LTC3882-1
Dual Output PolyPhase
Step-Down DC/DC Voltage Mode Controller
with Digital Power System Management
Description
Features
PMBus/I2C Compliant Serial Interface
– Monitor Voltage, Current, Temperature and Faults
– Program Voltage, Soft-Start/Stop, Sequencing, Margining, AVP and UV/OV/OC Limits
n 3V ≤ VINSNS ≤ 38V, 0.5V ≤ V
OUT ≤ 5.25V
n ±0.5% Output Voltage Error
n Programmable PWM Frequency or External Clock
Synchronization from 250kHz to 1.25MHz
n Accurate PolyPhase® Current Sharing
n Internal EEPROM with Fault Logging
n IC Supply Range: 3V to 13.2V
n Resistor or Inductor DCR Current Sensing
n Power Good Output Voltage Monitor
n Optional Resistor Programming for Key Parameters
n 40-Pin (6mm × 6mm) QFN Package
n
Applications
High Current Distributed Power Systems
Servers, Network and Storage Equipment
n Intelligent Energy Efficient Power Regulation
n
The LTC®3882-1 is a dual, PolyPhase DC/DC synchronous
step-down switching regulator controller with PMBus
compliant serial interface. It uses a constant frequency,
leading-edge modulation, voltage mode architecture for
excellent transient response and output regulation. Each
PWM channel can produce output voltages from 0.5V to
5.25V using a wide range of 3.3V compatible power stages,
including power blocks, DrMOS or discrete FET drivers.
Up to four LTC3882-1 devices can operate in parallel for
2-, 3-, 4-, 6- or 8-phase operation.
System configuration and monitoring is supported by the
LTpowerPlay™ software tool. The LTC3882-1 serial interface
can read back input voltage, output voltage and current,
temperature and fault status. Most operating parameters can
be set via the digital interface or stored in internal EEPROM
for use at power up. Switching frequency and phase, output
voltage and device address can also be set using external
configuration resistors.
n
L, LT, LTC, LTM, Linear Technology,the Linear logo and PolyPhase are registered trademarks
and LTpowerPlay is a trademark of Linear Technology Corporation. All other trademarks are the
property of their respective owners. Protected by U.S. Patents, including 5396245, 5859606,
6144194, 6937178, 7420359 and 7000125.
Typical Application
PWM ENABLE
TG/BG
HW WRITE
OUTPUT
CONTROL PROTECT
LTC3882
+
VCC VINSNS V
SENSE0
FB0
SDA
LTC3882-1
SCL
ALERT
•
VOUT0 & VOUT1
VOUT0 Only
•
VIN
VOUT
1V
70A
SW
Load Step Transient Current Sharing
(Using FDMF5820DC DrMOS)
GND
PWM0
TSNS0
PGOOD1
FAULT1
SYNC
SHARE_CLK
IAVG0
IAVG_GND
PWM
DIFFERENTIAL
VOUT SENSE
COMP1
RUN0
RUN1
IAVG1
FDMF5820DC
COMP0
PGOOD0
FAULT0
TO/FROM
EXTERNAL DEVICES
•
LTC3882-1
VIN
7V TO 13.2V
TO/FROM
MCU
•
DEDICATED
PGOOD OUTPUT
ISENSE0+
ISENSE0–
VSENSE1+
ISENSE1–
ISENSE1+
PWM1
IOUT
10A/DIV
FDMF5820DC
PWM
TSNS1
VSENSE0–
–
GND VSENSE1
VIN
SW
IL0, IL1
10A/DIV
GND
50µs/DIV
INDUCTORS: COOPER FP1007R1-R22
SOME DETAILS OMITTED FOR CLARITY
38821 TA01b
38821 TA01a
38821f
For more information www.linear.com/LTC3882-1
1
LTC3882-1
Table of Contents
Features...................................................... 1
Applications................................................. 1
Typical Application ......................................... 1
Description.................................................. 1
Absolute Maximum Ratings............................... 4
Order Information........................................... 4
Pin Configuration........................................... 4
Electrical Characteristics.................................. 5
Typical Performance Characteristics.................... 9
Pin Functions............................................... 13
Block Diagram.............................................. 15
Test Circuit.................................................. 16
Timing Diagram............................................ 16
Operation................................................... 16
Overview............................................................. 16
Main Control Loop............................................... 17
Power-Up and Initialization.................................. 19
Soft-Start............................................................20
Time-Based Output Sequencing..........................20
Output Ramping Control......................................20
Voltage-Based Output Sequencing......................20
Minimum Output Disable Times.......................... 21
Output Short Cycle.............................................. 21
Light Load Current Operation.............................. 21
Switching Frequency and Phase.......................... 21
PolyPhase Load Sharing......................................22
Active Voltage Positioning...................................22
Input Supply Monitoring......................................22
Output Voltage Sensing and Monitoring..............22
Output Current Sensing and Monitoring..............22
External and Internal Temperature Sense............23
Resistor Configuration Pins.................................23
Internal EEPROM with CRC................................. 24
Fault Detection..................................................... 24
Input Supply Faults.............................................. 24
Hardwired PWM Response to VOUT Faults.......... 24
Power Good Indication (Master)..........................25
Power Good Indication (Slave)............................25
Hardwired PWM Response to IOUT Faults................ 25
Hardwired PWM Response to Temperature Faults... 25
Hardwired PWM Response to Timing Faults.......26
External Faults.....................................................26
Fault Handling......................................................26
Status Registers and ALERT Masking..................26
FAULT Pin I/O....................................................... 28
Fault Logging....................................................... 28
Factory Default Operation.................................... 31
Serial Interface.................................................... 32
Serial Bus Addressing......................................... 32
Serial Bus Timeout..............................................36
2
Serial Communication Errors..............................36
PMBus Command Summary............................. 37
PMBus Commands.............................................. 37
Data Formats....................................................... 37
Applications Information................................. 42
Efficiency Considerations.................................... 42
PWM Frequency and Inductor Selection.............. 42
Power MOSFET Selection....................................43
MOSFET Driver Selection....................................44
Using PWM Protocols.........................................44
CIN Selection........................................................44
COUT Selection.....................................................45
Feedback Loop Compensation.............................46
PCB Layout Considerations................................. 47
Output Current Sensing.......................................48
Output Voltage Sensing.......................................50
Soft-Start and Stop............................................. 51
Time-Based Output Sequencing and Ramping.... 51
Voltage-Based Output Sequencing...................... 52
Using Output Voltage Servo................................54
Using AVP...........................................................54
PWM Frequency Synchronization........................55
PolyPhase Operation and Load Sharing..............56
External Temperature Sense................................60
Resistor Configuration Pins.................................60
Internal Regulator Outputs..................................62
IC Junction Temperature.....................................62
Derating EEPROM Retention at Temperature.......63
Configuring Open-Drain Pins...............................63
PMBus Communication and Command Processing.64
Status and Fault Log Management......................65
LTpowerPlay – An Interactive Digital Power GUI.65
Interfacing to the DC1613....................................66
Design Example...................................................66
PMBus COMMAND DETAILS.............................. 70
Addressing and Write Protect.................................. 70
PAGE................................................................... 70
PAGE_PLUS_WRITE........................................... 70
PAGE_PLUS_READ............................................. 71
WRITE_PROTECT................................................ 71
MFR_ADDRESS..................................................72
MFR_RAIL_ADDRESS........................................72
General Device Configuration.................................. 72
PMBUS_REVISION..............................................72
CAPABILITY.........................................................72
MFR_CONFIG_ALL_LTC3882-1..........................73
On, Off and Margin Control......................................73
ON_OFF_CONFIG.................................................73
OPERATION......................................................... 74
MFR_RESET........................................................ 74
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Table of Contents
PWM Configuration................................................. 75
FREQUENCY_SWITCH......................................... 75
MFR_PWM_CONFIG_LTC3882-1........................ 76
MFR_CHAN_CONFIG_LTC3882-1.......................77
MFR_PWM_MODE_LTC3882-1.......................... 78
Input Voltage and Limits.......................................... 79
VIN_ON............................................................... 79
VIN_OFF.............................................................. 79
VIN_OV_FAULT_LIMIT......................................... 79
VIN_UV_WARN_LIMIT........................................ 79
Output Voltage and Limits.......................................80
VOUT_MODE.......................................................80
VOUT_COMMAND...............................................80
MFR_VOUT_MAX................................................80
VOUT_MAX......................................................... 81
MFR_VOUT_AVP................................................. 81
VOUT_MARGIN_HIGH......................................... 81
VOUT_MARGIN_LOW......................................... 81
VOUT_OV_FAULT_LIMIT..................................... 81
VOUT_OV_WARN_LIMIT.....................................82
VOUT_UV_WARN_LIMIT.....................................82
VOUT_UV_FAULT_LIMIT.....................................82
Output Current and Limits.......................................83
IOUT_CAL_GAIN.................................................83
MFR_IOUT_CAL_GAIN_TC..................................83
IOUT_OC_FAULT_LIMIT......................................83
IOUT_OC_WARN_LIMIT......................................83
Output Timing, Delays, and Ramping......................84
MFR_RESTART_DELAY.......................................84
TON_DELAY........................................................84
TON_RISE...........................................................84
TON_MAX_FAULT_LIMIT....................................85
VOUT_TRANSITION_RATE..................................85
TOFF_DELAY.......................................................85
TOFF_FALL..........................................................85
TOFF_MAX_WARN_LIMIT...................................85
External Temperature and Limits.............................86
MFR_TEMP_1_GAIN............................................86
MFR_TEMP_1_OFFSET........................................86
OT_FAULT_LIMIT.................................................86
OT_WARN_LIMIT................................................86
UT_FAULT_LIMIT................................................. 87
Status Reporting...................................................... 87
STATUS_BYTE..................................................... 87
STATUS_WORD...................................................88
STATUS_VOUT....................................................88
STATUS_IOUT.....................................................89
STATUS_INPUT...................................................89
STATUS_TEMPERATURE.....................................89
STATUS_CML......................................................90
STATUS_MFR_SPECIFIC.....................................90
MFR_PADS_LTC3882-1...................................... 91
MFR_COMMON................................................... 91
CLEAR_FAULTS..................................................92
Telemetry.................................................................92
READ_VIN...........................................................92
MFR_VIN_PEAK..................................................93
READ_VOUT........................................................93
MFR_VOUT_PEAK...............................................93
READ_IOUT.........................................................93
MFR_IOUT_PEAK................................................93
READ_POUT........................................................93
READ_TEMPERATURE_1....................................94
MFR_TEMPERATURE_1_PEAK...........................94
READ_TEMPERATURE_2....................................94
MFR_TEMPERATURE_2_PEAK...........................94
READ_DUTY_CYCLE...........................................94
READ_FREQUENCY.............................................94
MFR_CLEAR_PEAKS..........................................94
Fault Response and Communication........................95
VIN_OV_FAULT_RESPONSE................................95
VOUT_OV_FAULT_RESPONSE.............................96
VOUT_UV_FAULT_RESPONSE.............................96
IOUT_OC_FAULT_RESPONSE.............................. 97
OT_FAULT_RESPONSE........................................98
UT_FAULT_RESPONSE........................................98
MFR_OT_FAULT_RESPONSE..............................98
TON_MAX_FAULT_RESPONSE...........................99
MFR_RETRY_DELAY...........................................99
SMBALERT_MASK............................................ 100
MFR_FAULT_PROPAGATE_LTC3882-1............. 101
MFR_FAULT_RESPONSE................................... 101
EEPROM User Access............................................ 102
MFR_FAULT_LOG.............................................. 102
MFR_FAULT_LOG_CLEAR................................. 102
STORE_USER_ALL........................................... 103
RESTORE_USER_ALL....................................... 103
MFR_COMPARE_USER_ALL............................ 103
MFR_FAULT_LOG_STORE................................ 103
MFR_EE_xxxx................................................... 103
USER_DATA_0x................................................ 103
Unit Identification.................................................. 104
MFR_ID............................................................. 104
MFR_MODEL..................................................... 104
MFR_SERIAL..................................................... 104
Typical Applications..................................... 105
Package Description.................................... 107
Typical Application...................................... 108
Related Parts............................................. 108
38821f
For more information www.linear.com/LTC3882-1
3
LTC3882-1
Absolute Maximum Ratings
Pin Configuration
(Note 1)
ISENSE1+
ISENSE1–
VSENSE1+
VSENSE1–
VSENSE0–
VSENSE0+
ISENSE0–
ISENSE0+
IAVG0
FB0
40 39 38 37 36 35 34 33 32 31
COMP0 1
30 IAVG1
TSNS0 2
29 FB1
TSNS1 3
28 COMP1
VINSNS 4
27 PGOOD1
41
GND
IAVG_GND 5
PGOOD0 6
26 PWM1
25 VCC
PWM0 7
24 VDD33
SYNC 8
23 SHARE_CLK
SCL 9
22 VDD25
SDA 10
21 PHAS_CFG
FREQ_CFG
VOUT1_CFG
ASEL1
ASEL0
RUN1
RUN0
FAULT1
ALERT
11 12 13 14 15 16 17 18 19 20
VOUT0_CFG
*See Derating EEPROM Retention at Temperature in the
Applications Information section for junction temperatures
in excess of 125°C.
TOP VIEW
FAULT0
VCC Supply Voltage..................................... –0.3V to 15V
VINSNS Voltage.......................................... –0.3V to 40V
VSENSEn –....................................................... –0.3V to 1V
VSENSEn+, ISENSEn+, ISENSEn –......................... –0.3V to 6V
FBn, COMPn, TSNSn, IAVG_GND, IAVGn....... –0.3V to 3.6V
SYNC, FAULTn, PGOODn, SHARE_CLK.........–0.3V to 3.6V
SCL, SDA, RUNn, ALERT............................ –0.3V to 5.5V
ASELn, VOUTn_CFG, FREQ_CFG,
PHAS_CFG............................................... –0.3V to 2.75V
PWMn, VDD25................................................... (Note 13)
VDD33................................................................ (Note 14)
Operating Junction Temperature
(Notes 2, 3)........................................... –40°C to 125°C*
Storage Temperature Range................. –65°C to 125°C*
UJ PACKAGE
40-LEAD (6mm × 6mm) PLASTIC QFN
TJMAX = 125°C, θJA = 33°C/W , θJC = 2.5°C/W
EXPOSED PAD (PIN 41) IS GND, MUST BE SOLDERED TO PCB
Order Information
LEAD FREE FINISH
TAPE AND REEL
PART MARKING*
PACKAGE DESCRIPTION
TEMPERATURE RANGE
LTC3882EUJ-1#PBF
LTC3882EUJ-1#TRPBF
LTC3882UJ-1
40-Lead (6mm × 6mm) Plastic QFN
–40°C to 125°C
LTC3882IUJ-1#PBF
LTC3882IUJ-1#TRPBF
LTC3882UJ-1
40-Lead (6mm × 6mm) Plastic QFN
–40°C to 125°C
Consult LTC Marketing for parts specified with wider operating temperature ranges. *The temperature grade is identified by a label on the shipping container.
Consult LTC Marketing for information on nonstandard lead based finish parts.
For more information on lead free part marking, go to: http://www.linear.com/leadfree/
For more information on tape and reel specifications, go to: http://www.linear.com/tapeandreel/
4
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Electrical Characteristics
The l denotes the specifications which apply over the specified operating
junction temperature range, otherwise specifications are at TJ = 25°C (Note 2). VCC = 5V, VSENSE0+ = VSENSE1+ = 1.8V, VSENSE0– =
VSENSE1– = IAVG_GND = GND = 0V, fSYNC = 500kHz (externally driven) unless otherwise specified.
SYMBOL
PARAMETER
CONDITIONS
MIN
TYP
MAX
UNITS
IC Supply
VCC
VCC Voltage Range
VDD33 = Internal LDO
VDD33_EXT
VDD33 Voltage Range
VCC = VDD33 (Note 6)
l
VUVLO
Undervoltage Lockout Threshold
VDD33 Rising
Hysteresis
l
IQ
IC Operating Current
tINIT
Controller Initialization Time
4.5
13.8
V
3
3.6
V
42
Delay from RESTORE_USER_ALL, MFR_RESET or
VDD33 > VUVLO Until TON_DELAY Can Begin
3
V
mV
32
mA
70
ms
VDD33 Linear Regulator
VDD33
VDD33 Regulator Output Voltage
VCC ≥ 4.5V
IDD33
VDD33 Current Limit
VDD33 = 2.8V
VDD33 = 0V
3.2
3.3
3.4
85
40
V
mA
mA
VDD25 Linear Regulator
VDD25
VDD25 Regulator Output Voltage
IDD25
VDD25 Current Limit
2.25
2.5
2.75
95
V
mA
PWM Control Loops
VINSNS
VIN Sense Voltage Range
RVINSNS
VINSNS Input Resistance
VOUT_R0
Range 0 Maximum VOUT
Range 0 Set Point Error (Note 7)
3
0.6V ≤ VOUT ≤ 5V
0.6V ≤ VOUT ≤ 5V
Range 0 Set Point Resolution
VOUT_R1
Range 1 Maximum VOUT
Range 1 Set Point Error (Note 7)
0.6V ≤ VOUT ≤ 2.5V
0.6V ≤ VOUT ≤ 2.5V
Range 1 Set Point Resolution
l
l
–0.5
–0.5
VSENSE+ = 5.5V
VSENSE– = 0V
IVSENSE
VSENSE Input Current
VLINEREG
VCC Line Regulation, No Output Servo
4.5V ≤ VCC ≤ 13.2V (See Test Circuit)
AVP
AVP ∆VOUT
AVP = 10%, VOUT_COMMAND = 1.8V,
ISENSE Differential Step 3mV to 12mV
with IOUT_OC_WARN_LIMIT = 15mV
38
kΩ
5.25
±0.2
V
%
%
mV
1.375
2.65
±0.2
0.6875
0.5
0.5
235
–335
–0.02
l
–118
V
278
–108
V
%
%
mV
µA
µA
0.02
%/V
–96
mV
AV(OL)
Error Amplifier Open-Loop Voltage Gain
87
dB
SR
Error Amplifier Slew Rate
9.5
V/µs
f0dB
Error Amplifier Bandwidth
(Note 12)
30
MHz
ICOMP
Error Amplifier Output Current
Sourcing
Sinking
–2.6
34
mA
mA
RVSFB
Resistance Between VSENSE+ and FB
Range 0
Range 1
VISENSE
ISENSE Differential Input Range
ISENSE± Input Current
0V ≤ VPIN ≤ 5.5V
IAVG_VOS
IAVG Current Sense Offset
Referred to ISENSE Inputs
fSYNC
Slave Current Sharing Offset
SYNC Frequency Error
52
37
67
49
–1
±0.1
83
61
±70
IISENSE
VSIOS
l
l
l
–600
l
–800
l
–10
Referred to ISENSE Inputs
250kHz ≤ fSYNC ≤ 1.25MHz
±175
±300
kΩ
kΩ
mV
1
µA
650
µV
µV
700
µV
µV
10
%
38821f
For more information www.linear.com/LTC3882-1
5
LTC3882-1
Electrical Characteristics
The l denotes the specifications which apply over the specified operating
junction temperature range, otherwise specifications are at TJ = 25°C (Note 2). VCC = 5V, VSENSE0+ = VSENSE1+ = 1.8V, VSENSE0– =
VSENSE1– = IAVG_GND = GND = 0V, fSYNC = 500kHz (externally driven) unless otherwise specified.
SYMBOL
PARAMETER
CONDITIONS
MIN
TYP
MAX
UNITS
Input Voltage Supervisor
VON_TOL
Input ON/OFF Threshold Error
NVON
Input ON/OFF Threshold Resolution
15V ≤ VIN_ON ≤ 35V
l
–2
2
143
%
mV
Output Voltage Supervisors
VUVOV_R0
Range 0 Maximum Threshold
Range 0 Error
Range 0 Threshold Resolution
Range 0 Threshold Hysteresis
VUVOV_R1
Range 1 Maximum Threshold
Range 1 Error
Range 1 Threshold Resolution
Range 1 Threshold Hysteresis
Output Current Supervisors
Output Current Limit Tolerance
VILIM_TOL
ISENSE+ – ISENSE–
2V ≤ VOUT ≤ 5V (Falling for UV and Rising for OV)
1V ≤ VOUT ≤ 2.5V (Falling for UV and Rising for OV)
15mV < ISENSE+ – ISENSE– ≤ 30mV
30mV < ISENSE+ – ISENSE– ≤ 50mV
50mV < ISENSE+ – ISENSE– ≤ 70mV
1LSB
l
l
l
l
l
–1
–1
5.5
11
2.75
5.5
–1.7
–2.5
–5.2
ISENSE+ – ISENSE– Threshold Resolution
NlLIM
ADC Readback Telemetry (Note 8)
VINSNS Readback Resolution
(Note 9)
NVIN
VINSNS Total Unadjusted Readback Error 4.5V ≤ VINSNS ≤ 38V
VIN_TUE
PWM Duty Cycle Resolution
PWM Duty Cycle Total Unadjusted
Readback Error
VOUT Readback Resolution
VOUT Total Unadjusted Readback Error
(Note 9)
PWM Duty Cycle = 12.5%
NISENSE
IOUT Readback Resolution
LSB Step Size (at ISENSE±)
ISENSE_TUE
ISENSE_OS
NTEMP
TEXT_TUE
IOUT Total Unadjusted Readback Error
IOUT Zero-Code Offset Voltage
Temperature Resolution
External Temperature Total Unadjusted
Readback Error
(Note 9)
0mV ≤ |ISENSE+ – ISENSE–| < 16mV
16mV ≤ |ISENSE+ – ISENSE–| < 32mV
32mV ≤ |ISENSE+ – ISENSE–| < 63.9mV
63.9mV ≤ |ISENSE+ – ISENSE–| ≤ 70mV
|ISENSE+ – ISENSE–| ≥ 6mV, 0V ≤ VOUT ≤ 5.5V
NVOUT
VOUT_TUE
Internal Temperature Total Unadjusted
Readback Error
Update Rate
tCONVERT
Internal EEPROM (Notes 4, 6)
Endurance
Number of Write Operations
Retention
Stored Data Retention
Mass Write Time STORE_USER_ALL Execution Duration
6
1
27
0.4
10
0.5
2
10
–2
0.6V ≤ VOUT ≤ 5.5V, Constant Load
l
TINT_TUE
54
1.7
2.5
5.2
l
NDC
DCTUE
1
–0.5
2
244
±0.2
0.5
10
15.625
31.25
62.5
125
l
–1
1
±32
0.25
TSNS0, TSNS1 ≤ 1.85V (Note 10)
MFR_PWM_MODE_LTC3882-1[6] = 0
MFR_PWM_MODE_LTC3882-1[6] = 1
Internal Diode (Note 10)
l
l
–3
–7
(Note 11)
0°C ≤ TJ ≤ 85°C During All Write Operations
TJ ≤ 125°C
0°C ≤ TJ ≤ 85°C During All Write Operations
3
7
V
%
mV
mV
V
%
mV
mV
mV
mV
mV
mV
Bits
%
%
Bits
%
µV
%
%
Bits
µV
µV
µV
µV
%
µV
°C
±1
°C
°C
°C
100
ms
0.2
Cycles
Years
s
10,000
10
2
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Electrical Characteristics
The l denotes the specifications which apply over the specified operating
junction temperature range, otherwise specifications are at TJ = 25°C (Note 2). VCC = 5V, VSENSE0+ = VSENSE1+ = 1.8V, VSENSE0– =
VSENSE1– = IAVG_GND = GND = 0V, fSYNC = 500kHz (externally driven) unless otherwise specified.
SYMBOL
PARAMETER
CONDITIONS
Digital Inputs (SCL, SDA, RUNn, FAULTn, SYNC, SHARE_CLOCK)
Input High Voltage
SCL, SDA, RUN0, RUN1, FAULT0, FAULT1
VIH
SYNC, SHARE_CLK
Input Low Voltage
SCL, SDA, RUN0, RUN1, FAULT0, FAULT1
VIL
SYNC, SHARE_CLK
VHYST
Input Hysteresis
SCL, SDA
CIN
Input Capacitance
SCL, SDA, RUN0, RUN1, FAULT0, FAULT1, SYNC,
SHARE_CLK (Note 12)
tFILT
Input Digital Filter Delay
FAULT0, FAULT1
RUN0, RUN1
MIN
l
l
TYP
MAX
2.0
1.8
1.4
0.6
l
l
80
UNITS
V
V
V
V
mV
10
3
10
pF
µs
µs
Digital Outputs (SCL, SDA, RUNn, FAULTn, SYNC, SHARE_CLOCK, ALERT, PWMn, PGOODn)
VOL
Output Low Voltage
ISINK = 3mA; SDA, SCL, RUN0, RUN1, FAULT0,
FAULT1, SYNC, SHARE_CLK, ALERT,
ISINK = 2mA; PWMn, PGOODn
l
l
0.2
l
0.4
V
0.3
V
1
5
µA
µA
5
µA
VOH
PWMn Output High Voltage
ISOURCE = 2mA
ILKG
Output Leakage Current
0V ≤ PWM0, PWM1, PGOOD0, PGOOD1 ≤ VDD33
0V ≤ FAULT0, FAULT1, SYNC, SHARE_CLK ≤ 3.6V
0V ≤ RUN0, RUN1 ≤ 5.5V
0V ≤ SCL, SDA, ALERT ≤ 5.5V
tRO
PWMn Output Rise Time
CLOAD = 30pF, 10% to 90%
5
ns
tFO
PWMn Output Fall Time
CLOAD = 30pF, 90% to 10%
4
ns
2.7
V
–1
–5
l
–5
Serial Bus Timing
fSMB
Serial Bus Operating Frequency
l
10
tBUF
Bus Free Time Between Stop and Start
l
1.3
µs
tHD,STA
Hold Time After (Repeated) Start
Condition. After This Period, the First
Clock Is Generated
l
0.6
µs
tSU,STA
Repeated Start Condition Setup Time
l
0.6
µs
tSU,STO
Stop Condition Setup Time
l
0.6
µs
tHD,DAT
Data Hold Time:
Receiving Data
Transmitting Data
l
l
0
0.3
tSU,DAT
Input Data Setup Time
l
100
tTIMEOUT
Clock Low Timeout
l
25
35
ms
tLOW
Serial Clock Low Period
l
1.3
10000
µs
tHIGH
Serial Clock High Period
l
0.6
400
0.9
kHz
ns
µs
ns
µs
38821f
For more information www.linear.com/LTC3882-1
7
LTC3882-1
Electrical Characteristics
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 device
reliability and lifetime.
Note 2: The LTC3882-1 is tested under pulsed load conditions such that
TJ ≈ TA. The LTC3882-1E is guaranteed to meet performance specifications
from 0°C to 85°C. Specifications over the –40°C to 125°C operating
junction temperature range are assured by design, characterization and
correlation with statistical process controls. The LTC3882-1I is guaranteed
over the full –40°C to 125°C operating junction temperature range.
Junction temperature TJ is calculated in °C from the ambient temperature
TA and power dissipation PD according to the formula:
TJ = TA + (PD • θJA)
where θJA is the package thermal impedance. Note that the maximum
ambient temperature consistent with these specifications is determined by
specific operating conditions in conjunction with board layout, the rated
package thermal impedance and other environmental factors. Refer to the
Applications Information section.
Note 3: This IC includes overtemperature protection that is intended to
protect the device during momentary overload conditions. The maximum
rated junction temperature will be exceeded when this protection is active.
Continuous operation above the specified absolute maximum operating
junction temperature may impair device reliability or permanently damage
the device.
Note 4: EEPROM endurance, retention and mass write times are
guaranteed by design, characterization and correlation with statistical
process controls. Minimum retention applies only for devices cycled less
than the minimum endurance specification. EEPROM read commands
(e.g. RESTORE_USER_ALL) are valid over the entire specified operating
junction temperature range.
8
Note 5: All currents into device pins are positive; all currents out of device
pins are negative. All voltages are referenced to GND unless otherwise
specified.
Note 6: Minimum EEPROM endurance, retention and mass write time
specifications apply when writing data with 3.15V ≤ VDD33 ≤ 3.45V.
EEPROM read commands are valid over the entire specified VDD33
operating range.
Note 7: Specified VOUT error with AVP = 0% requires servo mode to be
set with MFR_PWM_MODE_LTC3882-1 command bit 6. Performance is
guaranteed by testing the LTC3882-1 in a feedback loop that servos VOUT
to a specified value.
Note 8: ADC tested with PWMs disabled. Comparable capability
demonstrated by in-circuit evaluations. Total Unadjusted Error includes all
gain and linearity errors, as well as offsets.
Note 9: Internal 32-bit calculations using 16-bit ADC results are limited to
10-bit resolution by PMBus Linear 11-bit data format.
Note 10: Limits guaranteed by TSNS voltage and current measurements
during test, including ADC readback.
Note 11: Data conversion is done in round robin fashion. All inputs signals
are continuously scanned in sequence resulting in a typical conversion
latency of 100ms.
Note 12: Guaranteed by design.
Note 13: Do not apply a voltage or current source directly to these pins.
They should only be connected to passive RC loads, otherwise permanent
damage may occur.
Note 14: Do not apply a voltage source to this pin unless shorted to VCC.
See Electrical Characteristics for applicable limits beyond which permanent
damage may occur.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Typical Performance Characteristics
LTC3882-1 1.0V Regulated Output
vs Temperature
Typical LTC3882-1 Output Voltage
Distribution at 0°C
1000
1.0
0.9995
1200
1094 UNITS
900 VOUT_COMMAND = 1.0V
DIGITAL SERVO
800 ENGAGED
700
NUMBER OF CHANNELS
VOUT (V)
VIN = 12V
VOUT_COMMAND = 1.0V
DIGITAL SERVO ENGAGED
1.0005 I
OUT = 6.5A
600
500
400
300
0.999
1078 UNITS
VOUT_COMMAND = 1.0V
DIGITAL SERVO
ENGAGED
1000
NUMBER OF CHANNELS
1.001
Typical LTC3882-1 Output Voltage
Distribution at 105°C
800
600
400
200
200
100
0.9985
–5
15
35
55
TA (°C)
75
38821 G01
38821 G01a
Efficiency and Loss vs Load
(2-Phase Using FDMF5820DC
DRMOS)
Efficiency vs Load Current
(1-Phase Using D12S1R880A
Power Block)
92
13
95
7
86
85
84
5
83
85
3.3V
2.5V
1.8V
1.5V
1.2V
1.0V
80
VIN = 12V
VOUT = 1V
SYNC = 500kHz
82
81
0
10
20
30 40 50 60
LOAD CURRENT (A)
70
3
80
1
75
0
10
4000
3500
1500
1000
1000
–400 –300–200–100 0 100 200 300 400
CH1 ISENSE OFFSET TO IDEAL (µV)
38821 G06
0
30
50
20
40
LOAD CURRENT (A)
60
70
11783 UNITS
FROM 3 LOTS
TJ = 121°C
CHO MASTER
3000
2500
2000
1500
1000
500
500
10
Typical Distribution of Slave
IOUT Offset (Not Including DCR
Mismatch)
4500
2000
0
38821 G05
NUMBER OF ICs
NUMBER OF ICs
NUMBER OF ICs
80
8593 UNITS
FROM 3 LOTS
3000 T = 38°C
J
CHO MASTER
2500
1500
0
82
3500
2000
86
Typical Distribution of Slave
IOUT Offset (Not Including DCR
Mismatch)
9595 UNITS
3500 FROM 3 LOTS
TA = –40°C
3000 TJ = –22°C
CHO MASTER
2500
88
38821 G04
4000
38821 G02
84
40
20
30
LOAD CURRENT (A)
38821 G03
Typical Distribution of Slave
IOUT Offset (Not Including DCR
Mismatch)
2.5
90
EFFICIENCY (%)
87
POWERLOSS (W)
9
EFFICIENCY (%)
90
88
2
VIN = 12V
VOUT = 1.5V
92
89
EFFICIENCY (%)
94
11
90
1.5
Efficiency vs Load Current
(3-Phase Using D12S1R845A
Power Block)
VIN = 12V
91
80
0
–2.5 –2 –1.5 –1 –0.5 0 0.5 1
VOUT ERROR (mV)
0
–1.25 –1 –0.75–0.5–0.25 0 0.25 0.5 0.75 1 1.25
VOUT ERROR (mV)
95
500
–400 –300–200–100 0 100 200 300 400
CH1 ISENSE OFFSET TO IDEAL (µV)
38821 G07
0
–300 –200–100 0 100 200 300 400 500
CH1 ISENSE OFFSET TO IDEAL (µV)
38821 G08
38821f
For more information www.linear.com/LTC3882-1
9
LTC3882-1
Typical Performance Characteristics
3-Phase DC Output Current
Sharing (Using D12S1R845A
Power Block
20
Load Step Transient Current
Sharing (Using FDMF6707B
DrMOS)
CHANNEL 1
CHANNEL 2
CHANNEL 3
18
16
IOUT
20A/DIV
IOUT
20A/DIV
14
PHASE CURRENT (A)
Load Dump Transient Current
Sharing (Using FDMF6707B
DrMOS)
VOUT
20mV/DIV
VOUT
20mV/DIV
12
IL1, IL2
10A/DIV
10
8
IL1, IL2
10A/DIV
6
4
VOUT = 1V
VIN = 12V
SYNC = 500kHz
L = 320nH
2
0
0
10
20 30 40 50 60
TOTAL RAIL CURRENT (A)
80
70
38821 G10
5µs/DIV
VOUT = 1V
VIN = 12V
SYNC = 500kHz
L = 320nH
38821 G11
5µs/DIV
38821 G09
Efficiency and Power Loss vs
Input Voltage
(1-Phase Using LTC4449)
100
3.0
VO = 1.8V
98
EFFICIENCY (%)
2.0
92
90
1.5
88
1.0
86
84
10
15
20
VIN (V)
IOUT
(10A/DIV)
VSW
(10V/DIV)
VOUT
(10mV/DIV)
25mVP-P
0.5
POWER FET: BSC050N04LS G
SYNC FET: BSC010N04LS
5
POWERLOSS (W)
94
80
VOUT
(20mV/DIV)
2.5
96
82
25
30
100µs/DIV
0
38821 G12
3+1 Channel Crosstalk
(Using D12S1R845A Power
Blocks)
VOUT0
(1-PHASE)
20mV/DIV
25%
LOAD STEP
100µs/DIV
38821 G15
38821 G14
2µs/DIV
VOUT = 0.9V/90A
VIN = 12V
SYNC = 500kHz
L = 210nH
VOUT = 1V/25A
VIN = 12V
SYNC = 1MHz
L = 210nH
Load Step Transient Response
Using AVP
Line Step Transient Response
(1-phase Using LTC4449)
IO
10A/DIV
IOUT1
10A/DIV
IOUT
(10A/DIV)
38821 G13
VIN
2V/DIV
VOUT1
(3-PHASE)
20mV/DIV
10
1-Phase Single Cycle Response
(Using D12S1R860A Power Block
with COUT = 6 × 100µF X5R 1210)
3-Phase Transient Response
(Using D12S1R860A Power Block)
7V
1.8V
VOUT
50mV/DIV
VOUT
10mV/DIV
200µs/DIV
38821 G16
200µs/DIV
38821 G17
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Typical Performance Characteristics
Soft-Start Ramp
Start-Up Into a Prebiased Load
VOUT
0.5V/DIV
VOUT
0.5V/DIV
0V
IL1, IL2
10A/DIV
IL1, IL2
10A/DIV
VIN = 12V
1ms/DIV
38821 G18
1.7995
RUN
2V/DIV
VOUT
1V/DIV
VIN = 12V
Regulated Output vs Temperature
1.8000
Soft-Off Ramp
38821 G19
1ms/DIV
VOUT_COMMAND INL
VOUT_COMMAND DNL
1.5
VOUT_COMMAND = 1.8V
DIGITAL SERVO OFF
1.0
0.8
1.0
0.6
0.4
1.7985
0.5
DNL (LSB)
INL (LSB)
VOUT (V)
1.7990
0
1.7980
0.2
0
–0.2
–0.4
–0.5
1.7975
–0.6
1.7970
–40 –20
0
20 40 60 80
TEMPERATURE (°C)
–1.0
0.3
100 120
1.1
1.9
38821 G01a
Output Overvoltage Threshold
Error vs Temperature
0
–0.05
–0.10
VOUT_OV_FAULT_LIMIT = 2V
VOUT RANGE = 1
–0.15
–40 –20
0
20 40 60 80
TEMPERATURE (°C)
100 120
38821 G21
–0.8
0.3
5.1 5.5
4.3
1.1
500.2
1.0
500.1
0.8
0.6
0.4
0.2
0
4.3
5.1 5.5
38821 G02
500.0
499.9
499.8
499.7
–0.2
499.6
–0.4
–40 –20
499.5
–40 –20
20 40 60 80
TEMPERATURE (°C)
2.7
3.5
VOUT (V)
PWM Frequency vs Temperature
1.2
0
1.9
38821 G01
PWM FREQUENCY (kHz)
OUTPUT OC THRESHOLD ERROR (%)
0.05
2.7
3.5
VOUT (V)
Output Overcurrent Threshold
Error vs Temperature
0.10
VOUT OV THRESHOLD ERROR (%)
38821 G20
5ms/DIV
TOFF_DELAY = 10ms
TOFF_FALL = 5ms
100 120
38821 G22
FREQUENCY_SWITCH = 500kHz
0
20 40 60 80
TEMPERATURE (°C)
100 120
38821 G23
38821f
For more information www.linear.com/LTC3882-1
11
LTC3882-1
Typical Performance Characteristics
VIN(SNS) ADC TUE
0.40
IOUT ADC TUE
8
0.30
MEASUREMENT ERROR (mV)
MEASUREMENT ERROR (mV)
–1
VOUT ADC TUE
–2
–3
–4
–5
–6
–7
6
MEASUREMENT ERROR (mA)
0
0.20
0.10
0
–0.10
–0.20
–8
–0.30
–9
–0.40
0.5
0
5
10
15 20 25
VINSNS (V)
30
35
40
38821 G24
1.0
0
–2
–4
–6
1
1.5
2
2.5 3 3.5
VOUT (V)
4
4.5
5
–8
5.5
10
5
15
OUTPUT CURRENT (A)
0
38821 G25
20
38821 G26
IC Operating Current vs
Temperature
31.0
110
0.8
VCC = 14V
0.6
0.4
0.2
0
–0.2
–0.4
–0.6
ICC OPERATING CURRENT (mA)
30.8
SHARE_CLK FREQUENCY (kHz)
MEASUREMENT ERROR (°C)
2
SHARE_CLK Frequency vs
Temperature
Temperature ADC TUE
105
100
95
30.6
30.4
30.2
30.0
29.8
29.6
–0.8
–1.0
–45 –25 –5 15 35 55 75 95 115
ACTUAL TEMPERATURE (°C)
38821 G27
12
4
90
–50 –30 –10 10 30 50 70
TEMPERATURE (°C)
90
110
38821 G28
29.4
–40 –20
0
20 40 60 80
TEMPERATURE (°C)
100 120
38821 G29
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Pin Functions
COMP0/COMP1 (Pin 1/Pin 29): Error Amplifier Outputs.
PWM duty cycle increases with this control voltage. These
are true low impedance outputs and cannot be directly
connected together when active. For PolyPhase operation,
wiring FB to VDD33 will three-state the error amplifier output
of that channel, making it a slave. PolyPhase control is
then implemented in part by connecting all slave COMP
pins together to one master error amplifier output.
TSNS0/TSNS1 (Pin 2/Pin 3): External Temperature Sense
Inputs. The LTC3882-1 supports two methods of calculation of external temperature based on forward-biased P/N
junctions between these pins and GND.
VINSNS (Pin 4): VIN Supply Sense. Connect to the VIN
power supply to provide line feedforward compensation.
A change in VIN immediately modulates the input to the
PWM comparator and inversely changes the pulse width
to provide excellent transient line regulation and fixed
modulator voltage gain. An external lowpass filter can be
added to this pin to prevent noisy signals from affecting
the loop gain.
IAVG_GND (Pin 5): IAVG Ground Reference. The same
IAVG_GND should be shared between all channels of a
PolyPhase rail and connected to system ground at a single
point. IAVG_GND may be wired directly to GND on ICs that
do not share phases with other chips.
PGOOD/PGOOD1 (Pin 6/Pin 27): Power Good Indicator
Open-Drain Outputs. These outputs are driven low through
a 30µs filter when the respective channel output is below
its programmed UV fault limit or above its programmed
OV fault limit. If used, a pull-up resistor is required in the
application. Operating voltage range is GND to VDD33.
PWM0/PWM1 (Pin 7/Pin 26): PWM Three-State Control
Outputs. These pins provide single-wire PWM switching
control for each channel to an external gate driver, DrMOS
or power block. Operating voltage range is GND to VDD33.
SYNC (Pin 8): External Clock Synchronization Input and
Open-Drain Output. If desired, an external clock can be
applied to this pin to synchronize the internal PWM channels. If the LTC3882-1 is configured as a clock master, this
pin will also pull to ground at the selected PWM switching
frequency with a 125ns pulse width. A pull-up resistor to
3.3V is required in the application if SYNC is driven by
any LTC3882-1. Minimize the capacitance on this line to
ensure its time constant is fast enough for the application.
SCL (Pin 9): Serial Bus Clock Input. A pull-up resistor to
3.3V is required in the application.
SDA (Pin 10): Serial Bus Data Input and Output. A pull-up
resistor to 3.3V is required in the application.
ALERT (Pin 11): Open-Drain Status Output. This pin may
be connected to the system SMBALERT wire-AND interrupt signal and should be left open if not used. If used, a
pull-up resistor is required in the application. Operating
voltage range is GND to VDD33.
FAULT0/FAULT1 (Pin 12/Pin 13): Programmable Digital
Inputs and Open-Drain Outputs for Fault Sharing. Used
for channel-to-channel fault communication and propagation. These pins should be left open if not used. If used,
a pull-up resistor to 3.3V is required in the application.
RUN0/RUN1 (Pin 14/Pin 15): Run Control Inputs and
Open-Drain Outputs. A voltage above 2V is required on
these pins to enable the respective PWM channel. The
LTC3882-1 will drive these pins low under certain reset/
restart conditions regardless of any PMBus command
settings. A pull-up resistor to 3.3V is required in the application.
ASEL0/ASEL1 (Pin 16/Pin 17): Serial Bus Address Select
Pins. Connect optional 1% resistor dividers between VDD25
and GND to these pins to select the serial bus interface
address. Refer to the Applications Information section
for more detail.
VOUT0_CFG/VOUT1_CFG (Pin 18/Pin 19): Output Voltage
Configuration Pins. Connect optional 1% resistor dividers between VDD25 and GND to these pins to select the
output voltage for each channel. Refer to the Applications
Information section for more detail.
FREQ_CFG (Pin 20): Frequency Configuration Pin. Connect
an optional 1% resistor divider between VDD25 and GND
to this pin to configure PWM switching frequency. Refer
to the Applications Information section for more detail.
38821f
For more information www.linear.com/LTC3882-1
13
LTC3882-1
Pin Functions
PHAS_CFG (Pin 21): Phase Configuration Pin. Connect
an optional 1% resistor divider between VDD25 and GND
to this pin to configure the phase of each PWM channel
relative to SYNC. Refer to the Applications Information
section for more detail.
VDD25 (Pin 22): Internal 2.5V Regulator Output. Bypass
this pin to GND with a low ESR 1µF capacitor. Do not load
this pin with external current beyond that required for local
LTC3882-1 configuration pins, if any.
SHARE_CLK (Pin 23): Share Clock Open-Drain Output
(bussed). Share Clock, nominally 100kHz, is used to
sequence multiple rails in a power system utilizing more
than one LTC PSM controller. A pull-up resistor is required
in the application. Minimize the capacitance on this line to
ensure the time constant is fast enough for the application.
Operating voltage range is GND to VDD33.
VDD33 (Pin 24): Internal 3.3V Regulator Output. Bypass this
pin to GND with a low ESR 2.2µF capacitor. The LTC3882-1
may also be powered from an external 3.3V rail attached
to this pin, if also shorted to VCC. Do not overload this pin
with external system current. Local pull-up resistors for
the LTC3882-1 itself may be powered from VDD33. Refer
to the Applications Information section for more detail.
VCC (Pin 25): 3.3V Regulator Input. Bypass this pin to
GND with a capacitor (0.1µF to 1µF ceramic) in close
proximity to the IC.
–/V
–
VSENSE0 SENSE1 (Pin 35/Pin 34): Negative Output
Voltage Sense Inputs. These pins must still be properly
connected on slave channels for accurate output current
telemetry.
14
VSENSE0+/VSENSE1+ (Pin 36/Pin 33): Positive Output Voltage
Sense Inputs. These pins must still be properly connected
on slave channels for accurate output current telemetry.
ISENSE0–/ISENSE1– (Pin 37/Pin 32): Current Sense Amplifier Inputs. The (–) inputs to the amplifiers are normally
connected to the low side of a DCR sensing network or
output current sense resistor for each phase.
ISENSE0+/ISENSE1+ (Pin 38/Pin 31): Current Sense Amplifier Inputs. The (+) inputs are normally connected to the
high side of an output current sense resistor or the R-C
midpoint of a parallel DCR sense circuit.
IAVG0/IAVG1 (Pin 39/Pin 31): Average Current Control Pins.
A capacitor connected between these pins and IAVG_GND
stores a voltage proportional to the average output current
of the master channel. PolyPhase control is then implemented in part by connecting all slave IAVG pins together
to the master IAVG output. This pin should be left open
on channels that control single-phase outputs. Operating
voltage range is GND to 2.1V.
FB0/FB1 (Pin 40/Pin 30): Error Amplifier Inverting Inputs.
These pins provide an internally scaled version of the
output voltage for use in loop compensation. Refer to the
Applications Information section for additional details on
compensating the output voltage control loop with external
components.
GND (Exposed Pad Pin 41): Ground and VSENSE1–. All
small-signal and compensation components should
connect to this pad, which also serves as the negative
voltage sense input for channel 1. The exposed pad must
be soldered to a suitable PCB copper ground plane for
proper electrical operation and to obtain the specified
package thermal resistance.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Block Diagram
ROM
RAM
EEPROM
VINSNS
R_CONFIG
IAVG0
MCU AND
CUSTOM
LOGIC
SHARE_CLK
PMBus
±
ISENSE0
2.5V
REGULATOR
12-BIT
DAC
PWM0
VSENSE0±
PGOOD0
SYNC
PWM0
PLL
VOLTAGE
REFERENCE
VREF
IAVG_GND
3.3V
REGULATOR
VCC
PWM1
VDD33
PGOOD1
BIAS AND
HOUSEKEEPING
VSENSE1
IAVG1
INTERNAL DATA BUS
VSENSE0±
16-BIT
ADC
ISENSE0±
VSENSE1±
PWM1
ISENSE1±
VINSNS
PWM0
TSNS0
PWM1
12-BIT
DAC
±
VINSNS
ANALOG
MUX
INTERNAL
TEMPERATURE
ISENSE1±
TSNS1
3882-1 BD
38821f
For more information www.linear.com/LTC3882-1
15
LTC3882-1
Test Circuit
(Channel 0 Example)
LTC3882-1
1.024V
VR
12-BIT
D/A
DIGITAL
+
EA
–
VSENSE0–
VSENSE0+
35
FB0
36
COMP0
40
1
+
LTC1055
TARGET = VOUT_COMMAND
–
1V
38821 TC
Timing Diagram
SDA
tf
tLOW
tr
tSU(DAT)
tHD(SDA)
tf
tSP
tr
tBUF
SCL
tHD(STA)
tHD(DAT)
tHIGH
tSU(STA)
START
CONDITION
tSU(STO)
38821 TD
REPEATED START
CONDITION
STOP
CONDITION
START
CONDITION
Operation
Overview
Major features include:
The LTC3882-1 is a dual channel/dual phase, constant
frequency analog voltage mode controller for DC/DC stepdown applications. It features a PMBus compliant digital
interface for monitoring and control of important power
system parameters. The chip operates from an IC power
supply between 3V and 13.2V and is intended for conversion
from VIN between 3V and 38V to output voltages between
0.5V and 5.25V. It is designed to be used in a switching
architecture with external FET drivers, including higher
level integrations such as non-isolated power blocks.
• Digitally Programmable Output Voltage
16
• Digitally Programmable Output Current Limit
• Digitally Programmable Input Voltage Supervisor
• Digitally Programmable Output Voltage Supervisors
• Digitally Programmable Switching Frequency
• Digitally Programmable On and Off Delay Times
• Digitally Programmable Soft-Start/Stop
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
• Operating Condition Telemetry
Main Control Loop
• Phase Locked Loop for Synchronous PolyPhase Operation (2, 3, 4, 6, or 8 phases)
The LTC3882-1 utilizes constant frequency voltage mode
control with leading-edge modulation. This provides
improved response to a load step increase, especially at
larger VIN/VOUT ratios found in the low voltage, high current solutions demanded by modern digital subsystems.
The LTC3882-1 leading-edge modulation architecture
does not have a minimum on-time requirement. Minimum
duty cycle will be determined by performance limits of
the external power stage. The IC is also capable of active
voltage positioning (AVP) to afford the smallest output
capacitors possible for a given output voltage accuracy
over the anticipated full load range. The LTC3882-1 error
amplifiers have high bandwidth, low offset and low output impedance, allowing the control loop compensation
network to be optimized for very high crossover frequencies and excellent transient response. The controller also
achieves outstanding line transient response by using
input feedforward compensation to instantaneously adjust
PWM duty cycle and significantly reduce output under/
overshoot during supply voltage changes. This also has the
added advantage of making the DC loop gain independent
of input voltage.
• Fully Differential Load Sense
• Non-Volatile Configuration Memory
• Optional External Configuration Resistors for Key Operating Parameters
• Optional Time-Base Interconnect for Synchronization
Between Multiple Controllers
• Fault Event Data Logging
• Capable of Standalone Operation with Default Factory
Configuration
• PMBus Revision 1.2 Compliant Interface up to 400kHz
The PMBus interface provides access to important power
management data during system operation including:
• Average Input Voltage
• Average Output Voltages
• Average Output Currents
• Average PWM Duty Cycles
• Internal LTC3882-1 Temperature
• External Sensed Temperatures
• Warning and Fault Status, Including Input and Output
Undervoltage and Overvoltage
The LTC3882-1 supports four serial bus addressing
schemes to access the individual PWM channels separately
or jointly.
Fault communication, reporting and system response
behavior are fully configurable. Two fault I/Os are provided (FAULT0, FAULT1) that can be controlled independently. A separate ALERT pin also provides for a maskable
SMBALERT#. Fault responses for each channel may be
individually programmed, depending on the fault type.
PMBus status commands allow fault reporting over the
serial bus to identify a specific fault event.
The main PWM control loop used for each channel is
illustrated in Figure 1. During normal operation the top
MOSFET (power switch) driving choke L1 is commanded
off when the clock for that channel resets the RS latch.
The power switch is commanded back on when the main
PWM comparator VC, sets the RS latch. The error amplifier EA output (COMP) controls the PWM duty cycle to
match the FB voltage to the EA positive terminal voltage
in steady state. A patented circuit adjusts this output for
VINSNS line feedforward.
The positive terminal of the EA is connected to the output
of a 12-bit DAC with values ranging from 0V to 1.024V. The
DAC value is determined by the resistor configuration pins
detailed in application Table 8, by values retrieved from internal EEPROM, or by a combination of PMBus commands to
synthesize the desired output voltage. Refer to the following
PMBus Command Details section of this document for more
information. The LTC3882-1 supports two output ranges.
EA can regulate the output voltage to 5.5x the DAC output
(Range 0) or 2.75x the DAC output (Range 1).
38821f
For more information www.linear.com/LTC3882-1
17
LTC3882-1
Operation
LTC3882
MODE
OSCILLATOR
CLOCK
R
Q
PWM0
PWM
LOGIC
S
VIN
7
GATE
DRIVER
0V
VOC0
8-BIT DAC
IOUT_OC_FAULT_LIMIT
ILIM
RAMP
VC
VREV
IREV
4
VINSNS
FEED
FORWARD
+
+
CA
–
+
S
–
SLAVE
ENABLE
ISENSE0+
ISENSE0–
IAVG0
IAVG_GND
SLAVE
DETECT
RS
38
L1
CS
VOUT
37
39
COUT
5
MASTER
ENABLE
VOV0
9-BIT DAC
VOUT_OV_FAULT_LIMIT
9-BIT DAC
VOUT_UV_FAULT_LIMIT
OV
UV
VUV0
VSENSE0+
36
9R
(RANGE 0)
–
EA
VSP0
FB0
12-BIT DAC
VOUT_COMMAND
2R
VSENSE0–
+
COMP0
LOOP
COMPENSATION
NETWORK
40
35
1
38821 F01
Figure 1. LTC3882-1 PWM Control Loop Diagram
18
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
VC discriminates its positive input against an internally
generated PWM voltage ramp. The positive input is a composite control based on COMP voltage with line feedforward
compensation, and current sharing if the channel controls
a slave phase. When the ramp falls below this voltage the
comparator trips and sets the PWM latch.
If load current increases, VSENSE+ and FB will droop
slightly with respect to the 12-bit DAC output. This causes
the COMP voltage to increase until the average inductor
current matches the new load current and the desired
output voltage is restored. Programmable comparators
ILIM and IREV monitor peak instantaneous forward and
reverse inductor current for pulse-by-pulse protection.
The top power MOSFET is immediately commanded off if
the programmed positive limit is reached, and the bottom
MOSFET is immediately commanded off if the negative
limit is reached. Repeated peak overcurrent events cause
an overcurrent fault to be set.
When the top MOSFET is commanded off, the bottom
MOSFET is normally commanded on. In continuous conduction mode (CCM) the bottom MOSFET stays on until
comparator VC turns the top MOSFET back on. Otherwise
in discontinuous conduction mode (DCM, also known as
diode emulation) the bottom MOSFET is commanded off
if the IREV comparator detects that the inductor current
has decayed to approximately 0A. In any case the next
PWM cycle starts when the clock for that channel again
clears the RS latch.
Power-Up and Initialization
The LTC3882-1 is designed to provide stand-alone supply
sequencing with controlled turn-on and turn-off functions.
It operates from a single IC input supply of 3V to 13.2V
while two on-chip linear regulators generate internal
2.5V and 3.3V. If VCC is below 4.5V, the VCC and VDD33
pins must be shorted together and limited to a maximum
operating voltage of 3.6V. Controller configuration is
reset by the internal UVLO threshold, where VDD33 must
be at or above 3V and the internal 2.5V supply must be
within about 20% of its regulated value. At that point the
internal microcontroller begins initialization. A PMBus
RESTORE_USER_ALL or MFR_RESET command forces
this same initialization.
The LTC3882-1 features an internal RAM built-in self-test
(BIST) that runs during initialization. Should RAM BIST
fail, the following steps are taken.
• Device responds only at device address 0x7C and global
addresses 0x5A and 0x5B
• A persistent Memory Fault Detected is indicated by
STATUS_CML
• Internal EEPROM is not accessed
• RUNn and SHARE_CLK are driven low continuously
Normal operation can be restored if the RAM BIST subsequently passes, for instance as the result of another
MFR_RESET command issued to address 0x7C.
During initialization all PWM outputs are disabled. The
RUNn pins and SHARE_CLK are held low and FAULTn
pins are high impedance. External configuration resistors
are identified and the contents of the onboard EEPROM
are read into the controller command memory space. The
LTC3882-1 can determine key operating parameters from
external configuration resistors according to application
Table 8 through Table 11. See the following Resistor
Configuration Pins section for more detail. The resistor
configuration pins only determine some of the preset
values of the controller. The remaining values, retrieved
from internal EEPROM, are programmed at the factory or
with PMBus commands.
If the configuration resistor pins are all open, the LTC3882-1
will use only EEPROM contents to determine all operating
parameters. If Ignore Resistor Configuration Pins is set (bit
6 of MFR_CONFIG_ALL_LTC3882-1), the LTC3882-1 will
use only its EEPROM contents to determine all operating
parameters except device address. Unless both ASEL pins
are completely open, the LTC3882-1 will always determine
some portion of its device address from the resistors on
these pins. See Serial Bus Addressing later in this section.
The internal microcontroller typically requires 70ms to
complete initialization from VDD33 ≥ 3V. At that point, an
internal comparator monitors VINSNS, which must exceed
the VIN_ON threshold before output power sequencing
can begin (SHARE_CLK released, ready for TON_DELAY).
Accurate readback telemetry can then require an additional
100ms for initial round-robin A/D conversions.
38821f
For more information www.linear.com/LTC3882-1
19
LTC3882-1
Operation
Soft-Start
The RUN pins are released for external control after the
part initializes and VINSNS is greater than the VIN_ON
threshold. If multiple LTC3882-1 ICs are used in an application, shared RUN pins are held low until all units
initialize and VINSNS exceeds the VIN_ON threshold for
all devices. A common SHARE_CLK signal can also ensure
all connected devices use the same time reference for
initial start-up even if RUN pins cannot be shared due to
other design requirements. SHARE_CLK is not released
by each IC until the conditions for power sequencing have
been fully satisfied.
After a channel RUN pin rises above 2V and any specified
turn on delay (TON_DELAY) has expired, the LTC3882-1
performs an initial monotonic soft-start ramp on that channel. This is carried out with a digitally controlled ramp of
the regulated output voltage from 0V to the commanded
voltage set point over the programmed TON_RISE period,
allowing inrush current control. During the soft-start ramp,
the LTC3882-1 does not initiate PWM operation until the
commanded output exceeds the actual rail voltage. This
allows the regulator to start up into a pre-biased load
even when using gate drivers or power blocks that do not
support discontinuous operation. The soft-start feature
is disabled by setting the value of TON_RISE to any time
less than 0.25ms.
Time-Based Output Sequencing
The LTC3882-1 supports time-based on and off output
sequencing using a shared time reference (SHARE_CLK).
Following a valid qualified command to turn on, each output
is enabled after waiting its programmed TON_DELAY. This
can be used to sequence outputs in a prescribed order
that can be preprogrammed as needed without hardware
modification. Channel off-sequencing is accomplished in
a similar way with the TOFF_DELAY command.
Output Ramping Control
The LTC3882-1 supports synchronized output on and off
ramping control using a shared time reference (SHARE_
CLK). Power rail on and off relationships similar to those of
conventional analog tracking functions can be achieved by
using programmed delays and TON_RISE and TOFF_FALL
20
times. However, with LTC3882-1 digital control, on and
off ramping methods need not be the same, and ramping
configurations can be reprogrammed as needed without
hardware modification.
Programmable fault responses and fault sharing can
ensure that any desired time-based output sequencing
and ramping control is properly accomplished each time
the system powers up or down. Refer to the Applications
Information section for various LTC3882-1 hardware and
PMBus command configurations needed to fully support
synchronization for time-based sequencing and output
ramping when using multiple ICs.
Voltage-Based Output Sequencing
It is also possible to sequence outputs using cascaded
voltage events. To do this, the PGOOD status output from
one PWM channel can be used to control the RUN pin
of a downstream channel. This keeps the downstream
channel off unless acceptable output conditions exist on
the controlling channel.
Output Disable
Both PWM channels are disabled any time VINSNS is below
the VIN_OFF threshold. The power stages are immediately
shut off to stop the transfer of energy to the load(s) as
quickly as possible.
A PWM channel may also be disabled in response to certain
internal fault conditions, an external fault propagated into
a FAULT pin, or loss of SHARE_CLK. In these cases the
power stage is immediately shut off to stop the transfer
of energy to the load as quickly as possible. Refer to the
following Fault Detection and Handling section for additional details related to fault recovery.
Each PWM channel can be disabled with a PMBus OPERATION command at any time if enabled by ON_OFF_CONFIG.
This will force a controlled turn-off response with defined
delay (TOFF_DELAY) and ramp down rate (TOFF_FALL).
The controller will maintain the programmed mode of
operation for TOFF_FALL. In DCM, the controller will not
draw current from the load and fall time will be set by
output capacitance and load current.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
Finally, each PWM channel can be commanded off by
pulling the associated RUN pin low. Pulling the RUN pin
low can force the channel to perform a controlled turn off
or immediately disable the power stage, depending on the
programming of the ON_OFF_CONFIG command.
Minimum Output Disable Times
When a PMBus OPERATION command is used to turn off
an LTC3882-1 channel, a minimum output disable time of
120ms is imposed regardless of how quickly the channel
is commanded back on. If bit 4 of MFR_CHAN_CONFIG is
clear, a PMBus command to turn the channel off also pulses
the RUN pin low. Once the RUN pin is pulled low internally
or externally, a minimum output disable time (RUN forced
low) of TOFF_DELAY + TOFF_FALL + 136ms is enforced.
If MFR_RESTART_DELAY is greater than this mandatory
minimum, the larger value of MFR_RESTART_DELAY is
used. In either case the LTC3882-1 holds its own RUN
pin low during the entire disable period. These minimum
off times allow a consistent channel restart with coherent monitor ADC values and make the LTC3882-1 highly
compatible with other LTC PMBus digital power system
management products.
Output Short Cycle
An output short cycle condition is created when a master channel is commanded back on while waiting for
TOFF_DELAY or TOFF_FALL to expire. Any time this
occurs, the LTC3882-1 asserts the Short Cycle bit in
STATUS_MFR_SPECIFIC. Device response at that point
is governed by bits in MFR_CHAN_CONFIG_LTC3882-1
and SMBALERT_MASK. Refer to the detailed descriptions
of those commands for additional details. Generally, the
LTC3882-1 should be controlled so that short cycle conditions are not created during normal operation.
Light Load Current Operation
The LTC3882-1 has two modes of PWM operation: discontinuous conduction mode (DCM) and forced continuous
conduction mode (CCM). Mode selection is made with
the MFR_PWM_MODE command.
In DCM, the inductor current is not allowed to reverse.
The reverse current comparator IREV disables the external
bottom MOSFET (synchronous rectifier) when the induc-
tor current reaches approximately 0A, preventing it from
going substantially negative. The external gate driver or
power block must have short delays to a high impedance
output, relative to the PWM cycle, to support DCM.
Efficiency at light loads in CCM is lower than in DCM.
Continuous conduction mode exhibits less interference
with audio circuitry but may result in reverse inductor
current, for instance at light loads or under large transient
conditions.
Switching Frequency and Phase
There is a high degree of flexibility for setting the PWM
operating frequency of the LTC3882-1. The switching
frequency of the PWM can be established with an internal oscillator or an external time base. The internal
phase-locked loop (PLL) synchronizes PWM control to
this timing reference with proper phase relation, whether
the clock is provided internally or externally. The device
can also be configured to provide the master clock to
other ICs through PMBus command, EEPROM setting,
or external configuration resistors as outlined in application Table 10. For PMbus or EEPROM configuration, the
LTC3882-1 is designated as a clock master by clearing
bit 4 of MFR_CONFIG_ALL_LTC3882-1. As clock master,
the LTC3882-1 will drive its open-drain SYNC pin at the
selected rate with a pulse width of 125ns. An external
pull-up resistor between SYNC and VDD33 is required in
this case. Only one device connected to SYNC should be
designated to drive the pin. If more than one LTC3882-1
sharing SYNC is programmed as clock master, just one of
the devices is automatically elected to provide the clock.
The others disable their SYNC outputs and indicate this
with bit 10 of MFR_PADS_LTC3882-1.
The LTC3882-1 will automatically accept an external SYNC
input, disabling is own SYNC drive if necessary, as long
as the external clock frequency is greater than 1/2 of the
programmed internal oscillator. Whether configured to drive
SYNC or not, the LTC3882-1 can continue PWM operation
at the selected frequency (FREQUENCY_SWITCH) using
its own internal oscillator, if an external clock signal is
subsequently lost.
The MFR_PWM_CONFIG_LTC3882-1 command can be
used to configure the phase of each channel. Desired phase
38821f
For more information www.linear.com/LTC3882-1
21
LTC3882-1
Operation
can also be set from EEPROM or external configuration
resistors as outlined in Table 10. Phase designates the
relationship between the falling edge of SYNC and the
internal clock edge that resets the PWM latch. That reset
turns off the top power switch, producing a PWM falling
edge. Additional small propagation delays to the PWM
control pins will apply.
proportionally summed with the master error amplifier
COMP output to adjust the duty cycle and balance the
current contribution of that phase. Additional hardware
configuration and digital programming requirements apply
in PolyPhase systems. Refer to the Applications Information section for complete details on building PolyPhase
rails with the LTC3882-1.
The phase relationships and frequency are independent
of each other, providing numerous application options.
Multiple LTC3882-1 ICs can be synchronized to realize a
PolyPhase array. In this case the phases should be separated by 360/n degrees, where n is the number of phases
driving the output voltage rail.
Active Voltage Positioning
PolyPhase Load Sharing
Multiple LTC3882-1 ICs can be combined to provide a balanced load-share solution by configuring the necessary
pins. The SHARE_CLK and SYNC pins of all load-sharing
channels should be bussed together. Connecting the
SYNC pins synchronizes the PWM controllers with each
other. Bussing the SHARE_CLK pins together allows the
phases to start synchronously. Refer to the discussion in
the previous Power-Up and Initialization section. The last
device to see all start-up conditions satisfied controls the
initiation of power sequencing for all phases.
Due to the low output impedance of the LTC3882-1 error
amplifiers, PolyPhase applications should use the error
amplifier of only one phase as the master. The FB pins of
each slave channel must be wired to VDD33, and the COMP
pins of each slave phase must be connected to the master
error amplifier COMP output. This disables the slave error
amplifiers and provides a single point of voltage control
and loop stabilization for the PolyPhase output rail.
For PolyPhase load sharing the LTC3882-1 also incorporates
an auxiliary current sharing loop. Referring back to Figure 1,
the instantaneous current of each slave phase is sensed
by current amplifier CA and compared to the IAVG pin. The
IAVG and IAVG_GND pins of each phase are wired together,
and a small capacitor (50pF to 200pF) between IAVG and
IAVG_GND stores a voltage corresponding to the average
master phase output current. The difference in this average and the instantaneous phase current is integrated.
The output of integrator S of each slave phase is then
22
Load slope is programmable in the LTC3882-1 via the
MFR_VOUT_AVP PMBus command. The inductor current measured at the ISENSE pins is converted to a voltage
which is then subtracted from the voltage reference at the
positive input of the error amplifier. The final load slope
is defined by the inductor current sense element and the
bits set in the MFR_VOUT_AVP PMBus command. Setting
MFR_VOUT_AVP to a value greater than 0.0% automatically
disables output servo mode for that channel.
Input Supply Monitoring
The input supply voltage is sensed by the LTC3882-1 at the
VINSNS pin. Undervoltage, overvoltage, valid on and off
levels can be programmed for VIN. Refer to the following
PMBus Command Details section for more information on
programming the input supply thresholds. In addition, the
telemetry ADC monitors the VINSNS voltage relative to
GND. Conversion results are returned by the READ_VIN
PMBus command.
Output Voltage Sensing and Monitoring
Both PWM channels allow remote, differential sensing of the
load voltage with VSENSE pins. The channel 1 output sense
pin VSENSE1– is internally shorted to GND (the exposed
pad). The telemetry ADC is fully differential and makes its
measurements of the output voltages of channels 0 and 1
at VSENSE0± and VSENSE1±, respectively. Conversion results
are returned by the READ_VOUT PMBus command.
Output Current Sensing and Monitoring
Both channels allow differential sensing of the inductor
current using either the inductor DCR or a resistor in series
with the inductor across the ISENSE pins. When the ISENSE
pins for a channel are multiplexed to the differential inputs
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
of the LTC3882-1 monitor ADC, they have an input range
of approximately ±128mV and a noise floor of 7μVRMS.
Peak-peak noise is approximately 46.5μV. The internal ADC
anti-aliasing filter and conversion rate produce an average
reading of the ISENSE differential voltage. The resulting value
is returned by the READ_IOUT PMBus command. Refer to
the Applications Information section for details on sensing
output current using inductor DCR or discrete resistors.
External and Internal Temperature Sense
External temperature can best be measured using a remote,
diode-connected PNP transistor such as the MMBT3906.
The emitter should be connected to a TSNS pin while the
base and collector terminals of the PNP transistor must be
shorted together and returned directly to the LTC3882-1
GND pin. Two different currents are applied to the diode
(nominally 2μA and 32μA) and the temperature is calculated from a ΔVBE measurement made with the internal
16-bit monitor ADC.
The LTC3882-1 also supports direct VBE based external
temperature measurements. In this case the diode or diode network is trimmed to a specific voltage at a specific
current and temperature. In general this method does not
yield as accurate a result as the ΔVBE measurement. Refer
to MFR_PWM_MODE_LTC3882-1 in the PMBus Command
Details section for additional information on programming
the LTC3882-1 for these two external temperature sense
configurations.
The calculated temperature is returned by the PMBus
READ_TEMPERATURE_1 command. Refer to the Applications Information section for details on proper layout
of external temperature sense elements and PMBus
commands that can be used to improve the accuracy of
calculated temperatures.
ASEL1, VOUT0_CFG, VOUT1_CFG, FREQ_CFG, and PHAS_CFG.
If any of these pins are left open the value stored in the
corresponding EEPROM command is used. The resistor
configuration pins are only measured during power-up
and execution of RESTORE_USER_ALL or MFR_RESET
commands. If bit 6 of the MFR_CONFIG_ALL_LTC3882-1
command is set in EEPROM, all resistor inputs except
ASELn are ignored. Per the PMBus specification, all pinprogrammed parameters can be overridden at any time
by commands from the digital interface.
The ASELn pin settings are described in application
Table 11. These pins can be used to select the entire
LTC3882-1 device address. ASEL0 always programs the
bottom four bits of the device address for the LTC3882-1
unless left open. ASEL1 can be used to program the three
most-significant bits. Either portion of the address can also
be retrieved from the MFR_ADDRESS value in EEPROM.
If both pins are left open, the full 7-bit MFR_ADDRESS
value stored in EEPROM is used to determine the device
address. The LTC3882-1 always responds to 7-bit global
addresses 0x5A and 0x5B. MFR_ADDRESS should not be
set to either of these values.
The VOUTn_CFG pin settings are described in application
Table 8. These pins select the output voltages for the
related channel.
The following parameters are also set as a percentage of
the programmed VOUT if resistor configuration pins are
used to determined output voltage:
• VOUT_OV_FAULT_LIMIT: +10%
• VOUT_OV_WARN_LIMIT: +7.5%
• VOUT_MAX: +7.5%
• VOUT_MARGIN_HIGH: +5%
The READ_TEMPERATURE_2 command returns the
internal junction temperature of the LTC3882-1 using an
on-chip diode with a ΔVBE measurement and calculation.
• VOUT_MARGIN_LOW: –5%
Resistor Configuration Pins
The FREQ_CFG pin settings are described in application
Table 9. This pin selects the switching frequency of the
internal oscillator and enables the SYNC output if not left
open, shorted to GND or ignored by EEPROM setting.
Six input pins can be used to configure key operating parameters with selected 1% resistors arranged between VDD25
and GND as a divider to the pin(s). The pins are ASEL0,
• VOUT_UV_WARN_LIMIT: –6.5%
• VOUT_UV_FAULT_LIMIT: –7%
38821f
For more information www.linear.com/LTC3882-1
23
LTC3882-1
Operation
The PHAS_CFG pin settings are described in Table 10.
This pin selects the phase relationships between the two
channels and the selected clock source.
Internal EEPROM with CRC
The LTC3882-1 contains internal EEPROM to store user
configuration settings and fault log information. EEPROM
endurance and retention for user space and fault log
pages are specified in the Absolute Maximum Ratings and
Electrical Characteristics table. The LTC3882-1 EEPROM
also contains a manufacturing section that has internal
redundancy.
The integrity of the entire onboard EEPROM is checked with
a CRC calculation each time its data is to be read, such as
after a power-on reset or execution of a RESTORE_USER_
ALL command. If a CRC error occurs, the CML bit is set
in the STATUS_BYTE and STATUS_WORD commands, the
EEPROM CRC Error bit in the STATUS_MFR_SPECIFIC
command is set, and the ALERT and RUN pins pulled
low (PWM channels off). At that point the device will only
respond at special address 0x7C, which is activated only
after an invalid CRC has been detected. The chip will also
respond at the global addresses 0x5A and 0x5B, but use
of these addresses when attempting to recover from a
CRC issue is not recommended. All power supply rails
associated with either PWM channel of a device reporting
an invalid CRC should remain disabled until the issue is
resolved.
LTC recommends that the EEPROM not be written when
die temperature is greater than 85°C. If internal die temperature exceeds 130°C, all EEPROM operations except
RESTORE_USER_ALL and MFR_RESET are disabled. Full
EEPROM operation is not re-enabled until die temperature
falls below 125°C. Refer to the Applications Information
section for equations to predict retention degradation due
to elevated operating temperatures.
See the Applications Information section or contact the
factory for details on efficient in-system EEPROM programming, including bulk EEPROM programming, which the
LTC3882-1 also supports.
24
Fault Detection
A variety of fault and warning detection, reporting and
handling mechanisms are provided by the LTC3882-1.
Fault or warning detection capabilities include:
• Input Under/Overvoltage
• Output Under/Overvoltage
• Output Overcurrent (Peak and Average)
• Internal and External Overtemperature and External
Undertemperature
• CML Fault (Communication, Memory, or Logic)
• External Fault Detection via Bidirectional FAULT Pins
Reporting is covered in following sections on status commands (registers) and ALERT pin function. Fault handling
mechanisms include hardwired, low-level PWM safety
responses that always occur, and higher-level programmable event management. Both types are covered in the
following sections.
Input Supply Faults
Input undervoltage and overvoltage limits are determined
from multiplexed monitor ADC conversions. Therefore the
input UV/OV response is naturally deglitched by the 100ms
typical conversion cycle of the ADC. There is no hardwired
low-level PWM response for any input supply fault.
Hardwired PWM Response to VOUT Faults
VOUT undervoltage (UV) and overvoltage (OV) faults are
detected by supervisor comparators. The OV and UV fault
limits can be set in three ways:
• As a Percentage of VOUT if Using the Resistor Configuration Pins
• From Stored EEPROM Values
• By PMBus Command
The output overvoltage comparator guards against transient
overshoots as well as long term overvoltages at the output.
When an output OV fault is detected the top MOSFET for
that channel is commanded off and the bottom MOSFET is
commanded on until the overvoltage condition is cleared
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
or for PWM control protocol 0, reverse overcurrent is
detected. See IOUT faults below.
UV faults and warnings are masked if the channel has
been commanded off or until all of the following criteria
are achieved.
• TON_DELAY Has Expired
• TON_RISE Ramp Has Completed
• TON_MAX_FAULT_LIMIT Has Been Reached
• IOUT_OC_FAULT_LIMIT Has Not Been Reached
• TOFF_FALL Is Not in Progress
Output UV warnings are determined from multiplexed
monitor ADC conversions. The LTC3882-1 has no hardwired PWM response for output UV faults or warnings.
Power Good Indication (Master)
An LTC3882-1 master phase indicates Power Good on its
PGOOD pin and in PMBus commands STATUS_WORD
(paged) and MFR_PADS_LTC3882-1 based on programmed UV and OV fault limits. Power Good is indicated
on a master phase as long as it is enabled to run and VOUT
is between the UV and OV fault limits. If a master channel is off for any reason, its PGOOD pin is driven low and
Power Not Good is indicated in the status commands.
Power Good Indication (Slave)
As long as they are enabled, slave phases indicate Power
Good on PGOOD and in PMBus status commands, unless
a master error amplifier (EA) fault is detected. An EA fault
indicates the bussed COMP voltage appears to be too high.
When a slave detects an EA fault, its output is immedidately disabled and OV is indicated (see Figure 2). Any
valid higher-level OV fault response and propagation may
be set for a slave channel to handle a detected EA fault. If
the OV fault response is set to ignore, the slave output is
re-enabled when the EA/COMP condition clears.
A slave indicates Power Not Good with PMBus status commands during an EA fault, but its PGOOD pin remains high
impedance. If a slave phase is off for any other reason, its
PGOOD pin is also driven low.
Hardwired PWM Response to IOUT Faults
The LTC3882-1 measures average IOUT from the voltage
across the ISENSE pins, taking into account the sense resistor
or DCR value and its associated temperature coefficient.
Both are provided by PMBus command or EEPROM values.
An output overcurrent (OC) fault condition is detected by
a supervisor comparator for each PWM output when the
sensed instantaneous current for that channel reaches
its maximum allowed value. Refer to the IOUT_OC_
FAULT_LIMIT PMBus command for details. When an OC
fault is detected the controller immediately disables the
top FET, and the bottom FET is normally commanded on
for the remainder of that PWM cycle.
If programmed to operate in CCM, the LTC3882-1 also
uses the negative of IOUT_OC_FAULT_LIMIT to detect
a reverse overcurrent (ROC) fault. When an ROC fault
occurs the controller immediately disables both top and
bottom FETs, unless PWM output protocol 1 is selected
with MFR_PWM_MODE_LTC3882-1.
OC and ROC faults are both handled according to the
IOUT_OC_FAULT_RESPONSE for that channel. Either
hardware response can result in current-limited operation
using pulse truncation or skipping. Because the LTC3882-1
uses leading edge modulation, this will cause a shift in
average phase toward 0° on the faulted channel and an
increase in input ripple current
Output OC warnings are determined from multiplexed
monitor ADC conversions. The LTC3882-1 has no hardwired
PWM response if an output OC warning occurs.
Hardwired PWM Response to Temperature Faults
An internal temperature sensor measured by the monitor ADC protects against EEPROM and other IC damage.
When die temperature rises above 130°C, the LTC3882-1
will NACK any EEPROM-related command except RESTORE_USER_ALL and MFR_RESET and issue a CML
fault for Invalid/Unsupported Command. Normal EEPROM
access is re-enabled when die temperature drops below
125°C. Above 160°C, the part shuts down all PWM outputs
until die temperature is below 150°C. Internal temperature
fault limits cannot be adjusted. Writing to the EEPROM
above a die temperature of 85°C is strongly discouraged.
38821f
For more information www.linear.com/LTC3882-1
25
LTC3882-1
Operation
Refer to the Absolute Maximum Ratings for other important
temperature limitations on internal EEPROM use.
External temperature sensors may also be monitored by
the onboard ADC. There is no hardwired PWM response
for sensed external temperature faults or warnings.
Hardwired PWM Response to Timing Faults
There is no hardwired PWM response to any timing faults.
TON_MAX_FAULT_LIMIT is the time allowed for VOUT to
rise and settle at start-up. The TON_MAX_FAULT_LIMIT
timer, which has a resolution of 10µs, is started after
TON_DELAY has been reached and a soft-start sequence
is started. If the VOUT_UV_FAULT_LIMIT is not reached or
an OC remains within the specified time, fault response is
determined by the value of TON_MAX_FAULT_RESPONSE.
An internal watchdog detects if SHARE_CLK remains
low for more than 64µs. The part then actively holds
SHARE_CLK low for 120ms, ensuring all devices connected
to this shared control observe a minimum RETRY_DELAY
event.The LTC3882-1 sets the SHARE_CLK_LOW bit in
MFR_COMMON to indicate this fault condition.
The RUN pin must be released by any controlling external
application circuits for that channel to restart from the latch
off state. As the RUN pin for a given channel rises, associated internal fault indications are cleared automatically. The
LTC3882-1 can also be programmed to clear faults for both
outputs based solely on the RUN voltage of just one channel. See the MFR_CONFIG_ALL_LTC3882-1 command. The
CLEAR_FAULTS PMBus command can also be used to clear
all fault bits at any time, independent of PWM channel state.
Handling of some internally generated faults can be digitally
deglitched. See Table 12. External faults propagated into
the chip using FAULTn pins are not deglitched. Refer to
the following section on FAULT functions.
Status Registers and ALERT Masking
External Faults
There are no hardware-level responses to any external
faults propagated into the IC through the FAULTn pins.
Fault Handling
Higher-level input and output fault event handling (response)
can be programmed as described in the following PMBus
Command Details section. For most faults, the LTC3882-1
can manage response in one of three ways: ignore, autonomous recovery (hiccup), or latch off. The device takes no
additional action beyond previously discussed hardwarelevel responses when programmed to ignore a fault.
For autonomous recovery a new soft-start is attempted if
the fault condition is not present after the MFR_RETRY_
DELAY interval has elapsed. MFR_RETRY_DELAY can be
set from 120ms to 83 seconds in 1ms increments. If the
fault persists, the controller will continue to retry with an
interval specified by the MFR_RETRY_DELAY command.
This avoids damage to external regulator components
caused by repetitive, rapid power cycling.
26
No retry is attempted for a latch off fault response. In the
latch off state the gate drivers for the external MOSFETs
are immediately disabled to stop the transfer of energy
to the load as quickly as possible. The output remains
disabled until the channel is commanded off and then
on, or IC supply power is cycled. Commanding a PWM
channel off and on may require software and/or hardware
intervention depending on its programmed configuration.
Figure 2 summarizes the internal LTC3882-1 status registers accessible by PMBus command. These contain
indication of various faults, warnings and other important
operating conditions. As shown, the STATUS_BYTE and
STATUS_WORD commands also summarize contents of
other status registers. Refer to PMBus Command Details
for specific information.
NONE OF THE ABOVE in STATUS_BYTE indicates that
one or more of the bits in the most-significant nibble of
STATUS_WORD are also set.
In general, any asserted bit in a STATUS_x register also
pulls the ALERT pin low. Once set, ALERT will remain low
until one of the following occurs.
• A CLEAR_FAULTS, RESTORE_USER_ALL or MFR_RESET Command Is Issued
• The Related Status Bit Is Written to a One
• The Faulted Channel Is Properly Commanded Off and
Back On
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
STATUS_WORD
STATUS_VOUT*
7
6
5
4
3
2
1
0
VOUT_OV Fault
VOUT_OV Warning
VOUT_UV Warning
VOUT_UV Fault
VOUT_MAX Warning
TON_MAX Fault
TOFF_MAX Warning
(reads 0)
15
14
13
12
11
10
9
8
VOUT
IOUT
INPUT
MFR_SPECIFIC
POWER_GOOD#
(reads 0)
(reads 0)
(reads 0)
7
6
5
4
3
2
1
0
BUSY
OFF
VOUT_OV
IOUT_OC
(reads 0)
TEMPERATURE
CML
NONE OF THE ABOVE
STATUS_BYTE
(PAGED)
STATUS_IOUT
7
6
5
4
3
2
1
0
IOUT_OC Fault
(reads 0)
IOUT_OC Warning
(reads 0)
(reads 0)
(reads 0)
(reads 0)
(reads 0)
MFR_COMMON
7
6
5
4
3
2
1
0
STATUS_TEMPERATURE
OT Fault
OT Warning
(reads 0)
UT Fault
(reads 0)
(reads 0)
(reads 0)
(reads 0)
Chip Not Driving ALERT Low
Chip Not Busy
Internal Calculations Not Pending
Output Not In Transition
EEPROM Initialized
(reads 0)
SHARE_CLK_LOW
(reads 0)
DESCRIPTION
General Fault or Warning Event
Dynamic
Status Derived from Other Bits
MASKABLE GENERATES ALERT BIT CLEARABLE
Yes
No
No
Internal Temperature Fault
Internal Temperature Warning
EEPROM CRC Error
Internal PLL Unlocked
Fault Log Present
(reads 0)
VOUT Short Cycled
FAULT Low
MFR_PADS_LTC3882-1
STATUS_CML
Invalid/Unsupported Command
Invalid/Unsupported Data
Packet Error Check Failed
Memory Fault Detected
Processor Fault Detected
(reads 0)
Other Communication Fault
Other Memory or Logic Fault
7
6
5
4
3
2
1
0
(PAGED)
(PAGED)
7
6
5
4
3
2
1
0
VIN_OV Fault
(reads 0)
VIN_UV Warning
(reads 0)
Unit Off for Insuffcient VIN
(reads 0)
(reads 0)
(reads 0)
STATUS_MFR_SPECIFIC
(PAGED)
(PAGED)
7
6
5
4
3
2
1
0
STATUS_INPUT
7
6
5
4
3
2
1
0
Yes
No
Not Directly
Yes
No
No
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
Channel 1 is Slave
Channel 0 is Slave
(reads 0)
(reads 0)
Invalid ADC Result(s)
SYNC Output Disabled Externally
Channel 1 is POWER_GOOD
Channel 0 is POWER_GOOD
LTC3882-1 Forcing RUN1 Low
LTC3882-1 Forcing RUN0 Low
RUN1 Pin State
RUN0 Pin State
LTC3882-1 Forcing FAULT1 Low
LTC3882-1 Forcing FAULT0 Low
FAULT1 Pin State
FAULT0 Pin State
*IF THE CHANNEL IS CONFIGURED AS A SLAVE AS INDICATED BY MFR_PADS_LTC3882-1[15:14], VOUT_OV FAULT INDICATES A DETECTED MASTER ERROR
AMPLIFIER FAULT (COMP VOLTAGE TOO HIGH). NO OTHER BITS IN STATUS_VOUT ARE ACTIVE ON SLAVE CHANNELS
38821 F02
Figure 2. LTC3882-1 Status Register Summary
• The LTC3882-1 Successfully Transmits Its Address
During a PMBus Alert Response Address (ARA)
• IC Supply Power Is Cycled
With some exceptions, the SMBALERT_MASK command
can be used to prevent the LTC3882-1 from asserting
ALERT for bits in these registers on a bit-by-bit basis.
These mask settings are promoted to STATUS_WORD
and STATUS_BYTE in the same fashion as the status bits
themselves. For example, if ALERT is masked for all bits
in Channel 0 STATUS_VOUT, then ALERT is effectively
masked for the VOUT bit in STATUS_WORD for PAGE 0.
The BUSY bit in STATUS_BYTE also asserts ALERT low
and cannot be masked. This bit can be set as a result of
interaction between internal operation and PMBus com-
For more information www.linear.com/LTC3882-1
38821f
27
LTC3882-1
Operation
munication. This fault occurs when a command is received
that cannot be safely executed with one or both channels
enabled. As discussed in Application Information, BUSY
faults can be avoided by polling MFR_COMMON before
executing some commands.
Status information contained in MFR_COMMON and
MFR_PADS_LTC3882-1 can be used to clarify the contents of STATUS_BYTE or STATUS_WORD as shown,
but the contents of these registers do not affect the state
of the ALERT pin and may not directly influence bits in
STATUS_BYTE or STATUS_WORD.
FAULT Pin I/O
The LTC3882-1 can map various fault indicators to their
respective FAULT pin using the MFR_FAULT_PROPAGATE_LTC3882-1 command.
Channel-to-channel fault dependencies and communication can be created by connecting FAULT pins together. In
the event of an internal fault, one or more of the channels
is configured to pull the bussed FAULT pins low. All channels are then configured to shut down when the bussed
FAULT pins are pulled low (MFR_FAULT_RESPONSE set
to 0xc0). If latch off is the programmed response on the
faulted channel, the FAULT pin remains low until one of
the following occurs:
• A CLEAR_FAULTS, RESTORE_USER_ALL or MFR_RESET Command Is Issued
• The Related Status Bit Is Written to a One
• The Faulted Channel Is Properly Commanded Off and
Back On
• IC Supply Power Is Cycled
For autonomous group retry, the faulted channel is configured to release the FAULT pin(s) after a retry interval,
assuming the original fault has cleared. All the channels
in the group then begin a soft-start sequence.
As noted above, FAULT pins may be configured as inputs
to detect faults external to the controller that require an
immediate response. External faults propagated into the
chip using FAULT pins are not deglitched.
Fault Logging
The LTC3882-1 has fault logging capability. During normal
operation log data is continuously updated in internal RAM.
When a fault occurs that disables either PWM controller,
recording to internal memory is halted, the fault log information is made available from RAM via the MFR_FAULT_LOG
command, and the contents of the RAM log are copied
into EEPROM. EEPROM fault logging is allowed above
a die temperature of 85°C, but 10 years of retention is
not guaranteed. When die temperature exceeds 130°C
EEPROM fault logging is delayed until the temperature
drops below 125°C. Faults generating a log should be fully
cleared before the log is erased to prevent generation of
spurious fault logs. Faults propagated into the IC through
FAULTn pins do not trigger a fault logging event.
When the LTC3882-1 powers up it checks the EEPROM
for a valid fault log. If one is found the Valid Fault Log bit
in the STATUS_MFR_SPECIFIC PMBus command is set.
Additional fault logging will be disabled until the LTC3882-1
receives a CLEAR_FAULTS command. If the Memory Fault
Detected bit is also set in STATUS_CML, then the stored
fault log is partial. Data in one or more event records may
be incomplete or incorrect and MFR_CLEAR_FAULT_LOG
should also be commanded after all faults are cleared in
order to fully enable additional logging functions.
The MFR_FAULT_LOG command uses a block read protocol
with a fixed length of 147 bytes. The LTC3882-1 returns a
block byte count of zero if a fault log is not present.
Contents of a fault log are shown in Table 1 through Table 4.
Refer to Table 6 for an explanation of data formats. Each
event record represents one complete conversion cycle
through all multiplexed monitor ADC inputs and related
status. The six most recent event records are maintained
in internal memory in reverse chronological order. When
a fault log is created the present ADC input cycle is completed and the ADC input being converted at the time of
the fault is noted in the log header record.
Refer to the MFR_FAULT_PROPAGATE command for additional details.
28
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
Table 1. LTC3882-1 Fault Log Contents
STARTING
BYTE
ENDING
BYTE
COMMENTS
Header Information
0
26
See Table 2.
Fault Event Record
27
46
Fault may have occurred anywhere during this event record. See byte 4 of Table 2 and all of
Table 3 and Table 4.
Event Record N-1
47
66
Last complete cyclical data read before the fault was detected.
Event Record N-2
67
86
Older data records…
Event Record N-3
87
106
Event Record N-4
107
126
Event Record N-5
127
146
RECORD TYPE
Oldest recorded data.
Table 2. Fault Log Header Information
RECORD
BITS
FORMAT
Fault Log Preface
[7:0]
ASC
[7:0]
[15:8]
MFR_REAL_TIME
MFR_VOUT_PEAK (PAGE 0)
Reg
Returns LTxx beginning at byte 0 if a partial or complete fault log exists. Word xx is
a factory identifier that may vary part to part.
2
3
Reg
4
Refer to Table 3.
[7:0]
Reg
5
48 bit share-clock counter value when fault occurred (200µs resolution).
[15:8]
6
[23:16]
7
[31:24]
8
[39:32]
9
[47:40]
10
[15:8]
L16
[15:8]
MFR_IOUT_PEAK (PAGE 0)
[15:8]
MFR_IOUT_PEAK (PAGE 1)
[15:8]
MFR_VIN_PEAK
[15:8]
L16
L11
[15:8]
L11
Peak READ_IOUT on Channel 0 since last power-on or CLEAR_PEAKS command.
17
Peak READ_IOUT on Channel 1 since last power-on or CLEAR_PEAKS command.
19
Peak READ_VIN since last power-on or CLEAR_PEAKS command.
21
External temperature sensor 0 during last event.
22
L11
[7:0]
[7:0]
15
20
[7:0]
READ_TEMPERATURE2
Peak READ_VOUT on Channel 1 since last power-on or CLEAR_PEAKS command.
18
L11
[7:0]
[15:8]
13
16
L11
[7:0]
READ_TEMPERATURE1 (PAGE 1)
Peak READ_VOUT on Channel 0 since last power-on or CLEAR_PEAKS command.
14
[7:0]
[15:8]
11
12
[7:0]
READ_TEMPERATURE1 (PAGE 0)
DETAILS
[7:0]
[7:0]
MFR_VOUT_PEAK (PAGE 1)
0
1
[7:0]
Fault Source
BLOCK
BYTE
COUNT
23
External temperature sensor 1 during last event.
24
L11
25
Internal temperature sensor during last event.
26
38821f
For more information www.linear.com/LTC3882-1
29
LTC3882-1
Operation
Table 3. Fault Source Values
FAULT SOURCE VALUE
CAUSE OF FAULT LOG
0x00
TON_MAX
0x01
VOUT_OV
0x02
VOUT_UV
0x03
IOUT_OC
0x05
Over temperature
0x06
Under temperature
CHANNEL
0
0x07
VIN_OV
0x0A
Internal temperature
0x10
TON_MAX
0x11
VOUT_OV
0x12
VOUT_UV
0x13
IOUT_OC
0x15
Over temperature
0x16
Under temperature
0x17
VIN_OV
0x1A
Internal temperature
0xFF
MFR_FAULT_LOG_STORE
1
Table 4. Fault Log Event Record
DATA
BITS
FORMAT
RECORD BYTE INDEX
READ_VOUT (PAGE 0)
[15:8]
L16
0
[7:0]
READ_VOUT (PAGE 1)
[15:8]
READ_IOUT (PAGE 0)
[15:8]
READ_IOUT (PAGE 1)
[15:8]
1
L16
2
L11
4
L11
6
[7:0]
3
[7:0]
5
[7:0]
READ_VIN
[15:8]
7
L11
[7:0]
(Not used)
[15:8]
STATUS_VOUT (PAGE 0)
[7:0]
8
9
L11
10
Reg
12
[7:0]
11
STATUS_VOUT (PAGE 1)
[7:0]
Reg
13
STATUS_WORD (PAGE 0)
[15:8]
Reg
14
[7:0]
STATUS_WORD (PAGE 1)
[15:8]
15
Reg
[7:0]
16
17
STATUS_MFR_SPECIFIC (PAGE 0)
[7:0]
Reg
18
STATUS_MFR_SPECIFIC (PAGE 1)
[7:0]
Reg
19
30
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
Factory Default Operation
The LTC3882-1 ships from the factory with a default configuration stored in its non-volatile memory, unless custom
programming has been requested. These command values
are loaded into volatile RAM when the chip is initialized. Prior
to receiving any PMBus commands, a stock LTC3882-1 will
operate in the factory default mode. If a STORE_USER_
ALL command is executed, the contents of the non-volatile
memory are replaced with active command values from
internal RAM, and that will permanently overwrite the factory
defaults. Table 5 summarizes the default factory operation
settings of the LTC3882-1 if all resistor configuration pins
are left open. These defaults allow parameters listed in bold
text in the table to be overridden with configuration resistor
programming. Warning limits are given in Table 5 because
exceeding them will cause the ALERT pin to be asserted even
if the PMBus interface is not being utilized.
Table 5. Factory Default Operation Summary
PARAMETER*
DEFAULT SETTING
UNITS
PMBus Address
All writes enabled to Channel 0 at address 0x4F (no PEC).
–
Operation
OPERATION enabled with RUN pin control and soft-off.
–
Input Voltage OFF Threshold
6.0
V
Input Voltage UV Warning Limit
6.3
V
Input Voltage ON Threshold
6.5
V
Input Voltage OV Fault Limit
15.5
V
Input Voltage OV Fault Response
Latch off.
Soft-Start Time
8 (with no delay).
ms
–
Maximum Start-Up Time (TMAX)
10
ms
TMAX Fault Response
Retry every 350ms.
–
Output Voltage UV Fault/Warning Limits
0.900/0.925
V
Output Voltage UV Fault Response
Retry every 350ms.
–
Output Voltage
1.000
V
Active Voltage Positioning
Disabled.
–
Output Voltage OV Warning/Fault Limits
1.075/1.100
V
Output Voltage OV Fault Response
Retry every 350ms.
–
Shut Down
8ms soft-off.
–
Output Current Sense Element
0.63mΩ with 3930ppm/°C TC.
–
Output Current OC Warning/Fault Limits
20/29.75
A
Output Current OC Fault Response
Ignore
–
PWM Switching Mode
Continuous inductor current only.
–
PWM Control Protocol
Three-State PWM.
–
PWM Switching Frequency
500
Channel 0/1 Phase
0/180
kHz
Degrees
Internal Overtemperature Warning/Fault Limits
130/160
°C
Internal Overtemperature Responses
Warning: EEPROM disabled; Fault: PWM disabled.
–
External Undertemperature Fault Limit
–40
°C
External Undertemperature Fault Response
Retry every 350ms.
–
External Overtemperature Warning/Fault Limits
85/100
°C
External Overtemperature Fault Response
Retry every 350ms.
–
FAULT
Asserts low for the following faults: VOUT UV or OV, VIN OV, external or internal OT,
external UT, TON_MAX, or output short cycle.
–
ALERT Masking
ALERTs are masked for loss of PLL lock and external FAULT inputs.
–
*bold entries can be changed with external configuration resistors
38821f
For more information www.linear.com/LTC3882-1
31
LTC3882-1
Operation
Serial Interface
• Read Byte
The LTC3882-1 has a PMBus compliant serial interface that
can operate at any frequency between 10kHz and 400kHz.
The LTC3882-1 is a bus slave device that communicates
bidirectionally with a host (master) using standard PMBus
protocols. The Timing Diagram found earlier in this document, along with related Electrical Characteristics table
entries, define the timing relationships of the SDA and SCL
bus signals. SDA and SCL must be high when the bus is
not in use. External pull-up resistors or current sources
are required on these lines.
• Read Word
PMBus, an incremental extension of the SMBus standard,
offers more robust operation than a 2-wire I2C interface. In
addition to adding a protocol layer to improve interoperability and facilitate reuse, PMBus supports bus timeout
recovery for system reliability, optional packet error checking to ensure data integrity, and peripheral hardware alerts
for system fault management. In general, a programmable
device capable of functioning as an I2C bus master can be
configured for PMBus management with little or no change
to hardware. However, not all I2C controllers support repeat
start (restart) required for PMBus reads.
For a description of the minor extensions and exceptions
PMBus makes to the SMBus standard, refer to PMBus
Specification Part I Revision 1.2 Paragraph 5 on Transport.
For a description of the differences between SMBus and
I2C, refer to System Management Bus (SMBus) Specification Version 2.0 Appendix B on Differences Between
SMBus and I2C.
The user is encouraged to reference Part I of the latest
PMBus Power System Management Protocol Specification to understand how to interface the LTC3882-1 to a
PMBus system. This specification can be found at http://
www.pmbus.org/specs.html.
• Block Read
• Block Write – Block Read Process Call
• Alert Response Address
The LTC3882-1 does not require PEC for Quick Command
under any circumstances. The LTC3882-1 also supports
group command protocol (GCP) as required by PMBus
specification Part I, section 5.2.3. GCP is used to send commands to more than one PMBus device in one continuous
transmission. It should not be used with commands that
require the receiving device to respond with data, such as
a STATUS_BYTE command. Refer to Part I of the PMBus
specification for additional details on using GCP.
All LTC3882-1 message transmission types allow for packet
error checking. The later section on Serial Communication
Errors provides more detail on packet error checking.
Figure 4 to Figure 20 illustrate these protocols. Figure 3
provides a key to the protocol diagrams. Not all protocol
elements will be present in every data packet. For instance,
not all packets are required to include the packet error
code. A number shown above a field in these diagrams
indicates the number of bits in that field. All data transfers
are initiated by the present bus master regardless of how
many times data direction flow may change during the
subsequent transmission. The LTC3882-1 never functions
as a bus master.
This device includes handshaking features to ensure robust system communication. Please refer to the PMBus
Communication and Command Processing section in
Applications Information for more details.
Serial Bus Addressing
The LTC3882-1 uses the following standard serial interface
protocols defined in the SMBus and PMBus specifications:
The LTC3882-1 supports four types of serial bus addressing:
• Quick Command
• Global Bus Addressing
• Send Byte
• Power Rail Addressing
• Write Byte
• Individual Device Addressing
• Write Word
• Page+ Channel Addressing
32
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
Global addressing provides a means for the bus master
to communicate with all LTC3882-1 devices on the bus
simultaneously. The LTC3882-1 global addresses of 0x5A
and 0x5B cannot be changed or disabled. Commands sent
to address 0x5A are applied to both channels, as if the
PAGE command were set to 0xFF. Global address 0x5B is
paged, allowing channel-specific control of all LTC3882-1
devices on the bus. Other LTC device types may respond
at one or both of these global addresses. Reading from
global addresses is strongly discouraged.
PolyPhase rail. Different voltage rails should not attempt
to share a rail address. Reading from rail addresses is
also strongly discouraged.
Device addressing is the most common means used by a
bus master to communicate with an LTC3882-1. The value
of the device address is set by the combination of ASEL
pin programming and the MFR_ADDRESS command.
Refer to the previous section on Resistor Configuration
Pins for details.
Individual channel addressing allows the bus master to
communicate directly with a specific LTC3882-1 PWM
channel without first using a PAGE command. Refer to
the PAGE_PLUS commands for additional details.
Rail addressing provides a means for the bus master to
simultaneously communicate with all channels connected
together to produce a single output voltage (PolyPhase).
While similar to global addressing, the rail address can
be dynamically assigned with the paged MFR_RAIL_ADDRESS command, allowing for any logical grouping of
channels that might be required for reliable system control.
Rail addresses should be unique for each single-phase or
Use of any of the four types of addressing requires careful
planning to avoid address-related bus conflicts. Communication to LTC3882-1 devices at global and rail addresses
should be limited to command write operations.
S
START CONDITION
Sr
REPEATED START CONDITION
Rd
READ (BIT VALUE OF 1)
Wr
WRITE (BIT VALUE OF 0)
A
NA
ACKNOWLEDGE (BIT SHOULD BE 0), OR
NOT ACKNOWLEDGE (BIT SHOULD BE 1)
P
STOP CONDITION
PEC PACKET ERROR CODE
MASTER TO SLAVE
SLAVE TO MASTER
...
CONTINUATION OF PROTOCOL
38821 F03
Figure 3. PMBus Packet Protocol Diagram Element Key
1
7
S
1
1
SLAVE ADDRESS Rd/Wr A
1
P
38821 F04
Figure 4. Quick Command Protocol
1
S
1
1
SLAVE ADDRESS Wr A COMMAND CODE A
7
1
1
8
P
38821 F05
Figure 5. Send Byte Protocol
1
S
7
1
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A
8
1
1
PEC
A
P
38821 F06
Figure 6. Send Byte Protocol with PEC
38821f
For more information www.linear.com/LTC3882-1
33
LTC3882-1
Operation
1
S
7
1
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A
8
1
1
DATA BYTE
A
P
38821 F07
Figure 7. Write Byte Protocol
1
S
7
1
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A
8
1
8
1
1
DATA BYTE
A
PEC
A
P
38821 F08
Figure 8. Write Byte Protocol with PEC
1
S
7
1
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A
8
1
8
1
1
DATA BYTE LOW
A
DATA BYTE HIGH
A
P
38821 F09
Figure 9. Write Word Protocol
1
S
7
1
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A
8
1
8
1
8
1
1
DATA BYTE LOW
A
DATA BYTE HIGH
A
PEC
A
P
38821 F10
Figure 10. Write Word Protocol with PEC
1
S
7
1
1
8
1
1
7
1
1
8
SLAVE ADDRESS Wr A COMMAND CODE A Sr SLAVE ADDRESS Rd A
1
DATA BYTE
1
NA P
38821 F11
Figure 11. Read Byte Protocol
1
S
7
1
1
8
1
1
7
1
1
SLAVE ADDRESS Wr A COMMAND CODE A Sr SLAVE ADDRESS Rd A
8
1
8
1
1
DATA BYTE
A
PEC
A
P
38821 F12
Figure 12. Read Byte Protocol with PEC
1
S
7
1
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A
1
7
1
1
Sr SLAVE ADDRESS Rd A
8
1
DATA BYTE LOW
A
8
1
1
DATA BYTE HIGH NA P
38821 F13
Figure 13. Read Word Protocol
1
S
7
1
1
8
1
1
7
1
1
SLAVE ADDRESS Wr A COMMAND CODE A Sr SLAVE ADDRESS Rd A
8
1
DATA BYTE LOW
A
8
1
DATA BYTE HIGH A
8
1
1
PEC
A
P
38821 F14
Figure 14. Read Word Protocol with PEC
1
S
7
1
1
8
1
1
7
1
1
SLAVE ADDRESS Wr A COMMAND CODE A Sr SLAVE ADDRESS Rd A
8
1
8
DATA BYTE 1
A
DATA BYTE 2
1
…
A …
8
DATA BYTE N
8
1
BYTE COUNT = N A
1
…
1
NA P
38821 F15
Figure 15. Block Read Protocol
34
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Operation
1
S
7
1
1
8
1
1
7
1
1
8
SLAVE ADDRESS Wr A COMMAND CODE A Sr SLAVE ADDRESS Rd A
8
1
8
DATA BYTE 1
A
DATA BYTE 2
1
…
A …
1
BYTE COUNT = N A
8
1
8
DATA BYTE N
A
PEC
1
…
1
NA P
38821 F16
Figure 16. Block Read Protocol with PEC
1
S
7
1
1
8
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A BYTE COUNT = M A
8
1
DATA BYTE 2
1
7
1
8
A …
1
Sr SLAVE ADDRESS Rd A
1
8
A …
…
1
BYTE COUNT = N A
8
DATA BYTE 2
1
A
A …
DATA BYTE M
8
8
DATA BYTE 1
8
1
DATA BYTE 1
A
1
DATA BYTE N
…
1
NA P
38821 F17
Figure 17. Block Write – Block Read Process Call
1
S
7
1
1
8
1
8
1
SLAVE ADDRESS Wr A COMMAND CODE A BYTE COUNT = M A
8
DATA BYTE 2
1
7
1
1
Sr SLAVE ADDRESS Rd A
8
1
DATA BYTE 2
1
…
A …
A …
8
1
A
…
1
DATA BYTE M
8
8
DATA BYTE 1
A …
1
BYTE COUNT = N A
8
DATA BYTE 1
8
1
8
DATA BYTE N
A
PEC
A
1
…
1
NA P
38821 F18
Figure 18. Block Write – Block Read Process Call with PEC
1
7
1
1
8
1
1
S ALERT RESPONSE Rd A DEVICE ADDRESS NA P
ADDRESS
38821 F19
Figure 19. Alert Response Address Protocol
1
7
1
1
8
1
S ALERT RESPONSE Rd A DEVICE ADDRESS A
ADDRESS
8
PEC
1
1
NA P
38821 F20
Figure 20. Alert Response Address Protocol with PEC
38821f
For more information www.linear.com/LTC3882-1
35
LTC3882-1
Operation
Serial Bus Timeout
Serial Communication Errors
The LTC3882-1 implements a timeout feature to avoid
hanging the serial interface. The data packet timer begins running at the first START event before the SLAVE
ADDRESS write byte and ends with the STOP bit. Packet
transmission must be completed before the timer expires,
or the LTC3882-1 will tri-state the bus and ignore all message data. The data packet includes the SLAVE ADDRESS
byte, COMMAND CODE byte, repeated START and SLAVE
ADDRESS byte (if a read operation), all ACKNOWLEDGE
and flow control bits (R/W) and all data bytes.
The LTC3882-1 supports the optional PMBus packet error
checking protocol. This protocol appends a packet error
code (PEC) to the end of applicable message transfers to
improve communication reliability. The PEC is a CRC-8
error-checking byte calculated by the bus device sending
the last data byte. Refer to SMBus specification 1.2 or
higher for additional implementation details. All LTC3882-1
read operations will return a valid PEC if the bus master
requests it. If bit 2 in the MFR_CONFIG_ALL_LTC3882-1
command is set, the IC will not act in response to a bus
write operation unless a valid PEC is also received from
the host.
The packet timer is typically set to 30ms. If bit 3 of MFR_
CONFIG_ALL_LTC3882-1 is set, this period is extended
to 255ms. The LTC3882-1 automatically allows a packet
transmission time of 255ms for MFR_FAULT_LOG block
reads regardless of the setting of this bit. In no circumstances will the timeout period be less than the tTIMEOUT
specification (25ms minimum).
The LTC3882-1 supports the full PMBus frequency range
of 10kHz to 400kHz.
36
PEC errors on command writes, attempts to access unsupported commands, or writing invalid data to supported
commands all cause the LTC3882-1 to generate a CML
fault. The CML bit is then set in the STATUS_BYTE and
STATUS_WORD commands, and the appropriate bit is set
in the STATUS_CML command.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus Command Summary
PMBus Commands
Table 7 lists supported PMBus commands and manufacturer specific commands. Additional information about
these commands can be found in Revision 1.2 of Part II of
the PMBus Power System Management Protocol Specification that can be found at http://www.pmbus.org/specs.html.
Users are encouraged to reference that manual. Exceptions
or manufacturer-specific implementations are detailed in
the tables below. All standard PMBus commands from
0x00 through 0xCF not listed in this table are implicitly
not supported by the LTC3882-1. All commands from
0xD0 through 0xFF not listed in this table are implicitly
reserved by the manufacturer. The LTC3882-1 may execute
additional commands not listed in this table, and these
can change without notice. Reading these unlisted commands is harmless to the operation of the IC. Writes to any
unsupported or reserved command should be avoided, as
they may result in a CML fault and/or undesired operation
of the part.
If PMBus commands are received faster than they are being processed, the part may become too busy to handle
new commands. In these cases the LTC3882-1 follows the
protocols defined in the PMBus Specification V1.2, Part
II, Section 10.8.7, to communicate that it is busy. This
device includes handshaking features to eliminate busy
responses, simplify error handling software and ensure
robust communication and system behavior. Please refer
to PMBus Communication and Command Processing in
the Applications Information section for further details.
LTC has made an effort to establish PMBus command
compatibility and functional uniformity among its family
of parts. However, differences may occur due to specific
product requirements. Compatibility of PMBus commands
among any ICs should not be assumed based simply on
command name. Always refer to the manufacturer’s data
sheet of each device for a complete definition of a command function.
Data Formats
PMBus supports specific floating point number formats
and allows for a wide range of other data formats.
Table 6 describes the data formats used by the LTC3882-1.
Abbreviations of these formats appear throughout this
document.
Table 6. Abbreviations of Supported Data Formats
PMBus
TERMINOLOGY
SPECIFICATION
LTC
REFERENCE TERMINOLOGY DEFINITION
L11
Linear
Part II ¶7.1
Linear_5s_11s
L16
Linear VOUT_MODE
Part II ¶8.2
Linear_16u
CF
DIRECT
Part II ¶7.2
varies
Reg
register bits
Part II ¶10.3
Reg
ASC
text characters
Part II ¶22.2.1
ASCII
Floating point 16-bit data: value = Y • 2N,
where N = b[15:11] and Y = b[10:0], both
two’s compliment binary integers.
EXAMPLE
b[15:0] = 0x9807 = 10011_000_0000_0111
value = 7 • 2–13 = 854E-6
Floating point 16-bit data: value = Y • 2–12, b[15:0] = 0x4C00 = 0100_1100_0000_0000
where Y = b[15:0], an unsigned integer.
value = 19456 • 2–12 = 4.75
16-bit data with a custom format
defined in the detailed PMBus command
description.
Often an unsigned or two’s compliment
integer.
Per-bit meaning defined in detailed PMBus PMBus STATUS_BYTE command.
command description.
ISO/IEC 8859-1 [A05]
LTC (0x4C5443)
38821f
For more information www.linear.com/LTC3882-1
37
LTC3882-1
PMBus Command Summary
Table 7. PMBus Command Summary
COMMAND NAME
CMD
CODE DESCRIPTION
PAGE
0x00
Channel (page) presently selected for
any paged command.
R/W Byte
N
Reg
OPERATION
0x01
On, off and margin control.
R/W Byte
Y
Reg
ON_OFF_CONFIG
0x02
RUN pin and PMBus on/off command
configuration.
R/W Byte
Y
Reg
CLEAR_FAULTS
0x03
Clear all set fault bits.
Send Byte
N
92
PAGE_PLUS_WRITE
0x05
Write a command directly to a specified
page.
W Block
N
70
PAGE_PLUS_READ
0x06
Read a command directly from a
specified page.
Block R/W
Process
N
71
WRITE_PROTECT
0x10
Protect the device against unintended
PMBus modifications.
R/W Byte
N
STORE_USER_ALL
0x15
Store entire operating memory in
EEPROM.
Send Byte
N
103
RESTORE_USER_ALL
0x16
Restore entire operating memory from
EEPROM.
Send Byte
N
103
CAPABILITY
0x19
Summary of supported optional PMBus
features.
R Byte
N
Reg
SMBALERT_MASK
0x1B Mask ALERT activity
Block R/W
Y
Reg
VOUT_MODE
0x20
Voltage-related format (Linear) and
exponent.
R Byte
Y
Reg
VOUT_COMMAND
0x21
Nominal VOUT value.
R/W Word
Y
L16
V
VOUT_MAX
0x24
Maximum VOUT that can be set by any
command, including margin.
R/W Word
Y
L16
VOUT_MARGIN_HIGH
0x25
VOUT at high margin, must be greater
than VOUT_COMMAND.
R/W Word
Y
VOUT_MARGIN_LOW
0x26
VOUT at low margin, must be less than
VOUT_COMMAND.
R/W Word
VOUT_TRANSITION_RATE
0x27
VOUT slew rate for programmed output
changes.
FREQUENCY_SWITCH
0x33
VIN_ON
TYPE
DATA
PAGED FORMAT
DEFAULT
VALUE
SEE
PAGE
0x00
70
l
0x80
74
l
0x1E
73
UNITS NVM
Reg
l
0x00
71
0xB0
72
see CMD
details
100
0x14
2–12
80
l
1.0V
0x1000
80
V
l
5.5V
0x5800
81
L16
V
l
1.05V
0x10CD
81
Y
L16
V
l
0.95V
0x0F33
81
R/W Word
Y
L11
V/ms
l
0.25
0xAA00
85
PWM frequency control.
R/W Word
N
L11
kHz
l
500kHz
0xFBE8
75
0x35
Minimum input voltage to begin power
conversion.
R/W Word
N
L11
V
l
6.5V
0xCB40
79
VIN_OFF
0x36
Decreasing input voltage at which power R/W Word
conversion stops.
N
L11
V
l
6.0V
0xCB00
79
IOUT_CAL_GAIN
0x38
Ratio of ISENSE± voltage to sensed
current.
R/W Word
Y
L11
mΩ
l
0.63mΩ
0xB285
83
VOUT_OV_FAULT_LIMIT
0x40
VOUT overvoltage fault limit.
R/W Word
Y
L16
V
l
1.1V
0x119A
81
VOUT_OV_FAULT_RESPONSE
0x41
VOUT overvoltage fault response.
R/W Byte
Y
Reg
VOUT_OV_WARN_LIMIT
0x42
VOUT overvoltage warning limit.
R/W Word
Y
L16
38
l
V
l
0xB8
96
l
1.075V
0x1133
82
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus Command Summary
Table 7. PMBus Command Summary
COMMAND NAME
CMD
CODE DESCRIPTION
DEFAULT
VALUE
SEE
PAGE
VOUT_UV_WARN_LIMIT
0x43
VOUT undervoltage warning limit.
R/W Word
Y
L16
V
l
0.925V
0x0ECD
82
VOUT_UV_FAULT_LIMIT
0x44
VOUT undervoltage fault limit.
R/W Word
Y
L16
V
l
0.9V
0x0E66
82
VOUT_UV_FAULT_RESPONSE
0x45
VOUT undervoltage fault response.
R/W Byte
Y
Reg
l
0xB8
96
IOUT_OC_FAULT_LIMIT
0x46
Output overcurrent fault limit.
R/W Word
Y
L11
l
29.75A
0xDBB8
83
IOUT_OC_FAULT_RESPONSE
0x47
Output overcurrent fault response.
R/W Byte
Y
Reg
l
0x00
97
IOUT_OC_WARN_LIMIT
0x4A Output overcurrent warning limit.
R/W Word
Y
L11
A
l
20.0A
0xDA80
83
OT_FAULT_LIMIT
0x4F
External overtemperature fault limit.
R/W Word
Y
L11
°C
l
100.0°C
0xEB20
86
OT_FAULT_RESPONSE
0x50
External overtemperature fault response.
R/W Byte
Y
Reg
l
0xB8
98
°C
l
OT_WARN_LIMIT
0x51
External overtemperature warning limit.
R/W Word
Y
L11
85.0°C
0xEAA8
86
UT_FAULT_LIMIT
0x53
External undertemperature fault limit.
R/W Word
Y
L11
°C
l
–40.0°C
0xE580
87
UT_FAULT_RESPONSE
0x54
External undertemperature fault
response.
R/W Byte
Y
Reg
l
0xB8
98
VIN_OV_FAULT_LIMIT
0x55
VIN overvoltage fault limit.
R/W Word
N
L11
l
15.5V
0xD3E0
79
VIN_OV_FAULT_RESPONSE
0x56
VIN overvoltage fault response.
R/W Byte
Y
Reg
l
0x80
95
VIN_UV_WARN_LIMIT
0x58
VIN undervoltage warning limit.
R/W Word
N
L11
V
l
6.3V
0xCB26
79
TON_DELAY
0x60
Delay from RUN pin or OPERATION ON
command to TON_RISE ramp start.
R/W Word
Y
L11
ms
l
0.0ms
0x8000
84
TON_RISE
0x61
Time for VOUT to rise from 0.0V to
VOUT_COMMAND after TON_DELAY.
R/W Word
Y
L11
ms
l
8.0ms
0xD200
84
TON_MAX_FAULT_LIMIT
0x62
Maximum time for VOUT to rise above
VOUT_UV_FAULT_LIMIT after
TON_DELAY.
R/W Word
Y
L11
ms
l
10.0ms
0xD280
85
TON_MAX_FAULT_RESPONSE
0x63
Fault response when TON_MAX_FAULT_
LIMIT is exceeded.
R/W Byte
Y
Reg
l
0xB8
99
TOFF_DELAY
0x64
Delay from RUN pin or OPERATION OFF
command to TOFF_FALL ramp start.
R/W Word
Y
L11
ms
l
0.0ms
0x8000
85
TOFF_FALL
0x65
Time for VOUT to fall to 0.0V from
VOUT_COMMAND after TOFF_DELAY.
R/W Word
Y
L11
ms
l
8.0ms
0xD200
85
TOFF_MAX_WARN_ LIMIT
0x66
Maximum time for VOUT to decay below
12.5% of VOUT_COMMAND after
TOFF_FALL completes.
R/W Word
Y
L11
ms
l
150ms
0xF258
85
STATUS_BYTE
0x78
One-byte channel status summary.
R/W Byte
Y
Reg
87
STATUS_WORD
0x79
Two-byte channel status summary.
R/W Word
Y
Reg
88
STATUS_VOUT
0x7A VOUT fault and warning status.
R/W Byte
Y
Reg
88
STATUS_IOUT
0x7B IOUT fault and warning status.
R/W Byte
Y
Reg
89
STATUS_INPUT
0x7C Input supply fault and warning status.
R/W Byte
N
Reg
89
TYPE
DATA
PAGED FORMAT
UNITS NVM
A
V
38821f
For more information www.linear.com/LTC3882-1
39
LTC3882-1
PMBus Command Summary
Table 7. PMBus Command Summary
COMMAND NAME
CMD
CODE DESCRIPTION
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
SEE
PAGE
STATUS_TEMPERATURE
0x7D External temperature fault and warning
status.
R/W Byte
Y
Reg
89
STATUS_CML
0x7E
Communication, memory and logic fault
and warning status.
R/W Byte
N
Reg
90
STATUS_MFR_ SPECIFIC
0x80
LTC3882-1-specific status.
R/W Byte
Y
Reg
90
READ_VIN
0x88
Measured VIN.
R Word
N
L11
READ_VOUT
0x8B Measured VOUT.
R Word
Y
READ_IOUT
R Word
Y
READ_TEMPERATURE_1
0x8C Measured IOUT.
0x8D Measured external temperature.
R Word
Y
READ_TEMPERATURE_2
0x8E
Measured internal temperature.
R Word
N
READ_DUTY_CYCLE
0x94
Measured commanded PWM duty cycle.
R Word
READ_FREQUENCY
0x95
Measured PWM input clock frequency.
READ_POUT
0x96
Calculated output power.
PMBUS_REVISION
0x98
MFR_ID
0x99
MFR_MODEL
MFR_SERIAL
V
92
L16
V
93
L11
A
93
L11
°C
94
L11
°C
94
Y
L11
%
94
R Word
Y
L11
kHz
94
R Word
Y
L11
W
93
Supported PMBus version.
R Byte
N
Reg
0x22
V1.2
72
Manufacturer identification.
R String
N
ASC
LTC
104
0x9A LTC model number.
R String
N
ASC
LTC3882-1
104
0x9E
R Block
N
ASC
R Word
Y
L16
Device serial number.
104
LTC3882-1 Custom Commands
MFR_VOUT_MAX
0xA5 Maximum value of any VOUT related
command.
V
5.6V
0x599A
80
USER_DATA_00
0xB0 EEPROM word reserved for LTpowerPlay. R/W Word
N
Reg
l
103
USER_DATA_01
0xB1 EEPROM word reserved for LTpowerPlay. R/W Word
Y
Reg
l
103
USER_DATA_02
0xB2 EEPROM word reserved for OEM use.
R/W Word
N
Reg
l
USER_DATA_03
0xB3 EEPROM word available for general data
storage.
R/W Word
Y
Reg
l
0x0000
103
USER_DATA_04
0xB4 EEPROM word available for general data
storage.
R/W Word
N
Reg
l
0x0000
103
103
MFR_EE_UNLOCK
0xBD (contact the factory)
103
MFR_EE_ERASE
0xBE (contact the factory)
103
MFR_EE_DATA
0xBF
MFR_CHAN_CONFIG_
LTC3882-1
0xD0 LTC3882-1 channel-specific
configuration.
(contact the factory)
103
R/W Byte
Y
Reg
l
0x1D
77
R/W Byte
N
Reg
l
0x01
73
MFR_FAULT_PROPAGATE_
LTC3882-1
0xD2 Configure LTC3882-1 status propagation R/W Word
via FAULTn pins.
Y
Reg
l
0x6993
101
MFR_VOUT_AVP
0xD3 Specify VOUT load line.
R/W Word
Y
L11
l
0%
0x8000
81
MFR_PWM_MODE_LTC3882-1
0xD4 LTC3882-1 channel-specific PWM mode
control.
R/W Byte
Y
Reg
l
0xC8
78
MFR_FAULT_RESPONSE
0xD5 PWM response when FAULTn pin is low.
R/W Byte
Y
Reg
l
0xC0
101
MFR_CONFIG_ALL_LTC3882-1 0xD1 LTC3882-1 device-level configuration.
40
%
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus Command Summary
Table 7. PMBus Command Summary
COMMAND NAME
CMD
CODE DESCRIPTION
TYPE
DATA
PAGED FORMAT
MFR_OT_FAULT_RESPONSE
0xD6 Internal overtemperature fault response.
R Byte
N
Reg
MFR_IOUT_PEAK
0xD7 Maximum IOUT measurement since last
MFR_CLEAR_PEAKS.
R Word
Y
L11
A
MFR_RETRY_DELAY
0xDB Minimum time before retry after a fault.
R/W Word
Y
L11
ms
l
350ms
0xFABC
99
MFR_RESTART_DELAY
0xDC Minimum time RUN pin is held low by
the LTC3882-1.
R/W Word
Y
L11
ms
l
500ms
0xFBE8
84
MFR_VOUT_PEAK
0xDD Maximum VOUT measurement since last
MFR_CLEAR_PEAKS.
R Word
Y
L16
V
93
MFR_VIN_PEAK
0xDE Maximum VIN measurement since last
MFR_CLEAR_PEAKS.
R Word
N
L11
V
93
MFR_TEMPERATURE_1_PEAK
0xDF Maximum external temperature
measurement since last MFR_CLEAR_
PEAKS.
R Word
Y
L11
°C
94
MFR_CLEAR_PEAKS
0xE3
Clear all peak values.
Send Byte
N
MFR_PADS_LTC3882-1
0xE5
State of selected LTC3882-1 pads.
R Word
N
Reg
MFR_ADDRESS
0xE6
Specify right-justified 7-bit device
address.
R/W Byte
N
Reg
MFR_FAULT_LOG_STORE
0xEA Force transfer of fault log from operating Send Byte
memory to EEPROM.
N
103
MFR_FAULT_LOG_CLEAR
0xEC Clear existing EEPROM fault log.
Send Byte
N
102
MFR_FAULT_LOG
0xEE
Read fault log data.
R Block
N
Reg
MFR_COMMON
0xEF
LTC-generic device status reporting.
Reg
MFR_COMPARE_USER_ALL
0xF0
Compare operating memory with
EEPROM contents.
MFR_TEMPERATURE_2_PEAK
0xF4
Maximum internal temperature
measurement since last
MFR_CLEAR_PEAKS.
MFR_PWM_CONFIG_LTC3882-1
0xF5
MFR_IOUT_CAL_GAIN_TC
UNITS NVM
DEFAULT
VALUE
SEE
PAGE
0xC0
98
93
94
R Byte
N
Send Byte
N
R Word
N
L11
LTC3882-1 PWM configuration common
to both channels.
R/W Byte
N
Reg
0xF6
Output current sense element
temperature coefficient.
R/W Word
Y
CF
MFR_TEMP_1_GAIN
0xF8
Slope for external temperature
calculations.
R/W Word
Y
CF
MFR_TEMP_1_OFFSET
0xF9
Offset addend for external temperature
calculations.
R/W Word
Y
L11
MFR_RAIL_ADDRESS
0xFA
Specify unique right-justified 7-bit
address for channels comprising a
PolyPhase output.
R/W Byte
Y
Reg
MFR_RESET
0xFD Force full reset without removing power.
Send Byte
N
91
l
0x4F
72
102
91
103
°C
ppm/°C
°C or V
94
l
0x14
76
l
3900ppm/°C
0x0F3C
83
l
1.0
0x4000
86
l
0.0
0x8000
86
l
0x80
72
74
NVM l Indicates a command value stored to internal EEPROM using STORE_USER_ALL or restored to RAM from internal EEPROM at power-up or
execution of RESTORE_USER_ALL or MFR_RESET.
38821f
For more information www.linear.com/LTC3882-1
41
LTC3882-1
Applications Information
Efficiency Considerations
Normally, one of the primary goals of any LTC3882-1 application will be to obtain the highest practical conversion
efficiency. The efficiency of a switching regulator is equal
to the output power divided by the input power. It is often
useful to analyze individual losses to determine what is
limiting the efficiency and to ascertain which change would
produce the most improvement. Balancing or limiting these
individual losses plays a dominant role in the component
selection process outlined over the next few sections.
Percent efficiency can be expressed as:
%Efficiency = 100% – (L1 + L2 + L3 + …)
where L1, L2, et al, are the individual losses as a percentage of input power: 100 • PLn /PIN.
Although all dissipative elements in the system produce
losses, four main sources usually account for most of
the losses in LTC3882-1 applications: IC supply current,
I2R losses, topside power MOSFET transition losses and
total gate drive current.
1.The LTC3882-1 IC supply current is a DC value given
in the Electrical Characteristics table. The absolute loss
created by the IC itself is approximately this current
times the VCC supply voltage. IC supply current typically
results in a small loss (<0.1%).
2.I2R losses occur mainly in the DC resistances of the
MOSFET, inductor, PCB routing, and input and output
capacitor ESR. Since each MOSFET is only on for part
of the cycle, its on-resistance is effectively multiplied by
the percentage of the cycle it is on. Therefore the bottom MOSFET should have a much lower on-resistance
RDS(ON) than the top MOSFET in high step-down ratio
applications. It is crucial that careful attention is paid
to the layout of the power path on the PCB to minimize
its resistance. In a 2-phase 1.2V system, 1mΩ of PCB
resistance at the output costs 5% in efficiency with the
output running at 60A.
3. Transition losses apply only to the topside MOSFET and
become significant when operating at high input voltages
(typically above 12V). This loss can be minimized by
42
choosing a driver with very low drive resistance and a
MOSFET with low gate charge QG, gate resistance RG
and Miller capacitance CMILLER. Absolute transition loss
can be estimated by:
PTRANS = (1.7) • VIN2 • IOUT • CMILLER • fPWM
4. Gate drive current is equal to the sum of the top and bottom MOSFET gate charges multiplied by the frequency of
operation. These charges are based on the gate voltage
applied by the FET driver and can be determined from
manufacturer curves like the one shown in Figure 21.
Many driver ICs employ asymmetrical gate voltages for
top and bottom FETs.
Other sources of loss include body or Schottky diode
conduction during the driver dependent non-overlap time
and inductor core losses. These latter categories generally
account for less than 2% total additional loss.
PWM Frequency and Inductor Selection
The selection of the PWM switching frequency is a trade-off
between efficiency, transient response and component size.
High frequency operation reduces the size of the inductor
and output capacitor as well as increasing the maximum
practical control loop bandwidth. However, efficiency is
generally lower due to increased transition and switching losses. The inductor value is related to the switching
frequency fPWM and step-down ratio. It should be selected
to meet choke ripple current requirements. The inductor
value can be calculated using the following equation:
⎛ VOUT ⎞ ⎛ VOUT ⎞
L=⎜
• 1–
VIN ⎟⎠
⎝ f PWM • ΔIL ⎟⎠ ⎜⎝
Allowing a larger value of choke ripple current (ΔIL) leads
to smaller L, but results in greater core loss and higher
output voltage ripple for a given output capacitance and/
or ESR. A reasonable starting point for setting the ripple
current is 30% of the maximum output current.
The inductor saturation current rating needs to be higher
than the peak inductor current during transient conditions.
If IOUT is the maximum rated load current, then the maximum transient current IMAX would normally be chosen to
be some factor greater than IOUT (e.g., 1.6 • IOUT).
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
The minimum saturation current rating should be chosen
to allow margin due to manufacturing and temperature
variation in the sense resistor or inductor DCR. A reasonable ISAT value would be 2.2 • IOUT.
The programmed current limit IOUT_OC_FAULT_LIMIT
must be low enough to ensure that the inductor never
saturates and high enough to allow increased current
during transient conditions with margin for DCR variation.
For example, if
ISAT = 2.2 • IOUT, and
IMAX = 1.6 • IOUT
a reasonable output current limit would be
is normally less important for overall efficiency than its
input capacitance. MOSFET manufacturers have designed
special purpose devices that provide reasonably low onresistance with significantly reduced input capacitance
for the main switch application in switching regulators.
Selection criteria for the power MOSFETs include onresistance, gate charge, Miller capacitance, breakdown
voltage and maximum output current.
For maximum efficiency, RDS(ON) and QG should be minimized. Low RDS(ON) minimizes conduction losses and low
QG minimizes switching and transition losses. MOSFET
gate charge can be taken from the typical gate charge
curve included on most data sheets (Figure 21).
IOUT_OC_FAULT_LIMIT = 1.8 • IOUT
Once the value of L is known, the type of inductor must be
selected. High efficiency converters generally cannot afford
the core losses found in low cost powdered iron cores,
forcing the use of more expensive ferrite or molypermalloy
cores. Also, core losses decrease as inductance increases.
Unfortunately, increased inductance requires more turns
of wire, larger inductance and larger copper losses.
Ferrite designs have very low core loss and are preferred at
high switching frequencies. However, these core materials
exhibit hard saturation, causing an abrupt reduction in the
inductance when the peak current capability is exceeded.
Do not allow the core to saturate!
Power MOSFET Selection
The LTC3882-1 requires at least two external N-channel
power MOSFETs per channel, one for the top (main) switch
and one or more for the bottom (synchronous) switch. The
number, type and on-resistance of the MOSFETs selected
should take into account the voltage step-down ratio and
the FET circuit position (main or synchronous switch).
A much smaller and lower input capacitance MOSFET
should be used for the top MOSFET in applications that
have an output voltage that is less than one-third of the
input voltage. At operating frequencies above 300kHz
and where VIN >> VOUT, the top MOSFET on-resistance
MILLER EFFECT
VGS
QA
QB
QIN
CMILLER = (QB – QA)/VDS
38821 F21
Figure 21. Typical MOSFET Gate Charge Curve
CMILLER is the most important selection criteria for determining the transition loss term in the top MOSFET but is
not directly specified on MOSFET data sheets. CMILLER is
equal to the increase in gate charge along the horizontal
axis of Figure 21 while the curve is approximately flat,
divided by the specified change in VDS. This result is
then multiplied by the ratio of the actual application VDS
to the VDS specified on the gate charge curve. When the
controller is operating in continuous mode the duty cycles
for the top and bottom MOSFETs are given by:
Main Switch Duty Cycle =
VOUT
VIN
Synchronous Switch Duty Cycle =
VIN – VOUT
VIN
38821f
For more information www.linear.com/LTC3882-1
43
LTC3882-1
Applications Information
The power dissipation for the main and synchronous
MOSFETs at maximum output current are given by:
PMAIN =
VOUT
2
IMAX ) (1+ δ)RDS(ON) +
(
VIN
MOSFET Driver Selection
I
VIN 2 MAX (RDR ) (CMILLER ) •
2
⎡
1 ⎤
1
+
⎢
⎥ ( fPWM )
⎢⎣ VGG – VTH(IL) VTH(IL) ⎥⎦
V –V
2
PSYNC = IN OUT (IMAX ) (1+ δ)RDS(ON)
VIN
where δ is the temperature dependency of RDS(ON), RDR
is the effective top driver resistance, VIN is the drain potential and the change in drain potential in the particular
application. VGG is the applied gate voltage, VTH(IL) is
the typical gate threshold voltage specified in the power
MOSFET data sheet at the specified drain current, and
CMILLER is the capacitance calculated using the technique
previously described.
The term (1 + δ) is generally given for a MOSFET in the
form of a normalized RDS(ON) versus temperature curve.
Typical values for δ range from 0.005/°C to 0.01/°C depending on the particular MOSFET used.
Both MOSFETs have I2R losses while the topside N-channel
losses also include transition losses, which are highest
at high input voltages. For VIN < 20V the high current efficiency generally improves with larger MOSFETs, while
for VIN > 20V the transition losses rapidly increase to
the point that the use of a higher RDS(ON) device with
lower CMILLER actually provides higher efficiency. The
synchronous MOSFET losses are greatest at high input
voltage when the top switch duty factor is low or during
a short-circuit when the synchronous switch is on close
to 100% of the period.
Multiple MOSFETs can be used in parallel to lower RDS(ON)
and meet the current and thermal requirements if desired.
If using discrete drivers and MOSFETs, check the stress
on the MOSFETs by independently measuring the drainto-source voltages directly across the device terminals.
Beware of inductive ringing that could exceed the maximum
44
voltage rating of the MOSFET. If this ringing cannot be
avoided and exceeds the maximum rating of the device,
choose a higher voltage rated MOSFET.
Gate driver ICs, DrMOS devices and power blocks with an
interface compatible with the LTC3882-1 3.3V three-state
PWM control output(s) can be used. An external resistor
divider may be needed to set three-state control voltage
outputs to mid-rail while in the high impedance state, depending on the driver selected. These external driver/power
circuits do not typically present a heavy capacitive load to
the LTC3882-1 PWM outputs. Suitable drivers such as the
LTC4449 are capable of driving large gate capacitances at
high transition rates. In fact, when driving MOSFETs with
very low gate charge, it is sometimes helpful to slow down
the drivers by adding small gate resistors (5Ω or less) to
reduce noise and EMI caused by fast transitions.
Using PWM Protocols
For successful utilization of the driver selected, the
appropriate LT3882-1 PWM control protocol must be
programmed. The LTC3882-1 supports two three-state
PWM control protocols. See bit[1] of the MFR_PWM_
MODE_LTC3882-1 PMBus command.
The first of these protocols (bit[1]=0) is for drivers controlled by a single 3-state input that have sufficiently short
delay to the diode emulation state (both top and bottom
power MOSFETs disabled in a fraction of a PWM cycle),
such as the LTC4449. The second protocol (bit[1]=1)
handles all other 3.3V compatible drivers with a single
3-state control input.
CIN Selection
The input bypass capacitance for an LTC3882-1 circuit
needs to have ESR low enough to keep the supply drop
low as the top MOSFETs turn on, RMS current capability
adequate to withstand the ripple current at the input, and
a capacitance value large enough to maintain the input
voltage until the input supply can make up the difference.
Generally, a capacitor that meets the first two requirements (particularly a non-ceramic type) will have far more
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
capacitance than is required to keep capacitance-based
droop under control.
The input capacitance voltage rating should be at least 1.4
times the maximum input voltage. Power loss due to ESR
occurs as I2R dissipation in the capacitor itself. The input
capacitor RMS current and its impact on any preceding
input network is reduced by PolyPhase architecture. It can
be shown that the worst case RMS current occurs when
only one controller is operating. The controller with the
highest (VOUT)(IOUT) product should be used to determine
the maximum RMS current requirement. Increasing the
output current drawn from the other out-of-phase controller will decrease the input RMS ripple current from this
maximum value. Two channel out-of-phase operation
typically reduces the input capacitor RMS ripple current
by a factor of 30% to 70%.
In continuous inductor conduction mode, the source current of the top power MOSFET is approximately a square
wave of duty cycle VOUT/VIN. The maximum RMS capacitor
current in this case is given by:
IRMS ≈IOUT(MAX)
from higher inductance, larger case size and limited surface
mount applicability; and electrolytic capacitors have higher
ESR and can dry out. Sanyo OS-CON SVP(D) series, Sanyo
POSCAP TQC series, or Panasonic EE-FT series aluminum
electrolytic capacitors can be used in parallel with a couple
of high performance ceramic capacitors as an effective
means of achieving low ESR and high bulk capacitance.
In addition to PWM bulk input capacitance, a small (0.01μF
to 1μF) bypass capacitor between the chip VINSNS pin
and ground, placed close to the LTC3882-1, is also suggested. A small resistor placed between the bulk CIN and
the VINSNS pin provides further isolation between the two
channels. However, if the time constant of any such R-C
network on the VINSNS pin exceeds 30ns, dynamic line
transient response can be adversely affected.
COUT Selection
The selection of COUT is primarily determined by the ESR
required to minimize voltage ripple and load step transients.
The output ripple ΔVOUT is approximately bounded by:
VOUT ( VIN – VOUT )
VIN
This formula has a maximum at VIN = 2VOUT, where
where ΔIL is the inductor ripple current.
IRMS = IOUT/2
This simple worst-case condition is commonly used for
design because even significant deviations do not offer
much relief.
Note that manufacturer ripple current ratings for capacitors
are often based on only 2000 hours of life. This makes
it advisable to further derate the capacitor or to choose
a capacitor rated at a higher temperature than required.
Several capacitors may also be paralleled to meet size or
height requirements in the design. Always consult the
manufacturer if there is any question.
Ceramic, tantalum, semiconductor electrolyte (OS-CON),
hybrid conductive polymer (SUNCON) and switcher-rated
electrolytic capacitors can be used as input capacitors, but
each has drawbacks. Ceramics have high voltage coefficients of capacitance and may have audible piezoelectric
effects; tantalums need to be surge-rated; OS-CONs suffer
⎛
⎞
1
ΔVOUT ≤ ΔIL ⎜ ESR+
8 • f PWM •COUT ⎟⎠
⎝
ΔIL =
VOUT ⎛ VOUT ⎞
1–
L • f PWM ⎜⎝
VIN ⎟⎠
Since ΔIL increases with input voltage, the output ripple
voltage is highest at maximum input voltage. Typically
once the ESR requirement is satisfied, the capacitance is
adequate for filtering and has the necessary RMS current
rating.
Manufacturers such as Sanyo, Panasonic and Cornell Dubilier should be considered for high performance throughhole capacitors. The OS-CON semiconductor electrolyte
capacitor available from Sanyo has a good (ESR)(size)
product. An additional ceramic capacitor in parallel with
polarized capacitors is recommended to offset the effect
of lead inductance.
In surface mount applications, multiple capacitors may
have to be paralleled to meet the ESR or transient current
38821f
For more information www.linear.com/LTC3882-1
45
LTC3882-1
Applications Information
The LTC3882-1 is a voltage mode controller with a second,
dedicated current sharing loop to provide excellent phaseto-phase current sharing in PolyPhase applications. The
current sharing loop is internally compensated.
While Type 2 compensation for the voltage control loop
may be adequate in some applications (such as with the
use of high ESR bulk capacitors), Type 3 compensation
and ceramic capacitors are recommended for optimum
transient response.
Figure 22 shows a simplified view of the error amplifier EA
for one LTC3882-1 channel. The positive input of the error
amplifier is connected to the output of an internal 12-bit
DAC fed by a 1.024V reference, while the negative input is
connected to the FB pin and other internal circuits (not all
shown). R1 is internal to the IC with a value range given
by the RVSFB parameter in the Electrical Characteristics
table. The output is connected to COMP, from which the
PWM controller derives the required output duty cycle. To
speed up overshoot recovery time, the maximum potential
at the COMP pin is internally clamped.
Unlike many regulators that use a transconductance (gm)
amplifier, the LTC3882-1 is designed to use an inverting
summing amplifier topology with the FB pin configured
as a virtual ground. This allows feedback gain to be tightly
controlled by external components, which is not possible
with a simple gm amplifier. The voltage feedback amplifier
also provides flexibility in choosing pole and zero locations. In particular, it allows the use of Type 3 compensa-
46
C2
VOUT
C3
R1
R3
–
FB
INTERNAL
C1
R2
+
VDAC
EA
COMP
3882 F22
Figure 22. Type 3 Compensation Circuit
0
–1
GAIN
+1
–1
PHASE (DEG)
Feedback Loop Compensation
tion to provide phase boost at the LC pole frequency for
significantly improving the control loop phase margin,
as shown in Figure 23.
GAIN (dB)
handling requirements of the application. Aluminum electrolytic and dry tantalum capacitors are both available in
surface mount configurations. New special polymer surface
mount capacitors offer very low ESR also but have much
lower capacitive density per unit volume. In the case of
tantalum, it is critical that the capacitors are surge tested
for use in switching power supplies. Several excellent output capacitor choices include the Sanyo POSCAP TPD/E/F
series, the Kemet T520, T530 and A700 series, NEC/Tokin
NeoCapacitors and Panasonic SP series. Other suitable
capacitor types include Nichicon PL series and Sprague
595D series. Consult the manufacturer for other specific
recommendations.
FREQ
–90
PHASE
–180
BOOST
–270
–380
38821 F23
Figure 23. Type 3 Compensation Frequency Response
In a typical LTC3882-1 circuit, the feedback loop closed
around this control amplifier and compensation network
consists of the line feedforward circuit, the modulator,
the external inductor and the output capacitor. All these
components affect loop behavior and need to be accounted
for in the frequency compensation.
The modulator consists of the PWM generator, the output
MOSFET drivers and the external MOSFETs themselves.
Step-down modulator gain varies linearly with the input
voltage. The line feedforward circuit compensates for this
change in gain, and provides a constant gain AMOD of 4V/V
from the error amplifier output COMP to the inductor input
(average DC voltage) regardless of VIN. The combination
of the line feedforward circuit and the modulator looks
like a linear voltage transfer function from COMP to the
inductor input with a fairly benign AC behavior at typical
loop compensation frequencies. Significant phase shift
will not begin to occur in this transfer function until half
the switching frequency.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
The external inductor/output capacitor combination makes
a more significant contribution to loop behavior. These
components cause a 2nd order amplitude roll-off that filters
the PWM waveform, resulting in the desired DC output
voltage. But the additional 180° phase shift produced by
this filter causes stability issues in the feedback loop and
must be frequency compensated. At higher frequencies,
the reactance of the output capacitor will approach its
ESR, and the roll-off due to the capacitor will stop, leaving
–20dB/decade and 90° of phase shift.
The transfer function of the Type 3 circuit shown in
Figure 22 is given by the following equation:
VCOMP
–(1+ sC1R2)[1+ s(R1+R3)C3]
=
VOUT
sR1(C1+C2)[1+ s(C1//C2)R2](1+ sC3R3)
The RC network across the error amplifier and the feedforward components R3 and C3 introduce two pole-zero
pairs to obtain a phase boost at the system unity-gain
(crossover) frequency, fC. In theory, the zeros and poles are
placed symmetrically around fC, and the spread between the
zeros and the poles is adjusted to give the desired phase
boost at fC. However, in practice, if the crossover frequency
is much higher than the LC double-pole frequency, this
method of frequency compensation normally generates
a phase dip within the unity bandwidth and creates some
concern regarding conditional stability.
If conditional stability is a concern, move the error amplifier zero to a lower frequency to avoid excessive phase
dip. The following equations can be used to compute the
feedback compensation component values:
1
fLC =
2π LCOUT
fESR =
choose:
1
2πRESRCOUT
fC =crossover frequency =
fPWM
10
1
2πR2C1
f
1
fZ2(RES) = C =
5 2π(R1+R3)C3
1
fP1(ERR) = fESR =
2πR2(C1//C2)
1
fP2(RES) =5fC =
2πR3C3
fZ1(ERR) = fLC =
Required error amplifier gain at frequency fC is:
2
2
⎛ f ⎞
⎛ f ⎞
≈ 40log 1+ ⎜ C ⎟ – 20log 1+ ⎜ C ⎟ – 15.56
⎝ fLC ⎠
⎝ fESR ⎠
Once the value of resistor R1 (function of selected VOUT
range) and pole/zero locations have been decided, the
value of R2, C1, C2, R3 and C3 can be obtained from the
previous equations.
Compensating a switching power supply feedback loop is
a complex task. The applications shown in this data sheet
provide typical values, optimized for the power components
shown. Though similar power components should suffice,
substantially changing even one major power component
may degrade performance significantly. Stability also may
depend on circuit board layout. To verify the calculated
component values, all new circuit designs should be
prototyped and tested for stability.
The LTPowerCAD software tool can be used as a guide
through the entire power supply design process, including optimization of circuit component values according to
system requirements.
PCB Layout Considerations
To prevent magnetic and electrical field radiation or high
frequency resonant problems and to ensure correct IC
operation, proper layout of the components connected
to the LTC3882-1 is essential. Refer to Figure 24, which
also illustrates current waveforms typically present in the
circuit branches. RSENSE will be replaced with a dead short
if DCR sensing is used. For maximum efficiency, the switch
38821f
For more information www.linear.com/LTC3882-1
47
LTC3882-1
Applications Information
node rise and fall times should be minimized. The following
PCB design priority list will help ensure proper topology.
sensing is used, place the top resistor (R1, Figure 25)
close to the switch node.
1. Place a ground or DC voltage layer between a power
layer and a small-signal layer. Generally, power planes
should be placed on the top layer (4-layer PCB), or top
and bottom layer if more than 4 layers are used. Use
wide/short copper traces for power components and
avoid improper use of thermal relief around power
plane vias to minimize resistance and inductance.
6. Place low ESR output capacitors adjacent to the sense
resistor output and ground. Output capacitor ground
connections must feed into the same copper that connects to the input capacitor ground before connecting
back to system ground.
2. Low ESR input capacitors should be placed as close
as possible to switching FET supply and ground connections with the shortest copper traces possible. The
switching FETs must be on the same layer of copper
as the input capacitors with a common topside drain
connection at CIN. Do not attempt to split the input
decoupling for the two channels, as a large resonant
loop can result. Vias should not be used to make these
connections. Avoid blocking forced air flow to the
switching FETs with large size passive components.
3. If using a discrete FET driver, place that IC close to the
switching FET gate terminals, keeping the connecting
traces short to produce clean drive signals. This rule
also applies to driver IC supply and ground pins that
connect to the switching FET source pins. The driver
IC can be placed on the opposite side of the PCB from
the switching FETs.
4. Place the inductor input as close as possible to the
switching FETs. Minimize the surface area of the switch
node. Make the trace width the minimum needed to
support the maximum output current. Avoid copper
fills or pours. Avoid running the connection on multiple
copper layers in parallel. Minimize capacitance from
the switch node to any other trace or plane.
5. Place the output current sense resistor (if used) immediately adjacent to the inductor output. PCB traces
for remote voltage and current sense should be run
together back to the LTC3882-1 in pairs with the smallest spacing possible on any given layer on which they
are routed. Avoid high frequency switching signals
and ideally shield with ground planes. Locate any filter
component on these traces next to the LTC3882-1,
and not at the Kelvin sense location. However, if DCR
48
7. Connection of switching ground to system ground,
small-signal analog ground or any internal ground plane
should be single-point. If the system has an internal
system ground plane, a good way to do this is to cluster
vias into a single star point to make the connection. This
cluster should be located directly beneath the IC GND
paddle, which serves as both analog signal ground and
the negative sense for VOUT1. A useful CAD technique
is to make separate ground nets and use a 0Ω resistor
to connect them to system ground.
8. Place all small-signal components away from high
frequency switching nodes. Place decoupling capacitors for the LTC3882-1 immediately adjacent to the IC.
9. A good rule of thumb for via count in a given high current path is to use 0.5A per via. Be consistent when
applying this rule.
10.Copper fills or pours are good for all power connections except as noted above in rule 3. Copper planes
on multiple layers can also be used in parallel. This
helps with thermal management and lowers trace
inductance, which further improves EMI performance.
Output Current Sensing
The ISENSE+ and ISENSE– pins are high impedance inputs to
internal current comparators, the current-sharing loop and
telemetry ADC. The common mode range of the current
sense inputs is approximately 0V to 5.5V. Continuous linear
operation is provided throughout this range. Maximum
differential current sense input (ISENSE+ – ISENSE–) is 70mV,
including any variation over temperature. These inputs
must be properly connected in the application at all times.
To maximize efficiency at full load the LTC3882-1 is designed
to sense current through the inductor’s DCR, as shown in
Figure 25. The DCR of the inductor represents the small
amount of DC winding resistance of the copper, which
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
SW1
L1
D1
RSENSE1
VOUT1
COUT1
RL1
VIN
RIN
CIN
SW0
BOLD LINES INDICATE
HIGH SWITCHING
CURRENT. KEEP LINES
TO A MINIMUM LENGTH.
D0
L0
RSENSE0
VOUT0
COUT0
RL0
38821 F24
Figure 24. High Frequency Paths and Branch Current Waveforms
for most inductors suitable to LTC3882-1 applications, is
between 0.3mΩ and 1mΩ. If the filter RC time constant is
chosen to be exactly equal to the L/DCR time constant of
the inductor, the voltage drop across the external capacitor is equal to the voltage drop across the inductor DCR.
Check the manufacturer’s data sheet for specifications
regarding the inductor DCR in order to properly dimension
the external filter components. The DCR of the inductor
can also be measured using a good RLC meter.
Use the nominal or measured value of DCR to program
IOUT_CAL_GAIN (in mΩ). The temperature coefficient
of the inductor’s DCR is typically high, like copper. Again,
consult the manufacturer’s data sheet. The LTC3882-1 can
adjust for this non-ideality if the correct MFR_IOUT_CAL_
GAIN_TC value is programmed. Typically this coefficient
is around 3900ppm/°C.
Resistor R1 should be placed close to the switch node,
to prevent noise from coupling into sensitive small-signal
nodes. Capacitor C1 should be placed close to the IC pins.
An example of discrete resistor sensing of output current is
shown in Figure 26. Previously, the parasitic inductance of
the sense resistor could represent a relatively small error.
New high current density solutions may utilize low sense
resistor values producing sense voltages less than 20mV.
In addition, inductor ripple currents greater than 50%
with operation up to 1MHz are becoming more common.
Under these conditions, the voltage drop across the sense
resistor’s parasitic inductance is no longer negligible. An
RC filter can be used to extract the resistive component
of the current sense signal in the presence of parasitic
inductance. For example, Figure 27 illustrates the voltage
38821f
For more information www.linear.com/LTC3882-1
49
LTC3882-1
Applications Information
waveform across a 2mΩ resistor with a 2010 footprint.
The waveform is the superposition of a purely resistive
component and a purely inductive component. If the
RC time constant is chosen to be close to the parasitic
inductance divided by the sense resistor (L/R), the resultant waveform looks resistive, as shown in Figure 28.
If low value (<5mΩ) sense resistors are used, verify that
the signal across CF resembles the current through the
inductor, and reduce RF to eliminate any large step associated with the turn-on of the primary switch.
For applications using low maximum sense voltages,
check the sense resistor manufacturer’s data sheet for
information about parasitic inductance. In the absence of
data, measure the voltage drop directly across the sense
resistor to extract the magnitude of the ESL step and the
following equation to determine the ESL:
Accurate Kelvin sensing techniques should be used
to connect the output voltage differentially back to the
LTC3882-1 VSENSE± pins of the master channel for the best
output voltage regulation at the point of load. These pins
also provide the ADC inputs for output voltage telemetry.
ESL =
Output Voltage Sensing
VESL(STEP) tON • tOFF
•
∆IL
tON + tOFF
VIN
12V
VINSNS
5V
LTC3882-1
VCC
VDD33
PWM
VCC
BOOST
INDUCTOR
VLOGIC
TG
LTC4449
IN
TS
–
+
GND ISENSE ISENSE
GND
L
DCR
VOUT
BG
R1*
C1*
38821 F25
R1 • C1 =
L
*PLACE R1 NEAR INDUCTOR
DCR PLACE C1 NEAR I
+
–
SENSE , ISENSE PINS
Figure 25. Inductor DCR Output Current Sense
VIN
12V
VINSNS
LTC3882-1
5V
VCC
VCC
VDD33
VLOGIC
TG
LTC4449
IN
TS
PWM
–
+
GND ISENSE ISENSE
CF
SENSE RESISTOR
PLUS PARASITIC
INDUCTANCE
BOOST
GND
BG
L
RS
ESL
VOUT
CF • 2RF ≤ ESL/RS
POLE-ZERO
CANCELLATION
RF
RF
38821 F26
FILTER COMPONENTS PLACED NEAR SENSE PINS
Figure 26. Discrete Resistor Output Current Sense
50
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
While these considerations may or may not be important
for slave channels, VOUT must be connected back to the
slave channel VSENSE pin(s) in order for the IOUT telemetry
of those phases to be accurate. So in general, sound Kelvin
VOUT sensing techniques for all LTC3882-1 channels is
recommended.
VRSENSE
20mV/DIV
VESL(STEP)
500ns/DIV
38821 F27
Figure 27. Voltage Measured Directly Across RSENSE
VISENSE
20mV/DIV
500ns/DIV
38821 F28
Figure 28. Voltage Measured at ISENSE Pins
Soft-Start and Stop
The LTC3882-1 uses digital ramp control to create both
soft-start and soft-stop.
The LTC3882-1 must enter the run state prior to soft-start.
The RUN pins are released after the part initializes and
VINSNS is determined to be greater than the VIN_ON
threshold.
Once in the run state, soft-start is performed after any additional prescribed delay (next section) by actively regulating
the load voltage while digitally ramping the target voltage
from 0V to the commanded voltage set point. Rise time of
the voltage ramp can be programmed using the TON_RISE
command to minimize inrush currents associated with
start-up voltage ramp. The maximum rate at which the
LTC3882-1 can move the output in this fashion is 100µs/
step. Soft-start is disabled by setting TON_ RISE to any
value less than 0.250ms. The LTC3882-1 will perform the
necessary math internally to assure the voltage ramp is
controlled to the desired slope. However, the voltage slope
cannot be any faster than the fundamental limits of the
power stage. The smaller TON_RISE becomes, the more
noticeable an output voltage stair-step may become.
The LTC3882-1 also supports soft turn off in the same
manner it controls turn on. TOFF_FALL is processed when
the RUN pin goes low or if the part is commanded off.
If the part faults off or FAULT is pulled low and the part
is programmed to respond to this, the PWM instantly
commands the output off. The output will then decay as
a function of load current.
The LTC3882-1 can produce a controlled ramp off as long
as the power stage is configured to run in CCM and the
TOFF_FALL time is sufficiently slow that the power stage
can achieve the desired slope. The TOFF_FALL time can
only be met if the power stage can sink sufficient current
under closed loop control to assure the output is at 0V
by the end of the fall time. If TOFF_FALL is shorter than
the time required to discharge the load capacitance, the
output will not reach 0V. In this case, the power stage will
still be commanded off at the end of TOFF_FALL and VOUT
will decay at a rate determined by the load. If the controller is set to run DCM, the controller will not pull negative
current and the output will only be pulled low by the load,
not the power stage. The maximum fall time is limited to
1.3 seconds. The smaller TOFF_FALL becomes, the more
noticeable an output voltage stair-step may become.
Time-Based Output Sequencing and Ramping
The LTC3882-1 TON_DELAY and TOFF_DELAY commands
can be used in combination with the rise and fall time commands covered in the previous section to implement a wide
range of versatile sequencing and ramping schemes. The
key to time-based sequencing and ramping is the ability of
LTC3882-1 master phases to move their outputs up and
down according to PMBus command values as shown in
Figure 29 and Figure 30.
38821f
For more information www.linear.com/LTC3882-1
51
LTC3882-1
Applications Information
As shown in Figure 29, it is important to remember that the
hysteresis given in the Electrical Characteristics (EC) table
applies above VOUT_UV_FAULT_LIMIT, which specifies the
UV limit when the output is falling out of regulation. For
turn-on, the output must rise above this programmed limit
plus the hysteresis to avoid exceeding TON_MAX_FAULT_
LIMIT. PGOOD is indicated at that point. For this reason,
VOUT_UV_FAULT_LIMIT should be more than 27mV below
the programmed output voltage in low range and more
than 54mV below VOUT in high range. There is a fixed delay
and other timing uncertainty associated with all changes
in output voltage controlled by the LTC3882-1. A nominal
fixed timing delay of 270µs exists to process any change
in output voltage, including soft start/stop, margining and
general changes in VOUT_COMMAND value. The start of all
time-based output operations occur with an uncertainty of
±50µs and have a nominal step resolution of 100µs. This
means the minimum TON_DELAY or TOFF_DELAY that the
LTC3882-1 can produce will range from 220µs to 320µs,
not including basic oscillator tolerances. For softwarebased output changes (e.g., margining), this algorithmic
delay begins when the STOP bit is received on the serial
bus. An example of this minimum turn on/off delay and
step-wise output control can be seen in Figure 31, where
TON_DELAY = 0s and TON_RISE = 1ms.
To effectively implement sequencing and synchronized
ramping between rails controlled by LTC digital power
products, two signals should be shared between all
controlling ICs: SHARE_CLK and RUN (CONTROL pin
on LTC297x products). This facilitates synchronized rail
sequencing on or off based on shared input supply state
(VIN_ON threshold), external hardware control (RUN pin),
or PMBus commands (possibly using global addressing).
Figure 32 shows an example of output supply sequencing
using TON_DELAY.
Using this scheme, conventional coincident and ratiometric
tracking can also be emulated by setting equivalent turnon/off delays and appropriate rise and fall times as shown
in Figure 33 and Figure 34.
In addition, these schemes can easily be mixed and matched
to create any necessary ramping controls, some of which
might prove difficult to implement with conventional
analog-only controllers. These programmable features
52
RUN
PGOOD
TON_MAX_FAULT_LIMIT
DIGITAL SERVO
MODE ENABLED FINAL OUTPUT
VOLTAGE REACHED
VOUT_UV_FAULT_LIMIT
DAC VOLTAGE
ERROR (NOT
TO SCALE)
VOUT
TIME DELAY OF
<1S, TYPICAL
HYSTERESIS
(NOT TO SCALE)
TON_DELAY
TON_RISE
TIME
38821 F29
Figure 29. Time-Based VOUT Turn-On
RUN
VOUT
TOFF_DELAY
TOFF_FALL
TIME
38821 F30
Figure 30. Time-Based VOUT Turn-Off
greatly simplify actual system development because rails
can be re-sequenced without a hardware change as final
product requirements evolve. The LTpowerPlay GUI and
LTC3882-1 onboard EEPROM can be used for this task,
avoiding the need for firmware development to modify turn
on/off relationships between rails. Entire power systems
can then easily be scaled up or down, facilitating reuse of
proven hardware macro designs.
Voltage-Based Output Sequencing
The LTC3882-1 is capable of voltage-based output sequencing. For concatenated events between members of
the LTC388x family, it is possible to control one RUN pin
from a PGOOD pin of a different controller as shown in
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
VOUT
(1V/DIV)
RUN
(5V/DIV)
38821 F31
200µs/DIV
Figure 31. Example of 1ms TON_RISE
1V/DIV
VOUT4
Figure 35. This configuration hardware disables the next
downstream controller anytime the output is not within
the specified UV and OV limits, or the upstream controller
is disabled. When indicating power is not good, there is
a 30µs deglitching filter on the PGOOD output to assure
the signal does not toggle repeatedly at lower values of
TON_RISE/FALL due to noise on VOUT. If unwanted transitions still occur on PGOOD due to noise or longer rise/fall
settings, place a capacitor to ground on the PGOOD pin
to further filter the waveform. The RC time-constant of
the filter should be low enough to assure no appreciable
delay is incurred. A value of 300μs to 500μs will provide
some additional filtering without significantly delaying
the trigger event.
START
VOUT3
VOUT2
PGOOD0
RUN 0
LTC3882-1
RUN 1
PGOOD1
(PGOOD PULL-UP RESISTORS
TO 3.3V NOT SHOWN)
VOUT1
RUN 0
100ms/DIV
38821 F32
PGOOD0
LTC3882-1
PGOOD1
RUN 1
Figure 32. LTC3882-1 Time-Based Supply Sequencing
TO NEXT CHANNEL
IN THE SEQUENCE
38821 F35
Figure 35. Cascade Sequencing Configuration
1V/DIV
VOUT4
VOUT3
VOUT2
1V/DIV
VOUT1
VOUT4
VOUT3
100ms/DIV
38821 F33
Figure 33. LTC3882-1 Time-Based Coincident Supply Ramping
VOUT2
VOUT1
100ms/DIV
1V/DIV
38821 F36
Figure 36. Cascade Sequencing Waveforms
VOUT4
VOUT3
VOUT2
VOUT1
100ms/DIV
38821 F34
Figure 34. LTC3882-1 Time-Based Ratiometric Ramping
When the system is turned off, rails will shut down in the
same order as they turn on, as shown in Figure 36. If a
different sequence is required, the circuit must be rewired
or delays must be added by programming TON_DELAY or
TOFF_DELAY. A fundamental limitation of this application is
the inability of upstream rails to detect a start-up failure of
38821f
For more information www.linear.com/LTC3882-1
53
LTC3882-1
Applications Information
downstream rails. Due to this, cascade sequencing should
not be implemented without an external fast supervisor
to monitor downstream rails and assert a system fault if
problems occur.
AVP DISABLED
IO
(10A/DIV)
Using Output Voltage Servo
For best output voltage accuracy, enable digital
servo mode on the master phase by setting bit 6 of
MFR_PWM_MODE_LTC3882-1. In digital servo mode,
the LTC3882-1 will adjust the regulated output voltage
based on its related ADC voltage reading. Every 100ms the
digital servo loop will step the LSB of the DAC (nominally
1.375mV or 0.6875mV depending on the voltage range
bit) until the output is at the correct ADC reading.
173mV
VOUT
(50mV/DIV)
WITH AVP
IO
(10A/DIV)
When the master channel is turned on, digital servo is
enabled after all of the following conditions are satisfied.
• MFR_PWM_MODE_LTC3882-1 Bit 6 Is Set
VOUT
(50mV/DIV)
108mV
• The TON_RISE Sequence Is Complete
LOOP: BW = 118kHz, PM = 58°, GM = 7dB
• A VOUT_UV_FAULT Is Not Present
Figure 37. Active Voltage Positioning
• An IOUT_OC_FAULT Is Not Present
• MFR_AVP = 0%
Digital servo mode then engages after TON_MAX_
FAULT_LIMIT has expired as shown in Figure 29, unless
that limit is set to 0s (infinite). In that case, the mode is
engaged as soon as the above conditions are satisfied.
Using AVP
The LTC3882-1 features digitally programmable active
voltage positioning (AVP), where output voltage set point
is automatically adjusted as a function of output current at
the full bandwidth of the converter. AVP normally entails
specifying an output load line for a voltage mode switcher to
allow current sharing between master phases connected in
parallel. While AVP can be used to this effect in LTC3882-1
applications, use of the LTC3882-1 IAVG current sharing
control loop is recommended instead. This will produce
more accurate sharing across a wider number of phases
without degrading supply output impedance.
However, AVP can still be used to great benefit in LTC3882-1
applications. AVP can be applied to minimize the size of
54
38821 F37
output filter capacitance for some allowed output voltage
variation over the anticipated load range. An example of
AVP is shown in Figure 37. Refer to LTC Design Solution
10 for additional examples of using AVP to advantage.
MFR_VOUT_AVP specifies the percent reduction in programmed VOUT from no load at an output current value
equal to IOUT_WARN_LIMIT. LTC3882-1 AVP supports
a maximum reduction in VOUT of 15%, corresponding
to a ±7.5% tolerance about a nominal output voltage at
roughly 50% load. In order to effectively use AVP, apply
the following steps.
1.Set IOUT_OC_WARN_LIMIT. This specifies the master
phase output current at which the programmed AVP
level will apply. Generally this is above the 100% load
point to avoid spurious warnings at full load.
2. VOUT_COMMAND should be set to the value of VOUT desired with no load on the output. VOUT_MARGIN_HIGH/
LOW also specify no-load values when AVP is enabled.
AVP on the LTC3882-1 can only reduce the output from
these levels.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
3. Set MFR_VOUT_AVP to a percentage that produces the
desired output excursion as a function of current.
For example, if the goal is to allow a 2.5% output change
on a 3.3V 6-phase supply rated at 120A during an output
load step from 20% to 80%, the following parameters
should be programmed.
First set an output current warning level for the master (one
of six phases) just slightly higher than the rated full load
to avoid spurious warnings. Typically this same setting
would also be applied to the five slave phases.
IOUT_OC_WARN_LIMIT = 1.1 • 120A/6 = 22A
The open circuit output voltage calculation for the master
phase must reflect that the AVP specification in this case
only covers an output load swing of 60%.
VOUT_COMMAND= 3.3V(1 + 0.5 • 0.025/0.6)
= 3.3687V
The AVP calculation must then account for the fact that
IOUT_OC_WARN_LEVEL is set higher than the 100%
load point.
100% •1.1• 2 • ( 3.3687 – 3.3)
MFR_VOUT_AVP =
3.3687
= 4.487%
With the output voltage at 3.3V at 50% load these settings will move VOUT from approximately 3.34V to about
3.26V when the output load of the rail moves from 24A
to 96A. Note that VOUT will drop to 3.23V at full load in
this design example.
Digital output servo mode is automatically disabled if
AVP is enabled on a master phase. AVP is active during
all output ramping when enabled (e.g., a TON_RISE sequence). AVP is disabled on master phases by programming MFR_VOUT_AVP to 0.0% (factory default). AVP is
automatically disabled on phases configured as slaves
(FB tied to VDD33).
Because of related ISENSE input offsets, increased output
voltage error can occur at all operating currents when AVP
is engaged. To minimize this error a calibration offset can
be added to the master phase VOUT_COMMAND value
based on the READ_VOUT value obtained when operating at a known output current of at least 20% of full load
(READ_IOUT). The necessary correction, which will typically be less than several percent of the no load output
voltage, is calculated as:
VOS = VOUT_COMMAND
⎛ MFR_VOUT_AVP •READ_IOUT ⎞
• ⎜ 1–
⎟
⎝
100 •IOUT_OC_WARN_LIMIT ⎠
– READ_VOUT
PWM Frequency Synchronization
The LTC3882-1 incorporates an internal phase-locked
loop (PLL) which enables synchronization of both PWM
channels (falling edge PWM) to an external CMOS clock
from 250kHz to 1.25MHz. The PLL is locked to the falling
edge of the SYNC pin clock signal. This PLL also generates
very accurate channel phase relationships which can be
selected with the MFR_PWM_CONFIG_LTC3882-1 command. For PolyPhase applications, all phases should be
spaced evenly in the phase diagram for best results. For
instance, a 4-phase system should use a separation of
90° between channels.
The PLL has a lock detection circuit. If the PLL should lose
lock during operation, bit 4 of the STATUS_MFR_SPECIFIC
command is asserted and the ALERT pin is pulled low,
unless masked. The fault can be cleared by writing a 1 to
STATUS_MFR_SPECIFIC bit 4. A spurious ALERT for an
unlocked PLL may occur at start-up or during a reset if
this fault is not masked.
Neither PWM channel will transition from off to the RUN
state until PLL lock is indicated. When transitioning a
channel from off to RUN, bit 4 of STATUS_MFR_SPECIFIC
will be set if the PWM ramp generator for that channel
is not also locked to the desired PLL output frequency.
If the SYNC pin is not externally clocked in the application,
the PWMs will operate at the frequency specified by a nonzero FREQUENCY_SWITCH command. If that command
is set to 0x0000 (external clock only) in EEPROM or with
RCONFIG (FREQ_CFG pin grounded), then at power-up,
or MFR_RESET, or RESTORE_USER_ALL, the PWM will
not start without an external clock input. If the external
clock is lost while programmed for external clock only, or
if the PWM is simply switched to this setting under power
with no external clock present, the PLL will start/run at
For more information www.linear.com/LTC3882-1
38821f
55
LTC3882-1
Applications Information
the lowest free running frequency created by the internal
VCO. This can be well below the intended PWM frequency
of the application and may cause undesirable operation of
the converter. For this reason, it is generally recommended
that a useable PWM frequency be programmed for each
channel, regardless of whether that particiular LTC3882-1
unit serves as clock master, or not.
All channels of a PolyPhase rail are required to share SYNC
pins. Between rails and for other configurations, such synchronization is optional. If the SYNC pin is shared between
LTC3882-1s, only one LTC3882-1 should be programmed
to control the SYNC output.
PolyPhase Operation and Load Sharing
When the LTC3882-1 is used in a PolyPhase application,
the slave phases must be configured as such by connecting their FB pins to VDD33. Among other things, this
disables the error amplifiers of the slave phases. Five other
pins must then also be shared between all channels of a
PolyPhase rail:
• VINSNS
• COMP
• IAVG
• IAVG_GND
• SYNC
Using a common VINSNS connection reduces the dynamic
range required by the current loop and helps maintain
well-controlled master modulator gain.
The shared COMP signal allows the master phase error
amplifier to control the duty cycle of all slave phases to
produce the commanded output voltage.
Slave phases can detect system faults that cause the
master COMP (error amplifier) output to be too high. A
slave phase detecting this kind of error amplifier fault immediately shuts off its PWM output, indicates the fault on
its VOUT_OV Fault bit, and takes whatever additional action
may be indicated by VOUT_OV_FAULT_RESPONSE for that
channel. If this response is set to only provide hardwarelevel response (0x00), then normal channel operation will
automatically resume when the fault condition is cleared.
56
The shared IAVG and IAVG_GND signals actively balance the
amount of output current delivered from each channel using
a secondary current sharing loop. A capacitor with a value
between 100pF and 200pF should be placed between IAVG
and IAVG_GND. This capacitance can be distributed across
LTC3882-1 devices/pins for improved noise immunity. All
IAVG_GND pins for a PolyPhase rail should be tied together
and connected to a single ground point at or near the
package paddle of the master phase.
Load sharing accuracy is based primarily on the current
sense amplifier offset of each phase (IAVG_VOS) and the
offset of slave current error amplifiers (VSIOS). These are
given in the EC table. Current sense gain errors between
LTC3882-1 channels will be negligible. The secondary
current sharing loop acts to average any errors among the
phases. Because of this error averaging and the random
nature of these variables, the EC table limits ensure actual
per-phase offset will be less than or equal to ±300µV for
most designs over the full operating temperature range.
This signifies better than ±2% matching when ∆ISENSE =
15mV, not including external factors such as DCR make
tolerance.
It is necessary to properly connect VSENSE+ on a slave
phase for accurate IOUT telemetry, even though slave phases
do use need this information for PWM control. While not
strictly required, the VSENSE± lines of slave phases can
simply share with the master to provide additional output
voltage telemetry. If the only concern is accurate slave IOUT
telemetry, VSENSE+ for that channel may be locally wired
to ISENSE–. VSENSE– on a slave phase should always be
shorted to VSENSE– for its master channel. IOUT OC/ROC
function is not affected by VSENSE± wiring.
All phases must be synchronized to the same shared
SYNC clock and should be programmed to run at the
same default PWM frequency. Phases should be selected
to be evenly spaced around a 360° phasor diagram, and
all phases on a PolyPhase rail should be selected to have
the same maximum duty cycle. Refer to details for MFR_
PWM_CONFIG_LTC3882-1. Figure 38 shows an example
of connections for three phases and Figure 39 shows an
example of an 8-phase rail. Additional shared signals in
these figures highlight the ability of the LTC3882-1 to communicate fault status between phases and rails, perform
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
VIN
+
3-PHASE SENSE
–
VDD33
(0,120)
(60,240)
VINSNS
COMP0
FB0
LTC3882-1 COMP1
FB1
VSENSE0+
RUN0
VSENSE1+
RUN1
VSENSE0–
VSENSE1–
FAULT0
FAULT1
IAVG_GND
SYNC (ENABLED) SHARE_CLK
IAVG0 IAVG1
GND
COMP0
VINSNS
FB0
LTC3882-1 COMP1
FB1
VSENSE0+
RUN0
VSENSE1+
RUN1
VSENSE0–
VSENSE1–
FAULT0
FAULT1
IAVG_GND
SYNC (DISABLED) SHARE_CLK
IAVG0 IAVG1
GND
+
1-PHASE SENSE
–
SYSTEM ON/OFF
BOTH ICs SAME DEFAULT FREQUENCY SWITCH
38821 F38
Figure 38. 3+1-Phase Application
38821f
For more information www.linear.com/LTC3882-1
57
LTC3882-1
Applications Information
VIN
+
V
SENSE
– OUT
VDD33
(0,180, CLOCK MASTER)
(90,270)
VINSNS
COMP0
LTC3882-1
FB0
COMP1
FB1
VSENSE0+
RUN0
VSENSE1+
RUN1
VSENSE0–
VSENSE1–
FAULT0
IAVG_GND
FAULT1
SYNC
SHARE_CLK
IAVG0 IAVG1 GND
VINSNS
COMP0
LTC3882-1
FB0
COMP1
FB1
VSENSE0+
VSENSE1+
RUN0
RUN1
VSENSE0–
VSENSE1–
FAULT0
(45,225)
(135,315)
VINSNS
COMP0
LTC3882-1
FB0
COMP1
VSENSE0+
FB1
RUN0
VSENSE1+
RUN1
VSENSE0–
VSENSE1–
FAULT0
VINSNS
COMP0
LTC3882-1
FB0
COMP1
FB1
VSENSE0+
RUN0
VSENSE1+
RUN1
VSENSE0–
VSENSE1–
FAULT0
FAULT1
SYNC
IAVG_GND
SHARE_CLK
IAVG0 IAVG1 GND
FAULT1
SYNC
FAULT1
SYNC
IAVG_GND
SHARE_CLK
IAVG0 IAVG1 GND
IAVG_GND
SHARE_CLK
IAVG0 IAVG1 GND
SYSTEM ON/OFF
ALL ICs SAME DEFAULT FREQUENCY SWITCH
38821 F39
Figure 39. 8-Phase Application
58
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
SINGLE PHASE
SW1 V
DUAL PHASE
SW1 V
SW2 V
ICIN
IL1
ICOUT
IL2
ICIN
ICOUT
38821 F40
RIPPLE
Figure 40. Single and 2-Phase Current Waveforms
RMS INPUT RIPPLE CURRENT
DC LOAD CURRENT
0.6
1-PHASE
0.5
0.4
0.3
2-PHASE
0.2
0.1
0
0.1 0.2
0.3 0.4 0.5 0.6 0.7
DUTY FACTOR (VOUT/VIN)
0.8
0.9
38821 F41
Figure 41. Normalized RMS Input Ripple Current
1.0
0.9
0.8
1-PHASE
DIC(P-P)
VO/L
0.7
0.6
0.5
0.4
0.3
0.2
2-PHASE
0.1
0
0.1 0.2
0.3 0.4 0.5 0.6 0.7
DUTY FACTOR (VOUT/VIN)
0.8
0.9
38821 F42
Figure 42. Normalized Output Ripple Current [IRMS ~ 0.3(DIC(PP))]
synchronized time-based rail sequencing and ramping and
report accurate output current telemetry for all phases.
In general, only the PGOOD pin of the master phase needs
to be used for external Power Good indication. However,
PGOOD pins of slave phases may be shorted to a master
PGOOD bus to indicate full output power is available, unless the slave channel is used in active phase shedding.
In that case, the slave PGOOD should be left disconnected
or used only to indicate operating status for that phase.
Output current fault and warning limits should each be set
to the same values across all PolyPhase channels using
IOUT_FAULT_LIMIT and IOUT_WARN_LIMIT. The correct sense resistance and related temperature coefficient
should also be set for each phase (IOUT_CAL_GAIN,
MFR_IOUT_CAL_GAIN_TC) to achieve accurate IOUT
telemetry and consistent fault handling across phases.
Because the LTC3882-1 current sharing loop operates by
matching sensed voltage, it is important that well-matched
sense elements be used in the system. Current matching
parameters specified for the LTC3882-1 do not include these
external sources of error, such as inductor DCR tolerance.
Programming of VOUT related parameters is not required
for slave phases.
A PolyPhase power supply significantly reduces the amount
of ripple current in both the input and output capacitors.
The RMS input ripple current is divided by, and the effective ripple frequency is multiplied by, the number of
phases used as long as the input voltage is greater than
the number of phases used times the output voltage. The
output ripple amplitude is also reduced by the number of
phases used. Figure 40 graphically illustrates the principle.
The worst-case RMS ripple current for a single stage design peaks at an input voltage of twice the output voltage.
The worst case RMS ripple current for a 2-phase design
peaks at output voltages of 1/4 and 3/4 of the input voltage. When the RMS current is calculated, higher effective
duty factor results and the peak current levels are divided
as long as the current in each stage is balanced. Refer to
Application Note 19 at http://www.linear.com/designtools/
app_notes for a detailed description of how to calculate
RMS current for the single stage switching regulator. Figure 41 and Figure 42 illustrate how the input and output
currents are reduced by using an additional phase. For a
2-phase converter, the input current peaks drop in half and
the frequency is doubled. The input capacitor requirement
is then theoretically reduced by a factor of four.
For more information www.linear.com/LTC3882-1
38821f
59
LTC3882-1
Applications Information
high noise environments prevent use of the ΔVBE approach
with its lower signal levels.
TSNS
LTC3882-1
10nF
GND
MMBT3906
GND
38821 F43
Figure 43. External ΔVBE Temperature Sense
495µA
TSNS
LTC3882-1
GND
1nF
GND
1.35V AT 25°C
38821 F44
Figure 44. 2D+R Temperature Sense
External Temperature Sense
The LTC3882-1 facilitates external measurement of the
power stage temperature of each channel with several
silicon-junction-based means. The voltage produced by
the remote sense circuit is digitized by the internal ADC,
and the computed temperature value is returned by the
paged READ_TEMPERATURE_1 telemetry command.
The most accurate external temperature measurement
can be made using a diode-connected PNP transistor
such as the MMBT3906 as shown in Figure 43 with bit 5
of MFR_PWM_MODE_LTC3882-1 set to 0 (ΔVBE method).
The BJT should be placed in contact with or immediately
adjacent to the power stage inductor. Its emitter should be
connected to the TSNSn pin while the base and collector
terminals of the PNP transistor must be returned to the
LTC3882-1 GND paddle using a Kelvin connection. For
best noise immunity, the connections should be routed
differentially and a 10nF capacitor should be placed in
parallel with the diode-connected PNP.
The LTC3882-1 also supports direct junction voltage measurements when bit 5 of MFR_PWM_MODE_LTC3882-1 is
set to one. The factory defaults support a resistor-trimmed
dual diode network as shown in Figure 44. However, this
measurement method can be applied to simple single-diode
circuits of the type shown in Figure 43 with parameter
adjustments as described below. This second measurement method is not generally as accurate as the first, but
it supports legacy power blocks or may prove necessary if
60
For either method, the slope of the external temperature
sensor can be modified with the coefficient stored in
MFR_TEMP_1_GAIN. With the ΔVBE approach, typical
PNPs require temperature slope adjustments slightly
less than 1. The MMBT3906 has a recommended value
in this command of approximately MFR_TEMP_1_GAIN =
0.991 based on the ideality factor of 1.01. Simply invert
the ideality factor to calculate the MFR_TEMP_1_GAIN.
Different manufacturers and different lots may have different ideality factors. Consult with the manufacturer to
set this value. Bench characterization over temperature is
recommended when adjusting MFR_TEMP_1_GAIN for
the direct p-n junction measurement.
The offset of the external temperature sense can be adjusted by MFR_TEMP_1_OFFSET. For the ΔVBE method a
value of 0 in this register sets the temperature offset to
–273.15°C. For a direct p-n junction measurement, this
parameter adjusts the nominal circuit voltage at 25°C away
from that shown in Figure 44.
To ensure proper use of these temperature adjustment
parameters, refer to the specific formulas given for the two
methods by the MFR_PWM_MODE_LTC3882-1 command
in the later section covering PMBus command details, as
well as Application Note 137.
Resistor Configuration Pins
As a factory default, the LTC3882-1 is programmed to use
external resistor configuration, allowing output voltage,
PWM frequency and phasing, and the PMBus address to
be set without programming the part through its serial
interface or purchasing devices with custom EEPROM
contents. The RCONFIG pins all require a resistor divider
between VDD25 and GND. The RCONFIG pins are only
interrogated at initial power up and during a reset, so
modifying their values on the fly is not recommended.
RCONFIG pins on the same IC can be shared with a single
resistor divider if they require identical programming.
Resistors with a tolerance of 1% or better must be used
to assure proper operation. In the following tables, RTOP
is connected between VDD25 and the RCONFIG pin, while
RBOT is connected between the pin and GND. Noisy clock
signals should not be routed near these pins.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
Table 8. VOUTn_CFG Resistor Programming
Table 9. FREQ_CFG Resistor Programming
RTOP (kΩ)
RBOT (kΩ)
VOUT (V)
0 or Open
10
10
16.2
16.2
20
20
20
20
24.9
24.9
24.9
24.9
24.9
30.1
30.1
Open
23.2
15.8
20.5
17.4
17.8
15
12.7
11
11.3
9.09
7.32
5.76
4.32
3.57
1.96
From EEPROM
5.0
3.3
2.5
1.8
1.5
1.35
1.25
1.2
1.15
1.1
1.05
0.9
0.75
0.65
0.6
Open
0
Output OFF*
(VOUT from EEPROM)
*OPERATION value and RUNn pin must both command the channel to
start from this configuration.
Output voltage can be set as shown in Table 8. For example,
setting RTOP to 16.2kΩ and RBOT to 17.4kΩ is equivalent
to programming a VOUT_COMMAND value of 1.8V. Refer
to the Operations section for related parameters that are
also automatically set as a percentage of the programmed
VOUT if resistor configuration pins are used to determined
output voltage.
RTOP (kΩ)
RBOT (kΩ)
SWITCHING
FREQUENCY (kHz)
0 or Open
20
20
20
20
24.9
24.9
24.9
24.9
24.9
30.1
30.1
Open
Open
17.8
15
12.7
11
11.3
9.09
7.32
5.76
4.32
3.57
1.96
0
from EEPROM
1250
1000
900
750
600
500
450
400
350
300
250
External SYNC Only
Note that if SYNC pins are shared between LTC3882-1s,
only one SYNC output should be enabled. All other SYNC
outputs should be disabled. For example, if configuring
two LTC3882-1s as a 4-phase rail operating at a frequency
of 600kHz, both devices should have RTOP of 24.9kΩ and
RBOT of 11.3kΩ on the FREQ_CFG pin. In this case, selecting RTOP of 24.9kΩ and RBOT of 9.09kΩ for PHAS_CFG on
the first IC (clock master) affords 180° of phase separation
and enables the SYNC output. The second device should
have RTOP of 20kΩ and RBOT of 12.7kΩ on PHAS_CFG,
to )disable its SYNC output and run its phases with 180°
of separation in quadrature with the first IC. Only mix
phase selections that have the same maximum duty cycle
specified. Refer to Tables 9 and 10.
Table 10. PHAS_CFG Resistor Programming
RTOP (kΩ)
RBOT (kΩ)
θSYNC TO θ0
θSYNC TO θ1
MAXIMUM DUTY CYCLE
SYNC OUTPUT DISABLED
0 or Open
Open
From EEPROM
From EEPROM
See MFR_PWM_CONFIG
From EEPROM
20
15
135°
315°
87.5%
Yes
20
12.7
90°
270°
20
11
45°
225°
24.9
11.3
0°
180°
24.9
9.09
0°
180°
24.9
7.32
120°
300°
24.9
5.76
60°
240°
24.9
4.32
0°
180°
30.1
3.57
0°
120°
30.1
1.96
0°
180°
Open
0
0°
120°
No
83.3%
Yes
No
38821f
For more information www.linear.com/LTC3882-1
61
LTC3882-1
Applications Information
The LTC3882-1 address is selected based on the programming of the two configuration pins ASEL0 and ASEL1
according to Table 11. ASEL0 programs the bottom four
bits of the device address for the LTC3882-1, and ASEL1
programs the three most-significant bits. Either portion of
the address can also be retrieved from the MFR_ADDRESS
value in EEPROM. If both pins are left open, the full 7-bit
MFR_ADDRESS value stored in EEPROM is used to determine the device address. In the 4-phase example above,
it is recommended that one or both ASELn pins on both
parts be programmed to create two unique addresses. The
LTC3882-1 also responds to 7-bit global addresses 0x5A
and 0x5B. MFR_ADDRESS and MFR_RAIL_ADDRESS
should not be set to either of these values.
Table 11. ASELn Resistor Programming
RTOP (kΩ) RBOT (kΩ)
ASEL1
ASEL0
LTC3882-1 DEVICE
ADDRESS BITS[6:4]
LTC3882-1 DEVICE
ADDRESS BITS[3:0]
BINARY
HEX
from EEPROM
BINARY
HEX
0 or Open
Open
from EEPROM
10
23.2
1111
F
10
15.8
1110
E
16.2
20.5
1101
D
16.2
17.4
1100
C
20
17.8
1011
B
20
15
1010
A
20
12.7
1001
9
20
11
1000
8
24.9
11.3
111
7
0111
7
24.9
9.09
110
6
0110
6
24.9
7.32
101
5
0101
5
24.9
5.76
100
4
0100
4
24.9
4.32
011
3
0011
3
30.1
3.57
010
2
0010
2
30.1
1.96
001
1
0001
1
Open
0
000
0
0000
0
Do not draw more than 20mA from the internal 3.3V regulator for the host system, governed by IC power dissipation
as discussed in the next section. This limit includes current
required for external pull up resistors for the LTC3882-1
that are terminated to VDD33.
VDD33 powers a second internal 2.5V LDO whose output
is present on VDD25. This 2.5V supply provides power for
much of the internal processor logic on the LTC3882-1. The
VDD25 output should be bypassed directly to GND with a
low ESR X5R or X7R ceramic capacitor with a value of 1μF
or greater. Do not draw any external system current from
this supply beyond that required for LTC3882-1 specific
configuration resistor dividers.
IC Junction Temperature
The user must ensure that the maximum rated junction
temperature is not exceeded under all operating conditions.
The thermal resistance of the LTC3882-1 package (θJA)
is 33°C/W, provided the exposed pad is in good thermal
contact with the PCB. The actual thermal resistance in the
application will depend on forced air cooling and other heat
sinking means, especially the amount of copper on the
PCB to which the LTC3882-1 is attached. The following
formula may be used to estimate the maximum average
power dissipation PD (in watts) of the LTC3882-1 when
VCC is supplied externally.
PD = VCC(0.024 + fPWM • 1.6e-5 + IEXT + IRC25)
where:
Internal Regulator Outputs
The VDD33 pin provides supply current for much of the
internal LTC3882-1 analog circuitry at a nominal value of
3.3V. The LTC3882-1 features an internal linear regulator
that can be used to supply 3.3V to VDD33 from a higher
voltage VCC supply (up to 12V nominal). Use of this LDO is
optional. The LTC3882-1 will also accept an external 3.3V
62
supply attached to this pin if VCC and VDD33 are shorted. If
the internal 3.3V LDO is used, it can supply a peak current
of 85mA (including internal consumption), and the VDD33
regulator output must be bypassed to GND with a low ESR
X5R or X7R ceramic capacitor with a value of 2.2μF. If an
external source supplies VDD33, a local low ESR bypass
capacitor with a value between 0.01μF and 0.1μF should
be placed directly between the VDD33 and GND pins.
IEXT = total external load drawn from VDD33, including
local pull-up resistors, in amps
IRC25 = total current drawn from VDD25 by LTC3882-1
configuration resistor dividers, in amps
and the PWM frequency fPWM is given in kHz
For more information www.linear.com/LTC3882-1
38821f
LTC3882-1
Applications Information
If an external source supplies VDD33 directly, the following
formula may be used to estimate the maximum average
power dissipation PD (in watts) of the LTC3882-1
PD = VDD33(0.024 + fPWM • 1.6e-5 + IRC25)
The maximum junction temperature of the LTC3882-1 in
°C may then be found from the following equation
Configuring Open-Drain Pins
The LTC3882-1 has the following open-drain pins:
3.3V Pins
1. PGOODn
2. FAULTn
TJ = TA + 33 • PD
2. SYNC
with ambient temperature TA expressed in °C
3. SHARE_CLK
Derating EEPROM Retention at Temperature
EEPROM read operations between 85°C and 125°C will not
affect data storage. But retention will be degraded if the
EEPROM is written above 85°C or stored above 125°C. If
an occasional fault log is generated above 85°C, the slight
reduction in data retention in the EEPROM fault log area
will not affect the use of the function or other EEPROM
storage. See the Operation section for other high temperature EEPROM functional details. Degradation in data
can be approximated by calculating the dimensionless
acceleration factor using the following equation.
⎡ ⎛ Ea ⎞ ⎛
⎞⎤
1
1
−
⎢⎜ ⎟ •⎜
⎥
⎝ k ⎠ ⎝ TUSE +273 TSTRESS +273 ⎟⎠ ⎥
⎦
AF = e ⎢⎣
Where:
(These pins operate correctly when pulled to 3.3V.)
1. RUNn
2. ALERT
3. SCL
4. SDA
All of the above pins have on-chip pull-down transistors
that can sink 3mA at 0.4V. The low-state threshold on
these pins is 1.4V, so ample noise margin exists with 3mA
of current. For 3.3V pins, 3mA of current is produced by
a 1.1k pull-up resistor. Unless there are transient speed
issues associated with the RC time constant of the net, a
10k resistor or larger is generally recommended.
The pull-up resistor for PGOOD should be terminated to
the LTC3882-1 VDD33 pin or a separate bias supply under
3.6V that is up before the LTC3882-1 is enabled. Otherwise,
power-not-good may be falsely indicated after the PWM
outputs are running.
AF = acceleration factor
Ea = activation energy = 1.4eV
k = 8.617 • 10–5 eV/°K
TUSE = is the specified junction temperature
TSTRESS = actual junction temperature in °C
As an example, if the device is stored at 130°C for 10 hours,
TSTRESS = 130°C, and
5V-Capable Pins
⎡⎛
1 ⎞⎤
1.4
⎞ ⎛ 1
−
•
⎢ ⎜⎝
–5 ⎟⎠ ⎜⎝ 398 403 ⎟⎠ ⎥
⎦ = 1.66
AF = e ⎣ 8.617•10
indicating the effect is the same as operating the device at
125°C for 10 • 1.66 = 16.6 hours, resulting in a retention
derating of 6.6 hours.
The SYNC pin has an on-chip pull-down transistor with
the output held low for nominally 250ns when driven by
the LTC3882-1. If the internal oscillator is set for 500kHz
and the load is 100pF with a 1/3 rise time required, the
resistor calculation is as follows:
RPULLUP =
2µs – 250ns
= 5.83k
3 •100pF
The closest 1% resistor is 5.76k.
If timing errors are occurring or if the SYNC amplitude
is not as large as required, monitor the waveform and
determine if the RC time constant is too long for the
38821f
For more information www.linear.com/LTC3882-1
63
LTC3882-1
Applications Information
application. If possible reduce the parasitic capacitance.
Otherwise reduce the pull-up resistor sufficiently to assure
proper operation.
The SHARE_CLK output has a nominal period of 10μs
and is pulled low for about 1μs. If the system load on this
shared line is 100pF, the resistor calculation for this line
with a 1/3 rise time is:
RPULLUP =
9µs
= 30k
3 •100pF
The closest 1% resistor is 30.1k.
For high speed signals such as SDA, SCL and SYNC, a
lower value resistor may be required. The RC time constant
should be set to 1/3 to 1/5 the required rise time to avoid
timing issues. For a 100pF load and a 400kHz PMBus communication rate, the resistor pull-up on the SDA and SCL
pins with the time constant set to 1/3 the rise time equals
RPULLUP =
tRISE
= 1k
3 •100pF
The closest 1% resistor value is 1k.
Be careful to minimize parasitic capacitance on the SDA
and SCL lines to avoid communication problems. To
estimate the loading capacitance, monitor the signal in
question and measure how long it takes for the desired
signal to reach approximately 63% of the output value.
This is one time constant.
PMBus Communication and Command Processing
The LTC3882-1 has a one deep buffer to hold the last data
written for each supported command prior to processing,
as shown in Figure 45. Two distinct parallel sections of
the LTC3882-1 manage command buffering and command
processing to ensure the last data written to any command
is never lost. When the part receives a new command
from the bus, command data buffering copies the data
into the write command data buffer and indicates to the
internal processor that data for that command should be
handled. The internal processor runs in parallel and performs the sometimes slower task of fetching, converting
(to internal format) and executing commands so marked
for processing.
64
Some computationally intensive commands (e.g., timing
parameters, temperatures, voltages and currents) have
internal processor execution times that may be long
relative to PMBus timing. If the part is busy processing
a command, and a new command(s) arrives, execution
may be delayed or processed in a different order than
received. The part indicates when internal calculations
are in process with bit 5 of MFR_COMMON (LTC3882-1
Calculations Not Pending). When the internal processor is busy calculating, bit 5 is cleared. When this bit is
set, the part is ready for another command. An example
polling loop is provided in Figure 46, which ensures that
commands are processed in order while simplifying error
handling routines. MFR_COMMON always returns valid
data at PMBus speeds between 10kHz and 400kHz.
When the part receives a new command while it is busy,
it will communicate this condition using standard PMBus
protocol. Depending on part configuration it may either
NACK the command or return all ones (0xFF) for reads. It
may also generate a BUSY fault and ALERT notification,
or stretch the SCL clock low. For more information refer
to PMBus Specification V1.2, Part II, Section 10.8.7 and
SMBus V2.0 section 4.3.3. Clock stretching can be enabled
by asserting bit 1 of MFR_CONFIG_ALL_LTC3882-1.
Clock stretching will only occur if enabled and the bus
communication speed exceeds 100kHz.
PMBus protocols for busy devices are well accepted
standards but can make writing system level software
somewhat complex. The part provides three handshaking
status bits which reduce this complexity while enabling
robust system level communication. The three hand
CMD
PMBus
WRITE
WRITE COMMAND
DATA BUFFER
DECODER
PAGE
CMDS
DATA
MUX
CALCULATIONS
PENDING
S
R
0x00
INTERNAL
PROCESSOR
0x21
FETCH,
CONVERT
DATA
AND
EXECUTE
•
•
•
VOUT_COMMAND
•
•
•
MFR_RESET
0xFD
x1
38821 F45
Figure 45. Write Command Data Processing
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
shaking status bits are in the MFR_COMMON register.
When the part is busy executing an internal operation, it
will clear bit 6 of MFR_COMMON (LTC3882-1 Not BUSY).
When internal calculations are in process, the part will
clear bit 5 of MFR_COMMON (LTC3882-1 Calculations Not
Pending). When the part is busy specifically because it is
// wait until bits 6, 5, and 4 of MFR_COMMON are all set
do
{
mfrCommonValue = PMBUS_READ_BYTE(0xEF);
partReady = (mfrCommonValue & 0x68) == 0x68;
}while(!partReady)
// now the part is ready to receive the next command
PMBUS_WRITE_WORD(0x21, 0x2000); //write VOUT_COMMAND to 2V
Figure 46. Example of a Polling Loop to Write VOUT_COMMAND
transitioning VOUT (margining, off/on, moving to a new
VOUT_COMMAND, etc.) it will clear bit 4 of MFR_COMMON (LTC3882-1 Output Not In Transition). These three
status bits can be polled with a PMBus read byte of the
MFR_COMMON register until all three bits are set. A command immediately following all these status bits being set
will be accepted without a NACK, BUSY fault or ALERT
notification. The part can NACK commands for other
reasons, however, as required by the PMBus specification (e.g., an invalid command or data). An example of a
robust command write algorithm for the VOUT_COMMAND
register is provided in Figure 46.
It is recommended that all command writes be preceded
with such a polling loop to avoid the extra complexity of
dealing with busy behavior or unwanted ALERT notifications. A simple way to achieve this is to embed the polling
in subroutines to write command bytes and words. This
polling mechanism will allow system software to remain
clean and simple while robustly communicating with the
part. For a detailed discussion of these topics and other
special cases please refer to the application note section
located at www.linear.com/designtools/app_notes.
When communicating using bus speeds at or below
100kHz, the polling mechanism previously shown provides a simple solution that ensures robust communication without clock stretching. At bus speeds in excess
of 100kHz, it is strongly recommended that the part be
enabled to use clock stretching, requiring a PMBus mas-
ter that supports that function. Clock stretching does not
allow the LTC3882-1 to communicate reliably on busses
operating above 400kHz. Operating the LTC3882-1 with
PMBus SCL rates above 400kHz is not recommended.
System software that detects and properly recovers from
the standard PMBus NACK responses or BUSY faults
described in PMBus Specification V1.2, Part II, Section
10.8.7 is required to communicate above 100kHz without
clock stretching.
Refer to Application Note 135 for techniques that may
also apply to implement a robust PMBus interface to the
LTC3882-1.
Status and Fault Log Management
Due to internal operation, very infrequently the LS byte of
STATUS_WORD may be inconsistent with the state of bits
in the MS byte. This condition is quite transient and can
normally be resolved by simply re-reading STATUS_WORD.
If power is lost during an internal store of a fault log to
EEPROM, a partial write of the log can result. In this situation, the LTC3882-1 will not indicate that a fault log is
present the next time adequate supply voltage is applied
(bit 3 of STATUS_MFR_SPECIFIC). The existence of a
partial fault log can be detected by examining the header
of the log (MFR_FAULT_LOG). If the first two words of
the Fault Log Preface contain valid data as specified by
Table 2, and STATUS_MFR_SPECIFIC does not indicate the
presence of a complete fault log, then a partial log existed
in EEPROM at boot and has been retrieved to RAM. The
only way to then determine how much of the log is actually
valid is by subjective evaluation of the contents of each log
event record. MFR_CLEAR_FAULT_LOG will permanently
erase a partial fault log, allowing a subsequent log to be
written. It is a good practice to always check for a partial
fault log at power-up if fault logging is enabled (bit 7 of
MFR_CONFIG_ALL_LTC3882-1).
LTpowerPlay – An Interactive Digital Power GUI
LTpowerPlay is a powerful Windows-based development
environment that supports Linear Technology Power System Management ICs, including the LTC3882-1. LTpowerPlay can be used to evaluate LTC products by connecting
to a Linear Technology demo circuit or user application.
38821f
For more information www.linear.com/LTC3882-1
65
LTC3882-1
Applications Information
LTpowerPlay can also be used offline (no hardware present) to build multiple IC configuration files that can be
saved and later reloaded. LTpowerPlay uses the DC1613
USB-to-I2C/SMBus/PMBus controller to communicate
with a system for evaluation, development or debug. The
software also features automatic update to remain upto-date with the latest device drivers and documentation
available from Linear Technology. A great deal of context
sensitive help is available within LTpowerPlay, along with
several tutorials. Complete information is available at http://
www.linear.com/ltpowerplay.
Interfacing to the DC1613
The LTC DC1613 USB-to-I2C/SMBus/PMBus controller
can be interfaced to the LTC3882-1 on any board for
programming, telemetry and system debug. This includes
the DC1936 from Linear Technology, or any customer
target system. The controller, when used in conjunction
with LTpowerPlay, provides a powerful way to debug an
entire power system. Faults are quickly diagnosed using
telemetry, fault status registers and the fault log. The final
configuration can be quickly developed and stored to the
LTC3882-1 EEPROM and/or LTpowerPlay configuration file.
The DC1613 can communicate with, program and even
power one or more LTC3882-1s, regardless of whether
system supplies are up. The DC2086 Powered Programming Adapter can be used to extend the power sourcing
capability of the DC1613. Figure 47 illustrates an application schematic for in-system programming of multiple
LTC3882-1s normally powered from a VCC system supply
(5V to 12V). If VCC is applied, the DC1613 will not supply
the LTC3882-1s on the board. If the DC2086 is used, PFETs
with lower RDS(ON), such as the SiA907EDJT, should be
used in place of the TP0101K devices. Figure 48 shows an
example when the system normally provides 3.3V directly
to the LTC3882-1(s).
If system supplies are not up in either of these circuits, the
DC1613 will power the LTC3882-1 VDD33 supply, allowing
in-circuit configuration or manufacturing customization.
These circuits also facilitate remote diagnostics, control
and reprogramming of the LTC3882-1 while the host
system is fully operational, permitting very flexible insystem debugging.
66
If the system supply is restored while power is still applied
by the DC1613 or DC2086, the LTC3882-1 can often be
ready to initiate output soft-start before sufficient supply
bias for the power stage has been established. Create
additional LTC3882-1 delay with TON_DELAY or use a
common system RUN line to control both the LTC3882-1
and its related power stages based on acceptable operating parameters, as shown in Figure 55. The DC1613 I2C
connections are opto-isolated from the host PC USB. The
DC1613 3.3V current limit is only 100mA, so it should
only be used to power one or two LTC3882-1s in-system.
Because of this limited current sourcing capability, only
the LTC3882-1s, their associated pull-up resistors and the
I2C pull-up resistors should be powered from the isolated
3.3V supply provided by the DC1613. Using the DC2086
will enable in-system programming of several tens of
LTC3882-1 devices without normal system power applied.
Any other device sharing the I2C bus with the LTC3882-1
should not have body diodes between their SDA/SCL pins
and their respective logic supply, because this will interfere
with bus communication in the absence of system power.
Hold the RUN pins low externally to avoid providing power
to the load until the part is fully configured.
Design Example
As a design example, consider a 132W 2-phase application such as the one shown in Figure 53, where VIN = 36V,
VOUT = 3.3V, and IOUT = 40A. A fully discrete power stage
design is employed to allow better optimization given these
demanding requirements. Assume that a secondary 5V
supply will be available in the system for the LTC3882-1
VCC supply. The necessary local bypassing is then provided for the VDD33 (2.2µF) and VDD25 (1µF) LDO outputs.
These LDO outputs should not be shared with other ICs
that might have outputs of the same name, because they
have independent, internal control loops. When VDD33
is used as the LTC3882-1 supply input, it may be shared
with other ICs operating from that 3.3V supply. Local HF
bypassing of at least 0.1µF is still required on VDD33 in
this case.
First, the regulated output is established by programming
the VOUT_COMMAND stored in EEPROM to 3.3V.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
SYSTEM
VCC
TP0101K
102k
LT6703-2
–IN
VS
10k
VCC
TP0101K
–
OUT
10k
+
2N2002
10k
VDD33
2.2µF
10k
VDD25
LTC3882-1
1µF
SDA
10k
SCL
GND
400mV
REFERENCE
GND
LTC
CONTROLLER
HEADER
VCC
4.7µF
ISOLATED 5V
2.2µF
SDA
SCL
VDD33
VDD25
LTC3882-1
1µF
SDA
SCL
GND
TO LTC DC1613
USB TO I2C/SMBus/PMBus
CONTROLLER
38821 F47
Figure 47. DC1613 Connection (VCC Supply)
38821f
For more information www.linear.com/LTC3882-1
67
LTC3882-1
Applications Information
ENBA
STAT
GND
ENBB
INB
SYSTEM
3.3V
LTC4413
VCC
OUTB
VDD33
0.1µF
CONTROL CIRCUIT
10k
10k
INA
OUTA
VDD25
LTC3882-1
1µF
SDA
SCL
GND
LTC
CONTROLLER
HEADER
VCC
ISOLATED 3.3V
SDA
0.1µF
SCL
VDD33
VDD25
LTC3882-1
1µF
SDA
SCL
GND
TO LTC DC1613
USB TO I2C/SMBus/PMBus
CONTROLLER
38821 F48
Figure 48. DC1613 Connections (VDD33 Supply)
68
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
Applications Information
The frequency and phase are also set by EEPROM values.
Assume that solution footprint or vertical clearance is an
issue, so operating frequency will need to be increased
in an effort to minimize inductor value (size). This choice
could also result from the need to have above average
transient performance, although efficiency may be reduced slightly. FREQUENCY_SWITCH is set to 1.0MHz.
As a 2-phase system, MFR_PWM_CONFIG_LTC3882-1
is programmed to 0x14 to put Channel 0 phase at 0° and
Channel 1 phase at 180°. This produces the lowest input
ripple possible with this configuration and allows this
output to synchronize with other rails via SHARE_CLK.
The design will plan on a nominal output ripple of 70% of
IOUT to minimize the magnetics volume, and the inductance
value is chosen based on this assumption. Each channel
supplies an average 20A to the output at full load, resulting in a ripple of 14AP-P. A 200nH inductor per phase
would create this peak-to-peak ripple at 1.0MHz. A Pulse
PA0513.22LT 210nH inductor with a DCR of 0.32mΩ
typical is selected. Setting IOUT_FAULT_LIMIT to 35A per
phase leaves plenty of headroom for transient conditions
while still adequately protecting against the rated inductor
saturation current of 45A at temperature.
For top and bottom power FETs, the 40V rated Infineon
BSC050N04LSG and BSC010N04LS are chosen, respectively. These afford both low RDS(ON) and low gate charge
QG. Two of each of these could be paralleled to achieve
improved efficiency at full load, if desired.
The LTC4449 gate driver is chosen for its fast response
(13ns), suitable gate drive, VIN capability (38V) and the
ease with which it can be interfaced to the LTC3882-1. Basic
three-state control, CCM operation, fast boost refresh, low
VOUT range and digital output voltage servo are selected
by programming MFR_PWM_MODE_LTC3882-1 to 0xC0
for both channels.
For input filtering, a 47μF SUNCON capacitor and four
22μF ceramic capacitors are selected to provide acceptable AC impedance against the designed converter ripple
current. Four 470μF 9mΩ POSCAPs and two 100μF
ceramic capacitors are chosen for the output to maintain
supply regulation during severe transient conditions and
to minimize output voltage ripple.
A loop crossover frequency of 100kHz provides good
transient performance while still being well below the
switching frequency of the converter. The values of R29,
R30 and C25 to C27 were determined to produce a nominal system phase margin of about 65° at this bandwidth.
For the DCR sense filter network, R = 3.09k and C = 220nF
are chosen to match the L/DCR time constant of the inductor. PolyPhase connections (IAVG, et al) are shown in the
schematic to ensure good output current sharing between
the two power stages.
External temperature sense will employ an accurate ΔVBE
method, and Q1 and Q2 serve to sense the temperature of
L1 and L2, respectively. These components will be located
immediately adjacent to their chokes and the 10nF filter
capacitors placed with the BJTs.
Resistor configuration is used on the ASELn pins to program PMBus address (MFR_ADDRESS) to 0x4C. Each
LTC3882-1 must be configured for a unique address.
Using both ASELn pins to accomplish this programming
is recommended for simpliest in-system programming.
Check the selected address to avoid collision with global
addresses other any other specific devices. Identical
MFR_RAIL_ADDRESS can be set in EEPROM for both
channels to allow single-command control of common
rail parameters such as IOUT_OC_FAULT_LIMIT. The
LTC3882-1 also responds to 7-bit global addresses 0x5A
and 0x5B. MFR_ADDRESS and MFR_RAIL_ADDRESS
should not be set to either of these values.
PMBus connection (three signals), as well as shared
RUN control and fault propagation (FAULT) are provided.
SYNC can be used to synchronize other PWMs to this
rail if required.
Pull-ups are provided on all these shared open-drain signals
assuming a maximum 100pF line load and PMBus rate of
100kHz. These pins should not be left floating. Termination to 3.3V ensures the absolute maximum ratings for
the pins are not exceeded. All other operating parameters
such as soft start/stop and desired faults responses are
programmed via PMBus command values stored in internal
LTC3882-1 EEPROM.
38821f
For more information www.linear.com/LTC3882-1
69
LTC3882-1
PMBus COMMAND DETAILS
(by functional groups)
Addressing and Write Protect
COMMAND NAME
CMD
CODE DESCRIPTION
DATA
PAGED FORMAT
PAGE
0x00
Channel (page) presently selected for any
paged command.
PAGE_PLUS_WRITE
0x05
Write a command directly to a specified page.
PAGE_PLUS_READ
0x06
Read a command directly from a specified
page.
WRITE_PROTECT
0x10
Protect the device against unintended PMBus
modifications.
R/W Byte
N
MFR_ADDRESS
0xE6
Specify right-justified 7-bit device address.
R/W Byte
MFR_RAIL_ADDRESS
0xFA
Specify unique right-justified 7-bit address
for channels comprising a PolyPhase output.
R/W Byte
TYPE
R/W Byte
N
W Block
N
Block R/W
Process
N
UNITS NVM
Reg
DEFAULT
VALUE
0x00
0x00
Reg
l
N
Reg
l
0x4F
Y
Reg
l
0x80
Related commands: MFR_COMMON.
PAGE
The PAGE command provides the ability to configure, control and monitor both PWM channels through only one
physical address, either the MFR_ADDRESS or GLOBAL device address. Each PAGE contains the operating memory
for one PWM channel.
Pages 0x00 and 0x01 correspond to Channel 0 and Channel 1, respectively, in this device.
Setting PAGE to 0xFF applies any following paged commands to both outputs. With PAGE set to 0xFF the LTC3882-1
will respond to read commands as if PAGE were set to 0x00 (Channel 0 results).
This command has one data byte.
PAGE_PLUS_WRITE
The PAGE_PLUS_WRITE command provides a way to set the page within a device, send a command and then send
the data for the command, all in one communication packet. Commands allowed by the present write protection level
may be sent with PAGE_PLUS_WRITE.
The value stored in the PAGE command is not affected by PAGE_PLUS_WRITE. If PAGE_PLUS_WRITE is used to send
a non-paged command, the Page Number byte is ignored.
This command uses Write Block protocol. An example of the PAGE_PLUS_WRITE command with PEC sending a command that has two data bytes is shown in Figure 49.
1
7
1
1
S
SLAVE
ADDRESS
W
A
8
1
PAGE_PLUS
A
COMMAND CODE
8
LOWER DATA
BYTE
8
1
8
1
8
1
BLOCK COUNT
(= 4)
A
PAGE
NUMBER
A
COMMAND
CODE
A
1
8
1
8
1
1
A
UPPER DATA
BYTE
A
PEC BYTE
A
P
…
38821 F49
Figure 49. Example of PAGE_PLUS_WRITE
70
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Addressing and Write Protect)
PAGE_PLUS_READ
The PAGE_PLUS_READ command provides the ability to set the page within a device, send a command and then read
the data returned by the command, all in one communication packet .
The value stored in the PAGE command is not affected by PAGE_PLUS_READ. If PAGE_PLUS_READ is used to access
data from a non-paged command, the Page Number byte is ignored.
This command uses Block Write – Block Read Process Call protocol. An example of the PAGE_PLUS_READ command
with PEC is shown in Figure 50.
NOTE: PAGE_PLUS commands cannot be nested. A PAGE_PLUS command cannot be used to read or write another
PAGE_PLUS command. If this is attempted, the LTC3882-1 will NACK the entire PAGE_PLUS packet and issue a CML
fault for Invalid/Unsupported Data.
1
7
S
SLAVE
ADDRESS
1
7
Sr
SLAVE
ADDRESS
1
1
W
PAGE_PLUS
A
A
COMMAND CODE
1
R
8
1
8
A
BLOCK COUNT
(= 2)
1
8
BLOCK COUNT
(= 2)
1
8
A
LOWER DATA
BYTE
1
8
A
PAGE
NUMBER
1
8
1
A
COMMAND
CODE
A
1
8
1
8
A
UPPER DATA
BYTE
A
PEC BYTE
1
…
1
NA P
38821 F50
Figure 50. Example of PAGE_PLUS_READ
WRITE_PROTECT
The WRITE_PROTECT command is used to control PMBus write access to the LTC3882-1.
Supported Values:
VALUE
MEANING
0x80
Disable all writes except WRITE_PROTECT, PAGE, STORE_USER_ALL and MFR_EE_UNLOCK commands.
0x40
Disable all writes except WRITE_PROTECT, PAGE, STORE_USER_ALL, MFR_EE_UNLOCK, OPERATION, CLEAR_PEAKS and CLEAR_FAULTS
commands. Individual faults can also be cleared by writing a 1 to the respective status bit.
0x20
Disable all writes except WRITE_PROTECT, PAGE, STORE_USER_ ALL, MFR_EE_UNLOCK, OPERATION, CLEAR_PEAKS, CLEAR_FAULTS,
ON_OFF_CONFIG and VOUT_COMMAND commands. Individual faults can be cleared by writing a 1 to the respective status bit.
0x00
Enables writes to all commands.
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
71
LTC3882-1
PMBus COMMAND DETAILS
(Addressing and Write Protect)
MFR_ADDRESS
The MFR_ADDRESS command sets the seven bits of the PMBus device address for this unit (right justified).
Setting this command to a value of 0x80 disables device-level addressing. The GLOBAL device addresses 0x5A and
0x5B cannot be disabled. The LTC3882-1 always responds at these addresses. Even if bit 6 of MFR_CONFIG_ALL_
LTC3882-1 is set to ignore the device resistor configuration pins, any valid address, or portion of an address, specified
with external resistors on ASEL0 or ASEL1 is applied. If both of these pins are open, the device address is determined
strictly by the MFR_ADDRESS value stored in EEPROM. Refer to the Operation section on Resistor Configuration Pins
for additional details.
This command has one data byte.
MFR_RAIL_ADDRESS
The MFR_RAIL_ADDRESS command sets a direct 7-bit PMBus address (right justified) for the active channel(s) as
determined by the PAGE command. This address should be common to all channels attached to a single power supply
rail. Setting this command to a value of 0x80 disables rail device addressing for the selected channel. Only command
writes should be made to the rail address. If a read is performed from this address, a CML fault may result.
This command has one data byte.
General Device Configuration
COMMAND NAME
CMD
CODE DESCRIPTION
TYPE
DATA
PAGED FORMAT
PMBUS_REVISION
0x98
Supported PMBus version.
R Byte
Y
Reg
0x22
V1.2
CAPABILITY
0x19
Summary of supported optional PMBus
features.
R Byte
N
Reg
0xB0
MFR_CONFIG_ALL_LTC3882-1
0xD1 LTC3882-1 device-level configuration.
R/W Byte
N
Reg
UNITS NVM
l
DEFAULT
VALUE
0x01
PMBUS_REVISION
The PMBUS_REVISION command returns the revision of the PMBus Specification that the device supports. The
LTC3882-1 is compliant with PMBus Version 1.2, both Part I and Part II.
This read-only command has one data byte.
CAPABILITY
The CAPABILITY command reports some key LTC3882-1 features to the PMBus host device.
The LTC3882-1 supports packet error checking, 400kHz bus speeds and has an ALERT output.
This read-only command has one data byte.
72
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(General Device Configuration)
MFR_CONFIG_ALL_LTC3882-1
The MFR_CONFIG_ALL_LTC3882-1 command provides device-level configuration common to multiple LTC PMBus
products.
Bit Definitions:
BIT
MEANING
7
Enable fault logging.
6
Ignore resistor configuration pins. Does not apply to ASEL0 or ASEL1.
5
Disable CML fault for Quick Command message.
4
Disable SYNC output.
3
Enable 255ms PMBus timeout.
2
Require valid PEC for PMBus write.
1
Enable PMBus clock stretching.
0
Execute CLEAR_FAULTS on rising edge of either RUN pin.
If a legal command is received with an invalid PEC, the LTC3882-1 will not execute the command, regardless of the state
of bit 2. If clock stretching is enabled, the LTC3882-1 only uses it as required, generally above SCL rates of 100kHz.
This command has one data byte.
On, Off and Margin Control
COMMAND NAME
CMD
CODE DESCRIPTION
ON_OFF_CONFIG
0x02
RUN pin and PMBus on/off command
configuration.
OPERATION
0x01
On, off and margin control.
MFR_RESET
0xFD Force full reset without removing power.
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
Y
Reg
l
0x1E
R/W Byte
Y
Reg
l
0x80
Send Byte
N
R/W Byte
ON_OFF_CONFIG
The ON_OFF_CONFIG command specifies the combination of RUNn pin input state and PMBus commands needed to
turn the PWM channel on and off.
Supported Values:
VALUE
MEANING
0x1F
OPERATION value and RUNn pin must both command the device to start/run. Device executes immediate off when commanded off.
0x1E
OPERATION value and RUNn pin must both command the device to start/run. Device uses TOFF_ command values when commanded off.
0x17
RUNn pin control with immediate off when commanded off. OPERATION on/off control ignored.
0x16
RUNn pin control using TOFF_ command values when commanded off. OPERATION on/off control ignored.
Programming an unsupported ON_OFF_CONFIG value will generate a CML fault and the command will be ignored.
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
73
LTC3882-1
PMBus COMMAND DETAILS
(On, Off and Margin Control)
OPERATION
The OPERATION command is used to turn the PWM channel on and off in conjunction with RUN pin hardware control.
This command may also be used to move the output voltage to margin levels. VOUT changes commanded by OPERATION
margin commands occur at the programmed VOUT_TRANSITION_RATE. The unit stays in the commanded operating
state until an OPERATION command or RUN pin voltage instructs the device to change to another state.
Execution of margin commands is delayed until any on-going TON_RISE or TOFF_FALL output sequencing is completed. Margin values are affected by AVP function, if enabled. Margin operations that ignore faults are not supported
by the LTC3882-1.
Supported Values:
VALUE
MEANING
0xA8
Margin high.
0x98
Margin low.
0x80
On (VOUT back to nominal even if bit 3 of ON_OFF_CONFIG is not set).
0x40*
Soft off (with sequencing).
0x00*
Immediate off (no sequencing).
*Device does not respond to these commands if bit 3 of ON_OFF_CONFIG is not set.
Programming an unsupported OPERATION value will generate a CML fault and the command will be ignored.
This command has one data byte.
MFR_RESET
This command provides a means to reset the LTC3882-1 from the serial bus. This forces the LTC3882-1 to turn off
both PWM channels, load the operating memory from internal EEPROM, clear all faults and then perform a soft-start
of both PWM channels, if enabled.
This write-only command has no data bytes.
74
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(PWM Configuration)
PWM Configuration
COMMAND NAME
CMD
CODE DESCRIPTION
FREQUENCY_SWITCH
0x33
PWM frequency control.
R/W Word
N
L11
MFR_PWM_CONFIG_LTC3882-1
0xF5
LTC3882-1 PWM configuration common to
both channels.
R/W Byte
N
MFR_CHAN_CONFIG_LTC3882-1
0xD0 LTC3882-1 channel-specific configuration.
R/W Byte
MFR_PWM_MODE_LTC3882-1
0xD4 LTC3882-1 channel-specific PWM mode
control.
R/W Byte
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
l
500kHz
0xFBE8
Reg
l
0x14
Y
Reg
l
0x1D
Y
Reg
l
0xC8
kHz
Related commands MFR_TEMP_1_GAIN_ADJUST, MFR_TEMP_1_OFFSET.
FREQUENCY_SWITCH
The FREQUENCY_SWITCH command sets the switching frequency of both LTC3882-1 PWM channels in kilohertz. At
most only one IC sharing SYNC should be programmed as clock master. See bit 4 in MFR_CONFIG_ALL_LTC3882-1.
FREQUENCY_SWITCH value will determine the free-running frequency of PWM operation if an expected external clock
source is not present or the bussed SYNC line becomes stuck due an external fault or conflict. Both PWM channels
must be turned off by the RUNn pins, OPERATION command, or their combination, to process this command. If this
command is sent while either PWM controller is operating, the LTC3882-1 will NACK the command byte, ignore the
command and its data, and assert a BUSY fault. A PLL Unlocked status may be reported after changing the value of
this command until the new frequency is established.
Supported Frequencies:
VALUE
PWM FREQUENCY (TYPICAL)
0x0A71
1.25MHz
0x03E8
1MHz
0x0384
900kHz
0x02EE
750kHz
0x0258
600kHz
0xFBE8
500kHz
0xFB84
450kHz
0xFB20
400kHz
0xFABC
350kHz
0XFA58
300kHz
0xF3E8
250kHz
0x0000
External SYNC Only
This command has two data bytes in Linear_5s_11s format.
38821f
For more information www.linear.com/LTC3882-1
75
LTC3882-1
PMBus COMMAND DETAILS
(PWM Configuration)
MFR_PWM_CONFIG_LTC3882-1
The MFR_PWM_CONFIG_LTC3882-1 command controls PWM-related clocking for the LTC3882-1. Both PWM channels must be turned off by the RUNn pins, OPERATION command, or their combination, to process this command. If
this command is sent while either PWM controller is operating, the LTC3882-1 will NACK the command byte, ignore
the command and its data, and assert a BUSY fault.
Supported Values:
BIT
MEANING
7
(Reserved, must write as 0).
6
(Reserved, must write as 0).
5
(Reserved).
4
SHARE_CLK configuration:
0: SHARE_CLK continuously enabled once VINSNS ≥ VIN_ON after initialization.
1: SHARE_CLK always forced low if VINSNS ≤ VIN_OFF, then held low until VINSNS ≥ VIN_ON.
3
2:0
(Reserved).
Phase
Channel 0
Channel 1
Maximum Duty
Cycle
111b
135°
315°
87.5%
110b
90°
270°
101b
45°
225°
100b
0°
180°
011b
120°
300°
010b
60°
240°
001b
0°
180°
000b
0°
120°
Value
83.3%
Phase is expressed from the falling edge of SYNC to the falling edge of PWM.
This command has one data byte.
76
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(PWM Configuration)
MFR_CHAN_CONFIG_LTC3882-1
The MFR_CHAN_CONFIG_LTC3882-1 command provides per-channel configuration common to multiple LTC PMBus
products.
Bit Definitions:
BIT
MEANING
7:5
(Reserved).
4
RUN pin control:
0: When the channel is commanded off, the associated RUN pin is pulsed low for TOFF_DELAY + TOFF_FALL + 136ms
(or MFR_RESTART_DELAY, if longer) regardless of the state of bit 3.
1: RUN pin is not pulsed low if channel is commanded off.
3
Short cycle control:
0: No special control. Device attempts to follow on/off commands exactly as issued.
1: Output is immediately disabled if commanded back on while waiting for TOFF_DELAY or TOFF_FALL to expire. A minimum off time
of 120ms is then enforced before the channel is turned back on. Additional delay will apply if bit 4 is clear.
2
SHARE_CLK output control:
0: No special control.
1: Output disabled if SHARE_CLK is held low.
1
(Reserved, must write as 0).
0
MFR_RETRY_DELAY control:
0: Output decay to 12.5% of programmed value required for retry after ANY action that turns off the rail.
1: Output decay not required for retry.
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
77
LTC3882-1
PMBus COMMAND DETAILS
(PWM Configuration)
MFR_PWM_MODE_LTC3882-1
The MFR_PWM_MODE_LTC3882-1 command sets important PWM controls for each channel. The addressed channel(s)
must be turned off by its RUN pin, OPERATION command, or their combination, when this command is issued. Otherwise the LTC3882-1 will NACK the command byte, ignore the command and its data, and assert a BUSY fault.
When bit 5 is cleared, the LTC3882-1 computes temperature in °C from ΔVBE measured by the ADC at the TSNSn pin as
T = (G • ΔVBE • q/(K • ln(16))) – 273.15 + O
When bit 5 is set, the LTC3882-1 computes temperature in °C from TSNSn voltage measured by the ADC as
T = (G • (1.35 – VTSNSn + O)/4.3e-3) + 25
For both equations,
G = MFR_TEMP_1_GAIN • 2–14, and
O = MFR_TEMP_1_OFFSET
Supported Values:
BIT
7
MEANING
Output voltage range select:
0: Maximum VOUT = 5.25V.
1: Maximum VOUT = 2.65V.
6*
Enable VOUT servo.
5
External temperature sense:
0: ΔVBE measurement.
1: Direct voltage measurement.
4:3
BOOST refresh width:
11b: 250ns
10b: 125ns
01b: 50ns
00b: 25ns
2
(Reserved).
1
PWM control protocol:
0: 3-State PWM output.
1: 3-State PWM output with no DCM (including soft-start) or hardware ROC response (including OV).
0
PWM mode:
0: Forced continuous inductor current.
1: Discontinuous inductor current.
*This bit is ignored (servo disabled) if MFR_VOUT_AVP for this channel is programmed to a value greater than 0.0%.
This command has one data byte.
78
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Input Voltage and Limits)
Input Voltage and Limits
COMMAND NAME
CMD
CODE DESCRIPTION
VIN_ON
0x35
Minimum input voltage to begin power
conversion.
R/W Word
N
L11
V
l
6.5V
0xCB40
VIN_OFF
0x36
Decreasing input voltage at which power
conversion stops.
R/W Word
N
L11
V
l
6.0V
0xCB00
VIN_OV_FAULT_LIMIT
0x55
VIN overvoltage fault limit.
R/W Word
N
L11
V
l
15.5V
0xD3E0
VIN_UV_WARN_LIMIT
0x58
VIN undervoltage warning limit.
R/W Word
N
L11
V
l
6.3V
0xCB26
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
Related commands: STATUS_INPUT, SMBALERT_MASK, READ_VIN, VIN_OV_FAULT_RESPONSE
VIN_ON
The VIN_ON command sets the input voltage, in volts, required to start power conversion.
This command has two data bytes in Linear_5s_11s format.
VIN_OFF
The VIN_OFF command sets the minimum input voltage, in volts, at which power conversion stops.
This command has two data bytes in Linear_5s_11s format.
VIN_OV_FAULT_LIMIT
The VIN_OV_FAULT_LIMIT command sets the value of the input voltage measured by the ADC, in volts, that causes
an input overvoltage fault.
This command has two data bytes in Linear_5s_11s format.
VIN_UV_WARN_LIMIT
The VIN_UV_WARN_LIMIT command sets the value of input voltage measured by the ADC that causes an input undervoltage warning. This warning is disabled until the input exceeds the input startup threshold value set by the VIN_ON
command and the unit has been enabled. If the VIN_UV_WARN_LIMIT is then exceeded, the device:
• Sets the INPUT Bit in the STATUS_WORD
• Sets the VIN Undervoltage Warning Bit in the STATUS_INPUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
38821f
For more information www.linear.com/LTC3882-1
79
LTC3882-1
PMBus COMMAND DETAILS
(Output Voltage and Limits)
Output Voltage and Limits
COMMAND NAME
CMD
CODE DESCRIPTION
TYPE
VOUT_MODE
0x20
Output voltage format and exponent.
R Byte
Y
Reg
VOUT_COMMAND
0x21
Nominal VOUT value.
R/W Word
Y
L16
V
l
1.0V
0x1000
MFR_VOUT_MAX
0xA5 Maximum value of any VOUT-related
command.
R Word
Y
L16
V
l
5.6V
0x599A
VOUT_MAX
0x24
R/W Word
Y
L16
V
l
5.5V
0x5800
MFR_VOUT_AVP
0xD3 Specify VOUT load line.
R/W Word
Y
L11
%
l
0%
0x8000
VOUT_MARGIN_HIGH
0x25
VOUT at high margin, must be greater than
VOUT_COMMAND.
R/W Word
Y
L16
V
l
1.05V
0x10CD
VOUT_MARGIN_LOW
0x26
VOUT at low margin, must be less than
VOUT_COMMAND.
R/W Word
Y
L16
V
l
0.95V
0x0F33
VOUT_OV_FAULT_LIMIT
0x40
VOUT overvoltage fault limit.
R/W Word
Y
L16
V
l
1.1V
0x119A
VOUT_OV_WARN_LIMIT
0x42
VOUT overvoltage warning limit.
R/W Word
Y
L16
V
l
1.075V
0x1133
VOUT_UV_WARN_LIMIT
0x43
VOUT undervoltage warning limit.
R/W Word
Y
L16
V
l
0.925V
0x0ECD
VOUT_UV_FAULT_LIMIT
0x44
VOUT undervoltage fault limit.
R/W Word
Y
L16
V
l
0.9V
0x0E66
Maximum VOUT that can be set by any
command, including margin.
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
2–12
0x14
Related commands: OPERATION, STATUS_WORD, STATUS_VOUT, SMBALERT_MASK, READ_VOUT, MFR_VOUT_PEAK, READ_POUT, VOUT_OV_FAULT_
RESPONSE, VOUT_UV_FAULT_RESPONSE
VOUT_MODE
The VOUT_MODE command gives the format used by the LTC3882-1 for output voltage related commands. Only Linear
Mode is supported, with a mantissa expressed in microvolts. Sending the VOUT_MODE command to the LTC3882-1
using a write protocol will result in a CML fault.
This read-only command has one data byte.
VOUT_COMMAND
The VOUT_COMMAND is used to set the output voltage in volts (no load value if AVP is enabled). Execution of this
command is delayed until any on-going TON_RISE or TOFF_FALL output sequencing is completed, otherwise the
output voltage moves to a new value at VOUT_TRANSITION_RATE.
This command has two data bytes in Linear_16u format.
MFR_VOUT_MAX
The MFR_VOUT_MAX command returns the maximum value, in volts, allowed for any VOUT-related command, including VOUT_OV_FAULT_LIMIT. This value represents the maximum regulated voltage the selected channel could be
capable of producing.
This read-only command has two data bytes in Linear_16u format.
80
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Output Voltage and Limits)
VOUT_MAX
The VOUT_MAX command sets an upper limit, in volts, on the allowed value of any command that sets the output
voltage, including VOUT_MARGIN_HIGH. Setting VOUT_MAX to a value greater than MFR_VOUT_MAX will result in
a CML fault and VOUT_MAX will be set to the value of MFR_VOUT_MAX. A VOUT_MAX warning may also be generated if VOUT_MAX is set above 5.5V in output range 0 or above 2.75V in range 1. This command ensures that any
combination of commands attempting to set VOUT above VOUT_MAX will result in a warning with the output clamped
at VOUT_MAX. When a VOUT_MAX warning occurs, the device takes the following actions:
• Sets The Offending Command Value to the Voltage Specified by VOUT_MAX
• Sets the VOUT Bit in the STATUS_WORD
• Sets the VOUT_MAX Warning Bit in the STATUS_VOUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has two data bytes in Linear_16u format.
MFR_VOUT_AVP
The MFR_VOUT_AVP command sets the change in output voltage, in percent, for a full-scale change in output current.
MFR_VOUT_AVP can be used for active voltage positioning (AVP) requirements or passive current sharing schemes.
The LTC3882-1 interprets the IOUT_OC_WARN_LIMIT value as full-scale current for AVP. If MFR_VOUT_AVP is nonzero, VOUT_COMMAND sets the maximum, no-load output voltage and servo mode for that channel is automatically
disabled. Setting MFR_VOUT_AVP to 0.0% automatically disables the AVP function. Refer to the Applications Information section for additional details on range and resolution when using MFR_VOUT_AVP.
This command has two data bytes in Linear_5s_11s format.
VOUT_MARGIN_HIGH
The VOUT_MARGIN_HIGH command programs the output voltage, in volts, to be produced when Margin High is set
with the OPERATION command (no load value if AVP is enabled). The value must be greater than VOUT_ COMMAND.
This command has two data bytes in Linear_16u format.
VOUT_MARGIN_LOW
The VOUT_MARGIN_LOW command programs the output voltage, in volts, to be produced when Margin Low is set
by the OPERATION command (no load value if AVP is enabled). The value must be less than VOUT_COMMAND.
This command has two data bytes in Linear_16u format.
VOUT_OV_FAULT_LIMIT
The VOUT_OV_FAULT_LIMIT command sets the value of the output voltage measured by the OV supervisor at the VSENSE±
pins, in volts, which causes an output overvoltage fault. If VOUT_OV_FAULT_LIMIT is modified while the channel is on,
10ms should be allowed for the new value to take effect. Modifying VOUT during that time can result an erroneous OV
fault. The LTC3882-1 sets MFR_COMMON bits[6:5] low while it establishes the new VOUT_OV_FAULT_LIMIT value.
This command has two data bytes in Linear_16u format.
38821f
For more information www.linear.com/LTC3882-1
81
LTC3882-1
PMBus COMMAND DETAILS
(Output Voltage and Limits)
VOUT_OV_WARN_LIMIT
The VOUT_OV_WARN_LIMIT command sets the value, in volts, of the output voltage measured by the ADC at the
VSENSE± pins that causes an output overvoltage warning. If the VOUT_OV_WARN_LIMIT is exceeded, the device:
• Sets the VOUT Bit in the STATUS_WORD
• Sets the VOUT Overvoltage Warning Bit in the STATUS_VOUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has two data bytes in Linear_16u format.
VOUT_UV_WARN_LIMIT
The VOUT_UV_WARN_LIMIT command sets the value, in volts, of the output voltage measured by the ADC at the
VSENSE± pins that causes an output undervoltage warning. If the VOUT_UV_WARN_LIMIT is exceeded, the device:
• Sets the VOUT Bit in the STATUS_WORD
• Sets the VOUT Undervoltage Warning Bit in the STATUS_VOUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has two data bytes in Linear_16u format.
VOUT_UV_FAULT_LIMIT
The VOUT_UV_FAULT_LIMIT command sets the value of the output voltage measured by the UV supervisor at the
VSENSE± pins, in volts, which causes an output undervoltage fault.
This command has two data bytes in Linear_16u format.
82
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Output Current and Limits)
Output Current and Limits
COMMAND NAME
CMD
CODE DESCRIPTION
IOUT_CAL_GAIN
0x38
Ratio of ISENSE± voltage to sensed current.
R/W Word
Y
L11
mΩ
l
0.63mΩ
0xB285
MFR_IOUT_CAL_GAIN_TC
0xF6
Output current sense element temperature
coefficient.
R/W Word
Y
CF
ppm/°C
l
3900ppm/°C
0x0F3C
IOUT_OC_FAULT_LIMIT
0x46
Output overcurrent fault limit.
R/W Word
Y
L11
A
l
29.75A
0xDBB8
IOUT_OC_WARN_LIMIT
0x4A Output overcurrent warning limit.
R/W Word
Y
L11
A
l
20.0A
0xDA80
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
Related commands: STATUS_IOUT, SMBALERT_MASK, READ_IOUT, MFR_IOUT_PEAK, READ_POUT, IOUT_OC_FAULT_RESPONSE, MFR_VOUT_AVP
IOUT_CAL_GAIN
The IOUT_CAL_GAIN command is used to set the resistance value of the output current sense element in milliohms.
This command has two data bytes in Linear_5s_11s format.
MFR_IOUT_CAL_GAIN_TC
The MFR_IOUT_CAL_GAIN_TC command sets the temperature coefficient of the output current sense element in
ppm/°C. Effective sense resistance, in milliohms, is computed by the LTC3882-1 as
RSENSE = IOUT_CAL_GAIN • (1 + 1E-6 • MFR_IOUT_CAL_GAIN_TC • (READ_TEMPERATURE_1 – 27))
This command has two data bytes representing a 2’s compliment integer.
IOUT_OC_FAULT_LIMIT
The IOUT_OC_FAULT_LIMIT command sets the value of the instantaneous peak output current, in amperes, which
will cause the OC supervisor to detect an output overcurrent fault. The LTC3882-1 uses the computed effective sense
resistance and the voltage across the ISENSE± inputs to determine the output current. The programmed limit voltage is
rounded to the nearest 0.4mV in a range from 0.0mV to 80.0mV. Output overcurrent faults are ignored during TON_RISE
and TOFF_FALL output sequencing.
This command has two data bytes in Linear_5s_11s format.
IOUT_OC_WARN_LIMIT
The IOUT_OC_WARN_LIMIT command sets the value of the output current measured by the ADC, in amperes, that causes
an output overcurrent warning. To provide meaningful responses, this value should be set below IOUT_OC_FAULT_LIMIT
minus 1/2 of the maximum anticipated ripple current. If the IOUT_OC_WARN_LIMIT is exceeded, the device:
• Sets the IOUT Bit in the STATUS_WORD
• Sets the IOUT Overcurrent Warning Bit in the STATUS_IOUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
Output overcurrent warnings are ignored during TON_RISE and TOFF_FALL output sequencing.
This command has two data bytes in Linear_5s_11s format.
38821f
For more information www.linear.com/LTC3882-1
83
LTC3882-1
PMBus COMMAND DETAILS
(Output Timing, Delays, and Ramping)
Output Timing, Delays, and Ramping
COMMAND NAME
CMD
CODE DESCRIPTION
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
MFR_RESTART_DELAY
0xDC Minimum time RUN pin is held low by the
LTC3882-1.
R/W Word
Y
L11
ms
l
500ms
0xFBE8
TON_DELAY
0x60
Delay from RUN pin or OPERATION on
command to TON_RISE ramp start.
R/W Word
Y
L11
ms
l
0.0ms
0x8000
TON_RISE
0x61
Time for VOUT to rise from 0V to
VOUT_COMMAND after TON_DELAY.
R/W Word
Y
L11
ms
l
8.0ms
0xD200
TON_MAX_FAULT_LIMIT
0x62
Maximum time for VOUT to rise above
VOUT_UV_FAULT_LIMIT after TON_DELAY.
R/W Word
Y
L11
ms
l
10.0ms
0xD280
VOUT_TRANSITION_RATE
0x27
VOUT slew rate for programmed output
changes.
R/W Word
Y
L11
V/ms
l
0.25V/ms
0xAA00
TOFF_DELAY
0x64
Delay from RUN pin or OPERATION off
command to TOFF_FALL ramp start.
R/W Word
Y
L11
ms
l
0.0ms
0x8000
TOFF_FALL
0x65
Time for VOUT to fall to 0V from
VOUT_COMMAND after TOFF_DELAY.
R/W Word
Y
L11
ms
l
8.0ms
0xD200
TOFF_MAX_WARN_LIMIT
0x66
R/W Word
Maximum time for VOUT to decay below
12.5% of VOUT_COMMAND after TOFF_FALL
completes.
Y
L11
ms
l
150ms
0xF258
Related commands: MFR_RETRY_DELAY, STATUS_VOUT, SMBALERT_MASK, TON_MAX_FAULT_RESPONSE, MFR_CHAN_CONFIG_LTC3882-1,
MFR_PWM_MODE_LTC3882-1
These commands can be used to establish required on/off sequencing for any number of system power supply rails.
MFR_RESTART_DELAY
The MFR_RESTART DELAY command specifies the minimum PWM off time (RUN low) in milliseconds. The LTC3882-1
will actively hold its RUN pin low for this length of time if a falling RUN edge is detected. After this delay, a standard
start-up sequence can be initiated. A minimum of TOFF_DELAY + TOFF_FALL + 136ms is recommended for this command value. Valid value range is 136ms to 65.52 seconds. The LTC3882-1 uses a resolution of 16ms for this command
and will not produce delays outside of this range.
This command has two data bytes in Linear_5s_11s format.
TON_DELAY
The TON_DELAY command sets the delay, in milliseconds, between a start condition and the beginning of the output
voltage rise. Values from 0ms to 83 seconds are considered valid, and the LTC3882-1 will not produce delays outside
of this range.
This command has two data bytes in Linear_5s_11s format.
TON_RISE
The TON_RISE command sets the desired time, in milliseconds, from the point the output starts to rise until it enters
the regulation band. Values from 0 seconds to 1.3 seconds are considered valid, and the LTC3882-1 will not produce
rise times outside of this range. Values of TON_RISE less than 0.25ms or resulting slopes greater than 4V/ms will
result in an output step to the commanded voltage limited only by PWM analog loop response.
This command has two data bytes in Linear_5s_11s format.
84
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Output Timing, Delays and Ramping)
TON_MAX_FAULT_LIMIT
The TON_MAX_FAULT_LIMIT command sets the maximum time, in milliseconds, the unit is allowed from the beginning
of TON_RISE to power up the output without passing VOUT_UV_FAULT_LIMIT. A value of 0ms means there is no limit
and the unit can attempt to bring up the output voltage indefinitely. The maximum allowed TON_MAX is 8 seconds.
This command has two data bytes in Linear_5s_11s format.
VOUT_TRANSITION_RATE
The VOUT_TRANSITION_RATE command sets the rate at which the output voltage changes, in volts per millisecond
(or mV/µs), in response to a VOUT_COMMAND or OPERATION (margin) command. This rate of change does not apply to operations that fully turn the PWM channel on or off. Values from 1mV/ms to 4V/ms are considered valid. The
LTC3882-1 will not produce VOUT transitions slower than 1mV/ms, and values exceeding 4V/ms cause the device to
transition the output as quickly as possible, limited only by PWM analog loop response.
This command has two data bytes in Linear_5s_11s format.
TOFF_DELAY
The TOFF_DELAY command sets the delay, in milliseconds, between a stop condition and the beginning of the output
voltage fall. Values from 0s to 16s are considered valid.
This command has two data bytes in Linear_5s_11s format.
TOFF_FALL
The TOFF_FALL command sets the time, in milliseconds, from the end of TOFF_DELAY until the output voltage is commanded fully to zero. The part attempts to linearly reduce the commanded output voltage to zero during TOFF_FALL.
At the end of this period, the PWM output is disabled.
The part will maintain its programmed PWM operating mode during TOFF_FALL. Using continuous conduction mode
will produce a well defined VOUT ramp off but may result in negative output current. The minimum supported fall time
is 0.25ms, or any value that results in a rate of fall exceeding 4V/ms. Programmed values less than this will result in a
commanded 0.25ms ramp, possibly limited by PWM analog loop response. Maximum fall time is 1.3 seconds.
In discontinuous conduction mode, the controller will not be able to draw current from the load and fall time will be
set by output capacitance and load current.
This command has two data bytes in Linear_5s_11s format.
TOFF_MAX_WARN_LIMIT
The TOFF_MAX_WARN_LIMIT command sets the time, in milliseconds, the unit is allotted to have the output off after
TOFF_FALL completes before a warning is issued. The output is considered off when VOUT is less than 12.5% of the
VOUT_COMMAND value.
A data value of 0ms means there is no limit and the unit can attempt to turn the output off indefinitely. There is also
no limit enforced if bit 0 of MFR_CHAN_CONFIG_LTC3882-1 is set.
This command has two data bytes in Linear_5s_11s format.
38821f
For more information www.linear.com/LTC3882-1
85
LTC3882-1
PMBus COMMAND DETAILS
(External Temperature and Limits)
External Temperature and Limits
COMMAND NAME
CMD
CODE DESCRIPTION
MFR_TEMP_1_GAIN
0xF8
Set slope for external temperature
calculations.
R/W Word
Y
CF
MFR_TEMP_1_OFFSET
0xF9
Offset addend for external temperature
calculations.
R/W Word
Y
L11
OT_FAULT_LIMIT
0x4F
External overtemperature fault limit.
R/W Word
Y
OT_WARN_LIMIT
0x51
External overtemperature warning limit.
R/W Word
UT_FAULT_LIMIT
0x53
External undertemperature fault limit.
R/W Word
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
l
1.0
0x4000
°C or V
l
0.0
0x8000
L11
°C
l
100.0°C
0xEB20
Y
L11
°C
l
85.0°C
0xEAA8
Y
L11
°C
l
–40.0°C
0xE580
Related commands: STATUS_TEMPERATURE, SMBALERT_MASK, MFR_TEMPERATURE1_PEAK, OT_FAULT_RESPONSE, UT_FAULT_RESPONSE,
STATUS_MFR_SPECIFIC, READ_TEMPERATURE_2, MFR_OT_FAULT_RESPONSE, MFR_PWM_MODE_LTC3882-1
MFR_TEMP_1_GAIN
The MFR_TEMP_1_GAIN command sets the slope used in the calculation of external temperature to account for nonidealities in the element and remote sensing errors, if any. Refer to the MFR_PWM_MODE_LTC3882-1 command for
equation details.
This command has two data bytes representing a 2’s complement integer.
MFR_TEMP_1_OFFSET
The MFR_TEMP_1_OFFSET command sets the offset used in the calculation of external temperature to account for
non-idealities in the element and remote sensing errors, if any. The unit of measure for MFR_TEMP1_OFFSET depends
on bit 5 of MFR_PWM_MODE. MFR_TEMP1_ OFFSET is expressed volts if this bit is set and in °C otherwise. Refer to
the MFR_PWM_MODE_ LTC3882-1 command for equation details.
This command has two data bytes in Linear_5s_11s format.
OT_FAULT_LIMIT
The OT_FAULT_LIMIT command sets the value of sensed external temperature, in degrees Celsius, which causes an
overtemperature fault.
This command has two data bytes in Linear_5s_11s format.
OT_WARN_LIMIT
The OT_WARN_LIMIT command sets the value of sensed external temperature, in degrees Celsius, which causes an
overtemperature warning. If the OT_WARN_LIMIT is exceeded, the device:
• Sets the TEMPERATURE Bit in the STATUS_BYTE
• Sets the Overtemperature Warning Bit in the STATUS_TEMPERATURE Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has two data bytes in Linear_5s_11s format.
86
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(External Temperature Limits)
UT_FAULT_LIMIT
The UT_FAULT_LIMIT command sets the value of sensed external temperature, in degrees Celsius, which causes an
undertemperature fault.
This command has two data bytes in Linear_5s_11s format.
Status Reporting
COMMAND NAME
CMD
CODE DESCRIPTION
STATUS_BYTE
0x78
One-byte channel status summary.
R/W Byte
Y
Reg
STATUS_WORD
0x79
Two-byte channel status summary.
R/W Word
Y
Reg
STATUS_VOUT
0x7A VOUT fault and warning status.
R/W Byte
Y
Reg
TYPE
DATA
PAGED FORMAT
STATUS_IOUT
0x7B IOUT fault and warning status.
R/W Byte
Y
Reg
STATUS_INPUT
0x7C Input supply fault and warning status.
R/W Byte
N
Reg
STATUS_TEMPERATURE
0x7D External temperature fault and warning
status.
R/W Byte
Y
Reg
STATUS_CML
0x7E
Communication, memory and logic fault and
warning status.
R/W Byte
N
Reg
STATUS_MFR_SPECIFIC
0x80
LTC3882-1-specific status.
R/W Byte
Y
Reg
MFR_PADS_LTC3882-1
0xE5
State of selected LTC3882-1 pads.
R Word
N
Reg
MFR_COMMON
0xEF
LTC-generic device status reporting.
R Byte
N
Reg
CLEAR_FAULTS
0x03
Clear all set fault bits.
Send Byte
N
UNITS NVM
DEFAULT
VALUE
Refer to Figure 2 for a graphical depiction of these register contents and their relationships.
STATUS_BYTE
The STATUS_BYTE command returns a one-byte summary of the most critical faults.
STATUS_BYTE Message Contents:
BIT
STATUS BIT NAME
MEANING
7*
BUSY
6
OFF
5
VOUT_OV
An output overvoltage fault has occurred.
4
IOUT_OC
An output overcurrent fault has occurred.
3
VIN_UV
Not supported (LTC3882-1 returns 0).
2
TEMPERATURE
1
CML
0*
NONE OF THE ABOVE
A fault was declared because the LTC3882-1 was unable to respond.
This bit is set if the channel is not providing power to its output, regardless of the reason, including simply not
being enabled.
A temperature fault or warning has occurred.
A communications, memory or logic fault has occurred.
A fault Not listed in bits[7:1] has occurred.
*ALERT can be asserted if either of these bits is set. They may be cleared by writing a 1 to their bit position in the STATUS_BYTE, in lieu of a CLEAR_FAULTS
command.
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
87
LTC3882-1
PMBus COMMAND DETAILS
(Status Reporting)
STATUS_WORD
The STATUS_WORD command returns a two-byte summary of the channel's fault condition. The low byte of the STATUS_WORD is the same as the STATUS_BYTE command.
STATUS_WORD High Byte Message Contents:
BIT
STATUS BIT NAME
15
VOUT
An output voltage fault or warning has occurred.
MEANING
14
IOUT
An output current fault or warning has occurred.
13
INPUT
An input voltage fault or warning has occurred.
12
MFR_SPECIFIC
A fault or warning specific to the LTC3882-1 has occurred.
11
POWER_GOOD#
The POWER_GOOD state is false if this bit is set.
10
FANS
Not supported (LTC3882-1 returns 0).
9
OTHER
Not supported (LTC3882-1 returns 0).
8
UNKNOWN
Not supported (LTC3882-1 returns 0).
This command has two data bytes.
STATUS_VOUT
The STATUS_VOUT command returns one byte of VOUT status information.
STATUS_VOUT Message Contents:
BIT
MEANING
7
VOUT overvoltage fault.
6
VOUT overvoltage warning.
5
VOUT undervoltage warning.
4
VOUT undervoltage fault.
3
VOUT_MAX warning.
2
TON_MAX fault.
1
TOFF_MAX warning.
0
Not supported by the LTC3882-1 (returns 0).
ALERT can be asserted if any of bits[7:1] are set. These may be cleared by writing a 1 to their bit position in STATUS_VOUT, in lieu of a CLEAR_FAULTS
command.
This command has one data byte.
88
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Status Reporting)
STATUS_IOUT
The STATUS_IOUT command returns one byte of IOUT status information.
STATUS_IOUT Message Contents:
BIT
MEANING
7
IOUT overcurrent fault.
6
Not supported (LTC3882-1 returns 0).
5
IOUT overcurrent warning.
4:0
Not supported (LTC3882-1 returns 0).
ALERT can be asserted if any supported bits are set. Any supported bit may be cleared by writing a 1 to that bit position in STATUS_IOUT, in lieu of a
CLEAR_FAULTS command.
This command has one data byte.
STATUS_INPUT
The STATUS_INPUT command returns one byte of VIN (VINSNS) status information.
STATUS_INPUT Message Contents:
BIT
MEANING
7
VIN overvoltage fault.
6
Not supported (LTC3882-1 returns 0).
5
VIN undervoltage warning.
4
Not supported (LTC3882-1 returns 0).
3
2:0
Unit off for insufficient VIN.
Not supported (LTC3882-1 returns 0).
ALERT can be asserted if bit 7 is set. Bit 7 may be cleared by writing it to a 1, in lieu of a CLEAR_FAULTS command.
This command has one data byte.
STATUS_TEMPERATURE
The STATUS_TEMPERATURE command returns one byte of sensed external temperature status information.
STATUS_TEMPERATURE Message Contents:
BIT
MEANING
7
External overtemperature fault.
6
External overtemperature warning.
5
Not supported (LTC3882-1 returns 0).
4
External undertemperature fault.
3:0
Not supported (LTC3882-1 returns 0).
ALERT can be asserted if any supported bits are set. Any supported bit may be cleared by writing a 1 to that bit position in STATUS_TEMPERATURE, in lieu
of a CLEAR_FAULTS command.
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
89
LTC3882-1
PMBus COMMAND DETAILS
(Status Reporting)
STATUS_CML
The STATUS_CML command returns one byte of status information on received commands, internal memory and logic.
STATUS_CML Message Contents:
BIT
MEANING
7
Invalid or unsupported command received.
6
Invalid or unsupported data received.
5
Packet error check failed.
4
Memory fault detected.
3
Processor fault detected.
2
Reserved (LTC3882-1 returns 0).
1
Other communication fault.
0
Other memory or logic fault.
ALERT can be asserted if any supported bits are set. Any supported bit may be cleared by writing a 1 to that bit position in STATUS_CML, in lieu of a
CLEAR_FAULTS command.
This command has one data byte.
STATUS_MFR_SPECIFIC
The STATUS_MFR_SPECIFIC command returns one byte with LTC3882-1 specific status information.
STATUS_MFR_SPECIFIC Message Contents:
BIT
MEANING
7
Internal temperature fault (>160°C).
6
Internal temperature warning (>130°C).
5
EEPROM CRC error.
4
Internal PLL unlocked.
3
Fault log present.
2
Not supported (LTC3882-1 returns 0).
1
Output short cycled.
0
FAULT low.
If any supported bits are set, the MFR bit in the STATUS_WORD will be set and ALERT may be asserted. Any supported bit may be cleared by writing a 1 to
that bit position in STATUS_MFR_SPECIFIC, in lieu of a CLEAR_FAULTS command.
This command has one data byte.
90
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Status Reporting)
MFR_PADS_LTC3882-1
The MFR_PADS_LTC3882-1 command provides status of the LTC3882-1 digital I/O and control pins, in addition to
general output voltage conditions.
MFR_PADS_LTC3882-1 Message Contents:
BIT
MEANING
15
Channel 1 is a slave.
14
Channel 0 is a slave.
13:12
Not supported (LTC3882-1 returns 0).
11
ADC results for IOUT may be invalid.
10
SYNC output disabled externally.
9
Channel 1 POWER_GOOD (normally returns 1 if slave).
8
Channel 0 POWER_GOOD (normally returns 1 if slave).
7
LTC3882-1 forcing RUN1 low.
6
LTC3882-1 forcing RUN0 low.
5
RUN1 pin state.
4
RUN0 pin state.
3
LTC3882-1 forcing FAULT1 low.
2
LTC3882-1 forcing FAULT0 low.
1
FAULT1 pin state.
0
FAULT0 pin state.
This read-only command has two data bytes.
MFR_COMMON
The MFR_COMMON command contains status bits that are common to multiple LTC PMBus products.
MFR_COMMON Message Contents:
BIT
MEANING
7
LTC3882-1 not forcing ALERT low.
6
LTC3882-1 not BUSY.
5
LTC3882-1 calculations not pending.
4
LTC3882-1 output not in transition.
3
LTC3882-1 EEPROM initialized.
2
Not supported (LTC3882-1 returns 0).
1
SHARE_CLK timeout.
0
Not supported (LTC3882-1 returns 0).
This read-only command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
91
LTC3882-1
PMBus COMMAND DETAILS
(Status Reporting)
CLEAR_FAULTS
The CLEAR_FAULTS command clears any fault bits that have been set and deasserts (releases) the ALERT pin. This
command clears all bits in all status commands simultaneously.
CLEAR_FAULTS does not cause a channel that has latched off for a fault condition to restart. Channels that are latched
off for a fault condition are restarted when the output is commanded to turn off and then on through the OPERATION
command or RUN pins, or IC supply power is cycled.
If a fault is still present when CLEAR_FAULTS is commanded, that fault bit will immediately be set and ALERT again
asserted low.
This write-only command has no data bytes.
Telemetry
COMMAND NAME
CMD
CODE DESCRIPTION
READ_VIN
0x88
MFR_VIN_PEAK
READ_VOUT
MFR_VOUT_PEAK
TYPE
Measured VIN.
DATA
PAGED FORMAT
UNITS NVM
R Word
N
L11
V
0xDE Maximum VIN measurement since last
MFR_CLEAR_PEAKS.
R Word
N
L11
V
0x8B Measured VOUT.
0xDD Maximum VOUT measurement since last
MFR_CLEAR_PEAKS.
R Word
Y
L16
V
R Word
Y
L16
V
READ_IOUT
0x8C Measured IOUT.
R Word
Y
L11
A
MFR_IOUT_PEAK
0xD7 Maximum IOUT measurement since last
MFR_CLEAR_PEAKS.
R Word
Y
L11
A
READ_POUT
0x96
Calculated output power.
R Word
Y
L11
W
READ_TEMPERATURE_1
0x8D Measured external temperature.
R Word
Y
L11
°C
MFR_TEMPERATURE_1_PEAK
0xDF Maximum external temperature measurement
since last MFR_CLEAR_PEAKS.
R Word
Y
L11
°C
READ_TEMPERATURE_2
0x8E
Measured internal temperature.
R Word
N
L11
°C
MFR_TEMPERATURE_2_PEAK
0xF4
Maximum internal temperature measurement
since last MFR_CLEAR_PEAKS.
R Word
N
L11
°C
READ_DUTY_CYCLE
0x94
Measured commanded PWM duty cycle.
R Word
Y
L11
%
READ_FREQUENCY
0x95
Measured PWM input clock frequency.
R Word
Y
L11
kHz
MFR_CLEAR_PEAKS
0xE3
Clear all peak values.
Send Byte
N
DEFAULT
VALUE
Related commands: IOUT_CAL_GAIN, MFR_IOUT_CAL_GAIN_TC, MFR_PWM_MODE_LTC3882-1
READ_VIN
The READ_VIN command returns the input voltage measured between VINSNS and GND in volts.
This read-only command has two data bytes in Linear_5s_11s format.
92
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Telemetry)
MFR_VIN_PEAK
The MFR_VIN_PEAK command reports the highest voltage, in volts, measured for READ_VIN. This peak value can be
reset by a MFR_CLEAR_PEAKS command.
This read-only command has two data bytes in Linear_5s_11s format.
READ_VOUT
The READ_VOUT command returns the output voltage measured at the VSENSE± pins in volts.
This read-only command has two data bytes in Linear_16u format.
MFR_VOUT_PEAK
The MFR_VOUT_PEAK command reports the highest voltage, in volts, measured for READ_VOUT. This peak value can
be reset by a MFR_CLEAR_PEAKS command.
This read-only command has two data bytes in Linear_16u format.
READ_IOUT
The READ_IOUT command returns the output current in amperes. This value is computed from:
• The differential voltage measured across the ISENSE± pins
• The IOUT_CAL_GAIN value
• The MFR_IOUT_CAL_GAIN_TC value
• The READ_TEMPERATURE_1 value
• The MFR_TEMP_1_GAIN value
• The MFR_TEMP_1_OFFSET value
This read-only command has two data bytes in Linear_5s_11s format.
MFR_IOUT_PEAK
The MFR_IOUT_PEAK command reports the highest current, in amperes, calculated for READ_IOUT. This peak value
can be reset by a MFR_CLEAR_PEAKS command.
This read-only command has two data bytes in Linear_5s_11s format.
READ_POUT
The READ_POUT command reports the output power in watts. The value is calculated from the product of the most
recent correlated output voltage and current readings.
This read-only command has two data bytes in Linear_5s_11s format.
38821f
For more information www.linear.com/LTC3882-1
93
LTC3882-1
PMBus COMMAND DETAILS
(Telemetry)
READ_TEMPERATURE_1
The READ_TEMPERATURE_1 command returns the temperature, in degrees Celsius, of the external sense element.
This read-only command has two data bytes in Linear_5s_11s format.
MFR_TEMPERATURE_1_PEAK
The MFR_TEMPERATURE_1_PEAK command reports the highest temperature, in degrees Celsius, calculated for
READ_TEMPERATURE_1. This peak value can be reset by a MFR_CLEAR_PEAKS command.
This read-only command has two data bytes in Linear_5s_11s format.
READ_TEMPERATURE_2
The READ_TEMPERATURE_2 command returns the LTC3882-1 internal temperature in degrees Celsius.
This read-only command has two data bytes in Linear_5s_11s format.
MFR_TEMPERATURE_2_PEAK
The MFR_TEMPERATURE_2_PEAK command reports the highest temperature, in degrees Celsius, calculated for
READ_TEMPERATURE_2. This peak value can be reset by a MFR_CLEAR_PEAKS command.
This read-only command has two data bytes in Linear_5s_11s format.
READ_DUTY_CYCLE
The READ_DUTY_CYCLE command returns the duty cycle of the PWM control in percent. This will not be the exact duty
cycle of the PWM switch node due to efficiency losses in the power stage and current consumption of the LTC3882-1
itself.
This read-only command has two data bytes in Linear_5s_11s format.
READ_FREQUENCY
The READ_FREQUENCY command returns the switching frequency supplied to the internal PLL in kilohertz, whether
derived internally or provided by external clock on the SYNC pin. This may not be the actual PWM output switching
frequency during certain exception processing, such as an output overcurrent condition.
This read-only command has two data bytes in Linear_5s_11s format.
MFR_CLEAR_PEAKS
The MFR_CLEAR_PEAKS command resets all stored _PEAK values. The LTC3882-1 determines new peak values after
this command is received.
This write-only command has no data bytes.
94
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
Fault Response and Communication
COMMAND NAME
CMD
CODE DESCRIPTION
VIN_OV_FAULT_RESPONSE
0x56
VIN overvoltage fault response.
R/W Byte
Y
Reg
l
0x80
VOUT_OV_FAULT_RESPONSE
0x41
VOUT overvoltage fault response.
R/W Byte
Y
Reg
l
0xB8
VOUT_UV_FAULT_RESPONSE
0x45
VOUT undervoltage fault response.
R/W Byte
Y
Reg
l
0xB8
IOUT_OC_FAULT_RESPONSE
0x47
Output overcurrent fault response.
R/W Byte
Y
Reg
l
0x00
OT_FAULT_RESPONSE
0x50
External overtemperature fault response.
R/W Byte
Y
Reg
l
0xB8
UT_FAULT_RESPONSE
0x54
External undertemperature fault response.
R/W Byte
Y
Reg
l
0xB8
MFR_OT_FAULT_RESPONSE
0xD6 Internal overtemperature fault response.
R/W Byte
N
Reg
l
0xC0
TON_MAX_FAULT_RESPONSE
0x63
R/W Byte
Y
Reg
l
0xB8
MFR_RETRY_DELAY
0xDB Minimum time before retry after a fault.
R/W Word
N
L11
l
350ms
0xFABC
SMBALERT_MASK
0x1B Mask ALERT Activity.
Block R/W
Y
Reg
l
See CMD
Details
MFR_FAULT_PROPAGATE_
LTC3882-1
0xD2 Configure status propagation via FAULTn
pins.
R/W Word
Y
Reg
l
0x6993
MFR_FAULT_RESPONSE
0xD5 PWM response when FAULTn pin is low.
R/W Byte
Y
Reg
l
0xC0
R Block
N
Reg
Send Byte
N
Fault response when
TON_MAX_FAULT_LIMIT is exceeded.
MFR_FAULT_LOG
0xEE
MFR_FAULT_LOG_CLEAR
0xEC Clear existing EEPROM fault log.
Read fault log data.
TYPE
DATA
PAGED FORMAT
UNITS NVM
ms
DEFAULT
VALUE
Related commands: STATUS_BYTE, STATUS_WORD, MFR_PADS_LTC3882-1, MFR_RESTART_DELAY, MFR_CHAN_CONFIG_LTC3882-1, MFR_FAULT_
LOG_STORE, CLEAR_FAULTS
These commands detail programmable device responses for detected faults beyond the hardware-level actions described
in the Operations section. LTC3882-1 hardware-level fault responses cannot be modified. Refer to Table 1 to Table 4
for details of fault log contents. PMBus warning event responses are listed under _WARN_LIMIT command details.
VIN_OV_FAULT_RESPONSE
The VIN_OV_FAULT_RESPONSE command instructs the device on what action to take in response to an input overvoltage fault. The format for this command is given in Table 13. The device also:
• Sets the INPUT Bit in the STATUS_WORD
• Sets the VIN Overvoltage Fault Bit in the STATUS_INPUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
95
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
VOUT_OV_FAULT_RESPONSE
The VOUT_OV_FAULT_RESPONSE command instructs the device on what action to take in response to an output
overvoltage fault. The format for this command is given in Table 12. The device also:
• Sets the VOUT_OV Bit in the STATUS_BYTE
• Sets the VOUT Bit in the STATUS_WORD
• Sets the VOUT Overvoltage Fault Bit in the STATUS_VOUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has one data byte.
VOUT_UV_FAULT_RESPONSE
The VOUT_UV_FAULT_RESPONSE command instructs the device on what action to take in response to an output
undervoltage fault. The format for this command is given in Table 12. The device also:
• Sets the VOUT Bit in the STATUS_WORD
• Sets the VOUT Undervoltage Fault Bit in the STATUS_VOUT Command,
• Notifies the Host by Asserting ALERT, Unless Masked
This command has one data byte.
Table 12. Data Byte Contents for VOUT_OV_FAULT_RESPONSE and VOUT_UV_FAULT_RESPONSE
BITS
DESCRIPTION
[7:6]
For all values of bits [7:6], the LTC3882-1:
VALUE
• Sets the corresponding fault bits in the status commands.
• Notifies the host by asserting ALERT, unless masked.
The fault, once set, is cleared only when one or more of the
following events occurs:
• The device receives a CLEAR_FAULTS command.
• The corresponding STATUS_VOUT bit is written to a one.
• The output is commanded off, then on, by the RUN pin or
OPERATION command.
• The device receives a RESTORE_USER_ALL command.
• The device receives an MFR_RESET command.
• IC supply power is cycled.
[5:3]
[2:0]
96
Retry setting.
Delay time.
MEANING
00
The LTC3882-1 continues to operate indefinitely with the
normal hardware response described in the Operation
section.
01
The LTC3882-1 continues operating with the normal
hardware response for the delay time specified by bits [2:0].
If the fault is continuously present for the entire delay, the
unit then disables the output and does attempt to restart.
10
The LTC3882-1 immediately disables the output and
responds according to the retry setting in bits [5:3].
11
Not supported. Writing this value will generate a CML fault.
000-110
The LTC3882-1 does not attempt to restart. The output
remains disabled until the fault is cleared, the device is
commanded off and then on, or bias power (LTC3882-1
power supply input) is cycled.
111
The LTC3882-1 attempts to restart continuously without
limitation with an interval set by MFR_RETRY_DELAY.
This response persists until the unit is commanded off, or
bias power is removed, or another fault response forces
shutdown without retry.
xxx
Response delay time in 10µs increments. This delay
time determines how long the fault may have to persist
before the controller is disabled, depending on bits [7:6].
Hardware-level response, if any, will occur during this delay.
These bits always return zero if bits [7:6] are not set to 0x2.
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
IOUT_OC_FAULT_RESPONSE
The IOUT_OC_FAULT_RESPONSE command instructs the device on what action to take in response to an output
overcurrent fault. The device also:
• Sets the IOUT_OC Bit in the STATUS_BYTE
• Sets the IOUT Bit in the STATUS_WORD
• Sets the IOUT Overcurrent Fault Bit in the STATUS_IOUT Command
• Notifies the Host by Asserting ALERT
Output overcurrent faults are ignored during TON_RISE and TOFF_FALL output sequencing.
Data Byte Contents for IOUT_OC_FAULT_RESPONSE:
BITS
[7:6]
DESCRIPTION
VALUE
For all values of bits [7:6], the LTC3882-1:
• Sets the corresponding fault bits in the status commands.
• Notifies the host by asserting ALERT, unless masked.
The fault, once set, is cleared only when one or more of the
following events occurs:
• The device receives a CLEAR_FAULTS command.
• The corresponding STATUS_IOUT bit is written to a one.
• The output is commanded off, then on, by the RUN pin or
OPERATION command.
• The device receives a RESTORE_USER_ALL command.
• The device receives an MFR_RESET command.
• IC supply power is cycled.
[5:3]
[2:0]
Retry setting.
Delay time.
MEANING
0x
The LTC3882-1 continues to operate indefinitely with the
normal hardware response described in the Operation
section.
10
The LTC3882-1 continues operating with the normal
hardware response for the delay time specified by bits
[2:0]. If the fault is continuously present for the entire delay,
the unit then disables the output and does not attempt to
restart.
11
The LTC3882-1 shuts down (disables the output) and
responds according to the retry setting in bits [5:3].
000-110
The LTC3882-1 does not attempt to restart. The output
remains disabled until the fault is cleared, the device is
commanded off and then on, or bias power is cycled.
111
The LTC3882-1 attempts to restart continuously without
limitation with an interval set by MFR_RETRY_DELAY.
This response persists until the unit is commanded off,
bias power is removed, or another fault response forces
shutdown without retry.
xxx
Response delay time in 16ms increments. This delay time
determines how long the fault may have to persist before
the controller is disabled, depending on bits [7:6]. These
bits always return zero if bits [7:6] are not set to 0x2.
Programming an unsupported IOUT_OC_FAULT_RESPONSE value will generate a CML fault and the command will
be ignored.
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
97
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
OT_FAULT_RESPONSE
The OT_FAULT_RESPONSE command instructs the device on what action to take in response to an overtemperature
fault. The format for this command is given in Table 13. The device also:
• Sets the TEMPERATURE Bit in the STATUS_BYTE
• Sets the Overtemperature Fault Bit in the STATUS_TEMPERATURE Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has one data byte.
UT_FAULT_RESPONSE
The UT_FAULT_RESPONSE command instructs the device on what action to take in response to an undertemperature
fault. The format for this command is given in Table 13. The device also:
• Sets the TEMPERATURE Bit in the STATUS_BYTE
• Sets the Undertemperature Fault Bit in the STATUS_TEMPERATURE Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has one data byte.
MFR_OT_FAULT_RESPONSE
The MFR_OT_FAULT_RESPONSE command instructs the device on what action to take in response to an internal
overtemperature fault (150°C to 160°C). The device also:
• Sets the MFR Bit in the STATUS_WORD
• Sets the Overtemperature Fault Bit in the STATUS_MFR_SPECIFIC Command
• Notifies the Host by Asserting ALERT, Unless Masked
Supported Values:
VALUE
MEANING
0xC0
The LTC3882-1 continues to operate indefinitely with the normal hardware response described in the Operation section.
0x80
The LTC3882-1 shuts down immediately and does not attempt to restart. The output remains disabled until the fault is cleared and the unit
is commanded off and then on, or bias power (LTC3882-1 power supply input) is cycled.
Programming an unsupported MFR_OT_FAULT_RESPONSE value will generate a CML fault and the command will be
ignored.
This command has one data byte.
98
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
TON_MAX_FAULT_RESPONSE
The TON_MAX_FAULT_RESPONSE command instructs the device on what action to take in response to a TON_MAX
fault. The format for this command is given in Table 13. The device also:
• Sets the VOUT Bit in the STATUS_WORD
• Sets the TON_MAX_FAULT Bit in the STATUS_VOUT Command
• Notifies the Host by Asserting ALERT, Unless Masked
This command has one data byte.
Table 13. Data Byte Contents for the Following _FAULT_RESPONSE Commands: VIN_OV, OT, UT and TON_MAX
BITS
[7:6]
DESCRIPTION
VALUE
For all values of bits [7:6], the LTC3882-1:
• Sets the corresponding fault bits in the status commands.
• Notifies the host by asserting ALERT, unless masked.
The fault, once set, is cleared only when one or more of the
following events occurs:
• The device receives a CLEAR_FAULTS command.
• The corresponding fault bit is written to a one.
• The output is commanded off, then on, by the RUN pin or
OPERATION command.
• The device receives a RESTORE_USER_ALL command.
• The device receives an MFR_RESET command.
• IC supply power is cycled.
[5:3]
[2:0]
Retry setting.
Delay time.
MEANING
00
The LTC3882-1 continues operating without interruption.
01
Not supported. Writing this value will generate a CML fault.
10
The LTC3882-1 shuts down immediately (disables the
output) and responds according to the retry setting in bits
[5:3].
11
Not supported. Writing this value will generate a CML fault.
000-110
The LTC3882-1 does not attempt to restart. The output
remains disabled until the fault is cleared, the device is
commanded off and then on, or bias power is cycled.
111
The LTC3882-1 attempts to restart continuously without
limitation with an interval set by MFR_RETRY_DELAY.
This response persists until the unit is commanded off,
bias power is removed, or another fault response forces
shutdown without retry.
xxx
Not supported. Values ignored.
MFR_RETRY_DELAY
The MFR_RETRY_DELAY command sets the time in milliseconds between restart attempts for all retry fault responses.
The actual retry delay may be the longer of MFR_RETRY_DELAY or the time required for the output voltage to decay
below 12.5% of its programmed value. Decay qualification can be disabled using the MFR_CHAN_CONFIG_LTC3882-1
command. Retry delay starts once the fault is no longer detected by the LTC3882-1 or its FAULT pin is externally released. Legal values run from 120ms to 32.7 seconds.
This command has two data bytes in Linear_5s_11s format.
38821f
For more information www.linear.com/LTC3882-1
99
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
SMBALERT_MASK
The SMBALERT_MASK command can be used to prevent a particular status bit or bits from asserting ALERT as they
are asserted.
Figure 51 shows an example of the Write Word format used to set an ALERT mask, in this case without PEC. The bits
in the mask byte align with bits in the specified status register. For example, if the STATUS_TEMPERATURE command
code is sent in the first data byte, and the mask byte contains 0x40, then a subsequent External Overtemperature Warning would still set bit 6 of STATUS_TEMPERATURE but not assert ALERT. All other supported STATUS_TEMPERATURE
bits would continue to assert ALERT if set.
1
7
1
1
S
SLAVE
ADDRESS
W
A SMBALERT_MASK A
COMMAND CODE
8
1
8
8
1
1
MASK BYTE
A
P
1
STATUS_x
A
COMMAND CODE
38821 F51
Figure 51. Example of Setting SMBALERT_MASK
Figure 52 shows an example of the Block Write – Block Read Process Call protocol used to read back the present state
of any supported status register, again without PEC.
1
7
1
1
S
SLAVE
ADDRESS
W
A SMBALERT_MASK A
COMMAND CODE
1
7
Sr
SLAVE
ADDRESS
8
1
R
1
8
1
BLOCK COUNT
(= 1)
A
8
1
STATUS_x
A
COMMAND CODE
1
8
1
8
A
BLOCK COUNT
(= 1)
A
MASK BYTE
1
…
1
NA P
38821 F52
Figure 52. Example of Reading SMBALERT_MASK
SMBALERT_MASK cannot be applied to STATUS_BYTE, STATUS_WORD, MFR_COMMON or MFR_PADS_LTC3882-1.
Factory default masking for applicable status registers is shown below. Providing an unsupported command code to
SMBALERT_MASK will generate a CML for Invalid/Unsupported Data.
SMBALERT_MASK Default Setting: (Refer Also to Figure 2)
STATUS RESISTER
ALERT Mask Value MASKED BITS
STATUS_VOUT
0x00
None
STATUS_IOUT
0x00
None
STATUS_TEMPERATURE
0x00
None
STATUS_CML
0x00
None
STATUS_INPUT
0x00
None
STATUS_MFR_SPECIFIC
0x11
Bit 4 (internal PLL unlocked), bit 0 (FAULT low)
100
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
MFR_FAULT_PROPAGATE_LTC3882-1
The MFR_FAULT_PROPAGATE_LTC3882-1 command determines events that cause FAULT to be asserted. Setting a
bit in this register to a one allows the specified condition to assert the FAULT output for that channel. FAULT is not asserted by a fault, even if set to propagate, if that FAULT_RESPONSE is set to Ignore. The state of SMBLALERT_MASK
does not affect FAULT propagation.
Supported Values:
BIT
PROPAGATED CONDITION
15
Waiting for VOUT decay before restart.
14
VOUT short cycled (automatically deasserted 120ms after VOUT is fully OFF).
13
TON_MAX_FAULT_LIMIT exceeded.
12
(Reserved, must be set to 0).
11
MFR_OT_FAULT_LIMIT exceeded.
10
(Reserved, must be set to 0).
9
(Reserved, must be set to 0).
8
UT_FAULT_LIMIT exceeded.
7
OT_FAULT_LIMIT exceeded.
6
(Reserved).
5
(Reserved).
4
VIN_OV_FAULT_LIMIT exceeded.
3
(Reserved).
2
IOUT_OC_FAULT_LIMIT exceeded.
1
VOUT_UV_FAULT_LIMIT exceeded.
0
VOUT_OV_FAULT_LIMIT exceeded.
This command has two data bytes.
MFR_FAULT_RESPONSE
The MFR_FAULT_RESPONSE command instructs the device on what action to take in response to a FAULT pin being
pulled low by anything other than an internal fault.
Supported Values:
VALUE
MEANING
0xC0
Related PWM output is immediately disabled.
0x00
Input ignored, PWM operation continues without interruption.
When a FAULT pin is low, the device also:
• Sets the MFR_SPECIFIC Bit in the STATUS_WORD
• Sets Bit 0 in the STATUS_MFR_SPECIFIC Command to Indicate FAULT Is or Has Been Low
• Notifies the Host by Asserting ALERT, Unless Masked
This command has one data byte.
38821f
For more information www.linear.com/LTC3882-1
101
LTC3882-1
PMBus COMMAND DETAILS
(Fault Response and Communication)
MFR_FAULT_LOG
The MFR_FAULT_LOG command allows the contents of the fault log to be read. This log is created with a
MFR_FAULT_LOG_STORE command or at the first fault occurrence after a CLEAR_FAULTS or MFR_FAULT_LOG_CLEAR
command. If a fault occurs within the first second after applying power, some earlier pages in the log may not contain
valid data.
This read-only command uses block protocol with 147 bytes of data requiring an estimated data transfer time of 3.4ms
at 400kHz. The tTIMEOUT parameter is extended when this command is executed and a fault log is present.
MFR_FAULT_LOG_CLEAR
The MFR_FAULT_LOG_CLEAR command erases all stored fault log values. After a clear is issued, up to 8ms may be
required to clear related bit 3 in STATUS_MFR_SPECIFIC.
This write-only command has no data bytes.
EEPROM User Access
COMMAND NAME
CMD
CODE DESCRIPTION
STORE_USER_ALL
0x15
Store entire operating memory in EEPROM.
Send Byte
N
NA
RESTORE_USER_ALL
0x16
Restore entire operating memory from
EEPROM.
Send Byte
N
NA
MFR_COMPARE_USER_ALL
0xF0
Compare operating memory with EEPROM
contents.
Send Byte
N
MFR_FAULT_LOG_STORE
0xEA Force transfer of fault log from operating
memory to EEPROM.
Send Byte
N
MFR_EE_UNLOCK
0xBD (contact the factory)
MFR_EE_ERASE
0xBE (contact the factory)
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
MFR_EE_DATA
0xBF
USER_DATA_00
0xB0 EEPROM word reserved for LTpowerPlay.
(contact the factory)
R/W Word
N
Reg
l
USER_DATA_01
0xB1 EEPROM word reserved for LTpowerPlay.
R/W Word
Y
Reg
l
USER_DATA_02
0xB2 EEPROM word reserved for OEM use.
R/W Word
N
Reg
l
USER_DATA_03
0xB3 EEPROM word available for general data
storage.
R/W Word
Y
Reg
l
0x0000
USER_DATA_04
0xB4 EEPROM word available for general data
storage.
R/W Word
N
Reg
l
0x0000
Related commands: MFR_CONFIG_ALL_LTC3882-1
Note that if the LTC3882-1 die temperature exceeds 130°C, execution of any command in the above table except RESTORE_USER_ALL and MFR_FAULT_LOG_STORE will be disabled until the IC temperature drops below 125°C. RESTORE_USER_ALL is executed immediately, and MFR_FAULT_LOG_STORE is executed after the IC temperature drops
below 125°C. Refer to Table 4 for details of fault log contents. Using any command that writes data to the EEPROM is
strongly discouraged if bit 6 of STATUS_MFR_ SPECIFIC is set, indicating the internal die temperature is above 85°C.
Data retention of 10 years is not guaranteed if the EEPROM is written above a junction temperature of 85°C.
102
38821f
For more information www.linear.com/LTC3882-1
LTC3882-1
PMBus COMMAND DETAILS
(EEPROM User Access)
STORE_USER_ALL
The STORE_USER_ALL command instructs the PMBus device to copy the entire contents of the operating memory to
internal EEPROM PMBus configuration space.
This write-only command has no data bytes.
RESTORE_USER_ALL
The RESTORE_USER_ALL command instructs the PMBus device to copy the entire contents of the internal EEPROM
to matching locations in operating memory. The values in operating memory are overwritten by the values retrieved
from EEPROM. Both channels should be turned off prior to issuing this command. The LTC3882-1 ensures both PWM
channels are off, loads the operating memory from internal EEPROM, clears all faults, reads the resistor configuration
pins, and then performs a soft-start of both PWM channels, if enabled
This write-only command has no data bytes.
MFR_COMPARE_USER_ALL
The MFR_COMPARE_USER_ALL command instructs the LTC3882-1 to compare current operating memory (PMBus
command values in RAM) with the contents of the internal EEPROM. If the compared memories differ, a CML fault is
generated.
This write-only command has no data bytes.
MFR_FAULT_LOG_STORE
The MFR_FAULT_LOG_STORE command forces a data log to be written to internal EEPROM as if a fault event had occurred.
This command will generate a CML fault if the Enable Fault Logging bit is cleared in MFR_CONFIG_ALL_LTC3882-1.
This write-only command has no data bytes.
MFR_EE_xxxx
The MFR_EE_xxxx commands facilitate bulk programming of the LTC3882-1 internal EEPROM. Contact the factory
for details.
USER_DATA_0x
The USER_DATA_0x commands provide uncommitted EEPROM locations that may be applied as system scratchpad
space. USER_DATA_00 and USER_DATA_01 should not be modified when using the LTpowerPlay GUI. Some contract
manufacturers also reserve use of USER_DATA_02 for their own inventory control.
38821f
For more information www.linear.com/LTC3882-1
103
LTC3882-1
PMBus COMMAND DETAILS
(Unit Identification)
Unit Identification
COMMAND NAME
CMD
CODE DESCRIPTION
MFR_ID
0x99
MFR_MODEL
MFR_SERIAL
Manufacturer identification.
TYPE
DATA
PAGED FORMAT
UNITS NVM
DEFAULT
VALUE
R String
N
ASC
LTC
0x9A LTC model number.
R String
N
ASC
LTC3882-1
0x9E
R Block
N
Reg
Device serial number.
MFR_ID
The MFR_ID command returns the manufacturer ID of the LTC3882-1 using 8-bit ASCII characters.
This read-only command is in block format.
MFR_MODEL
The MFR_MODEL command returns the LTC part number using 8-bit ASCII characters.
This read-only command is in block format.
MFR_SERIAL
The MFR_SERIAL command returns the serial number of this specific device using a maximum of fourteen 8-bit ASCII
characters.
This read-only command is in block format.
104
38821f
For more information www.linear.com/LTC3882-1
R14
1.37k
R15
10k
For more information www.linear.com/LTC3882-1
R35
5.76k
R34
24.9k
22
VOUT0_CFG
ASEL1
ASEL0
C35
100pF
5
IAVG_GND
VOUT1_CFG
20
FREQ_CFG
21
PHAS_CFG
39
IAVG0
30
IAVG1
19
18
17
16
PGOOD1
PWM1
ISENSE1+
ISENSE1–
VSENSE1
27
26
31
32
2
TSNS0
3
TSNS1
28
COMP1
29
FB1
+ 33
COMP0
1
VDD33
C58
2.2µF
VDD33
R84
2.21Ω
VCC
C27
11pF
VDD33
VDD33
C26
270pF
R30
100k
R21
100k
C18
4.7µF
R85
2.21Ω
IN
VCC
BOOST
C28
10nF
D2
Q1
TS
TG
2
1
6
9
LTC4449
3
VLOGIC
BG
5
4
IN
GND
GND
7
8
4
3
2
1
C56
0.22µF
C29
10nF
C55
0.22µF
Q2
PLACE Q1,Q2 NEAR L1, L2 RESPECTIVELY
9
GND
BG
TS
TG
GND
LTC4449
VLOGIC
VCC
BOOST
R29
C25 1.07k
200pF
C57
2.2µF
5
6
7
8
D1
Q4
Q3
Q6
Q5
Figure 53. 36V Input 3.3V/40A 1.0MHz Converter with Discrete Gate Drivers
41
GND
VCC
25
VCC
VCC
+5V INPUT SUPPLY
7
PWM0
6
PGOOD0
+ 38
ISENSE0
37
ISENSE0–
36
VSENSE0+
VSENSE0– 35
34
VSENSE1–
40
FB0
VINSNS
C17
1µF
4
LTC3882-1
VDD33 VDD25
24
VDD25
SCL
10
SDA
11
ALERT
12
FAULT0
13
FAULT1
14
RUN0
15
RUN1
23
SHARE_CLK
8
SYNC
R20
4.99k 9
C16
2.2µF
C8
100nF
Q1, Q2: MMBT3906
Q3, Q5: BSC050N04LSG
Q4, Q6: BSC010N04LS
L1, L2: PULSE PA0513.221NLT
R31
17.4k
RUN
R18
10k
R19
4.99k
R32
16.2k
VDD25
R17
10k
+
VDD33
C4
47µF
R5
1Ω
VIN
24V TO 36V
VIN
VIN
CIN1
22µF
×2
C24
0.22µF
R8
3.09k
L1
0.2µH
PULSE
CIN3
22µF
×2
38821 F53
L2
0.2µH
PULSE
R55
3.09k
C66
0.22µF
C12
100µF
×2
+
C14
470µF
×4
GND
VOUT
3.3V
40A
LTC3882-1
Typical Applications
38821f
105
106
RUN
ALERT
SDA
SCL
R14
4.22k
VDD25
R15
10k
R18
10k
R34
24.9k
R17
10k
R35
5.76k
R19
4.99k
VOUT0_CFG
ASEL0
ASEL1
SYNC
SHARE_CLK
For more information www.linear.com/LTC3882-1
C35
100pF
30
39
41
GND
LTC3882-1
VCC
25
36
37
PGOOD1
PWM1
27
26
32
31
ISENSE1+
ISENSE1–
VSENSE1
1
2
TSNS0
3
TSNS1
28
COMP1
29
FB1
+ 33
COMP0
35
VSENSE0–
34
VSENSE1–
40
FB0
VSENSE0+
ISENSE0–
7
PWM0
6
PGOOD0
38
ISENSE0+
VINSNS
C17
1µF
4
VDD33
PWM1
VDD33
R30
60.4k
C27
18pF
C26
180pF
R21
100k
VIN
TSNS0
TSNS1
C130
0.22µF
VSENSE–
VSENSE+
C110
47µF
38821 F54
C119
0.22µF
R29
C25 4.12k
470pF
VDD33
+3.3V INPUT SUPPLY
PWM0
C18
4.7µF
C111
22µF
+7V
VOUTB
VOUTA
PWMB
PWMA
TEMPB
TEMPA
2
6
11
GND GND GND
+CSB
–CSB
D12S1R845A
–CSA
+CSA
VINB
VINA
4
16
15
5
3
10
12
C107
2.2µF
POWER BLOCK: DELTA D12S1R845A
TSNS1
TSNS0
C129
1.0µF
8
7
C124
1.0µF
1
14
9
13
+7V INPUT SUPPLY
Figure 54. High Density 1.5V/45A 650kHz Converter Using Dual Power Block
5
IAVG_GND
IAVG1
IAVG0
VOUT1_CFG
20
FREQ_CFG
21
PHAS_CFG
19
18
17
16
8
23
22
VDD25
C8
100nF
VDD33 VDD25
SCL
24
R5
1Ω
10
SDA
11
ALERT
12
FAULT0
13
FAULT1
14
RUN0
15
RUN1
R20
4.99k 9
C16
0.1µF
VIN
7V TO 13.2V
+
VSENSE–
C117
100µF
×2
VSENSE+
PWM1
PWM0
C120
470µF
×3
GND
VOUT
1.5V
45A
LTC3882-1
Typical Applications
38821f
LTC3882-1
Package Description
Please refer to http://www.linear.com/designtools/packaging/ for the most recent package drawings.
UJ Package
40-Lead Plastic QFN (6mm × 6mm)
(Reference LTC DWG # 05-08-1728 Rev Ø)
0.70 ±0.05
6.50 ±0.05
5.10 ±0.05
4.42 ±0.05
4.50 ±0.05
(4 SIDES)
4.42 ±0.05
PACKAGE OUTLINE
0.25 ±0.05
0.50 BSC
RECOMMENDED SOLDER PAD PITCH AND DIMENSIONS
APPLY SOLDER MASK TO AREAS THAT ARE NOT SOLDERED
6.00 ±0.10
(4 SIDES)
0.75 ±0.05
R = 0.10
TYP
R = 0.115
TYP
39 40
0.40 ±0.10
PIN 1 TOP MARK
(SEE NOTE 6)
1
4.50 REF
(4-SIDES)
4.42 ±0.10
2
PIN 1 NOTCH
R = 0.45 OR
0.35 × 45°
CHAMFER
4.42 ±0.10
(UJ40) QFN REV Ø 0406
0.200 REF
0.00 – 0.05
NOTE:
1. DRAWING IS A JEDEC PACKAGE OUTLINE VARIATION OF (WJJD-2)
2. DRAWING NOT TO SCALE
3. ALL DIMENSIONS ARE IN MILLIMETERS
4. DIMENSIONS OF EXPOSED PAD ON BOTTOM OF PACKAGE DO NOT INCLUDE
MOLD FLASH. MOLD FLASH, IF PRESENT, SHALL NOT EXCEED 0.20mm ON ANY SIDE, IF PRESENT
5. EXPOSED PAD SHALL BE SOLDER PLATED
6. SHADED AREA IS ONLY A REFERENCE FOR PIN 1 LOCATION ON THE TOP AND BOTTOM OF PACKAGE
0.25 ±0.05
0.50 BSC
BOTTOM VIEW—EXPOSED PAD
38821f
Information furnished by Linear Technology Corporation is believed to be accurate and reliable.
However, no responsibility is assumed for its use. Linear Technology Corporation makes no representation that the interconnection
of itsinformation
circuits as described
herein will not infringe on existing patent rights.
For more
www.linear.com/LTC3882-1
107
LTC3882-1
Typical Application
VIN
7V TO 14V
1Ω
330µF
×2
+
10k
4.99k
10k
10k
4.99k
10k
5VBIAS
100nF
VDD33
VDD25
VCC
2.2µF
9
VDD25
16.2k
17.4k
17
7
6
PGOOD0
38
ISENSE0+
ISENSE0
VSENSE0+
19
22µF
×2
100k
VDD33
COMP0
1
7.32k
IAVG_GND
GND
5
28
29
33
VSENSE1+
VSENSE1– 34
32
ISENSE1–
31
ISENSE1+
26
PWM1
27
PGOOD1
PWM
PWM0
DISB
RUN
220pF
1.21k
+
VOUT0
1.8V, 30A
470µF
×4
0.22µF
PLACE Q1,Q2 NEAR L1, L2 RESPECTIVELY
10nF
Q2
5VBIAS
220pF
10nF
100pF
100µF
×4
GND
137Ω
Q1
7.32k
L1
0.22µH
PULSE
VSWH
GL
10nF
100pF
10nF
VOUT1_CFG
20
FREQ_CFG
21
PHAS_CFG
39
IAVG0
30
IAVG1
FDMF6820A
CGND
2
TSNS0
3
TSNS1
FB1
VCIN SMOD
36
ASEL0
VOUT0_CFG
VIN PHASE GH CGND GND BOOT VDRV
PGND
35
VSENSE0–
40
FB0
COMP1
18
PWM0
– 37
ASEL1
2k
1µF
VCC
PWM0
LTC3882-1
16
+5V INPUT SUPPLY
10Ω
25
VINSNS
SCL
2.2µF
VIN
4.7µF
4
22
VDD33 VDD25
10
SDA
11
ALERT
12
FAULT0
13
FAULT1
14
RUN0
15
RUN1
23
SHARE_CLK
8
SYNC
RUN
INPUT SUPPLY +5V TO +12V
1µF
24
4.99k
1µF
1µF
137Ω
2.2µF
10Ω
2k
VIN
1µF
22µF
×2
PWM1
100k
0.22µF
VIN PHASE GH CGND GND BOOT VDRV
PGND
VDD33
41
FDMF6820A
VCIN SMOD
PWM
PWM1
DISB
RUN
VSWH
CGND
DrMOS: FAIRCHILD FD6802A
GL
1.21k
L2
0.22µH
PULSE
100µF
×4
+
VOUT1
1V, 40A
470µF
×4
GND
38821 F55
Figure 55. 1V/40A and 1.8V/30A 500kHz Converter with DrMOS Power Stage
Related Parts
PART NUMBER DESCRIPTION
COMMENTS
LTM4676A
Dual 13A or Single 26A Step-Down DC/DC µModule
Regulator with Digital Power Management
VIN Up to 26.5V; 0.5V ≤ VOUT (±0.5%) ≤ 5.4V, ±2% IOUT ADC Accuracy,
Fault Logging, I2C/PMBus Interface, 16mm × 16mm × 5mm, BGA
Package
LTC3887/
LTC3887-1
Dual Output PolyPhase Step-Down DC/DC Controller with
Digital Power System Management, 70ms Start-Up
VIN Up to 24V, 0.5V ≤ VOUT0,1 ≤ 5.5V, 70mS Start-Up, I2C/PMBus
Interface, with EEPROM and 16-Bit ADC. –1 Version Operates with
DrMOS and Power Blocks
LTC3886
60V Dual Output Step-Down Controller with Digital Power
System Management. 70ms Start-Up
VIN Up to 60V, 0.5V ≤ VOUT ≤ 13.8V, Analog Control Loop, I2C/
PMBus Interface with EEPROM and 16-Bit ADC, Programmable Loop
Compensation
LTC3880/
LTC3880-1
Dual Output PolyPhase Step-Down DC/DC Controller with
Digital Power System Management
VIN Up to 24V, 0.5V ≤ VOUT ≤ 5.5V, Analog Control Loop, I2C/PMBus
Interface with EEPROM and 16-Bit ADC
LTC3883/
LTC3883-1
Single Phase Step-Down DC/DC Controller with Digital Power VIN Up to 24V, 0.5V ≤ VOUT ≤ 5.5V, Input Current Sense Amplifier, I2C/
System Management
PMBus Interface with EEPROM and 16-Bit ADC
LTC2977
8-Channel PMBus Power System Manager Featuring
Accurate Output Voltage Measurement
Fault Logging to Internal EEPROM Monitors Eight Output Voltages, Input
Voltage and Die Temperature
LTC2974
Quad Digital Power Supply Manager with EEPROM
Controls and Monitors Four Outputs, 16-Bit ADC, Differential Inputs,
with Fault Logging
LTC3774
Dual, Multiphase Current Mode Synchronous Controller for
Sub-Milliohm DCR Sensing, with Remote Sense
Operates with Power Blocks, DrMOS Devices or External MOSFETs 4.5V
≤ VIN ≤ 38V
LTC3861
Dual, Multiphase, Synchronous Step-Down DC/DC Controller Operates with Power Blocks, DrMOS Devices or External MOSFETs 3V≤
with Accurate Current Share
VIN ≤ 24V
LTC4449
High Speed Synchronous N-Channel MOSFET Driver
108 Linear Technology Corporation
VIN Up to 38V, 4V ≤ VCC ≤ 6.5V, Adaptive Shoot-Through Protection,
2mm × 3mm DFN-8 Package
1630 McCarthy Blvd., Milpitas, CA
For95035-7417
more information www.linear.com/LTC3882-1
(408) 432-1900 ● FAX: (408) 434-0507
●
www.linear.com/LTC3882-1
38821f
LT 0915 • PRINTED IN USA
LINEAR TECHNOLOGY CORPORATION 2015