acquired during fast continuous mode are
storedinindividualFIFOsforuptofourfastinputchannels
(I1, I2, BAT via P1 or P2 and AUX).
since the whole FIFO was read (0x55, default), or has
been overwritten because the FIFO was filled without
being read (0xAA). The other FIFOs present the respec-
tive quantities accordingly. All tags are initialized to their
default value (0x55), if fast conversions are triggered by
an ADCV command (while FACONV = 0) or by setting
FACONV. Also leaving the continuous mode by resetting
CONT will clear the FIFOs.
CH1 and CH2 automatically restart in slow high precision
mode after completion of joint fast mode conversions. If
CH2 was in fast mode while CH1 continued in slow high
precisionmode, CH2willstopconvertingaftercompletion
offastmodeacquisitionsreadytoperformfurtherfastmode
conversionswhenrequestedeitherbyanADCVcommand
or by setting FACONV again. The auxiliary channel slow
mode Round-Robin (BAT, TEMP, VCC, SLOT1, SLOT2,
VREF)isautomaticallyrestartedafterallfastmeasurements
werestoppedfor300ms. Inapplicationswerecontinuous
high AUX measurement rates (faster than every 100ms)
are required, it is recommended to measure also VREF
and optionally VCC via external connections to V1-V12
and implement a manual Round-Robin by fast single shot
measurements in software.
The LSB sizes of the fast conversion results are the same
for the RDCV and FIFO readings and are listed in Table 28.
When CH1 and CH2 are both in fast mode, 128 conversion
results of IADC1 and IADC2 are averaged and stored in
theirrespectivenon-accumulatedresultsregistersCurrent1
and Current2 and are added to the Charge1 and Charge2
registers ensuring that battery charge and discharge is
monitored also in fast mode.
Similarly 128 conversion results of PADC1 and PADC2
with 11 bit resolution are averaged and stored in their
respective non-accumulated results registers Power1 at
and Power2 and are added to the Energy1 and Energy2
registersifthePADCsareinPowerMode(PasV=0). IfCH1
isinslowmodeandCH2infastmode, onlyCH1resultsare
reported in the Current and Power results registers, CH2
results can be accessed via RDCV or the respective FIFOs.
Accessing Fast Mode Results
The last results of fast conversions can be read out by
the RDCV command providing sequentially the results of
I1, I2, BAT via P1 or P2 and AUX followed by an indica-
tor if the data is new (0xF) or old (0x0). BAT is the result
of the power ADC, if the ADC is set in voltage mode (bit
PasV=1), otherwise, the power ADC result is 0. If both
channels are in fast mode with their power ADCs in volt-
age mode, BAT is obtained from PADC1. Please note that
to be compliant with LTCs Battery Stack Monitors, RDCV
reports LSByte first.
Recommended Configurations of Data Acquisition
Channels
Inatypicalapplicationcase,bothdataacquisitionchannels
offeredbyLTC2949couldmonitorthecurrentoverasingle
shunt resistor, where CH1 is used to do high precision
current and power measurements and charge and energy
integration, while channel two takes fast snapshots of cur-
rentandvoltageforexampleforimpedancemeasurement.
Furthermore, the last 1000 fast conversion results of I1,
I2, BAT via P1 or P2 and AUX are stored in First-In-First-
Out Registers accessible by FIFOI1, FIFOI2, FIFOBAT and
FIFOAUX. Reading continuously from FIFOI1 provides
successively 3 bytes for each sample of the first current
ADC: I1 MSB, I1 LSB and a qualifier (TAG) whether the
corresponding FIFO data is fine (0x00), has been already
read because no new data has been added to the FIFO
Alternatively, the two data acquisition channels offered
by LTC2949 might be used to monitor currents over two
different shunt resistors. In this application case, both
channelsmightbeusedineitherfastorslowmode.CHAUX
can be configured fully independently from CH1 and CH2.
The default mode of AUXADC is RR, which is deactivated
by enabling fast mode on CHAUX.
Rev A
18
For more information www.analog.com
LTC2949
OPERATION
Table 1. Acquisition Channels Configurations
Dual Shunt Fast Measurement Configuration
SINGLE SHUNT
Slow
DUAL SHUNT
CH1 and CH2 are both set to fast mode by setting bits
FACONV, FACH1 and FACH2. Charge is accumulated by
summing up 15-bit current results, energy by summing
up 11 bit power results.
CH1
CH2
Slow
Slow
Fast
Fast
Fast
CHAUX
RR/Fast
RR/Fast
RR/Fast
One conversion takes 782µs and a new RR cycle is started
afterevery100ms. Thisconfigurationfitstoanapplication
where two shunt resistors are used and fast current, volt-
age and power is required. E.g.: Fast impedance tracking
or measuring pre-charge voltage and current.
Single Shunt Configuration
If only one external shunt is used, CH1 can be used to
perform continuous high precision integrated charge and
energy measurements, while CH2 is used to perform fast
measurements synchronized with LTC's Battery Stack
Monitors. By setting bit CONT in the Operation Control
Register (OPCTRL) and bit FACH2 in the Fast Control
Register (FACTRL), CH1 will effectuate consecutive slow
measurements, while CH2 is available for fast measure-
ments triggered by an ADCV command. Measurements
of CH1 and its integrated quantities are updated every
100ms while CH2 results can be read out with an RDCV
command or obtained from the FIFO registers, in case of
fast continuous (FACONV) operation.
POWER MEASUREMENT
The LTC2949 measures power with additional ADCs that
multiply voltage (VBATP – VBATM) and current at the full
5.24MHzsamplingfrequency,priortoanyaveragingdueto
the analog-to-digital conversion. This maintains accuracy
even if current and voltage change in phase during the
100ms conversion time, which can happen if the power
is drawn from a battery with significant impedance. Figure
4 shows an example of the BAT voltage dropping from 4V
to 3V due to battery impedance when an AC current is
drawn by the load. In this example, the multiplication of
average current with average voltage would lead to a +8%
error in the calculated power as the voltage is significantly
lower than the average voltage at the moments where the
peak current is drawn. The scheme used by the LTC2949
avoids this error, maintaining specified accuracy with
signals up to 50kHz.
Dual Shunt High Precision Configuration
For dual shunt applications that require continuous un-
interrupted high precision coulomb counting and energy
measurement CH1 and CH2 should be both configured to
slow high precision mode by setting bit CONT in OPCTRL.
Conversions on CH1 and CH2 take 100ms and a new RR
cycleofCHAUXisstartedateverystartofconversionofCH1.
If fast voltage data is required, the AUXADC can be config-
ured for fast mode without interrupting charge and energy
accumulation on CH1 and CH2. After setting bits FACONV
and FACHA the AUXADC immediately stops RR mode and
continuously measures a single quantity selected by the
FastMUXControlRegisters.DataiswrittentotheFIFOAUX.
AfterclearingbitFACHAtheAUXADCautomaticallyreturns
to RR operation (after 300ms). Measurements of VREF,
internal die temperature and VCC are only available if RR
is enabled. Alternatively, external connections to V1-V12
can be applied to measure VREF and VCC (via an external
resistive divider) also in fast mode.
ꢑꢒꢈ
ꢓꢒꢑ
ꢓꢒꢈ
ꢔꢒꢑ
ꢔꢒꢈ
ꢕꢒꢑ
ꢕꢒꢈ
ꢖꢕꢈ
ꢖꢈꢈ
ꢗꢈ
ꢘꢈ
ꢓꢈ
ꢕꢈ
ꢈ
ꢀꢁꢂꢃAꢄꢅ ꢆꢀꢇAꢃꢈꢉꢀꢇAꢃꢊꢋ
ꢀꢁRRꢂꢃꢄ ꢅꢆꢇꢈꢆꢉꢊ
ꢈ
ꢈꢒꢖ ꢈꢒꢕ ꢈꢒꢔ ꢈꢒꢓ ꢈꢒꢑ ꢈꢒꢘ ꢈꢒꢛ ꢈꢒꢗ ꢈꢒꢙ
ꢀꢁꢂꢃ ꢄꢅꢆꢇ
ꢖ
ꢕꢙꢓꢙ ꢚꢈꢓ
Figure 4. Power Measurement of Transient Signals
Rev A
19
For more information www.analog.com
LTC2949
OPERATION
CHARGE, ENERGY AND TIME
ꢀꢁꢂꢃꢄꢅꢄ
The LTC2949 integrates the current and power measure-
ments over time to calculate charge and energy flowing
to the load. It also keeps track of total accumulated time
used for the integration.
ꢂꢀꢓꢔ
ꢂꢀꢓꢕ
ꢃꢄꢅꢄ ꢖꢍꢗ
ꢆꢇ
ꢘꢘꢙꢖ
ꢘꢘꢙꢖ
ForthequantitieschargeandenergytheLTC2949provides
three sets of registers each, for the quantity "time" four
register sets.
ꢆꢇꢈ Aꢉꢀꢊꢃꢋꢅꢌꢍꢍꢍꢎꢏꢐꢋꢑꢅꢒꢋꢁ
Charge1, Energy1 and Time1 contain accumulated quan-
tities of Channel1. Charge2, Energy2 and Time2 contain
accumulated quantities of Channel2. Charge3 and Time3
contain the sum of charges monitored by Channel1 and
Channel2 and the corresponding time. Similarly Energy4
and Time4 contain the sum of energies monitored by
Channel1 and Channel2 and the corresponding time. See
the Accumulated Result Registers section in the Register
Map description for more details.
Figure 5. Reference Clock with a 4MHz Crystal
nal clock to the external frequency and represents Time,
Charge, and Energy as multiples of the external clock pe-
riod. To accommodate the large range of allowed external
frequencies, an internal prescaler must be configured via
the Timebase Control Register.
The prescaler consists of 2 stages, with the first divid-
PRE
ing the external frequency f
by a factor 2 , and the
REF
second by a factor DIV. PRE is set between 0 and 5 with
bits [2:0] of the Timebase Control Register. PRE should
be configured such that the external frequency divided by
is less than 1MHz as shown in Table 2:
Table 2. Parameter PRE with External Clock
Each register set can be separately configured to accumu-
late current and power based on the sign of the measured
current.Aminimumcurrentthresholdcanalsobesetbelow
which integration is stopped. See the Control Registers
section in the Register Map description for more details.
PRE
2
fREF
PRE
2PRE
1
PRE[2:0]
000
0.1MHz ≤ f ≤ 1MHz
0
1
2
3
4
5
7
Time Base
REF
1MHz ≤ f ≤ 2MHz
2
001
REF
Accurately measuring charge and energy by integrating
current and power requires a precise integration period.
The LTC2949 uses either a trimmed internal oscillator or
an external clock as the time base for determining the in-
tegration period. It can use either an external square wave
clock in a frequency range between 100kHz and 25MHz
or a 4MHz crystal as external clock input. If an external
square wave is used, it should be connected to the CLKI
pin and the CLKO pin should be left unconnected.
2MHz ≤ f ≤ 4MHz
4
010
REF
4MHz ≤ f ≤ 8MHz
8
011
REF
8MHz ≤ f ≤ 16MHz
16
32
100
REF
16MHz ≤ f ≤ 25MHz
101
REF
Internal
111
The second stage of the prescaler then divides the result
by a factor DIV. DIV is set between 0 and 31 by bits [7:3]
of the Time Base Control Register. DIV should be set to the
next lower integer value of the ratio between the output
Figure5showstherecommendedcircuitifacrystalisused
to generate the reference clock. In case the internal clock
is used, tie CLKI to ground and leave CLKO unconnected.
PRE
of the first stage of the prescaler (f _1 = f /2 ) and
REF
REF
32768Hz or, in other terms:
⎛
⎞
fREF
DIV = floor
⎜
⎟
Timebase Control
PRE
⎝
⎠
2
•32768Hz
The LTC2949 uses the internal oscillator by default. If
an external clock or a crystal is used, the PRE and DIV
parameters in the Timebase Control Register need to be
set appropriately. The LTC2949 then compares its inter-
If a crystal is used, the values are: PRE = 2, DIV = 30. The
Quick Eval™ Software for the LTC2949 contains an easy
to use calculator for these parameters. Table 3 gives a few
examples for common frequencies.
Rev A
20
For more information www.analog.com
LTC2949
OPERATION
Table 3. Timebase Settings for Common Frequencies
Bits OCCxPOLx control the polarity sensitivity of the
comparator as in Table 5.
Table 5. OCC Polarity Configuration
TIME BASE
fREF [MHz] PRE 2PRE fREF_1[MHz] DIV CONTROL [7:0]
1
1.5
4
0
1
2
4
5
5
7
1
2
1
30
22
30
19
19
23
X
1111 0000
1011 0001
1111 0010
1001 1100
1001 1101
1011 1101
XXXX X111
OCCxPOL1
OCCxPOL0
POLARITY
Both Polarities
0.75
1
0
0
1
0
1
0
4
Positive Currents
Negative Currents
10
20
25
Int.
16
32
32
0.625
0.625
0.781
The thresholds of the overcurrent comparators can be
programmed individually by means of the OCCxDACx bits
between 0 and 310mV.
Table 6. OCC Thresholds
Threshold
OVERCURRENT COMPARATORS
The LTC2949 features two fast differential over-current
comparators with rail to rail input common mode and
programmable threshold followed by configurable filters
to suppress input glitches. Overcurrent comparator 1
(OCC1)isconnectedtopinsI1PandI1Mwhileovercurrent
comparator 2 (OCC2) supervises the differential voltage
between I2P and I2M. Both overcurrent comparators can
individually be configured to detect either only positive or
only negative overcurrents or overcurrents independent
of their polarity. When at least one of the overcurrent
comparators is enabled, GPIO5 turns into a heartbeat
signal toggling at 400kHz while currents are within the
desired range and staying low if any current exceeds its
programmed limit.
OCCxDAC2
OCCxDAC1
OCCxDAC0
[mV]
0
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
26
52
78
103
155
207
310
Similarly,thedurationofathresholdexceedingnotreported
bythecomparator(DeglitchTime)canbeprogrammedby
the bits OCCxDGLTx between 20µs to 1.28ms.
Table 7. OCC Deglitch Time
The overcurrent comparator 1 (OCC1) between I1P and
I1M is controlled by the control register OCC1CTRL
while overcurrent comparator 2 (OCC2) is controlled by
its control register at OCC2CTRL, both on page 0 of the
register map. Both OCC Control Registers are organized
as shown in Table 4.
OCCxDGLT1
OCCxDGLT0
DEGLITCH TIME [µs]
0
0
1
1
0
1
0
1
20
80
320
1280
Table 4. Overcurrent Comparator Control Registers
Overcurrents are reported by setting OCC1 and OCC2 of
VCC and OCC Status Register (STATVCC) and by stop-
ping the heartbeat signal on GPIO5. While the update of
the output register can take up to 100ms, the heartbeat
is stopped within 15µs after an overcurrent exceeded the
programmeddeglitchtime.Onceanovercurrentoccurred,
the result bits in the register remain set until they are read
by the host and subsequently cleared.
BIT #
NAME
OCCxEN
FUNCTION
Enable OCC
0
1
2
3
4
5
6
7
OCCxDAC0
OCCxDAC1
OCCxDAC2
OCCxDGLT0
OCCxDGLT1
OCCxPOL0
OCCxPOL1
Threshold DAC[0]
Threshold DAC[1]
Threshold DAC[2]
Deglitch [0]
Deglitch [1]
Polarity [0]
For diagnostic purposes, the overcurrent comparators
have a self-test built in using test input signals IPT and
IMT, see the Safety Manual for more information.
Polarity [1]
The overcurrent comparators are enabled by setting the
respective OCCxEN to 1.
Rev A
21
For more information www.analog.com
LTC2949
SERIAL INTERFACES
SERIAL INTERFACES OVERVIEW
A second interface composed by pins 14 and 15 is a mas-
2
ter I C interface, allowing to save and restore LTC2949’s
LTC2949hastwoserialinterfaces,oneforcommunication
withthehostandasecondtoaddressanexternalEEPROM.
The interface for host communication is composed by
pins 27 through 30 and can be configured to be either
a standard 4-wire serial peripheral interface (SPI) or a
2-wire isolated interface (isoSPI) based on the voltage
of the IOVCC pin. Regardless of which configuration is
selected, the LTC2949 acts as an SPI slave. The LTC2949
can be operated in addressable mode (SPI & isoSPI) or
as last element of a daisy chain of LTC68xx Cell Monitors
(isoSPI only).
register content to and from an external EEPROM see
the External EEPROM Control Register section for more
information.
4-WIRE SERIAL PERIPHERAL INTERFACE (SPI)
PHYSICAL LAYER
ConnectingpinIOVCCtoasupplyvoltage≥1.8Vconfigures
the serial port for 4-wire SPI. Logic input thresholds and
output swings are set by the voltage at the IOVCC pin,
which should be connected to the same supply as the SPI
master device. A 1µF bypass capacitor is recommended
from IOVCC to DGND. The pin SDI is often referred to
as MOSI, the pin SDO as MISO. The 4-wire serial port is
configured to operate in a SPI system using CPHA = 1 and
CPOL=1.Consequently,dataonSDImustbestableduring
the rising edge of SCK and data on SDO will be updated on
the falling edge of SCK. The timing is depicted in Figure 7.
The maximum data rate is 1Mbps. See Electrical Charac-
teristics. SDO is open drain and requires a pull-up.
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢃ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂ
ꢀꢁꢂꢃꢄꢅꢄ
ꢀꢁ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢀ
ꢀꢁꢂꢁ ꢃꢄꢅ
Figure 6. 4-Wire SPI External Connections
ꢉ
ꢉ
ꢉ
ꢉ
ꢉ
ꢍ
ꢋ
ꢇ
ꢈ
ꢌ
ꢉ
ꢎ
ꢀꢁꢂ
ꢀꢃꢄ
ꢃꢈ
ꢃꢎ
ꢃꢋ
ꢃꢐ
ꢃꢍꢑꢑꢑꢃꢇ
ꢃꢈ
ꢉ
ꢏ
ꢁꢀꢅ
ꢀꢃꢆ
ꢉ
ꢊ
ꢃꢇ
ꢃꢈ
ꢃꢎ
ꢃꢋ
ꢃꢐ
ꢃꢍꢑꢑꢑꢃꢇ
ꢃꢈ
ꢎꢙꢇꢙ ꢚꢐꢍ
ꢒRꢓꢔꢄꢆꢕꢀ ꢁꢆꢖꢖAꢗꢃ
ꢁꢕRRꢓꢗꢘ ꢁꢆꢖꢖAꢗꢃ
Figure 7. Timing Diagram of 4-Wire Serial Peripheral Interface
Rev A
22
For more information www.analog.com
LTC2949
SERIAL INTERFACES
2-WIRE ISOLATED INTERFACE (ISOSPI) PHYSICAL
LAYER
through an external transformer. Capacitive coupling with
10nF capacitors could also be used, but has a very limited
common mode noise rejection (only for low frequencies)
and is only recommended for short single PCB intercon-
nections with limited voltage transients at the isolation
barrier. Additional clamping Schottky diodes might be
necessary from IP and IM to VCC and GND. Standard SPI
signals are encoded into differential pulses. The strength
of the transmission pulse and the threshold level of the
receiver are set by two external resistors. The values of
the resistors allow the user to trade off power dissipation
for noise immunity. Figure 9 illustrates how the isoSPI
Tying IOVCC to local chip ground enables the isoSPI
2-wire interface which allows fully isolated operation of
theLTC2949.The2-wireinterfaceprovidesmeanstocom-
municate to LTC2949 using simple twisted pair cabling.
An LTC6820 should be used for translating standard SPI
signals from the SPI master into pulses that are sent over
an isolation barrier to the LTC2949.
Theinterfaceisdesignedforlowpacketerrorrateswhenthe
cabling is subjected to high RF fields. Isolation is achieved
ꢄꢅꢆꢇAꢈꢄꢆꢉ ꢊARRꢄꢋR
ꢂꢃꢃꢄꢅꢆ
ꢜꢂ
ꢂꢃꢃ
ꢂ
ꢃꢃ
ꢄꢏ
ꢅꢃꢘ
ꢄꢏ
ꢄꢎ
ꢄꢏ
ꢐꢗꢗ
ꢐꢗꢗ
ꢄꢎ
ꢃꢅꢊ
ꢃꢅꢊ
ꢇꢈꢃꢒꢓꢔꢓ
ꢄꢆꢂꢃꢃ
ꢇꢈꢃꢕꢖꢒꢗ
ꢌꢉꢍ
ꢙꢃ
ꢀꢔꢛꢜꢂꢝꢐꢔꢂ
ꢁ
ꢄꢊꢄAꢅ
ꢄꢊꢄAꢅ
ꢎꢆꢅꢄ
ꢎꢆꢅꢄ
ꢐꢑ
ꢐꢑ
ꢐꢑ
ꢄꢃꢎꢏ
ꢄꢃꢎꢏ
ꢎꢄꢅꢆ
ꢎꢄꢅꢆ
ꢐꢑ
ꢌꢉꢍ
ꢌꢉꢍꢄꢅꢆ
ꢒꢓꢔꢓ ꢚꢗꢖ
Figure 8. isoSPI Physical Layer
ꢓꢚꢊꢀꢁꢂꢁ
ꢉꢇꢙꢋꢀ
ꢍꢢꢣ
ꢍꢢꢣ
ꢎ
ꢉ
ꢅꢊꢆꢋ
ꢌꢍ ꢎ ꢏꢐꢑꢒꢉ
ꢞ
ꢡAꢠꢘꢟꢋ
ꢊꢅRꢊꢟꢅꢚ
ꢅꢗꢓꢘ
ꢅꢗꢓꢘ
ꢚꢛ ꢝ ꢀꢄ ꢝ ꢅ
ꢇ
ꢚꢛ ꢜ ꢎꢏ
ꢚꢛ ꢜ ꢄ
ꢅꢋ
ꢈꢗꢔ
ꢓꢔꢕꢅꢊ
Aꢖꢗ
ꢆꢘꢆꢔRꢙ
ꢚꢛ ꢜ ꢞꢏ
ꢈꢗꢅ
ꢈꢊꢠ
ꢊꢈꢇ
R
ꢆ
ꢋꢟꢓꢈꢘ
ꢅꢆ
ꢘꢖꢊꢔꢗꢘRꢌ
ꢗꢘꢊꢔꢗꢘR
Rꢛ ꢜ ꢎꢏ
Rꢛ ꢜ ꢄ
ꢎ
ꢞ
ꢅ
ꢅꢇꢅAꢈ
ꢅꢊꢆꢋ
Rꢛ ꢜ ꢞꢏ
ꢇ
ꢎ
ꢀꢉ
ꢞ
R
R
ꢇꢏ
ꢏꢉ ꢝ R
ꢇꢀ
ꢊꢔꢆꢋARAꢚꢔR ꢚꢤRꢘꢈꢤꢔꢓꢗ ꢜ
R
ꢇꢏ
ꢎ R
ꢇꢀ
ꢄꢥꢢꢛ
ꢇꢀ
ꢀꢁꢂꢁ ꢃꢄꢁ
Figure 9. isoSPI Interface
Rev A
23
For more information www.analog.com
LTC2949
SERIAL INTERFACES
circuit operates. A 2V reference drives the IBIAS pin. Ex-
IBIAS is held at 2V, causing a current I to flow out of the
B
ternal resistors R and R create the reference current
IBIAS pin. The IP and IM pin drive currents are 20 • I .
B1
B2
B
I . This current sets the drive strength of the transmitter.
B
As an example, if divider resistor R is 2.8k and resistor
B1
R
and R also form a voltage divider of the 2V refer-
B1
B2
R
B2
is 1.21k (so that R
= 4k), then:
BIAS
ence at the ICMP pin. This sets the threshold voltage of
the receiver circuit.
2V
RB1+RB2
IB =
=0.5mA
Waking Up the isoSPI Port
I
DRV =I =I =20•IB =10mA
IP IM
The isoSPI port has 3 modes of operation, IDLE, READY
and ACTIVE as described in the section isoSPI State De-
scription. In IDLE, the WAKEUP circuit monitors activity
on pins IP and IM. Differential activity on IP-IM wakes up
the isoSPI interface. The isoSPI port will return to the low
power IDLE state if there is no activity on IP/IM for a time
RB2
RB1+RB2
VICMP =2V •
=IB •RB2 =603mV
VTCMP =0.5•VICMP =302mV
Inthisexample,thepulsedrivecurrentIDRVwillbe10mA,
andthereceivercomparatorswilldetectpulseswithIP-IM
amplitudes greater than 302mV. If the isolation barrier
uses 1:1 transformers connected by a twisted pair and
terminated with 100Ω resistors on each end, then the
transmitted differential signal amplitude ( ) will be:
of t . The LTC2949 will be ready to communicate when
IDLE
the core is not in SLEEP and the isoSPI state changes
to READY (within t after wakeup) as illustrated in
READY
Figure 3. Common mode signals will not wake up the
serial interface. The interface is designed to wake up
after receiving a large signal single-ended pulse, or a
low-amplitude symmetric pulse. The differential signal
R
2
VA =IDRV
•
M =0.5V
|IP – IM|, must be at least V
= 200mV for a minimum
WAKE
duration of t
= 240ns to qualify as a wake up signal
(This result ignores transformer and cable losses, which
may reduce the amplitude).
DWELL
that powers up the serial interface.
"Long -1" or "Long +1" pulses (= drive CSB low, high)
generatedbyLTC6820andcompatibleisoSPIdevices(e.g.
LTC68xxCellMonitors)willalwaysmeetthisrequirement.
See following chapters for isoSPI pulse details.
isoSPI Pulse Detail
The transmitter can output three voltage levels: +V , 0V,
A
and–V .ApositiveoutputresultsfromIPsourcingcurrent
A
and IM sinking current across load resistor R . A nega-
M
Selecting Bias Resistors
tive voltage is developed by IP sinking and IM sourcing.
When both outputs are off, the load resistance forces
the differential output to 0V. To eliminate the DC signal
componentandenhancereliability,theisoSPIusesbipolar
pulses of two different pulse length. This allows for four
types of pulses to be transmitted, as shown in Table 8. A
+1 pulse will be transmitted as a positive pulse followed
by a negative pulse. A –1 pulse will be transmitted as a
negative pulse followed by a positive pulse. The duration
Theadjustablesignalamplitudeallowsthesystemtotrade
power consumption for communication robustness, and
the adjustable comparator threshold allows the system to
accountforsignallosses.TheisoSPItransmitterdrivecur-
rentandcomparatorvoltagethresholdaresetbyaresistor
divider (R
= R + R ) between the IBIAS and DGND.
BIAS
B1
B2
The divided voltage (V
) is connected to the ICMP pin
ICMP
which sets the comparator threshold (V
voltage. When the isoSPI interface is enabled (not IDLE)
) to 1/2 of this
ICMP
of each pulse is defined as t
, since each is half of the
1/2PW
required symmetric pair. (The total isoSPI pulse duration
is 2 • t ).
1/2PW
Rev A
24
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Table 8. isoSPI Pulse Types
This allows for multiple slave devices on a single cable
without risk of collisions (Multidrop). Figure 11 shows the
isoSPI timing diagram for a READ command.
FIRST LEVEL
(t
SECOND LEVEL
(t
PULSE TYPE
Long +1
)
)
1/2PW
ENDING LEVEL
1/2PW
+V (150ns)
A
–V (150ns)
0V
0V
0V
0V
A
Table 9. LTC2949 isoSPI Port Function
INTERNAL SPI
Long –1
–V (150ns)
A
+V (150ns)
A
Short +1
Short –1
+V (50ns)
–V (50ns)
A
A
RECEIVED PULSE PORT ACTION
RETURN PULSE
–V (50ns)
A
+V (50ns)
A
Long +1
Long –1
Short +1
Drive CSB High None
Drive CSB Low
An LTC6820 should be used to translate the SPI signals of
a micro controller into isoSPI pulses. On the other side of
the isolation barrier (i.e. at the other end of the cable), the
LTC2949 will have IOVCC tied to its local GND. It receives
transmitted pulses and reconstructs the SPI signals in-
ternally, as shown in Table 9. In addition, during a READ
command this port may transmit return data pulses which
1. Set SDI = 1
2. Pulse SCK
Short –1 Pulse if Reading a 0-bit
(No Return Pulse if Not in READ
Mode or if Reading a 1-bit)
Short –1
1. Set SDI = 0
2. Pulse SCK
Supported Bus Structures
The addressing feature of the LTC2949 and LTC68xx-2
Cell Monitors allows multiple devices with different ad-
dressestobeconnectedonasinglebusbymulti-dropping
them. Multi-dropping can be used in either SPI or isoSPI
(See Figure 12 (A)). The LTC2949 also operates in paral-
lel to or as last element of a daisy chain of LTC68xx Cell
Monitors (See Figure 12 (B&C)).
aretransmittedt afterthereceivedpulse. TheLTC2949
RTN
isoSPI port is a slave port and only transmits short –1
pulses, never long CSB pulses nor short +1 pulses. The
master port recognizes a null response as a logic 1 (for
this reason a no reply to a read command is equivalent to
read only 0xFF... on the MISO line, which will also cause a
PEC error, see chapters DATA LINK LAYER and following).
+1 PULSE
ꢇ
ꢍꢉꢔꢍ
ꢄARꢓꢈꢉ
ꢇ
ꢄARꢓꢈꢉ
ꢀꢁ
ꢐꢈꢒꢂ
ꢂꢃꢄꢅ
ꢇ
ꢄARꢓꢈꢉ
V
IP
– V
ꢇ
ꢐꢈꢒꢂ
IM
ꢊꢋꢌꢅꢍ
ꢇ
ꢇ
ꢆꢁ
ꢈꢉꢁ
ꢊꢋꢌꢅꢍ
ꢂꢃꢄꢅ
–1 PULSE
ꢇ
ꢇ
ꢀꢁ
ꢈꢉꢁ
ꢊꢋꢌꢅꢍ
ꢂꢃꢄꢅ
ꢇ
ꢇ
ꢄARꢓꢈꢉ
ꢄARꢓꢈꢉ
V
IP
– V
ꢊꢋꢌꢅꢍ
ꢐꢈꢒꢂ
IM
ꢇ
ꢄARꢓꢈꢉ
ꢆꢁ
ꢐꢈꢒꢂ
ꢂꢃꢄꢅ
ꢇ
ꢍꢉꢔꢍ
ꢌꢎꢏꢎ ꢐꢊꢑ
Figure 10. isoSPI Pulse Detail
Rev A
25
For more information www.analog.com
LTC2949
SERIAL INTERFACES
ꢀꢁ
ꢀ
ꢀ
ꢀ
ꢎ
ꢁ
ꢃꢈꢄ
ꢀ
ꢀ
ꢀ
ꢍ
ꢏ
ꢌ
ꢂꢃꢄ ꢅꢆꢇꢈ ꢉ ꢊꢋ
ꢂꢃꢄ ꢅꢆꢇꢈ ꢉ ꢓꢋ
ꢀ
ꢐ
ꢀ
ꢊ
ꢑꢇꢂꢕ
ꢑꢕꢂꢇ
ꢕꢂꢇ
ꢃꢂꢛ ꢉ ꢓ
ꢑ
ꢚ
ꢑ
ꢚꢜꢊ
ꢑ
ꢇ
ꢃꢂꢛ ꢉ ꢊ
ꢂ
ꢂ
ꢕꢝꢚꢇRꢒꢘ
ꢚꢜꢊ
ꢚꢜꢐ
ꢕꢂꢇ
ꢀ
Rꢔꢚ
ꢂꢈAꢠꢒ ꢘꢇꢒꢂ ꢚꢇꢔ
ꢔRAꢚꢂꢑꢕꢔ ꢡꢊ
ꢂꢘꢕꢙꢕꢚꢔ
ꢂꢘꢇꢙꢕꢚꢔ
ꢂAꢑꢆꢈꢒ
ꢂAꢑꢆꢈꢒ
ꢓ
ꢏꢓꢓ
ꢊꢓꢓꢓ
ꢊꢏꢓꢓ
ꢐꢓꢓꢓ
ꢐꢏꢓꢓ
ꢍꢓꢓꢓ
ꢍꢏꢓꢓ
ꢌꢓꢓꢓ
ꢌꢏꢓꢓ
ꢏꢓꢓꢓ
ꢐꢞꢌꢞ ꢟꢊꢊ
ꢔꢕꢑꢒ ꢅꢖꢗꢋ
Figure 11. isoSPI Timing Diagram
LTC2949 in Addressable/Multidrop Bus Configuration
LTC2949 Connected to Reversible isoSPI Ring
An LTC2949 can be directly connected to the master in SPI
or through an LTC6820 in isoSPI mode. When operating
together with LTC68xx Cell Monitors it is recommended to
use the LTC2949 in a multidrop bus configuration to take
advantage of the full feature set and minimize communica-
tion overhead. The LTC2949 can operate on the same SPI/
isoSPIwithotherLTC68xxCellMonitors(Figure12(A,B)).
The LTC2949 responds to address 0xF. This address is
hardwired and cannot be changed. Consequently, other
LTC68xx Cell Monitors on the same bus must be pin-
configured to different addresses. Simultaneous writing
to all devices on one bus is done by issuing a broadcast
command.ThisfeatureprovesusefulforsynchronousADC
conversions of LTC2949 and LTC68xx Cell Monitors. The
common SDO pin is open drain and requires a pull-up.
LTC2949 can be connected to one end of a reversible
isoSPI ring. In this scenario the default communication
to LTC2949 is done with direct commands, equivalent
to the scenario where LTC2949 is connected in parallel
to a daisychain. In case the direct link between the left
side LTC6820 and LTC2949 fails, the communication to
LTC2949canberoutedthroughthedaisychainviatheright
side LTC6820, equivalent to the scenario where LTC2949
is connected on top of a daisychain.
LTC2949 on Top of a Daisy Chain
ItisrecommendedtooperateLTC2949inparalleltoadaisy
chain of LTC68xx Cell Monitors in either isoSPI or normal
SPI by using its addressing feature. This is the mode with
the minimum communication overhead. However, also
operation as last element of a daisy chain is supported.
Rev A
26
For more information www.analog.com
LTC2949
SERIAL INTERFACES
ꢀ
ꢀ
ꢀ
ꢀ
•
• •
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
•
• •
ꢀ
ꢀ
•
• •
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
•
• •
•
• •
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Rev A
27
For more information www.analog.com
LTC2949
SERIAL INTERFACES
As the LTC2949 has only one isoSPI port it must be
written or read from followed by the register data – see
following section Direct Read/Write Command (DCMD).
BesidetheDCMD,theLTC2949supportstheLTC68xxCell
Monitor compatible commands ADCV and RDCV.
th
th
placed as the last (M ) element in a daisy chain. The 0
element communicates to the master via port A which can
be configured as isoSPI or normal SPI mode, depending
on the connection of the ISOMOD pin. The 0th element
connects via port B to the 1st element of the daisy chain
using isoSPI, and so on. When the LTC68xx Cell Monitor
When LTC2949 is the last element in an isoSPI daisy
chain, the DCMD can be used to write to LTC2949 but not
to read data from the device, because the DCMD is not
supported by other LTC68xx Cell Monitors and thus they
do not pass data from their port B to port A. Therefore, the
RDCV command must be used to read from LTC2949 as
element of a daisy chain. LTC2949 registers read by the
RDCVcommandcanbeconfiguredbyaprecedingDCMD.
–
is operating with port A as SPI (ISOMD = V ), the SPI de-
tects one of four communication events: CSB falling, CSB
rising, SCK rising with SDI = 0, and SCK rising with SDI =
1. Each event is converted into one of the four pulse types
for transmission through the daisy chain. Long pulses are
used to transmit CSB changes and short pulses are used
to transmit data, as explained in Table 8. When both ports
are operated in isoSPI mode, isoSPI pulses on port A are
passed to port B within a short delay.
Table 10 summarizes the possibilities to communicate
to LTC2949.
Table 10. Communication with LTC2949
PARALLEL TO DAISY
CHAIN OR LTC2949 SOLE
LTC2949 ON TOP OF
DAISY CHAIN
DATA LINK LAYER
Read of
DCMD
Broadcast RDCV
All data transfers on LTC2949 occur in byte groups. Every
byteconsistsof8-bits. Bytesaretransferredwiththemost
significant bit (MSB) first. CSB must remain low for the
entiredurationofacommandsequence,includingbetween
a command byte and subsequent data. A write command
takes effect after a correct PEC is processed.
Registers
(RDCVCONF = 0, BCREN = 1)
Write of
Registers
DCMD
Trigger of
Fast ADC
Conversion
Addressed ADCV
Broadcast ADCV
Read of Fast
Addressed RDCV
Broadcast RDCV
Conversion (RDCVCONF = 1, BCREN = 0) (RDCVCONF = 1, BCREN = 1)
Results
NETWORK LAYER
The LTC2949 registers can be accessed by a direct read/
write command (DCMD) containing the registers to be
ꢂꢇꢀꢏꢐꢌꢑ
ꢒꢄꢊꢇ ꢓꢔ
ꢂꢇꢀꢏꢐꢌꢑ
ꢈAꢇꢇꢁRꢉ ꢊꢇAꢀꢋ
ꢃꢄꢅꢆꢇꢄR ꢂꢇꢀꢌꢍꢎꢍ
ꢀꢁꢂꢂ
ꢃꢄꢅꢆꢇꢄR
ꢀꢁꢂꢂ
ꢃꢄꢅꢆꢇꢄR
ꢀꢁꢂꢂ
ꢃꢄꢅꢆꢇꢄR
ꢀꢁꢂꢂ
ꢃꢄꢅꢆꢇꢄR
ꢌꢍꢎꢍ ꢕꢖꢗ
Figure 13. LTC2949 Connected to a Reversible isoSPI Ring
Rev A
28
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Fast Measurement Timings
The following timing diagrams shows LTC2949 fast single shot timing in relation to LTC6810’s timing of the ADCV
command for measuring all six cells. Other LTC681x Cell Monitors have similar timing diagram, just with different
number of ADCs and cells. The LTC681x devices have several ADC modes with different filter bandwidth and accuracy.
Typically, in the 7kHz normal mode, all cells are converted within a time window very close to LTC2949’s fast conver-
sion time which is nominal 782µs. For example, the LTC6810 converts all cells within 815µs in the normal mode.
As an example, only three (I2, BAT, AUX) of maximum four ADCs of LTC2949 are configured for fast conversions. The
displayed timings are valid for any allowed combinations of fast channels.
t
CYCLE,LTC681x,7kHz = 0.8ms
SERIAL
INTERFACE
ADCV + PEC
LTC6810
MEASURE
C1 TO C0
MEASURE
C2 TO C1
MEASURE
C3 TO C2
MEASURE
C4 TO C3
MEASURE
C5 TO C4
MEASURE
C6 TO C5
ADC1
CALIBRATE
t
0
t
t
t
3M
t
t
t
6M
t
C
1M
2M
4M
5M
I2ADC
P2ADC
MEASURE CURRENT
LTC2949
MEASURE STACK VOLTAGE
MEASURE AUX VOLTAGE
AUXADC
t
CYCLE,2949 = 0.8ms
2949 F14
Figure 14. Timing for ADCV Command Measuring Cell Voltages and LTC2949's Current and Voltage Inputs
ADCV
RDCV
ADC
I2
I2
ADC
ADC
BAT
BAT
(AUX)
AUX
t
2
t
1*
or t
1
t
3*
or t
3
2949 F15
Figure 15. Timing for LTC2949’s ADCV Command Measuring Current and Voltage Inputs.
DESCRIPTION
PARAMETER
VALUE/TOLERANCE
6 to 8μs
ADVC's last PEC byte to start of conversion latency without AUX conversion
ADVC's last PEC byte to start of conversion latency with AUX conversion
Conversion time
t
1
t
6 to 170µs
1*
t
742 to 821µs
855 to 945µs
2
3
ADVC's last PEC byte to HS = 0x0F (hand shake byte read by RDCV indicating conversion
results ready, see note below) without AUX
t
ADVC's last PEC byte to HS = 0x0F (hand shake byte read by RDCV indicating conversion
results ready, see note below) with AUX
t
855 to 1260µs
3*
Rev A
29
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Note: If the first HS byte of the RDCV data is 0x0F, the conversion results of the very same data set are already valid. If
the first HS byte of the RDCV data is 0x00, the conversion results of the very same data set are not valid. A new RDCV
must be issued to check for updated conversion results. If the first HS byte of the RDCV data is 0x00 the host can
continue to read HS bytes until it changes to 0x0F. The following RDCV command for sure will have valid conversion
results, still, the HS byte of the very same data set will be 0x00, because it was already internally cleared after it was
read 0x0F in the previous RDCV.
The following diagram shows details on LTC2949’s fast continuous conversion timing. Fast continuous operation
is started by a direct write command that sets bit FACONV and at least one of the channel bits (CH1, CH2, AUX) in
register FACTRL.
FIFO READ
FACONV = 1
RDCV
ADC
I2
•
•
•
•
•
•
•
•
•
I2
I2
I2
I2
ADC
BAT
BAT
AUX
BAT
AUX
BAT
AUX
BAT
AUX
ADC
AUX
2949 F16
t
or t
1
t
t
3
1*
2
Figure 16. Timing for LTC2949’s Fast Continuous Current and Voltage Measurements.
PARAMETER
DESCRIPTION
VALUE/TOLERANCE
6 to 8μs
Direct write's last PEC byte to start of conversion latency without AUX conversion
t
1
Direct write's last PEC byte to start of conversion latency with AUX conversion
Conversion time
t
6 to 170µs
1*
t
742 to 821µs
0 to 315µs
2
3
Any sample's start of conversion to HS = 0x0F (hand shake byte read by RDCV indicating
conversion results ready) or to sample available via FIFO read operation
t
Rev A
30
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Fast AUX Measurements
and disable FACONV periodically or to perform repetitive
FSSHT measurements it is recommended to configure
CH1 for slow and CH2 for fast mode, which ensures the
slow channel being updated periodically.
It is not possible to change the AUX MUX configuration
during fast continuous mode (FCM) measurements.
FAMUXP, FAMUXN can be written and the same value
read back at any time, but the internal MUX will only be
set to the requested configuration when receiving a new
fastconversionrequest.ThisiseithertheADCVcommand
for fast single shot (FSSHT) or the transition from 0 to 1
of the FACONV bit for FCM measurements.
Fast AUX Round-Robin Measurements
FSSHT measurements shall be used in applications that
require different inputs via the AUX MUX at high update
rates (faster than 100 ms). Optionally, the FSSHT mea-
surements can be surrounded by FCM periods, during
which a single MUX input is converted continuously. The
conversion results of the FCM periods can be read via the
FIFO registers or via RDCV. Reading of conversion results
from FIFOs is possible at any time. Still, the FSSHT trigger
command (ADCV while CONT= 1 and FACTRL unequal
zero) will clear all FIFOs.
When leaving the FCM (FACONV = 0), the host must wait
for the last conversion being completed before any new
fast conversion is triggered. This can be achieved by
waiting at least 1.26 ms.
If both channels are configured fast, it must be ensured
that either the FCM stays active for at least 128 samples
or FSSHT measurements are inhibited for at least 100ms.
Only after every 128 samples / 100 ms, the slow channel
registers (including STATUS, FAULTS, EXTFAULTS) are
updated. In applications in which it is required to enable
Following table shows an example sequence of 4 FSSHT
measurements that interrupt a FCM period. If only FSSHT
measurements are required, the rows CONT0, CONT1 can
be omitted.
NAME
MOSI / MISO
DESCRIPTION
CONT0
MOSI:FEF5EB50400EE4C6
MISO:XXXXXXXXXXXXXXXX
Write to FACTRL to disable FCM (FACONV = 0)
MUX0
ADCV
RDCV
MUX1
ADCV
RDCV
MUX2
ADCV
RDCV
MOSI:FEF3C7984500013D6E
MISO:XXXXXXXXXXXXXXXXXX
Write two bytes to FAMUXN to select V1 vs. GND
(NTC)
MOSI:FB60FADE
MISO:XXXXXXXX
ADCV to trigger conversion
MOSI:F8040970FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
MISO:XXXXXXXX010000000000FE4AE8180F0F0F0FC602
RDCV to read conversion results
MOSI:FEF3C798450016C1BA
MISO:XXXXXXXXXXXXXXXXXX
Write two bytes to FAMUXN to select VREF2 vs.
GND
MOSI:FB60FADE
MISO:XXXXXXXX
ADCV to trigger conversion
MOSI:F8040970FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
MISO:XXXXXXXX000000000000C212610F0F0F0F0F76B6
RDCV to read conversion results
MOSI:FEF3C7984517007512
MISO:XXXXXXXXXXXXXXXXXX
Write two bytes to FAMUXN to select GND vs.
VREF2_250k
MOSI:FB60FADE
MISO:XXXXXXXX
ADCV to trigger conversion
MOSI:F8040970FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
MISO:XXXXXXXX000000000000C212E7180F0F0F0F1A78
RDCV to read conversion results
Rev A
31
For more information www.analog.com
LTC2949
SERIAL INTERFACES
NAME
MOSI / MISO
DESCRIPTION
MUX3
MOSI:FEF3C7984500174A88
MISO:XXXXXXXXXXXXXXXXXX
Write two bytes to FAMUXN to select
VREF2_250k vs. GND
ADCV
MOSI:FB60FADE
MISO:XXXXXXXX
ADCV to trigger conversion
RDCV
MOSI:F8040970FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
MISO:XXXXXXXX000000000000C2121AE70F0F0F0F7C26
RDCV to read conversion results
MUXCONT
MOSI:FEF3C79845111294A6
MISO:XXXXXXXXXXXXXXXXXX
Write two bytes to FAMUXN to select CF2P vs.
CF2M (necessary to do the open wire check
during the FCM)
CONT1
MOSI:FEF5EB50400F6FF4
MISO:XXXXXXXXXXXXXXXX
Write to FACTRL to enable FCM (FACONV = 1)
Note 1: MISO data is only shown as an example and the actual data may change. ‘X’ indicates don’t care data.
Note 2: A delay of ≥ 1.26 ms is mandatory between any ADCV and RDCV to ensure valid conversion result readings.
Note 3: Above MUX settings convert one external pin voltage V1 vs. GND and the internal VREF2 in three different ways. This example and
can be adjusted and extended to any MUX configurations required.
Note 4: CONT0 and MUX0 can be merged to a single 3-write command to FAMUXN, FAMUXP, FACTRL (MOSI: 0xFEF3C7984600010E9516)
Note 5: When LTC2949 is in STANDBY state it can be moved to MEASURE state at any time by writing CONT=1. From the write command
it takes t
until the MEASURE state is active and then worst case 140ms until the first slow channel conversion results can be read
IDLE_CORE
form LTC2949's registers. First fast measurements can be performed, if FACTRL was configured appropriately, already when MEASURE
state is active.
Note 6: The following sequence allows a fast and efficient initialization of LTC2949 before any fast conversions can be performed:
1. Wakeup, configure ADCCONF and perform ADJUPD as required.
2. Write 0x08 (CONT) to OPCTRL
3. 3-Byte-Write to FAMUXN (e.g. 0x00 = GND), FAMUXP (e.g. 0x16 = VREF2), FACTRL (e.g. 0x02 = FACHA or another allowed combination
to perform any fast conversion)
4. Send any ADxx (e.g. ADCV, see table 17)
5. Wait at least 1.2 ms
6. Send any RDxx (e.g. RDCVA, see table 18). If HS byte is unequal 0x0F go back to 3.
7. Initialization is done. LTC2949 is now in MEASURE mode and any fast conversions can be performed subsequently (e.g. the sequence
from above table)
Rev A
32
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Direct Read/Write Command (DCMD)
To access the full register map of the LTC2949, a special direct command (DCMD) is provided which is not used by
other LTC68xx Cell Monitors. DCMD allows to read/write arbitrary number of bytes from/to LTC2949's register map.
The LTC2949 auto-increments its address pointer after each data byte, so multiple registers can be written/read as part
of a single transaction. Data packets from one to up to 16 bytes are interleaved by a two-byte data PEC. Data packets
written without PEC are discarded. Read commands may stop at any byte.
Table 11. Direct Read/Write Command Format
Byte
0
1
2
3
4
5
…
N+4
N+5
N+6
N+7
…
2N+6
2N+7 2N+8
…
R/W
Master to slave (MOSI)
Write commands: Master to slave (MOSI)
Read commands: Slave to master (MISO)
Name DCMD RADDR PEC0 PEC1 ID
DATA
…
DATA
PEC0 PEC1 DATA
…
DATA2
N-1
PEC0 PEC1
…
0
N-1
N
ID byte is used to distinguish between read and write commands and it defines the number of data bytes per PEC (parameter N in above table). The ID
byte is not part of any PEC, instead it has intrinsic error detection via redundancy and error check bits. See following tables for details.
Table 12. Bit Definitions of Byte ID[7:0]
ID[7]
ID[6]
ID[5]
ID[4]
ID[3]
ID[2]
ID[1]
ID[0]
RW
NOT RW PECC[3] XOR PECC[2]
PECC[3] PECC[2] PECC[1] XOR PECC[0]
PECC[1]
PECC[0]
Table 13. ID[7:0] Byte Format Description
NAME
DESCRIPTION
RW
ID[7] = RW indicates read (RW=1) or write (RW=0) commands. For safe data transmission redundancy is added to ID[6] which
is the inverse of RW (NOT RW, meaning 0 for read and 1 for write).
PECC[3:0]
PEC Configure determines the number of data bytes after which a PEC is transmitted. Number of data bytes is
DECIMAL(PECC[3:0])+1 (parameter N in above table). Allowed values for PECC are 0 to 15 (1 to 16 data bytes per PEC). ID[5]
(= PECC[3] XOR PECC[2]) and ID[2] (=PECC[1] XOR PECC[0]) are error check bits for safe data transmission.
Table 14. Direct Read/Write Command Format Details
NAME
DESCRIPTION
DCMD[7:0]
RADDR[7:0]
PEC0,1
Direct command. It is fixed to the value 0xFE.
Starting register address from which data is read or to which data is written.
The Packet Error Code Bytes PEC0 and PEC1 hold a 15-bit CRC according to CAN BUS CRC15 which is right padded with
0 (see Table 22. Write/Read PEC Format). The PEC in bytes 2-3 is calculated on DCMD and RADDR. All following PECs are
calculated on preceding M data-bytes. The number of data bytes per PEC is defined by the ID byte, which is not part of any
PEC, see below.
For read commands the MOSI line is don’t care, the slave will send the data PEC on its MISO line and the master must
compare it to the PEC calculated on the data received. The read command was successful of both PECs match. For
write commands, the master must send the PEC on its MOSI line and LTC2949 will compare the received PEC with its
internal calculated PEC. In case of a PEC mismatch the data will be discarded and an external communication PEC error
(EXTCOMMERR) will be flagged in the FAULTS register.
DATA
Data bytes to be send to or read from LTC2949’s register map. The starting address is given by RADDR and is auto-
incremented for every data byte.
X
Rev A
33
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Commands Compatible with LTC68xx Cell Monitors
(RDCV, ADCV)
acquired samples can be read from the FIFO registers (FI-
FOI1, FIFOI2, FIFOBAT, FIFOAUX). Still, also in this mode,
it is possible to read the last fast conversion results via
RDCV(RDCVCONF=1).Iffastcontinuousmodeisdisabled
(FACONV=0), any ADCV command will clear all FIFOs.
If needed, samples within FIFOs should be read before
sending an ADCV command while FACONV is cleared.
The LTC2949 supports several RDCV and ADCV style
commands compatible with LTC68xx Cell Monitors.
RDCVA-RDCVF,RDAUXA-RDAUXD,RDCFGA,RDCFGB(all
referred as RDCV commands in this document), ADCV,
ADOW, ADOL, ADAX, ADAXD, ADCVAX and ADCVSC (all
referred as ADCV commands in this document) are sup-
portedasbroadcastandaddressedcommands.Addressed
commands are used to address LTC2949 exclusively. To
trigger actions on LTC2949 and LTC68xx Cell Monitors
simultaneously, broadcast commands are used. This is
forexampleusefultoinitiatingADCconversionsofseveral
devices at the same time. No matter if LTC2949 is con-
nected parallel to or on top of a daisy chain, addressed
ADCV commands can trigger measurements on LTC2949
only and broadcast ADCV commands can perform syn-
chronous ADC conversions on all devices connected to
the SPI / isoSPI bus. The conversion starts at the end of
the PEC for all devices.
In the configuration where LTC2949 is on top of a daisy
chain, after a broadcast RDCV command, the stacked
LTC68xxCellMonitorsturnintoacascadedshiftregister,in
whichdataisshiftedthrougheachdevicetothenextdevice
in the stack. In this scenario, LTC2949 responds to RDCV
commands only if the Broadcast Read Enable bit (BCREN)
in the Register Control Register (REGSCTRL) is set.
In the scenario where LTC2949 is placed in parallel to
a daisy chain of LTC68xx Cell Monitors, the bit BCREN
must be cleared (default) to avoid bus collisions. Still,
any broadcast RDCV may clear LTC2949's internal HS
byte (see below) if the bit RDCVCONF is set, no matter if
BCREN is set or not. For this reason, it is recommended
to read LTC2949’s fast conversion results before reading
from the LTC68xx Cell Monitors. Or, if for software tim-
ing reasons it is necessary to read from the LTC68xx Cell
Monitors first, the bit RDCVCONF must be cleared before
(it must be set again afterwards to read fast conversion
data from LTC2949).
In cases where LTC2949 is connected parallel to a daisy
chain, direct read commands (DCMD) must be used
to read register data and addressed RDCV commands
(with RDCVCONF=1) must be used to read the last fast
conversion results. If LTC2949 is connected on top of a
daisy chain, broadcast RDCV commands must be used
to read register data (RDCVCONF=0) or to read the last
fast conversion results (RDCVCONF=1) depending on
setting of REGSCTRL.RDCVCONF. Write to LTC2949’s
register map is always done with direct write commands
(DCMD) independent of the bus topology. See Table 10
(Communication with LTC2949) for a summary of all
possible communication scenarios.
The last results of any fast conversion can be read out by
the RDCV command, providing sequentially the results
of I1, I2, BAT and AUX (least significant bytes first to be
compatible with LTC68xx Cell Monitors) followed by one
or more hand shake (HS) bytes, indicating if the data is
new (0x0F) or old (0x00). Once LTC2949 sends the HS
byte, it will continue sending it as long as the master is
reading bytes. Still, a PEC will always be send for every 6
data bytes. Per transaction, the HS byte may only change
from 0x00 to 0x0F, once it is 0x0F it won’t change. This
allows the master to poll for conversion results being
ready by checking for a transition of the HS byte from
0x00 to 0x0F. If the first HS byte is 0x0F the conversion
data received with that command is already new and valid.
If the first HS byte is 0x00 the data received in the same
transaction was not yet updated. A subsequent RDCV
Before LTC2949 can react on ADCV commands it must be
runninginslowcontinuousmeasurementmode(bitCONT
in OPCTRL), at least one fast measurement channel must
be selected via the Fast Control Register (FACTRL) and
optionally, to make also fast BAT conversions via one of
the power ADCs (P1 or P2 in voltage mode), P1ASV and/
or P2ASV in the ADC configuration register (ADCCONF)
must be set. If LTC2949 is running in fast continuous
mode (FACONV=1) any ADCV command is ignored and
Rev A
34
For more information www.analog.com
LTC2949
SERIAL INTERFACES
command is necessary to read the updated conversion
results, still, the HS byte will be reported as 0x00 for this
RDCV command, as it was internally already cleared after
it was read 0x0F with the previous RDCV.
Forfastcontinuousmeasurement,thesamplerateisfixedto
1.25kspsandthementioneddelayisjustalatencybetween
the actual measurements and the time the samples are
available to be read. The master shall wait at least 1.26ms
between the time fast continuous mode is enabled and the
firsttimeanyFIFOregisterisread, toallowthefirstsample
to be read from the FIFO. Alternatively, it is also possible
to use a RDCV command to check when the first sample
is ready (see HS byte above) and then read all samples
from the FIFO registers periodically.
LTC2949’s fast conversion time is typically 0.8ms. Addi-
tional processing time is necessary for the results being
ready to be read by the master. Worst case, 1.26 ms after
a conversion was triggered, the results can be read via a
FIFO register (in case of fast continuous mode) or via the
RDCV command. For fast single shot measurements, this
limits the maximum guaranteed sample rate to ~0.8ksps.
Table 15. Format of ADCV/RDCV Style Commands. Only for RDCV Commands the Slave Sends Data to the Master on the MISO Line
BYTE
0
1
2
3
4
…
9
10
11
12
…
17
18
19
…
R/W
Master to slave (MOSI)
Slave to master (MISO)
DATA6 PEC0 PEC1 DATA7
Name CMD0 CMD1 PEC0 PEC1 DATA0
…
…
DATA11
PEC0 PEC1
…
CMD0 and CMD1 are the command bytes. The format for the commands is shown in Table 16. CC[10:0] is the 11-bit command code. A list of
supported command codes is shown in Table 17 and Table 18. Broadcast commands have a value 0, addressed commands have a value 1 for
CMD0[7] through CMD0[3]. The PEC must be computed on the entire 16-bit command (CMD0 and CMD1).
Table 16. CMD0, CMD1 Command Bytes Format. Data is Send from Master to Slave (MOSI). A/B is 0 for Broadcast and 1 for
Addressed Commands.
NAME
CMD0
CMD1
BIT 7
A/B
BIT 6
A/B
BIT 5
A/B
BIT 4
A/B
BIT 3
A/B
BIT 2
CC[10]
CC[2]
BIT 1
CC[9]
CC[1]
BIT 0
CC[8]
CC[0]
CC[7]
CC[6]
CC[5]
CC[4]
CC[3]
Rev A
35
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Table 17. ADCV Style Commands. LTC2949 Performs a Fast Conversion Depending on FACTRL/ADCCONF Registers Upon Those
Commands
NAME
ADCV
CMD0[7:3]
CC[10:0]
DESCRIPTION
Broadcast: all 0
Addressed: all 1
0 1 x x 1 1 x 0 x x x
0 1 x x x 1 x 1 x x x
0 1 x x 0 0 x 0 0 0 1
1 0 x x 1 1 0 0 x x x
1 0 x x 0 0 0 0 x x x
1 0 x x 1 1 x 1 1 1 1
1 0 x x 1 1 x 0 1 1 1
LTC2949 reacts to all commands in the same way by triggering a fast
conversion (if enabled via FACTRL register).
ADOW
ADOL
ADAX
ADAXD
ADCVAX
ADCVSC
Table 18. RDCV Style Commands. Used to Read Fast Conversion Results or for Indirect Memory Map Reads from LTC2949
NAME
RDCFGA
RDCFGB
RDCVA
CMD0[7:3]
CC[10:0]
DESCRIPTION
Broadcast: all 0
Addressed: all 1
0 0 0 0 0 0 0 0 0 1 0
0 0 0 0 0 1 0 0 1 1 0
0 0 0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 0 1 1 0
0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 1 0 1 0
0 0 0 0 0 0 0 1 0 0 1
0 0 0 0 0 0 0 1 0 1 1
0 0 0 0 0 0 0 1 1 0 0
0 0 0 0 0 0 0 1 1 1 0
0 0 0 0 0 0 0 1 1 0 1
0 0 0 0 0 0 0 1 1 1 1
LTC2949 reacts to all commands in the same way by either transmitting fast
conversion results (default, RDCVCONF=1) or by transmitting register data
(RDCVCONF=0)
RDCVB
RDCVC
RDCVD
RDCVE
RDCVF
RDAUXA
RDAUXB
RDAUXC
RDAUXD
Rev A
36
For more information www.analog.com
LTC2949
SERIAL INTERFACES
The RDCV command allows reading data packets much longer than the daisy chain shift register. The length of the
shift register is 64 • M bits, whereas M is the number of elements in the chain excluding LTC2949. Data from the
LTC2949 is received by the master after 64 • M data bits starting with bit I1[7] according to the figure below. A PEC
is calculated after every 6 data bytes. After the transmission of the conversions results, the LTC2949 continuously
sends the hand shake byte (HS) indicating if the data is new (0x0F) or old (0x00). A transition of the hand shake
byte from 0x00 to 0x0F indicates the arrival of new data, which can be read out by a subsequent RDCV command.
Table 19. RDCV Command Format. CMD0, CMD1 According to Table 18 (RDCV Style Commands). Requires RDCVCONF=1.
CMD0
BAT[15:8]
PEC0
CMD1
PEC0
PEC1
HS
PEC0
PEC1
HS
PEC1
AUX[7:0]
HS
I1[7:0]
AUX[15:8]
HS
I1[15:8]
I2[7:0]
I2[15:8]
BAT[7:0]
HS
HS
HS
HS
HS
HS
HS
PEC0
PEC1
HS
HS
HS
HS
HS
……
Note: The conversion results I1, I2, BAT, AUX are converted to volts by multiplication with the LSB size 7.60371 µV for the current and 375.183 µV for the
BAT and AUX channel.
Indirect Memory Access RDCV Command
DCMD is the recommended way of reading data from the LTC2949 in addressable mode. When LTC2949 is the last ele-
ment in a daisy chain, an RDCV has to be used to read data from the LTC2949, as DCMD is not supported by LTC68xx
Cell Monitors and therefore does not configure them as shift registers, In default, the LTC2949 will respond to RDCV
with the fast mode conversion results as described above (RDCVCONF = 1).
In order to gain access to the entire register map of LTC2949 an address pointer can be set causing the LTC2949 to
provide data starting at that register address on subsequent RDCV commands. To use this indirect memory access
RDCV command, the RDCV Configuration Bit (RDCVCONF) has to be reset and Broadcast Read Enable bit (BCREN) has
to be set in REGSCTRL and the starting pointer must be written to the RDCV Indirect Address Register (RDCVIADDR).
In this way any register can be read by subsequent RDCV commands. The address pointer is auto-incremented after
every data byte for reading data bursts of any length. A PEC is transmitted after every six data bytes. Once the REGSC-
TRL is written accordingly, only the Indirect Address Register must be updated to read from other memory locations.
PleasenotethattheRDCVIADDRcanalsobewrittenbeforetheREGSCTRLifasingleDCMDspanningfromRDCVIADDR
to REGSCTRL is used. For this single write burst, the two bytes between RDCVIADDR and REGSCTRL are don't care
and can be written 0x00.
ꢊꢆꢅ ꢋꢁRꢆꢌ ꢍ
Rꢆꢊꢆꢅ Rꢀꢁꢂꢁꢎꢌꢉ
Aꢅ ꢇꢈꢉ ꢏ ꢋꢄꢅ ꢐꢑꢒ ꢍ ꢐꢓꢒ
ꢃRꢄꢅꢆ
RꢀꢁꢂꢄAꢀꢀR
Aꢅ ꢇꢈꢉꢁ
Rꢀꢁꢂ
ꢑꢔꢕꢔ ꢉꢖꢓ
Figure 17. Indirect Memory Access Read Procedure Using RDCV Command
Table 20. Indirect Address RDCV Command Format, CMD0, CMD1 According to Table 18 (RDCV Style Commands), Requires
RDCVCONF = 0 and BCREN = 1
CMD0
CMD0
PEC0
PEC1
PEC0
PEC1
……
PEC1
DATA
DATA
DATA
DATA
DATA
0
1
2
3
4
DATA
DATA
DATA
DATA
DATA
DATA
DATA
5
6
7
8
9
10
11
PEC0
DATA is the content of the register at the starting address that was written to RDCVIADDR. DATA is the content of
0
1
the register at the following address and so on.
Rev A
37
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Packet Error Code
4. Go back to step 2 until all the data is shifted. The final
PEC (16-bits) is the 15-bit value in the PEC register
right padded with 0. An example to calculate the PEC
for a 16-bit word (0x0001) is listed in Table 21. The
PEC for 0x0001 is computed as 0x3D6E after stuffing
a 0-bit at the LSB. For longer data streams, the PEC is
valid at the end of the last bit of data sent to the PEC
register. LTC2949 calculates PEC for any command
or data received and compares it with the PEC follow-
ing the command or data. The command or data is
regarded as valid only if the PEC matches. LTC2949
also attaches the calculated PEC at the end of the data
it shifts out. Table 22 shows the format of PEC while
writing to or reading from LTC2949.
The packet error code (PEC) is a 15-bit cyclic redundancy
check (CRC) value calculated for all of the bits in a reg-
ister group in the order they are passed, using the initial
PEC seed value of 000000000010000 and the following
characteristic polynomial: x15 + x14 + x10 + x8 + x7 +
x4 + x3 + 1. To calculate the 15-bit PEC value, a simple
procedure can be established:
1. Initialize the PEC to 000000000010000 (PEC is a 15-
bit register group)
2. For each bit DIN coming into the PEC register group,
set
IN0 = DIN XOR PEC [14]
IN3 = IN0 XOR PEC [2]
IN4 = IN0 XOR PEC [3]
IN7 = IN0 XOR PEC [6]
IN8 = IN0 XOR PEC [7]
IN10 = IN0 XOR PEC [9]
IN14 = IN0 XOR PEC [13]
3. Update the 15-bit PEC as follows
PEC [14] = IN14,
PEC [13] = PEC [12],
PEC [12] = PEC [11],
PEC [11] = PEC [10],
PEC [10] = IN10,
PEC [9] = PEC [8],
PEC [8] = IN8,
PEC [7] = IN7,
PEC [6] = PEC [5],
PEC [5] = PEC [4],
PEC [4] = IN4,
PEC [3] = IN3,
PEC [2] = PEC [1],
PEC [1] = PEC [0],
PEC [0] = IN0
Rev A
38
For more information www.analog.com
LTC2949
SERIAL INTERFACES
Table 21. PEC Calculation for 0x0001
PEC[14]
PEC[13]
PEC[12]
PEC[11]
PEC[10]
PEC[9]
PEC[8]
PEC[7]
PEC[6]
PEC[5]
PEC[4]
PEC[3]
PEC[2]
PEC[1]
PEC[0]
IN14
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
3
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
4
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
5
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
9
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
0
10
1
0
0
0
1
0
1
1
0
0
1
1
0
0
1
1
1
0
1
0
1
1
0
11
1
0
0
1
1
1
0
1
0
1
0
1
0
1
1
1
0
0
1
0
1
1
0
12
1
0
1
1
0
0
0
1
1
0
0
1
1
1
1
1
1
0
0
0
0
1
0
13
1
1
1
0
1
0
0
0
0
0
0
0
1
1
1
0
1
1
1
1
0
1
0
14
0
1
0
1
1
0
1
1
0
0
1
0
1
1
1
0
1
0
1
1
0
1
1
15
0
0
1
1
1
1
0
1
0
1
1
0
1
1
1
0
0
1
1
1
1
0
1
0
1
1
0
1
1
1
0
IN10
PEC Word
IN8
IN7
IN4
IN3
IN0
DIN
Clock Cycle
16
Table 22. Write/Read PEC Format
NAME
PEC0
PEC1
RD/WR
RD/WR
RD/WR
Bit 7
Bit 6
Bit 5
Bit 4
Bit 3
Bit 2
Bit 1
Bit 0
PEC[7]
0
PEC[14]
PEC[6]
PEC[13]
PEC[5]
PEC[12]
PEC[4]
PEC[11]
PEC[3]
PEC[10]
PEC[2]
PEC[9]
PEC[1]
PEC[8]
PEC[0]
Improved PEC Calculation
derived PEC calculation method. There are two functions,
thefirstfunctioninit_PEC15_Table()shouldonlybecalled
once when the microcontroller starts and will initialize a
PEC15 table array called pec15Table[]. This table will be
used in all future PEC calculations. The pec15 table can
also be hard coded into the microcontroller rather than
running the init_PEC15_Table() function at startup. The
pec15() function calculates the PEC and will return the
correct 15-bit PEC for byte arrays of any given length.
The PEC allows the user to have confidence that the serial
data read from the LTC2949 is valid and has not been cor-
ruptedbyanyexternalnoisesource.Thisisacriticalfeature
for reliable communication and the LTC2949 requires that
aPECbecalculatedforalldatabeingreadfromandwritten
to the LTC2949. For this reason it is important to have an
efficient method for calculating the PEC. The code below
demonstrates a simple implementation of a lookup table
Rev A
39
For more information www.analog.com
LTC2949
SERIAL INTERFACES
/************************************
Copyright 2012 Linear Technology Corp. (LTC)
Permission to freely use, copy, modify, and distribute this software for any
purpose with or without fee is hereby granted, provided that the above
copyright notice and this permission notice appear in all copies:
THIS SOFTWARE IS PROVIDED “AS IS” AND LTC DISCLAIMS ALL WARRANTIES
INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO
EVENT SHALL LTC BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM ANY USE OF SAME, INCLUDING
ANY LOSS OF USE OR DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE
OR OTHER TORTUOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
PERFORMANCE OF THIS SOFTWARE.
***********************************************************/
int16 pec15Table[256];
int16 CRC15_POLY = 0x4599;
void init_PEC15_Table()
{
for (int i = 0; i < 256; i++)
{
remainder = i << 7;
for (int bit = 8; bit > 0; --bit)
{
if (remainder & 0x4000)
{
remainder = ((remainder << 1));
remainder = (remainder ^ CRC15poly)
}
else
{
remainder = ((remainder << 1));
}
}
pec15Table[i] = remainder&0xFFFF;
}
}
unsigned int16 pec15 (char *data , int len)
{
int16 remainder,address;
remainder = 16;//PEC seed
for (int i = 0; i < len; i++)
{
address = ((remainder >> 7) ^ data[i]) & 0xff;//calculate PEC table address
remainder = (remainder << 8 ) ^ pec15Table[address];
}
return(remainder*2);//TheCRC15hasa0intheLSBsothefinalvaluemustbemultipliedby2
}
Rev A
40
For more information www.analog.com
LTC2949
REGISTER MAP
Rev A
41
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Register Naming Conventions
RW Read-Write
DEF Default Value
RO
SO
Read Only
SI
UI
Signed Integer
Set Only (Note)
Unsigned Integer
Note: Write 1 to SO bits to request the associated action. Successful write can be checked by a read command directly following the write command. After
the action was performed the SO bit will be cleared automatically by LTC2949. Do not write to registers with SO bits, before all SO bits were read as 0. The
typical processing time per action is t
or t
depending on operation mode, see related bit description for details.
IDLE_CORE
CONT
MEMORY MAP AND PAGING MECHANISM
The memory map of the LTC2949 is organized in two pages, PAGE0 and PAGE1. PAGE0 contains all slow channel
result quantities, control and status registers while PAGE1 contains all threshold and configuration registers. Each
page has a register address space ranging from 0x00 to 0xEF, with each register consisting of one 8-bit byte of data.
Registers 0xF0 to 0xFF are common to both register pages. Register REGSCTRL at 0xFF is part of this range and is
used to switch between pages, see below. For clarity, all register addresses on PAGE1 are expressed as p1.0xYY in the
following while addresses on PAGE0 are simply referred to as 0xYY.
Multiple-byte data is stored with most significant byte at the lowest address (little-endian). For instance, the MSB
C1[47:40] of the quantity C1 is stored at address 0x00 in PAGE0.
Note that reading data from LTC2949's memory map (no matter if using direct command or indirect memory ac-
cess RDCV command with RDCVCONF=0) reports MSBytes first, while reading fast conversion results via RDCV
(RDCVCONF=1) reports LSBytes first.
Some addresses in the register map are not used and are reserved. Bits in non-reserved registers that are not explic-
itly described are also reserved. Writing to unused reserved registers or reserved bits in non-reserved registers may
result in unwanted behavior of the LTC2949, reserved bits in non-reserved registers should be written as 0; reading of
unused registers is generally harmless but will return random data. If software detection of device revision is neces-
sary, then contact the factory for details.
Register Control Register
The Register Control Register (0xFF) selects the active memory page, allows to configure the LTC2949 to respond to
broadcast read commands, configures the RDCV command to indirect memory access mode and provides a memory
locking mechanism.
Table 23. Control Register REGSCTRL (0xFF)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
0
PAGE
RW
0
Memory Map Page Select
0: PAGE0 of the memory map is selected.
1: PAGE1 of the memory map is selected.
2
BCREN
RW
RW
0
Broadcast Read Enable
0: LTC2949 does not respond to broadcast read commands
1: LTC2949 responds to broadcast read commands
[5:4]
MLK[1:0]
00
Memory Lock
00: Memory not locked
01: Memory lock requested by master
10: Memory locked
7
RDCVCONF
RW
1
RDCV Configuration Bit
0: Indirect memory access mode. RDCV will report data starting at address written to
RDCVIADDR (0xFC).
1: RDCV command will report latest fast channel conversion results.
Rev A
42
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
LTC2949 provides a mechanism to lock the memory to
keep coherency between registers when accessing the
memory.Amemorylockingcanberequestedbythemaster
by setting bits MLK[1:0] to 01. Now the LTC2949 updates
its registers, e.g. with measurement results. During this
time, read and write access to the LTC2949 memory apart
from registers REGSCTR and RDCVIADDR is blocked.
After all registers are updated by the LTC2949, it sets
MLK[1:0] to 10 to indicate to the Master that memory is
locked. The LTC2949 does not update the memory map
any more until the master unlocks the memory by writing
MLK[1:0] to 00.
ꢀRꢁꢂꢃ
ꢄꢅꢆꢇꢈꢉꢊꢋꢌꢊꢈ
ꢁꢍꢎꢂꢃAꢏ ꢐꢑ ꢒꢐꢅꢅꢁꢍꢓ ꢄꢅꢆ ꢁꢂ ꢁꢎ Aꢅꢎꢐ
ꢒꢐꢎꢎꢁꢔꢅꢃ ꢂꢐ ꢕꢖꢎꢂ ꢀAꢁꢂ ꢗꢊꢘꢙ ꢁꢍ
ꢚAꢎꢃ ꢐꢑ ꢎꢂAꢍꢏꢔꢛ ꢄꢐꢏꢃ ꢐR ꢈꢜꢊꢘꢙ
ꢁꢍ ꢚAꢎꢃ ꢐꢑ ꢄꢃAꢎꢖRꢃ ꢄꢐꢏꢃ
ꢒꢐꢅꢅ ꢖꢍꢂꢁꢅ
ꢄꢅꢆꢇꢈꢉꢊꢋꢌꢈꢊ
ꢄꢃꢄꢐRꢛ
ꢅꢐꢚꢆꢃꢏ
Aꢚꢚꢃꢎꢎ
ꢄꢃꢄꢐRꢛ
The LTC2949’s internal memory is still updated while the
memory is locked and thus accumulation of values and
updating of alerts is not interrupted. Once the memory is
locked the master can read consistent data even by single
byte access. Data does not change between those single
read transactions. It is also possible to manipulate result
parametervaluese.g.tosetacertainstartvalueforCharge
ꢀRꢁꢂꢃ
ꢄꢅꢆꢇꢈꢉꢊꢋꢌꢊꢊ
ꢝꢞꢗꢞ ꢑꢈꢞ
Figure 19. Memory Locking
or Energy etc. Coherency within a single row of 16 bytes
(register ranges 0x00 to 0x0F, 0x10 to 0x1F, ... and 0xE0
to 0xEF) is always guaranteed for multiple byte read and
write bursts and does not require memory locking.
Table 24. Operation Control OPCTRL (0xF0)
BIT
SYMBOL TYPE DEFAULT OPERATION
0
SLEEP
RW
SO
SO
0
0
0
0: Normal operation
1: SLEEP. The LTC2949 will exit SLEEP state if the pin CSB is pulled low in SPI mode or if a wake-up pulse
followed by another "Long -1" pulse is sent in isoSPI mode.
1
2
CLR
1: Clear. The Accumulation and Tracking (max/min) registers are cleared: C1, E1, TB1, C2, E2, TB2, C3, TB4, E4,
TB4 IMAX, IMIN, PMAX, PMIN, VMAX, VMIN, TEMPMAX, TEMPMIN, VDVCCMAX, VDVCCMIN, SLOTXMAX,
SLOTXMIN
SSHOT
1: Single Shot Measurement. LTC2949 enters state MEASURE and returns to state STANDBY after completion
of a single set of measurements of Current, Voltage, Power, Temperature, V , SLOT1, SLOT2, V . The result
CC
REF
registers are updated and SSHOT is cleared when returning to STANDBY. If CONT is set, it is cleared after
completion of any conversion cycle in progress and the single shot measurement is executed.
3
5
CONT
RW
SO
0
0
0: Continuous measurement is disabled
1: Continuous measurement is enabled after < t
. Measurement cycles run continuously. Charge, Energy
IDLE_CORE
and Time measurements are only active in Continuous mode.
ADJUPD
To insure data coherency, changes in the 2nd page's configuration registers, except threshold settings,
only become effective after an update procedure, which is initialized by setting bit ADJUPD. Once the new
configuration values are valid, LTC2949 resets the bit ADJUPD. Changes to the threshold registers may be done
at any time and don't require the ADJUPD procedure.
1: Request an update of configuration registers on 2nd memory page, except threshold settings.
0: Update done
7
RST
SO
0
0: Normal operation, 1: Reset device. As default the reset feature is locked and writing 1 to this bit has no effect.
See RSTUNLCK for procedure how to reset the device.
Note 9: ADJUPD shall be issued when in STANDBY mode only. The recommended implementation is to set ADJUPD as the last action of the initialization
routine, before entering continuous mode. Once ADJUPD is set it takes a maximum of 100ms until LTC2949 has finished the internal update process and the
bit ADJUPD is cleared automatically. Thus it is also possible to poll the bit ADJUPD for being cleared to indicate when the operation is done.
In cases where it is necessary to assert ADJUPD after CONT was enabled, it is necessary to clear the bit CONT before, wait 100ms to ensure that all
measurement cycles have completed and then assert ADJUPD. After ADJUPD was cleared automatically, the Continuous Mode may be entered again.
To avoid the 100ms wait time when going to STANDBY mode, it is also possible to clear CONT and set CLR at the same time and poll for the bit CLR being
cleared by LTC2949. This indicates also the end of any previously ongoing measurement cycle.
Rev A
43
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Reading of status (0x80), faults (0xDC-0xDD) and alert (0x81-0x87) registers may always be done without memory
locking. Still, once it is necessary to clear those registers, memory locking is mandatory to avoid missing any alert
and faults reporting. If an alert condition occurs while the memory is locked, LTC2949 will set the corresponding bit
after the memory is unlocked by the host. This rule must be followed independent of the core state, as certain faults
may also be raised in STANDBY mode.
Any fast conversions are not affected by the memory locking mechanism, thus it is still possible to change FAMUX
settings, trigger fast single shots, read results via RDCV. Also reading conversion results from the FIFOs during fast
continuous measurements is still possible.
Operation Control Register
The Operation Control Register OPCTRL (0xF0) controls LTC2949's transitions between its CORE States: SLEEP,
STANDBY, and MEASURE. Furthermore, it allows to clear accumulation and tracking registers, to validate changes of
the configuration registers and to reset the LTC2949.
REGISTER MAP PAGE0
Wake Up Acknowledge
The LTC2949 automatically returns to SLEEP state if no wake up confirmation command is received within 1 second
after entering STANDBY state. Wake up confirmation can be either writing 0x00 to register 0x70 or starting of a mea-
surement. Before wake up confirmation, WKUPACK will report a countdown from 0xFF to 0x00 within approximately
1 second. Countdown will stop and WKUPACK will statically report 0x00 after wake up confirmation.
ꢀAꢁꢂꢃꢄꢅ ꢀꢆꢇꢈ ꢉꢆRꢂꢊꢇ
RꢂAꢉ ꢊꢋꢌꢌAꢍꢉꢎ
ꢀAꢁꢂꢃꢄꢅ ꢀꢆꢇꢈ ꢆꢉꢊꢆRꢂꢋꢇ
RꢂAꢊ ꢋꢌꢍꢍAꢉꢊꢎ
ꢀAꢁꢂꢃꢄꢅ ꢅRꢆꢇꢂꢈꢄRꢂ
ꢀꢁꢂꢃ ꢃꢄꢅꢅꢆ
ꢇꢆꢈꢁ ꢈꢉ ꢊAꢋꢁꢌꢄꢍ ꢎꢏꢐꢀꢍꢑ
ꢀꢁꢁꢂ
ꢀꢁꢁꢂ
ꢀꢁꢂꢃ ꢃꢄꢅꢅꢆ
ꢇꢆꢈꢁ ꢈꢉ ꢊAꢋꢁꢌꢄꢍ ꢎꢉRꢁ
ꢀꢁꢁꢂ
ꢀꢁꢁꢂ
ꢀRꢁꢂꢃ ꢄꢅꢆꢄ ꢂꢇ
RꢈꢉꢊꢁAꢈꢈR Aꢋꢈ ꢄꢅꢄꢌ ꢂꢇ
RꢃꢍꢎꢉꢂRꢏ ꢊꢁA ꢌꢐꢑꢒꢂꢃ
ꢀRꢁꢂꢃ ꢉꢇꢓꢓAꢋꢈ
ꢀRꢁꢂꢃ RꢃꢄꢅꢆꢂRꢇ ꢀꢁꢂꢈ ꢉAꢄꢃ ꢊ ꢋ
Aꢌꢍ RꢍꢆꢎꢆꢏꢌꢐꢑꢒꢆRꢃꢌ
AꢆꢆꢏRꢍꢁꢌꢄ ꢂꢏ ꢍꢁRꢃꢆꢂꢑꢁꢌꢍꢁRꢃꢆꢂ
RꢃAꢍ ꢓꢏꢍꢃ
ꢀꢁꢂꢁꢁꢃꢂꢄꢅꢆꢄꢅꢄꢇꢆꢈꢇꢉ
ꢀꢁꢂꢁꢃꢄꢅꢁꢆꢇꢃꢃꢃꢈꢃꢃꢃꢉ
ꢀꢁꢂꢁꢃꢄꢄꢅꢃꢆꢇꢁꢈꢈꢈꢈꢈꢈꢆꢇꢉAꢉꢊ
ꢀꢁꢁꢁꢂꢁꢃꢄꢅꢁꢁꢁꢁꢁꢁꢁꢁꢁꢁꢁꢁꢄꢅꢆꢅꢇ
ꢀꢁꢂꢁꢃꢄꢅꢁꢆꢇꢃꢃꢅAꢈꢉꢊꢋ
ꢀꢁꢁꢁꢂꢁꢃꢄꢅꢆꢆꢆꢆꢆꢆꢆꢆꢆꢆꢆꢆꢆꢆꢆꢆꢇ
RꢀAꢁ ꢂꢃꢄꢅRꢆ
ꢀꢁ ꢂꢃꢄꢁ ꢅAꢆA ꢇRꢈAꢅ Aꢄ ꢉ ꢊꢊꢊꢋ
ꢌꢍꢃꢎꢈ ꢏꢐRꢈꢀ ꢑ ꢒ ꢇꢅꢈꢉAꢓꢎꢆꢋ
ꢀꢁꢂ
ꢀ ꢁ ꢂꢃ ꢄꢅꢆ ꢅR ꢇꢄꢅꢈꢉ ꢊꢋꢌꢃꢍꢎ ꢍꢏꢄꢃꢐ
ꢑ ꢁ ꢂꢃ ꢒꢎꢉꢒ ꢅR ꢓꢄꢅꢈꢉ ꢊꢋꢌꢃꢍꢎ ꢍꢏꢄꢃꢐ
ꢔꢐ ꢕꢕꢕ ꢁ ꢖꢅꢃꢎ ꢗꢖAꢃꢘꢐRꢙꢘꢅꢙꢃꢄAꢚꢐꢛ ꢜAꢘA
ꢝꢞAꢕꢕꢕ ꢁ ꢖꢎꢃꢅ ꢗꢃꢄAꢚꢐꢙꢘꢅꢙꢖAꢃꢘꢐRꢛ ꢜAꢘA
ꢀꢁꢂꢂꢃ ꢄ ꢅ
ꢀꢁꢂꢃꢄꢅꢆꢇAꢈꢄꢄꢄꢉꢄꢄꢄꢊ
ꢀꢁꢂꢃꢄꢅꢆꢇAꢈꢄꢄꢄꢉꢄꢄꢄꢊ
ꢀRꢁꢂꢃ ꢄꢅꢄꢄ ꢂꢆ
ꢀꢇꢈꢉAꢊꢇ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂꢁ ꢃꢀꢄꢅ
ꢀꢁꢂ
Note: When operating on top of a daisy chain, more than one dummy byte has to be send every time the isoSPI chain has to be woken up. See also
'Waking a Daisy Chain' chapters in latest cell monitor datasheets (e.g. LTC6812) for recommended procedures.
Figure 20a. Flow Chart of Recommended Wake-Up Procedure and Example SPI Transactions for Direct and Indirect Read Scenarios
Rev A
44
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
ꢀAꢁꢂꢃꢄꢅ ꢀꢆꢇꢈ ꢉꢆRꢂꢊꢇ
RꢂAꢉ ꢊꢋꢌꢌAꢍꢉꢎ
ꢀAꢁꢂꢃꢄꢅ ꢅRꢆꢇꢂꢈꢄRꢂ
ꢀꢁꢁꢂ
ꢀꢁꢁꢂ
ꢀꢁꢂꢃ ꢃꢄꢅꢅꢆ
ꢇꢆꢈꢁ ꢈꢉ ꢊAꢋꢁꢌꢄꢍ ꢎꢏꢐꢀꢍꢑ
ꢀꢁꢂꢃ ꢃꢄꢅꢅꢆ
ꢇꢆꢈꢁ ꢈꢉ ꢊAꢋꢁꢌꢄꢍ ꢎꢉRꢁ
ꢀꢁꢂꢁꢁꢃꢂꢄꢅꢆꢄꢅꢄꢇꢆꢈꢇꢉ
ꢀRꢁꢂꢃ RꢃꢄꢅꢆꢂRꢇ ꢀꢁꢂꢈ ꢉAꢄꢃ ꢊ ꢋ
Aꢌꢍ RꢍꢆꢎꢆꢏꢌꢐꢑꢒꢆRꢃꢌ
AꢆꢆꢏRꢍꢁꢌꢄ ꢂꢏ ꢍꢁRꢃꢆꢂꢑꢁꢌꢍꢁRꢃꢆꢂ
RꢃAꢍ ꢓꢏꢍꢃ
Aꢀ
ꢀꢁ
ꢀꢁꢂꢁꢃꢄꢅꢁꢆꢇꢃꢃꢃꢈꢃꢃꢃꢉ
ꢀꢁꢂꢁꢃꢄꢅꢁꢆꢇꢃꢃꢅAꢈꢉꢊꢋ
RꢀAꢁ ꢂꢃꢄꢅRꢆ
ꢀꢁꢂꢃꢄꢅꢆꢇAꢈꢄꢄꢄꢉꢄꢄꢄꢊ
ꢀRꢁꢂꢃ ꢄꢅꢄꢄ ꢂꢆ
ꢀꢇꢈꢉAꢊꢇ
ꢀꢁꢂ ꢃꢄ ꢅꢆꢇꢈꢈꢉ ꢊ ꢋꢄ
Aꢀ ꢁꢂꢃꢄꢄꢅ ꢆ ꢇꢀ
ꢀꢁꢂ
ꢀꢁꢂꢃRꢄꢅꢆꢄꢇꢇꢁ ꢈ ꢉ
ꢀ ꢁ ꢂꢃ ꢄꢅꢆ ꢅR ꢇꢄꢅꢈꢉ ꢊꢋꢌꢃꢍꢎ ꢍꢏꢄꢃꢐ
ꢑ ꢁ ꢂꢃ ꢒꢎꢉꢒ ꢅR ꢓꢄꢅꢈꢉ ꢊꢋꢌꢃꢍꢎ ꢍꢏꢄꢃꢐ
ꢔꢐ ꢕꢕꢕ ꢁ ꢖꢅꢃꢎ ꢗꢖAꢃꢘꢐRꢙꢘꢅꢙꢃꢄAꢚꢐꢛ ꢜAꢘA
ꢝꢞAꢕꢕꢕ ꢁ ꢖꢎꢃꢅ ꢗꢃꢄAꢚꢐꢙꢘꢅꢙꢖAꢃꢘꢐRꢛ ꢜAꢘA
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂꢁ ꢃꢀꢄꢅ
Figure 20b. Flow Chart of Recommended Wake-Up Procedure Implemented in a Sequence Without Polling
Rev A
45
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 25. Wake Up Acknowledge Register (0x70)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
[7:0]
WKUPACK
RW
see above 0x00: Wake up is confirmed. Part will not return to SLEEP.
Anything else: Wake up is not confirmed. Part will return to SLEEP.
In applications where LTC2949 is connected on top of a daisychain, the bit BCREN in REGSCTRL must be set before
being able to read from LTC2949. This is only possible after the boot sequence has finished. Polling of the SLEEP bit
being cleared can still be implemented by interleaving the read of OPCTRL with a write to REGSCTRL with BCREN=1 (all
other bits 0). As this scenario also implies indirect register reads via RDCV commands (for this reason RDCVCONF=0),
the write to REGSCTRL can be done in a single four byte burst together with RDCVIADDR (set to 0xF0); the two bytes
between RDCVIADDR and REGSCTRL set to 0x00. Thus the four bytes to be written are 0xF0,0x00,0x00,0x04 and the
following RDCV command will report the content of OPCTRL. Still, if the boot sequence is not completed, LTC2949
will ignore the broadcast RDCV command and thus the master will read always 0xFF (which is also compatible with
SLEEP=1), including PEC=0xFFFF if 8 bytes are read (its not necessary to read till the PEC in this case). After the boot
OPCTRL will be read as 0x00.
Accumulated Result Registers
The registers in Table 26 and Table 27 contain the accumulated quantities Charge, Energy and Time. The Time regis-
ters are unsigned integer values while the Charge and Energy registers are two’s complement signed integer values.
The value of each accumulated quantity can be determined by multiplying the respective register value with the cor-
responding LSB value from Table 26 (if the internal clock or crystal is used as a reference clock) or Table 27 (if an
external reference clock is used).
Table 26. Accumulated Results Register Parameters for Use with Crystal or Internal Clock
LSB (4MHz CRYSTAL OR PRE (4MHz DIV (4MHz
ADDRESS
NAME
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
PARAMETER
INTERNAL CLOCK)
LSBC1 = 377.887e-12
LSBE1 = 2.32175e-09
CRYSTAL)
CRYSTAL)
UNIT
SI/UI
SI
0x00
C1[47:0]
E1[47:0]
TB1[31:0]
C2[47:0]
E2[47:0]
TB2[31:0]
C3[63:0]
TB3[31:0]
E4[63:0]
TB4[31:0]
Charge1 = C1 • LSBC1
Energy1 = E1 • LSBE1
2
2
2
2
2
2
2
2
2
2
30
30
30
30
30
30
30
30
30
30
Vs
2
0x06
V s
SI
0x0C
0x10
Time1 = TB1 • LSBTB1 LSBTB1 = 397.777E-06
s
UI
SI
Charge2 = C2 • LSBC2
Energy2 = E2 • LSBE2
LSBC2 = 377.887e-12
LSBE2 = 2.32175e-09
Vs
2
0x16
V s
SI
0x1C
0x24
Time2 = TB2 • LSBTB2 LSBTB2 = 397.777E-06
Charge3 = C3 • LSBC3 LSBC3 = 377.887e-12
Time3 = TB3 • LSBTB3 LSBTB3 = 397.777E-06
Energy4 = E4 • LSBE4 LSBE4 = 2.32175e-09
Time4 = TB4 • LSBTB4 LSBTB4 = 397.777E-06
s
Vs
s
UI
SI
0x2C
0x34
UI
SI
2
V s
0x3C
s
UI
Charge1, Energy1 and Time1 contain accumulated quantities of Channel1. Charge2, Energy2 and Time2 contain accu-
mulated quantities of Channel2. Charge3 and Time3 contain the weighted sum of charges monitored by Channel1 and
Channel2 and the corresponding time. Similarly Energy4 and Time4 contain the weighted sum of energies monitored
by Channel1 and Channel2 and the corresponding time.
If different sense resistors are used on CH1 and CH2, the LTC2949 uses the ratio of the sense resistors (RSRATIO) set
in the Gain Configuration Registers to compute the correct weighted sums Charge3 and Energy4.
When the internal clock is used, PRE and DIV should be set to their default values which is done by writing 0x07 to register
(0xE9), otherwise the values of PRE(0xE9)[2:0] and DIV(0xE9)[7:3] should be set according to section Timebase Control.
Rev A
46
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 27. Accumulated Results Register Parameters for Use with External Clock
ADDRESS
0x00
NAME
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
PARAMETER
LSB
UNIT
SI/UI
SI
PRE
C1[47:0]
E1[47:0]
TB1[31:0]
C2[47:0]
E2[47:0]
TB2[31:0]
C3[63:0]
TB3[31:0]
E4[63:0]
TB4[31:0]
Charge1 = C1 • LSBC1
Energy1 = E1 • LSBE1
Time1 = TB1 • LSBTB1
Charge2 = C2 • LSBC2
Energy2 = E2 • LSBE2
Time2 = TB2 • LSBTB2
Charge3 = C3 • LSBC3
Time3 = TB3 • LSBTB3
Energy4 = E4 • LSBE4
Time4 = TB4 • LSBTB4
LSBC1 = 1.21899e–5 • 1/fext • 2
• (DIV+1)
Vs
PRE
2
0x06
LSBE1 = 7.4895e–5 • 1/fext • 2
• (DIV+1)
• (DIV+1)
V s
SI
PRE
0x0C
0x10
LSBTB1 = 12.8315 • 1/fext • 2
LSBC2 = 1.21899e–5 • 1/fext • 2
s
UI
SI
PRE
• (DIV+1)
• (DIV+1)
• (DIV+1)
Vs
PRE
2
0x16
LSBE2 = 7.4895e–5 • 1/fext • 2
V s
SI
PRE
0x1C
0x24
LSBTB2 = 12.8315 • 1/fext • 2
LSBC3 = 1.21899e–5 • 1/fext • 2
s
Vs
s
UI
SI
PRE
• (DIV+1)
• (DIV+1)
PRE
0x2C
0x34
LSBTB3 = 12.8315 • 1/fext • 2
LSBE4 = 7.4895e–5 • 1/fext • 2
UI
SI
PRE
2
• (DIV+1)
V s
PRE
0x3C
LSBTB4 = 12.8315 • 1/fext • 2
• (DIV+1)
s
UI
Note: Values of PRE and DIV should be calculated according to Timebase Control section.
For instance, an external clock frequency of 10MHz would require values PRE to be set to 4 and DIV to be set to 19.
With f =10MHz, LSBC1 is calculated as 390.078e-12 VS. To get the Charge1 value, the register content of C1 is
EXT
multiplied with LSBC1. In this case, a C1 register value of 0x 75 5A 10 or 7690768 and the resulting Charge1 is 0.003
VS. For a sense resistor of 300µΩ this corresponds to 10As.
++
LSB values may be calculated easily using the Quick Eval software for the LTC2949 or by using the C/C header files
provided in the code section of the LTC2949 (see LTC2949 evaluation board DC2732A manual for details). The registers
for Charge, Energy and Time can be preset to a non-zero initial value. In Continuous Mode all bytes of the respective
quantity must be written in the same multi-byte transaction or while the memory is locked.
Non-Accumulated Result Registers
Registers in Table 28 contain measured values of Currents, Powers, Voltages, Temperatures, VCC and VREF. All quanti-
ties are represented as two's complement signed integer values.
Current 1 represents the differential voltage sensed between CF1P and CF1M. Current 2 represents the differential volt-
age sensed between CF2P and CF2M. Battery voltage (BAT) is the differential voltage across pins VBATP and VBATM.
Power 1 is the instantaneous multiplication of BAT and current 1. Power 2 is the instantaneous multiplication of BAT
and current 2. Temperature is the temperature of the on-silicon temperature sensor. VCC is the voltage across pins
A/DVCC and AGND. Registers SLOT1 and SLOT2 contain the results of the two multiplexer inputs choosen ac-
cording to Table 58 and can be configured to give out voltage or temperature by means of the NTC Configuration
Registers (Table 70). VREF is the voltage across pins VREF and AGND. The moving average of the four preced-
ing current measurements are stored in I1AVG and I2AVG. The values of these four current measurements are
retained in the Current History registers; Current 1 History 1 is the result prior to Current 1, Current 1 History 2 is
the current result prior to Current 1 History 1, and so on. All measured values are scaled with the LSB values from
. To calculate the physical value of the measured parameter, multiply the register value by the appropriate LSB value.
Table 28. Non-Accumulated Results Register Parameters
ADDRESS
NAME
TYPE
DEFAULT
PARAMETER
Current 1
LSB
950
UNIT
SI/UI
SI
0x90
I1[23:0]
RO
0x00
nV
2
Power 1 (Power, P1ASV =0)
Power 1 (Voltage, P1ASV=1)
Current 2
5.8368
46.875
950
μ[V ]
SI
0x93
0x96
P1[23:0]
I2[23:0]
RO
RO
0x00
0x00
µV
nV
SI
SI
Rev A
47
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 28. Non-Accumulated Results Register Parameters (continued)
ADDRESS
NAME
TYPE
DEFAULT
PARAMETER
Power 2 (Power, P2ASV =0)
Power 2 (Voltage, P2ASV=1)
Current 1 Moving Average
Battery Voltage
LSB
5.8368
46.875
237.5
375
UNIT
SI/UI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
2
μ[V ]
0x99
P2[23:0]
RO
0x00
µV
nV
μV
°C
0x9C
0xA0
0xA2
0xA4
0xA6
I1AVG[23:0]
BAT[15:0]
RO
RO
RO
RO
RO
0x00
0x00
0x00
0x00
0x00
TEMP [15:0]
VCC[15:0]
Temperature
0.2
Voltage at A/DVCC
SLOT 1 (Voltage)
2.26
mV
μV
°C
SLOT1[15:0]
375
SLOT 1 (Temp.)
0.2
0xA8
SLOT2[15:0]
RO
0x00
SLOT 2 (Voltage)
375
μV
°C
SLOT 2 (Temp)
0.2
0xAA
0xAC
0xB3
0xB6
0xB9
0xBC
0xC3
0xC6
0xC9
0xCC
0xF7
0xF8
0xF9
0xFA
VREF[15:0]
I2AVG[23:0]
I1H1[23:0]
I1H2[23:0]
I1H3[23:0]
I1H4[23:0]
I2H1[23:0]
I2H2[23:0]
I2H3[23:0]
I2H4[23:0]
FIFOI1[15:0]
FIFOI2[15:0]
FIFOBAT[15:0]
FIFOAUX[15:0]
RO
RO
RO
RO
RO
RO
RO
RO
RO
RO
RO
RO
RO
RO
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
Voltage at VREF
375
μV
nV
nV
nV
nV
nV
nV
nV
nV
nV
μV
μV
μV
μV
Current 2 Moving Average
Current 1 History 1
Current 1 History 2
Current 1 History 3
Current 1 History 4
Current 2 History 1
Current 2 History 2
Current 2 History 3
Current 2 History 4
Fast Current 1
237.5
950
950
950
950
950
950
950
950
7.60371
7.60371
375.183
375.183
Fast Current 2
Fast Battery Voltage
Fast AUX HVMUX
The FIFO registers 0xF7 to 0xFA allow to read conversion results in fast continuous mode. When reading from those
registers, the internal address auto-increment stops, allowing to read any number of bytes the fixed addresses. For
each sample, three bytes have to be read which are MSB, LSB and TAG, see also fast mode section and below descrip-
tion for more details.
In Fast Mode, ADC conversion results turn negative when exceeding positive and clip when exceeding negative full-scale
values. In Slow Modes, ADC conversion results clip when exceeding positive and negative full-scale values. Full-scale
values are always beyond the specified input range.
Registers I1AVG (0x9C) and I2AVG (0xAC) are copied to registers 0xB0 and 0xC0, respectively. Thus, 0x9C-0x9E will
report the same values as 0xB0-0xB2 and 0xAC-0xAE will report the same values as 0xC0-0xC2.
Rev A
48
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 29. FIFO Register Read Format.
DATA BYTE
NAME
S [15:8]
DESCRIPTION
0
MSB of measured sample N
LSB of measured sample N
Tag of sample N
N
1
S [7:0]
N
2
TAG
N
3
S
S
[15:8]
MSB of measured sample N+1
LSB of measured sample N+1
Tag of sample N+1
N+1
N+1
4
[7:0]
5
TAG
…
N+1
…
…
3*M
3*M+1
3*M+2
S
S
[15:8]
[7:0]
MSB of measured sample N+M
LSB of measured sample N+M
Tag of sample N+M
N+M
N+M
TAG
N+M
The column Data Byte does not count the PEC bytes, which depend on the setting of data bytes per PEC (ID-byte for
DCMD or fixed to six for RDCV). For maximum data throughput it is recommended to read the FIFOs with DCMDs with
16 bytes per PEC and multiples of 16 samples (= 48 data bytes + 3 • 2 PEC bytes).
Table 30. FIFO TAG Definitions.
TAG
NAME
OK
DESCRIPTION
0x00
0x55
0xAA
Valid, new sample
RDOVR
WROVR
Read overrun, sample was already read
Write overrun, FIFO was filled completely and at least one sample was already overwritten
It is recommended to connect LTC2949 always in parallel to a daisy chain. Also, in the scenario where LTC2949 is con-
nected at one end of a reversible isoSPI chain, the default communication should be done with direct read commands.
Only in case the direct isoSPI link to LTC2949 fails, the communication would be routed through the daisy chain.
InthisconfigurationwhereLTC2949isontopofadaisychain, afterthebroadcastRDCVcommand, thestackedLTC68xx
Cell Monitors turn into a cascaded shift register, in which data is shifted through each device to the next device in the
stack. In this scenario, at the end of the read transaction, there is always a fixed number of samples stuck in the shift
register, that is never transmitted to the master.
Several approaches are possible to avoid or minimize the loss of samples. Either fast continuous mode is stopped
(FACONV=0), the FIFO is read until empty and then fast continuous measurement is repeated (FACONV=1). This results
in a time window without measurements, equal to the time it takes to read the samples. For example, at 1Mbit/s serial
clock rate, it takes 0.8 ms (approximately one fast conversion time) to read 24 samples (four byte command and PEC,
72 data bytes, 24 data PEC bytes equals in total 100 bytes or 800 bits). For lower clock speeds or longer cycle times,
where more samples are read per cycle, the following approach is more efficient.
Alternatively, the FIFO read burst must be long enough to always empty the FIFO and read at least one sample with
TAG RDOVR. Still, also here, one sample with TAG OK can arrive after the RDOVR and would then be stuck in the daisy
chain’s shift register. If the latency of the shift register (number of devices in the daisy chain times eight bytes times
eight bit per byte divided by SPI clock speed) is longer than one fast conversion time, even more than one sample can
get stuck in the daisy chain.
Rev A
49
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Tracking Registers
The tracking registers keep track of the maximum and minimum values of previous conversions since the last reset.
Value scaling is done in the same manner as the non-accumulated register values, using LSB values from Table 31.
Negative values are treated as smaller (more minimum) than positive values as the minimum registers are updated.
For example: A register value I1MAX(0x40,0x41) of 0000 0001 1111 0100b = 01 F4h = 500d indicates a resulting
maximum sense resistor signal of 500 • 3.8µV = 1.9mV. A register value I1MIN(0x42,0x43) of 1111 1010 0010 0100b
= FA 24h = –1500d indicates a minimum sense resistor signal of –1500 • 3.8µV = –5.7mV. The calculation of the other
tracked parameter values is done the same way with their corresponding LSB values.
Table 31. Tracking Registers
ADDRESS
0x40
0x48
0x42
0x4A
0x44
0x4C
0x46
0x4E
0x50
0x52
0x54
0x56
0x58
0x5A
0x5C
0x5E
0x60
0x62
NAME
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x8000
0x8000
0x7FFF
0x7FFF
0x8000
0x8000
0x7FFF
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
PARAMETER
Max I1 Current
LSB
3.8
UNIT
μV
SI/UI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
SI
I1MAX[15:0]
I2MAX[15:0]
Max I2 Current
3.8
μV
I1MIN[15:0]
Min I1 Current
3.8
μV
I2MIN[15:0]
Min I2 Current
3.8
μV
2
P1MAX[15:0]
P2MAX[15:0]
P1MIN[15:0]
Max Power 1 (or Battery Voltage for P1ASV=1)
Max Power 2 (or Battery Voltage for P2ASV=1)
Min Power 1 (or Battery Voltage for P1ASV=1)
Min Power 2 (or Battery Voltage for P2ASV=1)
Max Battery Voltage
23.347 (187.5)
23.347 (187.5)
23.347 (187.5)
23.347 (187.5)
375
μ[V ] (µV)
2
μ[V ] (µV)
2
μ[V ] (µV)
2
P2MIN[15:0]
μ[V ] (µV)
BATMAX[15:0]
BATMIN[15:0]
TEMPMAX[15:0]
TEMPMIN[15:0]
VCCMAX[15:0]
VCCMIN[15:0]
SLOT1MAX[15:0]
SLOT1MIN[15:0]
SLOT2MAX[15:0]
SLOT2MIN[15:0]
μV
μV
Min Battery Voltage
375
Max Temperature
0.2
°C
Min Temperature
0.2
°C
Max Voltage at A/DVCC
2.26
mV
Min Voltage at A/DVCC
2.26
mV
Max Voltage (or Temperature) at SLOT1
Min Voltage (or Temperature) at SLOT1
Max Voltage (or Temperature) at SLOT2
Min Voltage (or Temperature) at SLOT2
375 (0.2)
375 (0.2)
375 (0.2)
375 (0.2)
µV (°C)
µV (°C)
µV (°C)
µV (°C)
Note that the tracking registers for current and power report only the 16MSBs of the respective 18-bit result registers, leading to a four times larger LSB
value.
STATUS, (EXT)FAULTS, Threshold and Overflow Alert Registers
The registers described in the following chapters are used to signal certain events, like a voltage threshold violation,
charger or energy overflow, supply undervoltage events and others. After power-up it is recommended to lock the
memory, read STATUS, FAULTS and EXTFAULTS registers, check for the default values, clear them and finally unlock the
memory. During normal operation those registers shall be checked periodically to be all zero, except for the UPDATE
bit. Any other value indicates a failure.
Any bits in the STATUS, alerts (0x81-0x87), FAULTS and EXTFAULTS registers are only set to 1 by LTC2949 in case
of an event, but never cleared automatically. After the master reads some bits being set, actions should be taken (e.g.
to clear the charge register in case of a charge overflow), the memory must be locked (see REGSCTRL), the registers
Rev A
50
For more information www.analog.com
LTC2949
should be read again as other events may have occurred in the meantime, registers that or not 0x00 must be written
to 0x00 and finally the memory must be unlocked to proceed with normal operation.
It is not recommended to write those registers without usage of the memory lock, as this may lead to loss of failure
and alert reports.
Status and Fault Registers
The STATUS register reports the status of register updates, undervoltage lockout, and reference clock errors. On power
up, all undervoltage lockouts and the power-on reset are set to 1. After exit from shutdown, bits UVLOA and UVLOD
are set. UPDATE is set to 1 when the LTC2949 has finished a measurement cycle and updated the result registers, the
accumulation registers, and the tracking registers.
ADCERR is set to 1 if the supply voltage at AVCC is too low for a proper operation of the ADCs. The values in the result
registers are not valid and should be discarded if ADCERR is set. TBERR is set to 1 if the internal time base overflows.
This indicates an incorrect setting of the values PRE and DIV with respect to the external clock at CLKI. The values of
accumulated results registers should be discarded if TBERR is set.
Table 32. STATUS (0x80)
BIT
0
SYMBOL
UVLOA
TYPE
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
0
0
0
1: Undervoltage in the analog domain or ADCs during a conversion
1: Power-on reset has occurred due to undervoltage in the analog domain
1: Undervoltage in the standby domain
1
PORA
2
UVLOSTBY
UVLOD
3
1: Undervoltage in the digital domain
4
UPDATE
ADCERR
TBERR
1: Result registers have been updated
5
1: The ADC conversion is not valid due to undervoltage during a conversion
1: Overflow of the internal timebase register. The values of accumulated result registers are invalid
6
LTC2949 has several types of internal memory that are checked during boot-up via built-in self-test (BIST) routines.
Errors in individual blocks are reported by the EXTFAULTS and FAULTS registers. The FAULTS register additionally
reports error in the internal (on-chip) and external (from master to slave on the SPI/isoSPI interface) communication
and indicates thermal shutdown and fast channel error events.
Table 33. EXTFAULTS (0xDC)
BIT
0
SYMBOL
HD1BITERR
ROMERR
TYPE
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
0
0
0
0
0
0
1
1: Hamming decoder 1-bit error
1: ROM CRC error
1: Memory error
1
2
MEMERR
3
FCAERR
1: Fast channel error
1: XRAM error
4
XRAMERR
IRAMERR
5
1: IRAM error
7
HWMBISTEXEC
1: Memory BIST was executed. If SDA is externally pulled low during power-up, the memory
BIST will be skipped and this bit will be zero. Connect a 4.7k-10k pull-up resistor from SDA to
BYP1 to ensure correct operation of memory BIST.
Rev A
51
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 34. FAULTS (0xDD)
BIT
0
SYMBOL
PROMERR
TSD
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1: Error in trim values stored in internal PROM
0
0
0
0
0
0
0
0
1
1: Shutdown due to over temperature
2
INTCOMMERR
EXTCOMMERR
FAERR
1: Parity check of internal communication failed
1: PEC error in external communication (SPI/isoSPI) occurred
1: Fast mode error, see Safety Manual for more information
1: Error during hardware BIST.
3
4
5
HWBIST
6
CRCCFG
1: Internal RAM gain coefficient CRC Error
1: User accessible register CRC Error
7
CRCMEM
Threshold and Overflow Alert Registers
Threshold and overflow alert registers are set when the respective threshold values are exceeded or when registers
overflow. Thresholds are set in the Threshold Registers section.
The accumulated quantities are continuously checked against guard values to warn that a register is nearing overflow,
nominally set to 90% of each register’s maximum value. When any quantity crosses its guard threshold, the LTC2949
sets the corresponding overflow bit in the status register, generates an alert (if enabled) and continues accumulation.
At the maximum voltage inputs, rollover typically happens several hours after an overflow alert is signaled, allowing
the host time to take action to avoid data loss, by reading and clearing the concerned accumulators using the memory
locking procedure. The overflow thresholds for 32-bit quantities (time) are 3865470565 LSB; for 48-bit quantities
(charge and energy) is 126663739519794 LSB.
The threshold and overflow comparators for accumulated quantities charge, energy and time use a floating point
format internally. This can appear to cause slight bit-level comparison discrepancies, but the comparisons between
accumulatedresultregistersandtheirrespectivethresholdregisterswillalwayshaveanaccuracyofbetterthan0.001%.
An alert condition needs to be present for at least for 200ms to be reported by the alert registers (0x81-0x87). Only
overcurrent conditions detected by either OCC1 or OCC2 can be signaled to the host within some µs by stopping heart
beat on GPIO5.
To acknowledge and release an overcurrent alert the following steps can be performed:
1. Stop the overcurrent event
2. Lock the memory
3. Read STATUS-STATVCC, (EXT)FAULTS for any newly arrived error flags (including the asserted OCCx bits)
4. Clear STATUS-STATVCC, (EXT)FAULTS
5. Unlock the memory
6. Heartbeat on GPIO5 will start again
Rev A
52
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 35. Voltage, Temperature Threshold Alerts STATVT (0x81)
BIT
0
SYMBOL
BATH
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1: Voltage (VBATP – VBATM) high threshold exceeded
1: Voltage (VBATP – VBATM) low threshold exceeded
1: Temperature high threshold exceeded
1: Temperature low threshold exceeded
1: SLOT1 high threshold exceeded
0
0
0
0
0
0
0
0
1
BATL
2
TEMPH
TEMPL
SLOT1H
SLOT1L
SLOT2H
SLOT2L
3
4
5
1: SLOT1 low threshold exceeded
6
1: SLOT2 high threshold exceeded
7
1: SLOT2 low threshold exceeded
Table 36. Current, Power Threshold Alerts STATIP (0x82)
BIT
0
SYMBOL
I1H
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
0
0
0
0
0
0
0
0
1: Current1 high threshold exceeded
1: Current1 low threshold exceeded
1: Power1 high threshold exceeded
1: Power1 low threshold exceeded
1: Current2 high threshold exceeded
1: Current2 low threshold exceeded
1: Power2 high threshold exceeded
1: Power2 low threshold exceeded
1
I1L
2
P1H
P1L
I2H
3
4
5
I2L
6
P2H
P2L
7
Table 37. Charge Threshold Alerts STATC (0x83)
BIT
0
SYMBOL
C1H
TYPE
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
OPERATION
OPERATION
0
0
0
0
0
0
1: Charge1 high threshold exceeded
1: Charge1 low threshold exceeded
1: Charge2 high threshold exceeded
1: Charge2 low threshold exceeded
1: Charge3 high threshold exceeded
1: Charge3 low threshold exceeded
1
C1L
2
C2H
3
C2L
4
C3H
5
C3L
Table 38. Energy Threshold Alerts STATE (0x84)
BIT
0
SYMBOL
E1H
TYPE
RW
RW
RW
RW
RW
RW
DEFAULT
0
0
0
0
0
0
1: Energy1 high threshold exceeded
1: Energy1 low threshold exceeded
1: Energy2 high threshold exceeded
1: Energy2 low threshold exceeded
1: Energy4 high threshold exceeded
1: Energy4 low threshold exceeded
1
E1L
2
E2H
3
E2L
6
E4H
7
E4L
Table 39. Charge, Energy Overflow Alerts STATCEOF (0x85)
BIT
0
SYMBOL
C1OVF
C2OVF
C3OVF
E1OVF
E2OVF
E4OVF
TYPE
RW
RW
RW
RW
RW
RW
DEFAULT
0
0
0
0
0
0
1: Charge1 overflow alert
1: Charge2 overflow alert
1: Charge3 overflow alert
1: Energy1 overflow alert
1: Energy2 overflow alert
1: Energy4 overflow alert
1
2
4
5
7
Rev A
53
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 40. Time Base Alerts STATTB (0x86)
BIT
0
SYMBOL
T1TH
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
0
0
0
0
0
0
0
0
1: Time1 threshold exceeded
1: Time2 threshold exceeded
1: Time3 threshold exceeded
1: Time4 threshold exceeded
1: Time1 overflow
1
T2TH
2
T3TH
3
T4TH
4
T1OVF
T2OVF
T3OVF
T4OVF
5
1: Time2 overflow
6
1: Time3 overflow
7
1: Time4 overflow
Table 41. VCCOCC Threshold Alerts STATVCC (0x87)
BIT
0
SYMBOL
VCCH
TYPE
RW
RW
RW
RW
DEFAULT
OPERATION
0
0
0
0
1: VCC high threshold exceeded
1: VCC low threshold exceeded
1
VCCL
2
OCC1H
OCC2H
1: Current1 above OCC1 threshold for more than deglitch time
1: Current2 above OCC2 threshold for more than deglitch time
3
Note: The memory must be locked to clear status (0x80), faults (0xDC-0xDD) and alert (0x81-0x87) registers to avoid missing of any failure and alert
reports. See REGSCTRL description for details.
Mask Registers
The mask registers control which alerts stop heartbeat. If a mask register bit is reset to 0, exceeding of the respective
threshold causes the heartbeat on GPIO4 pin to stop, if the latter is configured accordingly in the GPIO4HBCTRL register.
When a bit of the Status Mask Register (STATUSM) is set to 0, the corresponding bits of register STATUS (0x80) will stop
heartbeat. When a bit of the status mask register is set to 1, heartbeat is unaffected by the corresponding bits of register
STATUS (0x80).
Table 42. Status Mask STATUSM (0x88)
BIT
0
SYMBOL
UVLOAM
UVLODM
UPDATEM
ADCERRM
TBCERRM
TYPE
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
1
Mask UVLOA of STATUS(0x80)
Mask UVLOD of STATUS(0x80)
Mask UPDATE of STATUS(0x80)
Mask ADCERR of STATUS(0x80)
Mask TBCERR of STATUS(0x80)
3
4
5
6
When bits of STATVTM are set to 0, corresponding bits of register STATVT (0x81) will stop heartbeat. When a bit of
STATVTM is set to 1, heartbeat is unaffected by the corresponding bits of register STATVT (0x81).
Table 43. Voltage, Temperature Threshold Alert Mask STATVTM (0x89)
BIT
0
SYMBOL
BATHM
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
1
1
1
1
Mask BATH of STATVT(0x81)
Mask BATL of STATVT(0x81)
Mask TEMPH of STATVT(0x81)
Mask TEMPL of STATVT(0x81)
Mask SLOT1H of STATVT(0x81)
Mask SLOT1L of STATVT(0x81)
Mask SLOT2H of STATVT(0x81)
Mask SLOT2L of STATVT(0x81)
1
BATLM
2
TEMPHM
TEMPLM
SLOT1HM
SLOT1LM
SLOT2HM
SLOT2LM
3
4
5
6
7
Rev A
54
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
When bits from STATIPM are set to 0, bits from register STATIP(0x82) will stop heartbeat. When a bit of STATIPM
register is set to 1, heartbeat is unaffected by the corresponding bits of register STATIP(0x82).
Table 44. Current, Power Threshold Alert Mask STATIPM (0x8A)
BIT
0
SYMBOL
I1HM
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
1
1
1
1
Mask I1H of STATIP (0x82)
Mask I1L of STATIP (0x82)
Mask P1H of STATIP (0x82)
Mask P1L of STATIP (0x82)
Mask I2H of STATIP (0x82)
Mask I2L of STATIP (0x82)
Mask P2H of STATIP (0x82)
Mask P2L of STATIP (0x82)
1
I1LM
2
P1HM
P1LM
I2HM
3
4
5
I2LM
6
P2HM
P2LM
7
When bits from STATCM are set to 0, bits from register STATC(0x83) will stop heartbeat. When a bit of STATCM register
is set to 1, heartbeat is unaffected by the corresponding bits of register STATC(0x83).
Table 45. Charge Threshold Alerts Mask STATCM (0x8B)
BIT
0
SYMBOL
C1HM
C1LM
C2HM
C2LM
C3HM
C3LM
TYPE
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
1
1
Mask C1H of STATC (0x83)
Mask C1L of STATC (0x83)
Mask C2H of STATC (0x83)
Mask C2L of STATC (0x83)
Mask C3H of STATC (0x83)
Mask C3L of STATC (0x83)
1
2
3
4
5
When bits from STATEM are set to 0, bits from register STATE(0x84) will stop heartbeat. When a bit of STATEM register
is set to 1, heartbeat is unaffected by the corresponding bits of register STATE(0x84).
Table 46. Energy Threshold Alerts Mask STATEM (0x8C)
BIT
0
SYMBOL
E1HM
E1LM
TYPE
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
1
1
Mask E1H of STATE (0x84)
Mask E1L of STATE (0x84)
Mask E2H of STATE (0x84)
Mask E2L of STATE (0x84)
Mask E4H of STATE (0x84)
Mask E4L of STATE (0x84)
1
2
E2HM
E2LM
3
6
E4HM
E4LM
7
When bits from STATCEOFM are set to 0, bits from register STATCEOF(0x85) will stop heartbeat. When a bit of STAT-
CEOFM register is set to 1, heartbeat is unaffected by the corresponding bits of register STATCEOF(0x85).
Table 47. Charge, Energy Overflow Alerts Mask STATCEOFM (0x8D)
BIT
0
SYMBOL
C1OFM
C2OFM
C3OFM
E1OFM
E2OFM
E4OFM
TYPE
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
1
1
Mask C1OVF of STATCEOF(0x85)
Mask C2OVF of STATCEOF(0x85)
Mask C3OVF of STATCEOF(0x85)
Mask E1OVF of STATCEOF(0x85)
Mask E2OVF of STATCEOF(0x85)
Mask E4OVF of STATCEOF(0x85)
1
2
4
5
7
Rev A
55
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
When bits from STATTBM are set to 0, bits from register STATTB (0x86) will stop heartbeat. When a bit of STATTBM
register is set to 1, heartbeat is unaffected by the corresponding bits of register STATTB (0x86).
Table 48. Time Base Alerts STATTBM (0x8E)
BIT
0
SYMBOL
T1THM
T2THM
T3THM
T4THM
T1OFM
T2OFM
T3OFM
T4OFM
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
OPERATION
1
1
1
1
1
1
1
1
Mask T1TH of STATTB(0x86)
Mask T2TH of STATTB(0x86)
Mask T3TH of STATTB(0x86)
Mask T4TH of STATTB(0x86)
Mask T1OVF of STATTB(0x86)
Mask T2OVF of STATTB(0x86)
Mask T3OVF of STATTB(0x86)
Mask T4OVF of STATTB(0x86)
1
2
3
4
5
6
7
When bits from STATVCCM are set to 0, bits from register STATVCC (0x87) will stop heartbeat. When a bit of STATVCCM
register is set to 1, heartbeat is unaffected by the corresponding bits of register STATVCC (0x87).
Table 49. VCC Threshold Alerts STATVCCM (0x8F)
BIT
0
SYMBOL
VCCH
TYPE
RW
DEFAULT
OPERATION
1
1
Mask VCCH of STATVCC (0x87)
Mask VCCL of STATVCC (0x87)
1
VCCL
RW
Note that the bits reporting the result of the overcurrent comparators OCC1 and OCC2 in STATVCC (0x87) do not have a
corresponding mask bit in STATVCCM. Therefore the impact of OCC1 and OCC2 on heartbeat either on GPIO4 or GPIO5
cannot be masked. If overcurrent comparison is not desired the enable bit in OCCxCTRL has to be cleared (default).
Control Registers
The control registers select the multiplexer input, control the accumulation of charge, energy and time, configure the
GPIO pins, set the overcurrent comparator thresholds, and setup the timebase if an external clock is used.
The Overcurrent Control Registers allow to set the overcurrent comparator thresholds and deglitch filters, see also the
Overcurrent Comparators section.
Table 50. OCC1CTRL (0xDE)
BIT
0
SYMBOL
OCC1EN
TYPE
RW
RW
RW
RW
DEFAULT
OPERATION
OCC1 enable bit. GPIO5 is configured as heartbeat
0
[3:1]
[5:4]
[7:6]
OCC1DAC [2:0]
OCC1DGLT [1:0]
OCC1POL [1:0]
000
00
OCC1 threshold setting bits, see Table 6 (OCC Thresholds)
OCC1 deglitch time setting bits, see Table 7 (OCC Deglitch Time)
OCC1 polarity setting bits, see Table 5 (OCC Polarity Configuration)
00
Table 51. OCC2CTRL (0xDF)
BIT
0
SYMBOL
OCC2EN
TYPE
RW
RW
RW
RW
DEFAULT
OPERATION
OCC2 enable bit. GPIO5 is configured as heartbeat
0
[3:1]
[5:4]
[7:6]
OCC2DAC [2:0]
OCC2DGLT [1:0]
OCC2POL [1:0]
000
00
OCC2 threshold setting bits, see Table 6 (OCC Thresholds)
OCC2 deglitch time setting bits, see Table 7 (OCC Deglitch Time)
OCC2 polarity setting bits, see Table 5 (OCC Polarity Configuration)
00
Rev A
56
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
The Accumulator Control and Deadband Registers allow to control the accumulation of Charge1, Energy1, Charge2,
Energy2, Charge3 and Energy4 (C1, E1, C2, E2, C3, E4). Accumulation can be enabled, disabled or conditionally enabled
based on the sign and absolute value of a measured current. C1 contains accumulated I1, C2 contains accumulated I2,
E1 contains accumulated P1 and E2 contains accumulated P2. C3 contains the accumulated sum of I1 and I2 weighted
by the gain setting parameters (see Gain Configuration Registers section) and E4 the accumulated weighted sum of
P1 and P2.
For example, by setting bit 0 of the ACCCTRL1 and bit 1 of ACCCTRL2, C1 contains the accumulation of positive cur-
rents I1 and C3 contains the accumulation of negative currents.
Table 52. Accumulator Control ACCCTRL1(0xE1)
BIT
SYMBOL
TYPE
RW
DEFAULT
OPERATION
[1:0]
[3:2]
ACC1I1[1:0]
ACC2I2[1:0]
00
00
Accumulation control of Charge1/Charge2 and Energy1/Energy2.
00: Accumulation takes place always,
01: Only if the current is positive,
RW
10: Only if the current is negative,
11: No accumulation takes place.
Table 53. Accumulator Control ACCCTRL2(0xE2)
BIT
SYMBOL
TYPE
RW
RW
RW
RW
DEFAULT
OPERATION
[1:0]
[3:2]
[5:4]
[7:6]
ACC3I1[1:0]
ACC3I2[1:0]
ACC4I1[1:0]
ACC4I2[1:0]
00
00
00
00
Accumulation control of Charge3 and Energy4.
00: Accumulation takes place always,
01: Only if the current is positive,
10: Only if the current is negative,
11: No accumulation takes place.
Small offset voltages in the current measurement path lead to large charge errors after a long integration time. The
accumulator dead band registers allow to set a minimum absolute value of I1 and I2 before being accumulated.
Table 54. Accumulation Dead Band ACCIDB1(0xE4)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
[7:0]
ACCIDB1
RW
0x0
Current 1 deadband for accumulation. If the absolute value of I1 is higher than or equal this
value, I1 and P1 are accumulated and C1, E1, C3, E4 are updated. A comparison to the respective
thresholds takes place right after an update. If the absolute value of I1 is lower, I1 and P1 are not
accumulated and C1 and E1 are not updated. Unit is the same as LSB of current I1(0x90).
Table 55. Accumulation Dead Band ACCIDB2(0xE5)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
[7:0]
ACCIDB2
RW
0x0
Current 2 deadband for accumulation .If the absolute value of I2 is higher than or equal this
value, I2 and P2 are accumulated and C2, E2, C3, E4 are updated. A comparison to the respective
thresholds takes place right after an update. If the absolute value of I2 is lower, I2 and P2 are not
accumulated and C2 and E2 are not updated. Unit is the same as LSB of current I2(0x96).
Rev A
57
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
The Time Base Control Register selects between the internal and an external reference clock, and sets the time base
parameters when an external reference clock is used. Set PRE[2:0] = 111b or 7d (default) to enable the internal refer-
ence clock. To use an external reference clock, set the values of PRE[2:0] and DIV[4:0] according to the external clock
frequency; see the Timebase Control section for more information.
Table 56. Timebase Control TBCTRL (0xE9), DEFAULT VALUE: 0x07
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
[2:0]
PRE [2:0]
RW
111
Prescaler value [2:0],
binary coded,
see Table 2. Parameter PRE with External Clock
[7:3]
DIV [4:0]
RW
00000
Divider value [4:0],
binary coded,
see Table 3
The multiplexer inputs are selected by the Multiplexer Control Registers. For each of the multiplexer outputs MUXP
and MUXN, a 5-bit word selects the connected input.
Table 57. MUX Settings
MUXP/MUXN Setting
Selected Input
Binary [4:0]
11XXX
10111
10110
10101
10100
10011
10010
10001
10000
01111
01110
01101
01100
01011
01010
01001
01000
00111
00110
00101
00100
00011
00010
00001
00000
Dec
31-24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
Reserved
VREF2 via 250k, see Safety Manual for more information
VREF2
Reserved
CF1P
CF1M
CF2P
CF2M
VBATP
VBATM
IPT, see Safety Manual for more information
IMT, see Safety Manual for more information
V12
V11
V10
V9
8
V8
7
V7
6
V6
5
V5
4
V4
3
V3
2
V2
1
V1
0
AGND
Rev A
58
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
In slow, high precision mode the auxiliary channel (CHAUX) converts in Round Robin mode two differential inputs
signals. For each of the two slots, the inputs multiplexed to MUXP and MUXN can be chosen by programming the
corresponding 5-bit setting in the following four registers.
In fast mode, the auxiliary channel (CHAUX) converts only one differential input signal, which can be chosen by choos-
ing the corresponding 5-bit setting in the Fast MUXP and MUXN Control registers.
During fast continuous measurements it is not possible to change the AUX MUX configuration. See note in chapter
'Fast Measurement Timings'.
Table 58. Multiplexer Control Registers in Slow Mode
ADDR
0xEB
0xEC
0xED
0xEE
BIT
SYMBOL
TYPE
RW
RW
RW
RW
DEFAULT
00000
00000
00000
00000
OPERATION
[4:0]
[4:0]
[4:0]
[4:0]
SLOT1MUXN [4:0]
SLOT1MUXP [4:0]
SLOT2MUXN [4:0]
SLOT2MUXP [4:0]
Slot1 MUXN [4:0] setting, binary coded, see Table 57
Slot1 MUXP [4:0] setting, binary coded, see Table 57
Slot2 MUXN [4:0] setting, binary coded, see Table 57
Slot2 MUXP [4:0] setting, binary coded, see Table 57
Table 59. Multiplexer Control Registers for Fast Mode
ADDR
0xF3
0xF4
BIT
SYMBOL
TYPE
RW
DEFAULT
00000
OPERATION
[4:0]
[4:0]
FAMUXN [4:0]
FAMUXP [4:0]
Fast Mode MUXN [4:0] setting, binary coded, see Table 57
Fast Mode MUXP[4:0] setting, binary coded, see Table 57
RW
00000
The Fast Control Register allows to configure and trigger fast conversions.
There is a timing constraint when writing to FGPIOCTRL and FACTRL, see description for FGPIOCTRL.
Table 60. Fast Control Register FACTRL (0xF5)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
0
FACONV
RW
0
Continuous fast conversion enable (1) / disable (0).
Conversion results are written to a FIFO if at least one of
bits FACHA, FACH1, FACH2 is set.
1
2
3
FACHA
FACH1
FACH2
RW
RW
RW
0
0
0
Channel AUX fast mode configuration bit. Fast conversion
starts within some us after rising edge of FACONV or
when issuing the ADCV command.
Channel 1 fast mode configuration bit. Fast conversion
starts within some us after rising edge of FACONV or
when issuing the ADCV command.
Channel 2 fast mode configuration bit. Fast conversion
starts within some us after rising edge of FACONV or
when issuing the ADCV command.
Rev A
59
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
The GPIO control registers allow to configure the GPIO pins to be either tristate, low, high or toggling at 400kHz by
setting the corresponding GPIO CTRL bits in 0xF1 and 0xF2.
GPIO4 and GPIO5 can be used as heartbeat pins toggling with a frequency of 400kHz and become static low upon an
alert. While GPIO5 is activated by enabling any of the overcurrent comparators and is dedicated to alerts issued by
these comparators, GPIO4 can be configured by the GPIO4 heartbeat control register (0xE8) to respond on any alert
not masked by the mask registers. The setting in the GPIO4 heartbeat control register overrules the GPIO4CTRL setting
in 0xF2. Similarly GPIO5CTRL setting in 0xF1 is overruled by enabling the overcurrent comparators.
Table 61. GPIO4 Heartbeat Control GPIO4HBCTRL (0xE8)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
0
GPIO4HBEN
RW
0
GPIO4 heartbeat master enable control
0: GPIO4 is not configured as heart beat.
1: GPIO4 is configured as heart beat and unmasked alerts
(see mask registers) stop heart beat of GPIO4
Table 62. Current Source and GPIO5 Control FCURGPIOCTRL (0xF1)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
[1:0]
GPIO5CTRL [1:0]
RW
00
GPIO5 CTRL:
[00]=Tristate
[01]=LOW(DGND),
[10]=Toggle at 400kHz
[11] =HIGH(DVCC)
4
5
6
7
MUXPCURPOL
MUXPCUREN
MUXNCURPOL
MUXNCUREN
RW
RW
RW
RW
0
0
0
0
0: MUXP sinks 250µA of current
1: MUXP sources 250µA of current
0: MUXP current source off
1: MUXP current source on
0: MUXN sinks 250µA of current
1: MUXN sources 250µA of current
0: MUXN current source off
1: MUXN current source on
The 250μA current sources at MUXP and MUXN allow open wire detection at all multiplexer inputs. If enabled, the
current sources connect to the inputs during the time the ADC is connected and performs the conversion, no matter
if it is the fast or slow channel. Typically, open wire detection is performed using the fast channel by first writing cur-
rent source, GPIO and MUX control (4-byte write to 0xF1-0xF4), triggering the fast conversion by sending the ADCV,
reading the results with RDCV and finally by writing again 4 bytes to 0xF1-0xF4 to set the next MUX input and current
source configuration. This way, the time when current sources are enabled on a dedicated pin, is precisely timed. See
the safety manual for more information.
Enabled current source on MUXP will alter the internal measurement of VREF (~4V/~0.7V when MUXPCURPOL = 1 /
0) and this change won't be visible on the external VREF pin. If enabled, also the NTC temperature measurement and
the compensation of a shunt resistor's temperature drift will be altered, as they depend on the internal VREF measure-
ment. The correct VREF voltage can always be measured via an external connection to one Vx pin, see also section
Unused Input Pins V1-V12.
Rev A
60
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 63. GPIO Control FGPIOCTRL (0xF2)
BIT
SYMBOL
TYPE
RW
RW
RW
RW
DEFAULT
OPERATION
[1:0]
[3:2]
[5:4]
[7:6]
GPIO1CTRL [1:0]
GPIO2CTRL [1:0]
GPIO3CTRL [1:0]
GPIO4CTRL [1:0]
00
00
00
00
GPIO1-4 CTRL:
[00]=Tristate
[01]=LOW(DGND),
[10]=Toggle at 400kHz
[11] =HIGH(DVCC)
If set to tristate, the GPIO pins can be used as analog inputs (V8-V12) to the auxiliary channel by choosing the
corresponding multiplexer settings according to Table 57.
Write to register FCURGPIOCTRL does not take effect immediately. Instead any changes will only become active once
FGPIOCTRL is written. Thus it is recommended to always write both registers in a single burst.
Additionally there is a timing constraint when writing to FGPIOCTRL and FACTRL. Writing to those registers must
always be delayed by a minimum of 1ms. Thus, its not allowed to write FGPIOCTRL and FACTRL in a single burst.
Usually this can be easily achieved by partitioning code sections properly that control those registers.
Table 64. RDCV Indirect Address (0xFC)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
[7:0]
RDCVIADDR
RW
0
Address pointer for indirect memory access via RDCV
command, see section Indirectly Addressed RDCV
Command
The following register is provided for software debug purposes only. It allows synchronization of the host controller
to the CORE system tick. DBGCNT is reset by the CORE after the last register of memory page 0 was updated. This
happens typically 23ms after the end of a slow mode current/power conversion (EOC of I1, I2, P1, P2).
Table 65. DBGCNT (0xD5)
SYMBOL
SI/UI
OPERATION
DBGCNT
UI
Debug counter, counts milliseconds, resets once per core cycle. In MEASURE mode it typically
counts 0 to 100; in STANDBY mode it typically counts 0 to 17.
Rev A
61
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
REGISTER MAP PAGE1
PAGE1 of the LTC2949 register map contains threshold and configuration registers. The threshold registers allow to set
threshold values for each measured quantity. The configuration registers allow to store application and board specific
parameters and settings, which usually do not need to be modified during operation.
Software Reset
The LTC2949 has a software reset feature, described in following table.
Table 66. RSTUNLCK (p1.0xA9)
SYMBOL
OPERATION
RSTUNLCK
Writing 0x55 to this register will unlock the RESET function within OPCTRL. After putting LTC2949 into SLEEP mode a
writing command to OPCTR with the value 0x80 will issue the reset. Detailed steps to reset LTC2949 are:
1. Write 0 to OPCTRL (0xF0)
2. Write 0x55 to RSTUNLCK (0xA9 on page 1)
3. Wait 130ms
4. Go to SLEEP (set bit SLEEP in OPCTRL)
5. Wait 20ms
6. Write 0x80 to OPCTRL (0xF0)
Threshold Registers
The threshold registers set threshold values for each measured quantity. When a measured value exceeds its threshold,
an alert is triggered and the corresponding bits in the threshold and overflow alert registers (0x81 to 0x87) are set.
When GPIO4 heartbeat is enabled in register GPIO4HBCTRL (0xE8), unmasked alerts in registers (0x88 to 0x8F) stop
heartbeat on GPIO4. Value scaling is done in the same manner as in the corresponding result register values, using
LSB values from Table 26, Table 27 and Table 28. Non-Accumulated Results Register Parameters.
Table 67. Threshold Registers
ADDRESS
p1.0x00
p1.0x06
p1.0x0C
p1.0x10
p1.0x16
p1.0x20
p1.0x26
p1.0x2C
p1.0x30
p1.0x36
p1.0x44
p1.0x4C
p1.0x54
NAME
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x7FFF FFFF FFFF
0x8000 0000 0000
0xFFFF FFFF
0x7FFF FFFF FFFF
0x8000 0000 0000
0x7FFF FFFF FFFF
0x8000 0000 0000
0xFFFF FFFF
0x7FFF FFFF FFFF
0x8000 0000 0000
0x7FFF FFFF FFFF FFFF
0xFFFF FFFF
PARAMETER
C1TH[47:0]
C1TL[47:0]
TB1TH[31:0]
E1TH[47:0]
E1TL[47:0]
C2TH[47:0]
C2TL[47:0]
TB2TH[31:0]
E2TH[47:0]
E2TL[47:0]
C3TH[63:0]
TB3TH[31:0]
C3TL[63:0]
Charge1 threshold high
Charge1 threshold low
Timebase1 threshold high
Energy1 threshold high
Energy1 threshold low
Charge2 threshold high
Charge2 threshold high
Timebase2 threshold high
Energy2 threshold high
Energy2 threshold low
Charge3 threshold high
Timebase3 threshold high
Charge3 threshold low
0x8000 0000 0000 0000
Rev A
62
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 67. Threshold Registers (continued)
ADDRESS
p1.0x64
p1.0x6C
p1.0x74
p1.0x80
p1.0x82
p1.0x84
p1.0x86
p1.0x88
p1.0x8A
p1.0x8C
p1.0x8E
p1.0x90
p1.0x92
p1.0x94
p1.0x96
p1.0x98
p1.0x9A
p1.0xA0
p1.0xA2
p1.0xA4
p1.0xA6
NAME
E4TH[63:0]
TB4TH[31:0]
E4TL[63:0]
I1TH[15:0]
I1TL[15:0]
P1TH[15:0]
P1TL[15:0]
I2TH[15:0]
I2TL[15:0]
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x7FFF FFFF FFFF FFFF
0xFFFF FFFF
0x8000 0000 0000 0000
0x7FFF
PARAMETER
Energy4 threshold high
Timebase4 threshold high
Energy4 threshold low
Current1 threshold high
Current1 threshold low
Power1 threshold high
Power1 threshold low
Current2 threshold high
Current2 threshold low
Power2 threshold high
Power2 threshold low
BAT threshold high
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
0x8000
0x7FFF
P2TH[15:0]
P2TL[15:0]
BATTH[15:0]
BATTL[15:0]
TEMPTH[15:0]
TEMPTL[15:0]
VCCTH[15:0]
VCCTL[15:0]
SLOT1TH[15:0]
SLOT1TL[15:0]
SLOT2TH[15:0]
SLOT2TL[15:0]
BAT threshold low
Die temperature threshold high
Die temperature threshold low
VCC threshold high
VCC threshold low
Slot1 threshold high
Slot1 threshold low
Slot2 threshold high
0x8000
Slot2 threshold low
FLOAT24 Format
The NTC configuration parameters and the gain correction factors described in the following paragraphs are stored as
floating point numbers represented by the FLOAT24 format according to IEEE 754 standard. The LTC2949 implemen-
tation uses 1-bit for sign, a 7-bit exponent in two’s complement format with 63 as bias and 16-bits for the mantissa,
with an implicit leading bit of value 1 unless the exponent is stored with all zeros. As an example the value of 0.95 is
represented by the 3 bytes number 0x3EE666 as shown below:
Table 68. FLOAT24 Example
3E
E6
66
0
0
1
1
1
1
1
0
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
Sign
1 x
Exponent
2^(62-63)
Mantissa MSB
Mantissa LSB
x
(1+0.899994)
= 0.5 • 1.899994 = 0.94999
The GUI controlling the LTC2949 demo board supports converting numbers to FLOAT24. The code section of the
LTC2949 (https://www.analog.com/en/products/ltc2949.html#product-tools) also provides conversion functions
++
written in C/C for this purpose.
LTC2949.cpp contains following conversion functions:
void LTC2949_FloatToF24Bytes(float f32, byte* bytes)
void LTC2949_F24BytesToFloat(byte* bytes, float* f32)
Rev A
63
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Configuration Registers
The following registers allow to configure LTC2949 application specific. Please note that LTC2949 must be in STANDBY
and an update of these registers must be requested by setting the ADJUPD bit in the Operation Control Register (OPC-
TRL) to make changes effective.
ADC Configuration Register
The ADC Configuration Register allows to disable power multiplication on P1ADC and P2ADC, and to turn on NTC
linearization of Slot1/2 measurements.
Table 69. ADC Configuration ADCCONF (p1.0xDF)
BIT
SYMBOL
TYPE
DEFAULT
OPERATION
0
P1ASV
RW
0
0: P1ADC configured to power mode
1: P1ADC configured to voltage mode
1
3
P2ASV
NTC1
RW
RW
0
0
0: P2ADC configured to power mode
1: P2ADC configured to voltage mode
0: SLOT1 voltage measurement.
SLOT1 output LSB = 375µV
1: SLOT1 NTC temperature measurement.
SLOT1 output LSB = 0.2˚C
4
6
NTC2
RW
RW
0
0
0: SLOT2 voltage measurement.
SLOT2 output LSB = 375µV
1: SLOT2 NTC temperature measurement.
SLOT2 output LSB = 0.2˚C
NTCSLOT1
0: TC compensation of shunt for I1 (I2) is linked to temperature measurement via SLOT1 (SLOT2).
1: TC compensation of shunt for I1 and I2 is linked to temperature measurement via SLOT1 only.
See NTC Configuration Registers and section Temperature Measurement for more information regarding NTC tem-
perature measurement.
See Table 28 and Table 31 for effect of P1ASV/P2ASV on LSB sizes or Power-ADC related measurements/settings.
Table 70. Allowed Combinations of FACTRL and ADCVCONF (PxASV) Configurations.
FACONV
0 or 1
0 or 1
0 or 1
0 or 1
0 or 1
0 or 1
FACH1
FACH2
P1ASV
0 or 1
1
P2ASV
0 or 1
1
FAST BAT (b)
SLOW P1
Not Supported!
SLOW P2
1
0
0
0
0
0
1
1
1
1
1
1
1
1
0
1
0
1
1
1
1
1
1
1
1
1
1
1
0 or 1
0 or 1
N/A
N/A
P2
N/A
N/A
P2
P1
P1
N/A
P2
P1
18-Bit Power or Voltage
0
0
1
0
0
1
1
0
0
1
1
0
1
0
0
1
0
1
0
1
0
1
18-Bit Power
18-Bit Power
18-Bit Voltage
11-Bit Power (c)
11-Bit Power (c)
11-Bit Voltage (c)
11-Bit Voltage (c)
18-Bit Power (a)
18-Bit Power (a)
18-Bit Voltage (a)
N/A
N/A
N/A
1
1
1
1
0
0
0
0
11-Bit Voltage (c)
11-Bit Power (c)
18-Bit Voltage (a)
18-Bit Power (a)
P1
18-Bit Voltage (a)
a) Any fast single shot measurement will interrupt the slow channel measurements. Slow channel will only resume to be updated if there was no fast
single shot measurement for at least 160ms.
b) 15-bit Fast BAT (VBATP-VBATM) voltage measurement via one power ADC. It is always possible to measure VBATP-VBATM via AUX ADC in fast
(FACHA=1) or slow mode.
c) Power ADCs (PxASV=0) of channels configured to fast mode will provide 11-bit conversions at an update rate of 100ms via P1/P2 registers. The LSB
size stays the same, but the resolution in this case is only guaranteed to be 11-bit.
Rev A
64
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
NTC Configuration Registers
When bits NTC1 and NTC2 in the ADC Configuration Register are set, LTC2949 reports the result of the corresponding
CHAUX slot in temperature by comparing NTC resistance to reference resistors and solving Steinhart-Hart equations.
The NTC Configuration Registers allow to set values of Steinhart-Hart coefficients (A,B,C) and reference resistors.
Furthermore, linear temperature coefficients of up to two external sense resistors can be compensated by thermally
closely coupled NTC thermistors. Temperature compensation is enabled by writing the respective temperature coef-
ficient and its reference temperature (T ) in the NTC Configuration Register. See section Temperature Measurement
0
and Sense Resistor Temperature Compensation for more details.
Table 71. NTC Configuration Registers
ADDRESS
p1.0xAA
p1.0xAD
p1.0xD0
p1.0xD3
p1.0xD6
p1.0xD9
p1.0xDC
p1.0xE0
p1.0xE3
p1.0xE6
p1.0xE9
p1.0xEC
p1.0x5C
p1.0x7C
NAME
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x000000
0x000000
0x000000
0x000000
0x000000
0x000000
0x0000
PARAMETER
NTC1 reference resistor
UNIT
Ω
FORMAT
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
RREF1[23:0]
RREF2[23:0]
NTC1A[23:0]
NTC1B[23:0]
NTC1C[23:0]
RS1TC[23:0]
RS1T0[16:0]
NTC2A[23:0]
NTC2B[23:0]
NTC2C[23:0]
RS2TC[23:0]
RS2T0[16:0]
RS1TC2[23:0]
RS2TC2[23:0]
NTC2 reference resistor
Ω
NTC1 Coefficient A
NTC1 Coefficient B
NTC1 Coefficient C
Sense resistor 1 (RS1) temperature coefficient (TC)
Reference temperature for TC compensation of RS1
NTC2 Coefficient A
1/K
°C
0x000000
0x000000
0x000000
0x000000
0x0000
NTC2 Coefficient B
NTC2 Coefficient C
Sense resistor 2 (RS2) temperature coefficient (TC)
Reference temperature for TC compensation of RS2
2nd order TC of sense resistor 1
2nd order TC of sense resistor 2
1/K
°C
2
0x000000
0x000000
1/K
2
1/K
Note that for the two 16-bit registers storing reference temperature RS1T0 and RS2T0 the least significant byte (LSB) of the mantissa is implicitly 0
Gain Configuration Registers
The LTC2949 can store gain correction factors for two current sense resistor, the battery voltage divider and four mul-
tiplexer inputs to compensate tolerances of external components in its Gain Setting Registers located between 0xB0
and 0xCF on measured page 1 of the register map. When these factors are set different than their default value of 1.0,
the LTC2949 corrects the computed values of current, voltage, power, charge and energy accordingly. As the LTC2949
accumulators can be configured to compute the sums of charge and energy flown through two sense resistors, also
the nominal ratio between the two sense resistors RSRATIO = R /R can be stored. The LTC2949 then multiplies the
S1 S2
measurements of CH2 (I2, P2) by RSRATIO before adding them to the sum of charge (C3) or energy (E4). All these
factors are stored in the Float24 format according to IEEE 754 standard described earlier.
Table 72. Gain Configuration Registers
ADDRESS
p1.0xB0
p1.0xB3
p1.0xB6
p1.0xB9
p1.0xC0
p1.0xC3
p1.0xC6
p1.0xC9
NAME
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x3F00 00
0x3F00 00
0x3F00 00
0x3F00 00
0x3F00 00
0x3F00 00
0x3F00 00
0x3F00 00
PARAMETER
Sense resistor 1(R ) gain correction factor
FORMAT
Float24
Float24
Float24
Float24
Float24
Float24
Float24
Float24
RS1GC[23:0]
RS2GC[23:0]
RSRATIO[23:0]
BATGC[23:0]
MUX1GC[23:0]
MUX2GC[23:0]
MUX3GC[23:0]
MUX4GC[23:0]
S1
Sense resistor 2(R ) gain correction factor
S2
Nominal ratio of R to R
S1
S2
BAT gain correction factor
Gain correction factor for MUX setting 1
Gain correction factor for MUX setting 2
Gain correction factor for MUX setting 3
Gain correction factor for MUX setting 4
Rev A
65
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
Table 72. Gain Configuration Registers (continued)
ADDRESS
p1.0xBC
p1.0xBD
p1.0xBE
p1.0xBF
p1.0xCC
p1.0xCD
p1.0xCE
p1.0xCF
NAME
TYPE
RW
RW
RW
RW
RW
RW
RW
RW
DEFAULT
0x00
0x00
0x00
0x00
0x00
0x00
0x00
0x00
PARAMETER
MUXN gain correction setting 1
MUXP gain correction setting 1
MUXN gain correction setting 2
MUXP gain correction setting 2
MUXN gain correction setting 3
MUXP gain correction setting 3
MUXN gain correction setting 4
MUXP gain correction setting 4
FORMAT
MUXNSET1[4:0]
MUXPSET1[4:0]
MUXNSET2[4:0]
MUXPSET2[4:0]
MUXNSET3[4:0]
MUXPSET3[4:0]
MUXNSET4[4:0]
MUXPSET4[4:0]
MUXN binary coded, see Table 57
MUXP binary coded, see Table 57
MUXN binary coded, see Table 57
MUXP binary coded, see Table 57
MUXN binary coded, see Table 57
MUXP binary coded, see Table 57
MUXN binary coded, see Table 57
MUXP binary coded, see Table 57
As an example, if a sense resistor of nominal 100µΩ is used for CH1, but a board calibration reveals the sense resistor
value to be 102µΩ, a factor of 100/102 = 0.9804 should be written to RS1GC [23:0] = float24(0.9804) = 0x3EF5F5.
Assuming a sense resistor of nominal value 10mΩ but real value of 9.8mΩ is used for CH2 the factor 1.024 should be
programmed in RS2GC [23:0] = float24(1.024) = 0x3F0624 and a factor 0.01 should be programmed in RSRATIO[23:0]
= float24(0.01) = 0x3847AE.
In many applications the LTC2949 measures high voltages using external resistor dividers which suffer from gain
errors due to resistor tolerances. The LTC2949 allows to store gain correction factors for the measurement of battery
voltage and four programmable MUX settings. E.g.: The LTC2949 will apply gain correction of 0.9 to differential mea-
surements between V1 and V2 if registers MUX1GC[23:0] = 0x3ECCCC, MUXPSET1[7:0] = 0x01 and MUXNSET1[7:0]
= 0x02, see also Table 57.
The assignment between MUX[1-4]GC and MUX[P,N ]SET[1-4] is independent of the polarity of the MUX settings.
Related to the example above the same gain correction is applied for measurements V1-V2 and V2-V1. Also swap-
ping the register values of MUXPSET1 and MUXNSET1 leads to the same behaviour. The link between gain correction
parameters and measurements is summarized in Table 73.
All Gain Configuration Registers can be copied to an external EEPROM, enabling a modular approach to factory cali-
bration of application boards.
Table 73. Relation Between Gain Correction Parameters and Measurements
NAME
AFFECTED MEASUREMENTS
BATGC
BAT for slow and fast measurements
P1, P2, E1, E2, E4
SLOT[1,2] AUX measurements if SLOT[1,2]MUX[P, N ] is set to VBATP, VBATM of any polarity
Fast AUX conversions if FAMUX[P, N ] is set to VBATP, VBATM of any polarity
Slow and fast AUX measurements if SLOT[1,2]MUX[P, N ] or FAMUX[P, N ] matches MUX[P, N ]SET[1-4] of any polarity
I1, I2, P1, P2, C1, C2, C3, E1, E2, E4
MUX[1-4]GC
RS[1,2]GC
RSRATIO
C3, E4 (Note: The charge that is accumulated to C3 is (I1 + RSRATIO • I2) • dT, the energy that is accumulated to E4 is
(P1 + RSRATIO • P2) • dT.)
Rev A
66
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
External EEPROM Control Register
To prevent data loss when the LTC2949 is not powered, it can store its entire register content in an external EEPROM
2
via its dedicated I C interface. The communication to an EEPROM is controlled by the EEPROM Control Register
(EEPROMCTRL).
Table 74. EEPROM Control Register EEPROMCTRL (p1.0x50)
BIT
0
SYMBOL
INIT
TYPE
SO
DEFAULT
OPERATION
Write signature to EEPROM
DURATION
40ms
0
0
0
0
0
0
0
0
1
CHECK
SO
Check the signature in EEPROM
Save MEM (but special row) to EEPROM
Restore EEPROM to MEM (but special row)
Result of INIT
25ms
2
SAVE
SO
1100ms
1250ms
3
RESTORE
INITRSL
CHECKRSL
SAVERSL
RESTORERSL
SO
4
RW
RW
RW
RW
5
Result of CHECK
6
Result of SAVE
7
Result of RESTORE
The lower 4-bits of the EEPROMCTRL are SET ONLY and trigger dedicated communications with the EEPROM, while
the higher 4-bits are read/write and are set by the LTC2949 after the typical time given in the column Duration upon
a successful termination of a communication and must be reset by a write command from the host before the next
communication is requested.
Before any other interaction, the EEPROM must be initialized by setting the INIT bit, which causes LTC2949 to write a
defined signature to the EEPROM. Once the signature is written, LTC2949 will reset the INIT bit and set the INITRSL
2
bit if the EEPROM has acknowledged according to the I C protocol. Any other interaction with the EEPROM will be
preceded by reading and checking this signature – additionally the signature can be checked by setting bit CHECK and
verifying that bit CHECKRSL is set after bit CHECK was reset by LTC2949.
Setting bit SAVE causes the LTC2949 to save its entire memory except the last common row of both register pages
to the EEPROM together with a CRC calculated from the entire register content. LTC2949 signals a successful save
operation by setting bit SAVERSL, based on the check of the signature and the acknowledge of the EEPROM.
Setting bit RESTORE causes the LTC2949 to copy the content of the EEPROM to its internal RAM, calculate the CRC
and set bit RESTORERSL if the CRC was found correct.
Any communication with the external EEPROM requires the LTC2949 to be in STANDBY mode to prevent data loss or
data corruption.
Many applications require to store additional custom data, e.g. serial number. The reserved registers p1.0x1C - p1.0x1F
(4 bytes) and p1.0x3C - p1.0x3F (4 bytes) can be used for that purpose. Additionally, registers of LTC2949 that do not
require a board specific initialization and thus are initialized by the host controller, can be used to store custom data.
For example, all accumulators are either just initialized with zero or via some state-of-charge algorithm. All threshold
values are typically hard coded or initialized by some higher instance in the system. Mux setting registers are adjusted
programmatically. Min/maxtrackingregisterscanbeinitializedbythehostcontroller. TheregistermapshowninFigure18
gives an overview of those registers. Using all of them allows to increase the number of custom data bytes to 240 bytes.
To access this data the host commands an EEPROM RESTORE, reads all registers with custom data and initializes the
registers afterwards to their desired value.
Rev A
67
For more information www.analog.com
LTC2949
REGISTER DESCRIPTION
ꢄꢅꢂꢆ
ꢓꢔꢍꢀ
ꢂꢃ
ꢂꢃ
ꢏꢄꢋꢉꢐꢋꢑAꢀꢄꢂ
ꢒꢉꢙ
ꢇꢈꢉꢄꢊꢋꢊ
ꢒꢉꢇ
ꢒꢕA
ꢖꢗꢕ
ꢀꢁ
ꢌꢌꢍRꢎꢏ
ꢒꢕA
ꢄꢊꢋꢊ ꢘꢄꢀ
Figure 21. EEPROM Connection. Recommended 4k Bit Automotive Qualified EEPROM with Internal EEC:
M24C04-A125 from STMicroelectronics or without Internal EEC: AT24C04C or 24AA04 from Microchip.
APPLICATION INFORMATION
TEMPERATURE MEASUREMENT
The following table shows the relevant registers from the
NTC configuration register for the application shown in
Figure 22.
TheLTC2949shighimpedanceinputsV1-V12canbeused
to measure temperature by means of thermistors and a
reference resistor as shown below.
As VREF is measured with the same ADC as V1-V12,
imperfections of VREF and gain error of the ADC do not
impact the temperature measurement accuracy.
WhenbitsNTC1orNTC2intheADCConfigurationRegister
are set, LTC2949 reports the result of the corresponding
CHAUXslotinslowhighprecisionmodeintemperatureby
comparing the resistance of a thermistor (NTC) R
referenceresistorandsolvingtheSteinhart-Hartequation.
to a
NTC
ꢍRꢇꢎ
R
ꢊꢈꢑ
Rꢇꢎ
1
T
3
ꢀꢁꢂꢃꢄꢅꢄ
= A+B•InRNTC +C•(InRNTC
)
ꢍꢊ
ꢆꢁꢂ
ꢆꢁꢂꢀꢇꢃꢈꢉꢇꢉꢊꢈꢉꢋꢌꢈ
The value of the reference resistor R and the Steinhart-
REF
Hart coefficients (A, B, C) need to be stored in the NTC
Configuration Registers. Steinhart-Hart coefficients are
commonly specified parameters provided by thermistor
manufacturers or can be deduced from provided resis-
tance tables.
Aꢏꢆꢐ
ꢃꢄꢅꢄ ꢎꢃꢃ
Figure 22. Connecting Thermistors
Table 75. NTC1 Values in NTC Configuration Register
PARAMETER
NTC1 reference resistor
NTC1 Coefficient A
NTC1 Coefficient B
NTC1 Coefficient C
VALUE
10k
UNIT
ADDRESS
p1.0xAA
p1.0xD0
p1.0xD3
p1.0xD6
NAME
VALUE
FORMAT
Float24
Float24
Float24
Float24
Ω
RREF1[23:0]
NTC1A[23:0]
NTC1B[23:0]
NTC1C[23:0]
0x4C3880
0x352A5F
0x32E7F1
0x279079
1.1382e-3
2.3267e-4
0.93243e-7
Rev A
68
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
SENSE RESISTOR TEMPERATURE COMPENSATION
The following linear equation system has to be solved to
calculateA,B,Cusingthreevalues,R1(T1),R2(T2),R3(T3)
from a R versus T (in K) resistor table:
The LTC2949 can be configured to compensate the tem-
perature dependency of the used current sense resistors
upto2ndorderbasedontemperaturemeasurementswith
external NTCs. The compensation is enabled by writing
thetemperaturecoefficients(TC,TC2)ofthesenseresistor
−1
T1
1 lnR1 ln3 R1
⎛
⎜
⎜
⎜
⎞ ⎛
⎟ ⎜
⎟ ⎜
⎞
⎟
⎟
⎟
⎛ ⎞
A
⎜ ⎟
−1
T2 = 1 lnR2 ln3 R2
B
⎜ ⎟
−1⎟ ⎜
⎜ ⎟
and the reference temperature (T ) in the NTC configura-
3
0
C
⎜
⎟ ⎜
⎟
⎝ ⎠
T3
1 lnR3 ln R3
⎝
⎠ ⎝
⎠
tion register. The LTC2949 will then compensate for the
temperature induced deviation of the sense resistor from
After solving the linear equation system, A, B, C can be
expressed as:
its nominal value R according to:
0
2
RSENSE =R •[1+TC •(T−T )+TC2• T−T0 ]
(
)
O
O
l1=lnR1
l2 =lnR2
l3 =lnR3
The temperature coefficient and reference temperature
can be set for each sense resistor individually in the NTC
configuration register. Table 76 shows the programmed
coefficientsforacoppersenseresistorwithatemperature
coefficient of 3900ppm/K whose nominal value R was
measured at 20°C.
T2−1−T1−1
m2 =
l2 −l1
0
T3−1−T1−1
m3 =
If temperature compensation is enabled for one or both
channels, thesense resistor connected to CH1 is compen-
sated with the temperature of NTC1 measured during the
first slot of CHAUX, while the sense resistor connected to
CH2 is compensated with the temperature of NTC2 mea-
sured during the second slot of CHAUX. The multiplexer
settings must be set such that the input pin connected
to the corresponding NTC is selected during the respec-
tive slot. If CHAUX is changed to fast mode, the last NTC
temperature measurements taken in Round Robin mode
are used for temperature compensation.
l3 −l1
m3 −m2
l3 −l2
l3 +l2 +l1
C=
B=m2 −C•(l12 +l1•l2 +l22 )
A =T1−1−(B+l12•C)•l1
Rev A
69
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
For single shunt scenarios the temperature measured by
one NTC with SLOT1 can be used to compensate both
channels. See ADCCONF, bit NTCSLOT1.
setting I to 0.5mA is a good compromise between power
B
consumption and noise immunity. Using this I setting
B
with a 1:1 transformer and R = 100Ω, R should be
M
B1
set to 2.8k and R set to 1.2k. In a typical CAT5 twisted
B2
isoSPI Setup
pair these settings will allow for communication up to
50m. For applications that require cables longer than
The LTC2949 allows the isoSPI link in each application
to be optimized for power consumption or for noise
immunity. The power and noise immunity of an isoSPI
50m it is recommended to increase the I to 1mA. This
B
compensates for the increased insertion loss in the cable
and maintains high noise immunity. So when using cables
over 50m and, again, using a transformer with a 1:1 turns
system is determined by the programmed I current.
B
The I current can range from 100μA to 1mA. A low I
B
B
ratio and R = 100Ω, R would be 1.4k and R would
M
B1
B2
reduces the isoSPI power consumption in the READY and
ACTIVE states, while a high I increases the amplitude
be 600Ω. Other I settings can be used to reduce power
B
B
consumption or increase the noise immunity as required
of the differential signal voltage V across the matching
A
by the application. In these cases when setting threshold
termination resistor, R . I is programmed by the sum
M
B
voltage V
and choosing R and R resistor values
B1 B2
ICMP
of the R and R resistors connected between the I
B1
B2
BIAS
the following rules should be used:
pin and GND as shown in Figure 23. For most applications
Table 76. Sense Resistor Temperature Coefficients
PARAMETER
VALUE
0.0039
20
UNIT
1/K
°C
ADDRESS
p1.0xD9
p1.0xDC
NAME
VALUE
FORMAT
Float24
Sense resistor 1 (RS1) temperature coefficient (TC)
Reference temperature for TC compensation of RS1
RS1TC[23:0]
RS1T0[16:0]
0x36FF2E
0x4340
Float24, 2 bytes
Note: For copper only the first order temperature coefficient is relevant and thus TC2=0.0 (default).
Table 77. Procedure to Enable Temperature Compensation of Sense Resistor
WHAT
HOW AND WHERE
CONT= 0 in OPCTRL
ADDRESS
0xF0
Step 1
Step 2
Step 3
Step 4
Step 5
Step 6
Set LTC2949 to Standby Mode
Write NTC and R
DATA
NTC configuration Registers
NTCx = 1 in ADC Configuration Register
Mux Setting Registers
See Table 71.
p1.0xDF
0xEB-0xEE
0xF0
SENSE
Set AUX Slot to Temperature Mode
Select MUX input channel with connected NTC
Update Configuration Registers
ADJUP = 1 in OPCTRL
Set LTC2949 to Continuous Mode
CONT = 1 in OPCTRL
0xF0
ꢁꢄꢎꢋAꢌꢁꢎꢘ ꢃARRꢁꢕR
ꢙꢀAꢚ ꢛꢄꢕ ꢎꢘꢕ ꢎR ꢌꢔꢎ ꢌRAꢘꢉꢎRꢀꢕRꢄꢜ
ꢁꢄꢎꢀꢏ ꢁꢂꢃ
ꢁꢂ
ꢝ
ꢝ
ꢝ
ꢝ
ꢀAꢄꢌꢕR
R
R
ꢋꢌꢅꢑꢒꢓꢓ
ꢀꢎꢄꢁ ꢁꢀꢃ
ꢀꢁꢄꢎ ꢁꢃꢁAꢄ
ꢄꢅꢐ
ꢋꢌꢅꢆꢇꢈꢇ
ꢀ
ꢀ
ꢄꢏꢎ
ꢄꢏꢁ
ꢄꢅꢐ
ꢀꢁ
ꢁꢀ
ꢁꢃꢁAꢄ
ꢌꢔꢁꢄꢌꢕꢏꢖꢂAꢁR ꢅAꢃꢋꢕ
ꢔꢁꢌꢗ ꢅꢗARAꢅꢌꢕRꢁꢄꢌꢁꢅ ꢁꢀꢂꢕꢏAꢘꢅꢕ R
ꢀ
R
R
R
R
ꢃꢍ
ꢃꢆ
ꢃꢍ
ꢃꢆ
ꢀꢁ
ꢁꢅꢀꢂ
ꢁꢅꢀꢂ
ꢆꢇꢈꢇ ꢉꢆꢊ
Figure 23. isoSPI Circuit
Rev A
70
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
For cables under 50m:
inductances above 60µH and a 1:1 turns ratio. It is also
necessary to use a transformer with less than 2.5µH of
leakage inductance. In terms of pulse shape the primary
inductance will mostly effect the pulse droop of the 50ns
and 150ns pulses. If the primary inductance is too low,
the pulse amplitude will begin to droop and decay over
the pulse period. When the pulse droop is severe enough,
the effective pulse width seen by the receiver will drop
substantially, reducing noise margin. Some droop is ac-
ceptable as long as it is a relatively small percentage of
thetotalpulseamplitude.Theleakageinductanceprimarily
affects the rise and fall times of the pulses. Slower rise
and fall times will effectively reduce the pulse width. Pulse
width is determined by the receiver as the time the signal
is above the threshold set at the ICMP pin. Slow rise and
fall times cut into the timing margins. Generally it is best
to keep pulse edges as fast as possible. When evaluating
transformers, it is also worth noting the parallel winding
capacitance. While transformers have very good CMRR at
lowfrequency,thisrejectionwilldegradeathigherfrequen-
cies, largely due to the winding to winding capacitance.
When choosing a transformer, it is best to pick one with
less parallel winding capacitance when possible.
I = 0.5mA
B
V = (20 • I ) • (R /2)
A
B
M
V
TCMP
V
ICMP
= ½ • V
A
= 2 • V
TCMP
R
B2
R
B1
= V
/I
ICMP B
= (2/I ) – R
B
B2
For cables over 50m:
I = 1mA
B
V = (20 • I ) • (R /2)
A
B
M
V
TCMP
V
ICMP
= 1/4 • V
A
= 2 • V
TCMP
R
B2
R
B1
= V
/I
ICMP B
= (2/I ) – R
B
B2
The maximum data rate of an isoSPI link is determined by
the length of the cable used. For cables 10 meters or less
the maximum 1MHz SPI clock frequency is possible. As
the length of the cable increases the maximum possible
SPI clock rate decreases. This is a result of the increased
propagation delays through the cable creating possible
timingviolations.Figure24showshowthemaximumdata
rate is reduced as the cable length increases when using a
CAT 5 twisted pair. Cable delay affects three timing speci-
fications, t , t and t . In the Electrical Characteristics
ꢌꢑꢒ
ꢀAꢆꢚꢛ Aꢊꢊꢜꢉꢃꢍ
ꢌꢑꢓ
ꢓꢑꢔ
ꢓꢑꢖ
ꢓꢑꢕ
ꢓꢑꢒ
ꢓ
CLK
6
7
table, each is derated by 100ns to allow for 50ns of cable
delay. For longer cables, the minimum timing parameters
may be calculated as shown below:
t , t6 and t7 > 0.9μs + 2 • t
CLK CABLE
ꢌ
ꢌꢓ
ꢌꢓꢓ
ꢀAꢁꢂꢃ ꢂꢃꢄꢅꢆꢇ ꢈꢉꢃꢆꢃRꢊꢋ
Transformer Selection Guide
ꢒꢗꢕꢗ ꢘꢒꢙ
As shown in Figure 23, a transformer or pair of transform-
ers isolates the isoSPI signals between two isoSPI ports.
The isoSPI signals have programmable pulse amplitudes
Figure 24. Data Rate vs Cable Length
up to 1.6V and pulse widths of 50ns and 150ns. To be
P-P
able to transmit these pulses with the necessary fidelity
the system requires that the transformers have primary
Rev A
71
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
When choosing a transformer, it is equally important to
pick a part that has an adequate isolation rating for the
application. The working voltage rating of a transformer
is a key spec when selecting a part for an application. In-
terconnecting daisy-chain links, devices see <60V stress
in typical applications and ordinary pulse and LAN type
transformers will suffice. Multi-drop connections and
connections to the LTC6820, in general, may need much
higherworkingvoltageratingsforgoodlong-termreliabil-
ity. Usually, matching the working voltage to the voltage
of the entire battery stack is conservative. Unfortunately,
transformervendorswilloftenonlyspecifyone-secondHV
testing, and this is not equal to the long-term (permanent)
rating of the part. For example, according to most safety
standards a 1.5kV rated transformer is expected to handle
230V continuously, and a 3kV device is capable of 1100V
long-term, though manufacturers may not always certify
to those levels (refer to actual vendor data for specifics).
Usually, the higher voltage transformers are called high
isolation or reinforced insulation types by the suppliers.
Table 78 shows a list of transformers that have been
evaluated in isoSPI links.
current rejection a common mode choke should also be
placed in series with the IP and IM lines of the LTC2949.
ThecommonmodechokewillbothincreaseEMIimmunity
and reduce EMI emission. When choosing a common
mode choke, the differential mode impedance should be
20Ω or less for signals 50MHz and below. Common mode
chokes similar to what is used in Ethernet applications are
recommended.
Layout of the isoSPI signal lines also plays a significant
role in maximizing the immunity of a circuit. The following
layout guidelines should be followed:
1. The transformer should be placed as close to the isoSPI
cable connector as possible. The distance should be
kept less than 2cm. The LTC2949 should be placed at
least 1cm to 2cm away from the transformer to help
isolate the IC from magnetic field coupling.
2. On the top component layer, no ground plane should
be placed under the transformer, the isoSPI connec-
tor, or in between the transformer and the connector.
3. The isoSPI signal traces should be isolated from sur-
roundingcircuitsandtracesbygroundmetalorspace.
No traces should cross the isoSPI signal lines, unless
separated by a ground plane on an inner layer.
EMC
For the best electromagnetic compatibility (EMC) perfor-
mance, it is recommended to use one of the circuits in
Figure 25 and Figure 26. The center tap of the transformer
should be bypassed with a 10nF capacitor. The center tap
capacitorwillhelpattenuatecommonmodesignals.Large
center tap capacitors greater than 10nF should be avoided
astheywillpreventtheisoSPItransmitterscommonmode
voltage from settling. If a transformer without a center tap
is used, the termination resistor should be split into two
equal halves and connected in series across the IP and IM
lines. The center of the two resistors should be bypassed
with a capacitor as shown. To improve common mode
The isoSPI drive currents are programmable and allow
for a trade-off between power consumption and noise
immunity. The noise immunity of the LTC2949 has been
evaluated using a bulk current injection (BCI) test. The
BCI test injects current into the twisted-pair lines at set
levels over a frequency range of 1MHz to 400MHz. With
the minimum I current, 100μA, the isoSPI serial link was
B
capable of passing a 40mA BCI test with no bit errors. A
40mABCItestlevelissufficientforindustrialapplications.
Automotive applications have a much higher BCI require-
mentsotheLTC2949I issetto1mA,themaximumpower
B
level. The isoSPI system is capable of passing a 200mA
BCI test with no transmitted bit errors. The 200mA test
level is typical for automotive requirements.
Rev A
72
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
Table 78. Recommended Transformers
W
AEC-
Q200
SUPPLIER
PART NUMBER
TEMP RANGE
V
V
/60S CT CMC
HIPOT
H
L
(W/ LEADS) PINS
WORKING
Recommended Dual Transformers
l
l
–
l
l
l
l
l
–
Pulse
Pulse
Pulse
Pulse
Sumida
Sumida
Wurth
Wurth
Wurth
Halo
HX1188FNL
HX0068ANL
–40°C to 85°C
–40°C to 85°C
–40°C to 105°C
–40°C to 125°C
60V (est)
60V (est)
1000V
1.5kV
1.5kV
6.0mm 12.7mm
2.1mm 12.7mm
9.7mm
9.7mm
16SMT
16SMT
10SMT
12SMT
12SMT
16SMT
16SMT
12SMT
12SMT
16SMT
–
–
RMS
RMS
l
l
–
HM2100NL
4.3kVDC
4.3kVDC
3.4mm 14.7mm 14.9mm
4.9mm 14.8mm 14.7mm
9mm 17.5mm 15.1mm
l
l
l
l
l
l
l
HM2112ZNL
1000V
CLP178-C20114
CLP0612-C20115
7490140110
–40°C to 125°C 1000V (est) 3.75kV
RMS
RMS
600V
250V
3.75kV
4kV
5.7mm 12.7mm
9.4mm
–
RMS
RMS
l
–40°C to 85°C
0°C to 70°C
0°C to 70°C
10.9mm 24.6mm 17.0mm
8.4mm 17.1mm 15.2mm
8.4mm 17.1mm 15.2mm
–
RMS
7490140111
1000V (est) 4.5kV
–
–
RMS
RMS
l
l
749014018
250V
4kV
–
RMS
l
TG110-AE050N5LF
–40°C to 85/125°C 60V (est)
1.5kV
6.4mm 12.7mm
9.5mm
RMS
Recommended Single Transformers
Pulse
Pulse
Pulse
Wurth
Wurth
Halo
PE-68386NL
HM2101NL
–40°C to 130°C
–40°C to 105°C
–40°C to 125°C
–40°C to 105°C
–40°C to 125°C
–40°C to 125°C
–40°C to 125°C
–40°C to 105°C
–40°C to 105°C
60V (est)
1000V
1600V
250V
1.5kVDC
4.3kVDC
4.3kVDC
–
–
–
l
l
–
–
–
–
–
–
–
–
–
–
2.5mm 6.7mm
5.7mm 7.6mm
8.6mm
9.3mm
15.5mm
9.1mm
6SMT
6SMT
6SMT
4SMT
6SMT
6SMT
6SMT
6TH
–
l
l
–
l
HM2113ZNL
3.5mm
9mm
750340848
3kV
3kV
3kV
3kV
5kV
–
2.2mm 4.4mm
RMS
RMS
RMS
RMS
RMS
l
l
l
l
l
–
750317011
800V
7.62mm 9.14mm 12.95mm
–
TGR04-6506V6LF
TGR04-A6506NA6NL
TDR04-A550ALLF
ALT4532V-201-T001
300V
10mm 9.5mm
9.4mm 8.9mm
6.4mm 8.9mm
2.9mm 3.2mm
12.1mm
12.1mm
16.6mm
4.5mm
–
l
l
l
–
Halo
300V
Halo
1000V
60V (est)
600V
TDK
~1kV
6SMT
4SMT
8SMT
4SMT
8SMT
Sumida
Sumida
Sumida
TDK
CEEH96BNP-LTC6804/11 –40°C to 125°C
2.5kV
2.5kV
7mm
9.2mm
12.0mm
12.0mm
9.1mm
RMS
RMS
RMS
l
CEP99NP-LTC6804
ESMIT-4180/A
–40°C to 125°C
–40°C to 105°C
–40°C to 125°C
600V
10mm 9.2mm
3.5mm 5.2mm
–
l
250V
3kV
–
RMS
l
VGT10/9EE-204S2P4
250V (est)
2.8kV
10.6mm 10.4mm 12.7mm
–
RMS
COMMON
MODE
CHOKE
TRANSFORMER WITH
COMMON MODE CHOKE
IP
LTC2949
IM
51Ω
51Ω
10nF
IP
LTC2949
IM
100Ω
2949 F26
Figure 26. Recommended isoSPI Circuit for Best EMC
Performance when Using a Transformer without Center Tap
10nF
2949 F25
Figure 25. Recommended isoSPI Circuit for EMC
Table 79. Recommended Common Mode Chokes
MANUFACTURER
TDK
PART NUMBER
ACT45B-220-2P
DLW43SH510XK2
744232102
Murata
Wurth
Rev A
73
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
R
ꢎꢉ
c. VREF can be used to bias resistive dividers to al-
low measurement of voltages below GND. In those
scenarios the differential voltage between Vx and
VREF must be measured which is only possible if
VREF is connected to one of V1-V12 pins.
ꢉꢏꢏꢐ
ꢉꢐ
ꢏꢑꢉꢐ
ꢏꢑꢉꢐ
Aꢋꢌꢍ ꢂꢆꢉꢔ ꢈꢉꢔ
2. Connect divided supply voltage AVCC/DVCC (e.g. 10k
betweenGND/Vxand20kbetweenVxandAVCC/DVCC)
to one of V1-V12
ꢈꢉꢊ ꢂꢆꢉꢊ
ꢕꢖAꢁꢊ
ꢀꢁꢂꢃꢄꢅꢄ
ꢖAꢁꢁ
ꢉꢏꢓ
ꢉꢔ
ꢅꢑꢇꢒ
a. Same as 1.a. Register VCC is not updated while
performing fast AUX conversions.
ꢕꢖAꢁꢔ
3. Connect NTC to one of V1-V12
ꢃꢄꢅꢄ ꢆꢃꢇ
a. See section Temperature Measurement
Figure 27. Input Filtering
4. Connect pin VREF via an external 4k resistor to one of
V1-V12
CURRENT AND VOLTAGE INPUT FILTERING
To ensure the full electrical performance of the ADCs for
current, power and voltage, apply the input filtering cir-
cuitry as depicted in Figure 27 to pins CFP, CFM, VBATP
andVBATM.Thesecomponentsensureanoptimizedinput
filteringfornoisereduction.Equaltimeconstantsatcurrent
andvoltageinputsminimizeerrorsinpowermeasurement
of transient signals due to different delays in both paths.
a. Allows to diagnose the internal current sources.
5. Connect pins BYP1 and BYP2 to separate unused pins
of V1-V12.
a. Allows to diagnose internal generated supply volt-
ages additionally to built-in self-tests.
High-Ohmic Resistive Dividers
Unused Input Pins V1-V12
Any high voltage to be measured by LTC2949 must be
divided down and optionally biased via VREF to move it
into the LTC2949’s supply voltage rails 100mV and into
theADC’sdifferentialinputvoltagerange( 4.8V).Typically,
several resistors are connected in series, to minimize the
voltage drop and power dissipation of individual resistors
(Figure 28). LTC2949 Operation State Diagram shows a
typical schematic, as an example the signal is measured
via V1 by the AUX ADC. The measurement is performed
similarly via the BAT inputs (VBATP, VBATM) by the AUX
or the Power ADCs (P1/P2 as voltage).
InputpinsV1-V12canbeleftfloatingorconnectedtoGND
if not used. Still, itis notrecommendedtoconnectV8-V12
to GND as those pins can be driven high optionally when
used as digital outputs.
Depending on the application and the configuration of
LTC2949, it might be beneficial or sometimes even nec-
essary to apply following external connections to inputs,
which were unused.
1. Connect pin VREF to one of V1-V12
a. SeelastparagraphofsectionFastModeConfigura-
tionsforeffectoffastAUXmeasurementoninternal
slow AUX Round-Robin. Register VREF is not
updated while performing fast AUX conversions.
AllvoltageinputsofLTC2949arebuffered. Thedifferential
input impedance during measurements is guaranteed to
be>50Meg. Fortypicalresistivedividers, likethoseshown
in the typical application, the effect of this equivalent
>50Meg resistor is negligible. The error can be calculated
as following:
b. The internal VREF measurement is altered while the
current sources (see register FCURGPIOCTRL) are
enabled. Connecting VREF to one Vx pin always mea-
surement of the correct VREF pin voltage at any time.
Nominal gain factor: g = R /(R +R
)
low
low high
Rev A
74
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
Effective low side resistor: R
= R • R /
optional external EEPROM can be used as a nonvolatile
storage for those calibration factors.
lowd
low
d
(R +R ), R = 50e5
low
d
d
Effective gain factor: g = R
/(R +R )
lowd high
LTC2949’s gain correction registers are not limited to
values around 1.0, for example it is possible to write a
value of 10.0 and this factor would be applied just as any
other. Still, there is a limitation by the size of the result
registerswhichis16-bit(includingsign)foranyAUX-ADC
measurementlikeBAT,SLOT1/2andthefastAUXmeasure-
ment results. This leads to an absolute maximum register
d
lowd
3
Example values: R =30 , R =5 • 1.3e6
low
high
Gain factor error: err = g /g–1 = –0.06%
d
The differential input impedance during measurements
is nonlinear, it increases with decreasing input signal and
it may also change over temperature. For this reason, it
can be calibrated only partially during a post-production
board test procedure. Still, its guaranteed to be >50Meg
over the full-scale input range and the whole operating
temperature range and thus the error calculated above
gives the worst-case error.
15
valueofroughly12.3V(375µV•[2 -1]).Toavoidclipping
or overflowing of the results, it is always recommended
to use gain correction factors that correct the deviation
fromnominalfactors(e.g.deviationfromnominalresistor
dividerratio)likeintheexampleabove. Thehostcontroller
software then holds the hard-coded nominal factor (e.g.
6.53Meg/30k)andLTC2949isapplyingthefine-trimbased
on the values that were stored into the external EEPROM
during board calibration.
The high voltage is calculated from the ADC measurement
the following way:
Divider connected to GND: V
= V /g
ADC
HVa
Divider connected to VREF: V
= V /g+VREF
ADC
HVb
POWERING THE LTC2949
As any resistive divider is affected by static tolerances, it
can be calibrated by applying a known input signal and
calculating greal from the ADC measurement. LTC2949
can compensate for this error by writing the gain correc-
tion factor GC=gnominal/greal to one of the related gain
correction registers (BATGC, MUX1GC-MUX4GC). The
The LTC2949 requires a single supply voltage of 4.5 to
14V. The maximum supply current is 20mA when active
and 120µA when sleeping. If isoSPI is selected up to 7mA
are required additionally during communication. Any cur-
rent that is necessary to drive circuits connected to the
GPIO pins must also be considered when selecting and
designing a suitable power supply.
R
ꢁꢂꢁ
ꢂ ꢃꢄꢅꢆꢇ
ꢃ ꢄꢅꢆꢇꢈ
ꢀꢁꢀ
ꢀ
R
ꢀꢁꢂꢃ
ꢀ
ꢀꢁꢁꢂ
ꢀꢁꢁꢂ
ꢀꢁꢁꢂ
ꢀꢁꢁꢂ
ꢀꢁꢁꢂ
R
ꢀ
R
ꢀꢁꢂꢀ
ꢃ ꢄꢅRꢀ
ꢀꢁꢂꢃ
R
ꢀ
R
ꢀꢁꢂꢃ
ꢀꢁ
ꢀꢁꢂꢀ
Aꢀꢁꢂꢃꢃ
ꢀꢁꢂꢃꢄꢅꢄ
R ꢁ ꢂꢃꢄ
ꢀ
ꢀꢁ ꢂꢃꢃꢄ
ꢀꢁ ꢂꢃꢄꢅ
R
ꢀ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂ
R
ꢀꢁꢂ
ꢀꢁꢂꢃꢄꢅꢄ
ꢁ ꢂꢃꢄ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
R
ꢀ
ꢀꢁꢂꢃ
ꢀꢁꢂ
ꢀꢁꢁꢁꢂ
ꢀRꢁꢂ
R
ꢀ
ꢀꢁ
ꢀꢁ
R
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂ
ꢀꢁꢂꢃ
Aꢀꢁꢂꢃꢁ
Aꢀꢁꢂꢃꢁ
ꢀꢁꢂꢁ ꢃꢀꢄ
Figure 28. Left: High-Ohmic Resistive Divider for Measuring 1000V Battery Voltage Via the AUX ADC Input V1.
Right: Resistive Divider Connected to VREF Allows to Measure 1000V.
Rev A
75
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
Non-Isolated Supply
PART
POWER
SWITCH
NUMBER
V
RANGE
MAX. P
2W
PACKAGE
SOT23-5
SOT23-5
SOT23-5
SO-8E
IN
OUT
LTC provides a broad spectrum of non-isolated power
supply solutions including LDOs, switched mode power
suppliesandμModuleregulators.AstheLTC2949ismainly
targetinghighvoltagebatteryapplicationstheLT8315can
be considered. Using the LT8315 it is possible to supply
the LTC2949 directly out of a high voltage battery of up
to 560V as shown in Figure 29.
LT8300
LT8303
6V to 100V 0.26A/150V
5.5V to 100V 0.45A/150V
5W
LT8301
2.7V to 42V
2.8V to 42V
3V to 100V
1.2A/65V
3.6A/65V
2A/150V
6W
LT8302
18W
24W
LT8304/-1
SO-8E
Minimum load requirement of the flyback converters
must be considered which is typically much higher than
the sleep current of the LTC2949. To prevent the output
voltage from rising above LTC2949’s operating ratings
a 12V Zener diode (e.g. NXP: BZX384-B12,115) should
be placed.
Isolated supply
MosthighvoltagebatteryapplicationswheretheLTC2949
cannot be supplied directly out of the battery require an
isolatedpowersupply.AsimpleDC/DCconverterisshown
in Figure 30 using Linear Technology’s LT3999 DC/DC
converter and a high isolation-rated transformer.
Transformer specification and design is possibly the most
critical part of successfully applying the above mentioned
DC/DC converters. Data sheets of the suggested parts
give detailed information about critical parameters like
saturation current, inductance, isolation voltage rating
and creepage distance.
The LT830X family of flyback converters with a suitable
transformer is also a possible choice.
ꢀ
ꢀꢁ
ꢀꢁꢂ ꢃꢄ ꢅꢆꢁꢂ
ꢀꢁꢂ ꢃꢄꢅꢆRAꢇ ꢃꢈꢀꢉꢊꢋꢌꢍ
ꢀꢁꢂ ꢀꢃꢄ ꢅꢆꢇꢈRAꢉ ꢅꢊꢊRꢋꢌꢍꢎꢏ
ꢀꢁꢂ ꢃꢄꢅRꢆ ꢅꢆꢃꢃꢇRꢅꢄAꢈ ꢉꢃꢊꢋꢌꢍꢌꢎꢊꢏꢐꢑ
ꢀꢁꢂꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁ
ꢀꢁAꢂ
ꢀꢁꢂ
ꢀꢁ
ꢀꢀꢁ
ꢀꢁꢂꢃꢄꢅꢆ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢀ
ꢀꢁ
ꢀꢁꢂꢃꢄꢅ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄ
ꢀRAꢁꢂ
ꢀꢁꢂꢃꢄ
ꢀꢁꢂꢃ
ꢀ
ꢀ
ꢀꢁꢂRꢃꢄ
ꢀꢀꢁꢂꢃ
ꢀꢀꢁꢂꢃ
ꢀRꢁꢂꢃꢄꢄ
ꢀꢁꢂ
330mΩ
ꢀꢁꢁꢂꢃ
ꢀꢁꢂ
ꢀ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁA ꢂꢃ ꢄꢅꢆꢁA
ꢀꢀꢁꢂ
ꢀꢁ
ꢀꢁ
ꢀꢁꢂꢁ ꢃꢀꢁ
Figure 29. Nonisolated Supply Generation
Rev A
76
For more information www.analog.com
LTC2949
APPLICATION INFORMATION
ꢠ
ꢀ
ꢁꢂ
ꢞꢀ
ꢜ
ꢖꢛꢁ
ꢇꢅ
ꢀꢅ
ꢅꢆꢅ
ꢘꢜ
ꢈ
ꢢꢥꢑꢒ
ꢏꢐꢀ
ꢜ
ꢁꢂ
ꢗꢎ
ꢔꢝꢞꢜ ꢁꢖ ꢅꢟꢜ
ꢅꢡA ꢁꢖ ꢔꢄꢄꢡA ꢢꢜ ꢣ ꢃꢝꢃꢜꢤ
ꢀ
ꢗꢎ
ꢊꢏ
ꢁꢂ
ꢀ
ꢅꢃꢋ
ꢋꢏ
ꢅꢡA ꢁꢖ ꢃꢄꢄꢡA ꢢꢜ ꢣ ꢘꢜꢤ
ꢗꢎ
ꢅꢄꢌꢈ
ꢨ
ꢞꢀ
ꢃꢀꢄꢅ
ꢆꢎA
ꢅꢡA ꢁꢖ ꢥꢘꢄꢡA ꢢꢜ ꢣ ꢅꢔꢜꢤ
ꢗꢎ
ꢟ
ꢟ
ꢟ
ꢟ
ꢅꢄꢌꢍ
ꢘꢅꢝꢅꢧ
ꢅꢄꢌꢍ
ꢈ
ꢏꢐꢑꢒ
ꢏꢐꢀ
ꢜ
ꢐꢘꢝA
ꢅꢃꢋ
ꢗꢎ
ꢒꢎꢚꢛꢜꢀꢖ
ꢀꢁꢂꢃꢄꢅ
ꢆꢇꢂꢈ
ꢨ
ꢥꢞꢌꢈ
ꢅꢀꢄꢅꢉꢊꢈ
ꢊꢋ
ꢦ
ꢄꢋꢌꢍꢍꢍ
ꢜ
ꢖꢛꢁ
Rꢊꢈ
ꢊꢔ
R
ꢇꢅꢆ ꢎꢏꢐ ꢐꢑꢒꢓꢃꢄꢔꢄꢒꢍ
ꢀꢅꢆ ꢕꢖꢗꢀꢕRAꢈꢁ ꢀꢐꢇꢘꢄꢃꢄꢙꢅꢄꢃꢑꢒꢉ
ꢈꢉ
Rꢋ
ꢆꢎꢠ
ꢓꢎꢇ
ꢁꢄꢁꢗꢉꢆꢆ
RꢠꢁAꢆ
ꢔꢩꢥꢩ ꢈꢃꢅ
ꢊꢏꢓ ꢊꢔꢕ ꢈꢖꢂꢋRAꢄ ꢆꢖꢗꢁꢘ ꢈꢗꢆꢙꢏꢚꢔꢐꢗ
ꢋꢏꢕ ꢈꢅꢁꢄꢈRAꢒꢋ ꢛAꢜꢌꢝꢌ
R
ꢏꢔꢘꢏꢣ
ꢏꢗꢙꢤ
ꢋ
ꢡꢂꢊ
R
ꢠꢁAꢆ
ꢢꢍꢘꢍꢣ
ꢔꢍꢢꢍ ꢒꢌꢐ
Figure 31. Isolated Supply Generation with Flyback Converter
Figure 30. Isolated Supply Generation with Push-Pull
Transformer
R
ꢀꢁꢂꢀꢁꢃ
ꢀꢁꢂ ꢃꢄRRꢅꢆꢇ ꢈAꢇꢉ ꢊAꢃꢋ ꢉꢅAꢇꢅRꢋ ꢌꢃꢍꢌꢃ ꢎꢏꢐ ꢅꢇꢃꢑꢒ
ꢀꢁꢂ
R
ꢀꢁꢂꢀꢁꢃ
ꢀꢁꢂꢀ ꢃꢄRRꢅꢆꢇ ꢈAꢇꢀ ꢉꢊꢋꢇꢋR ꢅꢇꢃꢌꢍ
ꢀꢁꢁꢂ
R
ꢀRꢁꢂꢃ
ꢀꢁꢂ
ꢀꢁꢂ
Aꢀꢁꢂꢂ
Aꢀꢁꢁ
ꢀꢁꢂ
ꢀꢁ
ꢀꢁꢂꢂ
ꢀꢁꢂꢃꢄꢅ
ꢀꢁꢂ
ꢀꢁ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄꢅꢆꢇ
ꢀꢁAꢂꢃ
ꢀꢁꢂ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁAꢂꢃ
ꢀꢁ
ꢀꢁꢂꢃꢄꢅꢆꢇ
ꢀꢁꢂꢃꢄꢅꢄ
ꢀꢁ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
Aꢀꢁꢂꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁ
ꢀꢁꢂꢃꢄꢅꢆꢇꢈ
ꢀꢁꢂꢀꢃ
ꢀꢁ
ꢀAꢁꢁꢂRꢃ
ꢀꢁꢁꢂꢀꢁꢃꢄꢅꢆꢇꢈꢉꢂꢊ
ꢀꢁꢁ
ꢀꢁꢂꢃꢄꢅ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀRꢁꢂ
ꢀꢁꢁ
ꢀꢁꢁ
ꢀꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁꢂꢃꢄꢅꢆ
ꢀꢁ
ꢀꢁꢂꢃꢄꢅ
ꢀꢁ
ꢀꢁꢀAꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂꢃꢄꢅꢄAꢀ
ꢀꢁ
ꢀꢁ
ꢀꢁ
ꢀꢁꢂꢃꢄꢅꢆꢇ
ꢀꢁꢂꢃꢃ
ꢀꢁꢂꢃ
ꢀꢁꢂ
ꢀꢁꢂ
ꢀꢁ
ꢀꢁ
ꢀꢁꢂ
ꢀꢁꢂꢃ
ꢀꢁꢂꢁ ꢃꢄꢀ
Figure 32. Battery Monitoring with High Side Current Sensing Over Two Sense Resistors, Battery Voltage Measurement, Temperature
Measurement and Pre-Charge Voltage Measurement Using Isolated Supply and Isolated NMOS Gate Drive as Well as PhotoMos Relay.
Rev A
77
For more information www.analog.com
LTC2949
PACKAGE DESCRIPTION
LXE Package
48-Lead Plastic Exposed Pad LQFP (7mm × 7mm)
ꢅReꢪeꢫeꢬꢭe ꢀꢞꢙ ꢠꢥꢚ ꢮꢊꢓꢨꢊꢄꢨꢌꢖꢗꢑ Rev ꢎꢆ
Exposed Pad Variation AA
ꢖꢒꢊꢊ ꢎꢘꢙ
ꢑꢒꢊꢊ ꢎꢘꢙ
ꢃꢒꢌꢓ ꢊꢒꢗꢊ
ꢃꢄ
ꢋꢑ
ꢋꢑ
ꢃꢄ
ꢘꢂꢂ ꢜꢝꢞꢂꢟ ꢋ
ꢌ
ꢋꢔ
ꢋꢔ
ꢌ
ꢖꢒꢊꢊ ꢎꢘꢙ
ꢑꢒꢊꢊ ꢎꢘꢙ
ꢃꢒꢌꢓ ꢊꢒꢗꢊ
A
A
ꢗꢓ
ꢌꢗ
ꢌꢗ
ꢗꢓ
ꢙꢊꢒꢋꢊ ꢐ ꢊꢒꢓꢊ
ꢗꢃ
ꢌꢋ
ꢌꢋ
ꢗꢃ
ꢎꢝꢞꢞꢝꢕ ꢝꢈ ꢉAꢙꢢAꢚꢂꢩꢂꢁꢉꢝꢘꢂꢠ ꢉAꢠ ꢅꢘꢣAꢠꢂꢠ ARꢂAꢆ
ꢌꢒꢔꢊ
ꢌꢌꢏ ꢐ ꢌꢋꢏ
ꢌꢒꢋꢓ ꢐ ꢌꢒꢃꢓ ꢕAꢁ
Rꢊꢒꢊꢄ ꢐ ꢊꢒꢗꢊ
ꢚAꢛꢚꢂ ꢉꢀAꢜꢂ
ꢊꢒꢗꢓ
ꢊꢏ ꢐ ꢑꢏ
ꢀꢁꢂꢃꢄ ꢅAAꢆ ꢀꢇꢈꢉ ꢊꢋꢌꢄ Rꢂꢍ ꢎ
ꢌꢌꢏ ꢐ ꢌꢋꢏ
ꢌꢒꢊꢊ Rꢂꢈ
ꢊꢒꢓꢊ
ꢎꢘꢙ
ꢊꢒꢊꢖ ꢐ ꢊꢒꢗꢊ
ꢊꢒꢌꢑ ꢐ ꢊꢒꢗꢑ
ꢘꢡꢠꢂ ꢍꢡꢂꢥ
ꢊꢒꢊꢓ ꢐ ꢊꢒꢌꢓ
ꢊꢒꢃꢓ ꢐ ꢊꢒꢑꢓ
ꢘꢂꢙꢞꢡꢝꢜ A ꢐ A
ꢜꢝꢞꢂꢟ
ꢌꢒ ꢠꢡꢕꢂꢜꢘꢡꢝꢜꢘ ARꢂ ꢡꢜ ꢕꢡꢀꢀꢡꢕꢂꢞꢂRꢘ
ꢗꢒ ꢠꢡꢕꢂꢜꢘꢡꢝꢜꢘ ꢝꢈ ꢉAꢙꢢAꢚꢂ ꢠꢝ ꢜꢝꢞ ꢡꢜꢙꢀꢛꢠꢂ ꢕꢝꢀꢠ ꢈꢀAꢘꢣꢒ ꢕꢝꢀꢠ ꢈꢀAꢘꢣ
ꢘꢣAꢀꢀ ꢜꢝꢞ ꢂꢁꢙꢂꢂꢠ ꢊꢒꢗꢓꢤꢤ ꢅꢌꢊ ꢕꢡꢀꢘꢆ ꢎꢂꢞꢥꢂꢂꢜ ꢞꢣꢂ ꢀꢂAꢠꢘ Aꢜꢠ ꢝꢜ Aꢜꢦ
ꢘꢡꢠꢂ ꢝꢈ ꢂꢁꢉꢝꢘꢂꢠ ꢉAꢠꢧ ꢕAꢁ ꢊꢒꢓꢊꢤꢤ ꢅꢗꢊ ꢕꢡꢀꢘꢆ Aꢞ ꢙꢝRꢜꢂR ꢝꢈ ꢂꢁꢉꢝꢘꢂꢠ
ꢉAꢠꢧ ꢡꢈ ꢉRꢂꢘꢂꢜꢞ
ꢋꢒ ꢉꢡꢜꢨꢌ ꢡꢜꢠꢂꢜꢞꢡꢈꢡꢂR ꢡꢘ A ꢕꢝꢀꢠꢂꢠ ꢡꢜꢠꢂꢜꢞAꢞꢡꢝꢜ
ꢃꢒ ꢠRAꢥꢡꢜꢚ ꢡꢘ ꢜꢝꢞ ꢞꢝ ꢘꢙAꢀꢂ
ꢑꢒꢌꢓ ꢐ ꢑꢒꢗꢓ
ꢓꢒꢓꢊ Rꢂꢈ
ꢃꢄ
ꢋꢑ
ꢋꢔ
ꢌ
ꢊꢒꢓꢊ ꢎꢘꢙ
ꢓꢒꢓꢊ Rꢂꢈ
ꢑꢒꢌꢓ ꢐ ꢑꢒꢗꢓ
ꢊꢒꢗꢊ ꢐ ꢊꢒꢋꢊ
ꢃꢒꢌꢓ ꢊꢒꢊꢓ
ꢃꢒꢌꢓ ꢊꢒꢊꢓ
ꢌꢗ
ꢌꢋ
ꢗꢓ
ꢉAꢙꢢAꢚꢂ ꢝꢛꢞꢀꢡꢜꢂ
ꢗꢃ
ꢙꢝꢕꢉꢝꢜꢂꢜꢞ
ꢉꢡꢜ ꢯAꢌꢰ
ꢌꢒꢋꢊ ꢕꢡꢜ
ꢞRAꢦ ꢉꢡꢜ ꢌ
ꢎꢂꢍꢂꢀ
Rꢂꢙꢝꢕꢕꢂꢜꢠꢂꢠ ꢘꢝꢀꢠꢂR ꢉAꢠ ꢀAꢦꢝꢛꢞ
Aꢉꢉꢀꢦ ꢘꢝꢀꢠꢂR ꢕAꢘꢢ ꢞꢝ ARꢂAꢘ ꢞꢣAꢞ ARꢂ ꢜꢝꢞ ꢘꢝꢀꢠꢂRꢂꢠ
ꢉAꢙꢢAꢚꢂ ꢡꢜ ꢞRAꢦ ꢀꢝAꢠꢡꢜꢚ ꢝRꢡꢂꢜꢞAꢞꢡꢝꢜ
Rev A
78
For more information www.analog.com
LTC2949
REVISION HISTORY
REV
DATE
11/20 Changed GPOx to GPIOx.
Moved Temperature Measurement into sub-section Application Information.
Adapted Abs Max of SDO.
Changed TYP value of V
DESCRIPTION
PAGE NUMBER
A
Throughout
2
3
in EC-table.
4
BYP2
Changed Input Leakage current for CFPx from 40nA to 60nA.
4
Adapted I Average Supply Current in EC table and in text Modes of Operation.
4,15
5
CC
Changed Input Leakage current for Voltage Measurement by Power ADC from 10nA to 60nA.
Symbol names corrections in EC table.
6
Changed Input Leakage current for Voltage Measurement by AUXILIARY ADC from 10nA to 60nA.
EC-Table External Clock Frequency adapted MIN value.
6
7
Adapted t
MAX value.
8
lDLE
Adapted V Transmitter Pulse Amplitude MAX value.
8
A
Corrected symbol A
Corrected symbol V
in V
.
8
TCMP
ICMP
in A
wrong.
8
ICM
TCMP
Corrected Pin 6 missing in list of DNC-pins.
Core states vs fast conversions: Update of Figure 1.
Corrected wrong Figure 29.
12
15
16
20
21
41
41
48
65, 66
66
74
77, 80
Corrected wrong minimum frequency of 200kHz.
Corrected missing unit at first column of Table 3.
ISORPT still in register map table: Removed.
Register name typos: Corrected Register Map.
ADC limit behavior: Added sentence after Table 28.
Register name typos: Corrected Tables 71 and 72.
Adapted Description of MUX[1-4]GC in Table 73.
Corrected Wrong Figure 1.
MOSFET polarities: Update of several transistor symbols.
Rev A
Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog
Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications
79
subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices.
LTC2949
TYPICAL APPLICATION
Battery Metering with Low-Side Current & Charge Sensing, Battery Voltage Measurement,
Temperature Measurement, Contactor Supervision and Isolated Gate Drive of Pre-Charge FET.
ꢀ
ꢂ
ꢃAꢄ
ꢛꢞAꢖ ꢞR ꢜꢠARꢑꢗR
RꢒRꢗꢜꢠ ꢊꢄꢃꢋꢌꢍꢎꢈꢏ
ꢊꢄꢃꢋꢌꢍꢎꢈꢏ
ꢅꢆ
ꢇꢧ
ꢅꢈꢉ
ꢇꢈꢧ
ꢁ
ꢅꢆ
ꢊꢄꢃꢋꢌꢍꢎꢈꢏ
ꢂꢎꢐꢑꢒꢈꢇ
ꢂꢃAꢄꢒ
ꢂꢋꢐꢑꢒꢈꢅ
ꢂ
ꢃAꢄ
ꢂꢇꢈꢐꢑꢒꢈꢓ
ꢂꢅ
ꢇꢈꢉ
ꢊꢄꢃꢋꢌꢍꢎꢈꢏ
ꢂꢃAꢄꢆ
ꢝꢞꢂꢜꢜ
ꢊꢜꢍꢐꢝꢒ
ꢛꢄꢜꢅꢋꢕꢋ
ꢂꢕꢙꢂꢚ
ꢂꢇꢅꢐꢑꢒꢈꢔ
ꢊꢒꢝꢐꢡꢢꢣꢊꢒꢝ
ꢝꢌꢄꢗRꢘAꢜꢗ
ꢜꢊꢃꢐꢝꢆ
ꢇꢈꢉ
ꢊꢖꢞꢐꢝꢃꢝAꢊ
ꢊꢖꢝꢐꢂꢜꢆꢒ
Aꢑꢌꢖ
ꢃAꢄꢄꢗRꢟ
ꢁ
ꢂ
ꢃAꢄ
ꢖꢂꢜꢜ
Aꢂꢜꢜ
ꢌꢄꢜ
ꢌꢄꢜꢛꢗꢅꢈꢓꢗ
ꢀ
ꢇꢧ
ꢇꢧ
ꢂ
ꢤ
ꢜꢜ
ꢁ
Aꢑꢌꢖ
ꢖꢑꢌꢖ
ꢃꢟꢒꢇ
ꢂꢇ
ꢇꢈꢉ
ꢇꢧ
ꢇꢧ
ꢂRꢗꢘ
ꢇꢈꢉ
ꢃꢟꢒꢅ
ꢜꢛꢍꢞ
ꢓꢓꢨ
ꢓꢓꢨ
ꢂꢓ
ꢂꢇꢇꢐꢑꢒꢈꢕ
ꢕꢆꢠꢩ
ꢊꢄꢃꢋꢌꢍꢎꢈꢏ
ꢅꢆ
ꢜꢛꢍꢝ
ꢝꢅꢆ ꢜꢘꢅꢆ ꢜꢘꢅꢒ ꢝꢅꢒ
ꢝꢇꢆ ꢜꢘꢇꢆ ꢜꢘꢇꢒ ꢝꢇꢒ
ꢇꢧ
ꢇꢧ
ꢁ
R
ꢊꢗꢌꢊꢗ
ꢂ
ꢃAꢄ
ꢅꢋꢕꢋ ꢄAꢈꢅ
ꢇꢈꢈꢧ
ꢤꢂ ꢊꢥꢒꢒꢛꢟ ꢂꢞꢛꢄAꢑꢗ ꢆꢥꢊꢄ ꢃꢗ ꢜꢠꢞꢊꢗꢌ ꢄꢞ ꢗꢌꢊꢥRꢗ ꢆꢝꢌꢝꢆꢥꢆ ꢂ ꢞꢘ ꢥꢊꢗꢖ ꢆꢞꢊꢘꢗꢄꢊ ꢥꢌꢖꢗR Aꢛꢛ ꢜꢞꢌꢖꢝꢄꢝꢞꢌꢊꢦ
ꢜꢜ ꢑꢊ
RELATED PARTS
PART NUMBER
DESCRIPTION
COMMENTS
LTC6810-1/
LTC6810-2
4th Generation 6-Cell Battery Stack
Monitor and Balancing IC
Measures Cell Voltages for Up to 6 Series Battery Cells. The isoSPI Daisy-Chain
Capability Allows Multiple Devices to Be Interconnected for Measuring Many Battery
Cells Simultaneously. The isoSPI Bus Can Operate Up to 1MHz and Can Be Operated
Bidirectionally for Fault Conditions, Such as a Broken Wire or Connector. Includes
Internal Passive Cell Balancing Capability of Up to 150mA.
LTC6811-1/
LTC6811-2
4th Generation 12-Cell Battery Stack
Monitor and Balancing IC
Measures Cell Voltages for Up to 12 Series Battery Cells. Daisy-Chain Capability Allows
Multiple Devices to Be Connected to Measure Many Battery Cells Simultaneously Via the
Built-In 1MHz, 2-Wire Isolated Communication (isoSPI). Includes Capability for Passive
Cell Balancing.
LTC6820
isoSPI Isolated Communications Interface Provides an Isolated Interface for SPI Communication Up to 100 Meters, Using a Twisted
Pair. Companion to the LTC6804, LTC6806, LTC6811, LTC6812 and LTC6813
LTC6812-1
4th Generation 15-Cell Battery Stack
Monitor and Balancing IC
Measures Cell Voltages for Up to 15 Series Battery Cells. The isoSPI Daisy-Chain
Capability Allows Multiple Devices to Be Connected to Measuring Many Battery
Cells Simultaneously. The isoSPI Bus Can Operate Up to 1MHz and Can Be Operated
Bidirectionally for Fault Conditions, Such as a Broken Wire or Connector. Includes
Internal Passive Cell Balancing Capability of Up to 200mA.
LTC6813-1
4th Generation 18-Cell Battery Stack
Monitor and Balancing IC
Measures Cell Voltages for Up to 18 Series Battery Cells. The isoSPI Daisy-Chain
Capability Allows Multiple Devices to Be Interconnected for Measuring Many Battery
Cells Simultaneously. The isoSPI Bus Can Operate Up to 1MHz and Can Be Operated
Bidirectionally for Fault Conditions, Such as a Broken Wire or Connector. Includes
Internal Passive Cell Balancing Capability of Up to 200mA.
Rev A
11/20
www.analog.com
ANALOG DEVICES, INC. 2019-2020
80
相关型号:
LTC6820HMS#PBF
LTC6820 - isoSPI Isolated Communications Interface; Package: MSOP; Pins: 16; Temperature Range: -40°C to 125°C
Linear
©2020 ICPDF网 联系我们和版权申明