MAXIM MAX16067ETJ+

19-5028; Rev 1; 2/10
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
The MAX16067 flash-configurable system manager
monitors and sequences multiple system voltages. The
MAX16067 manages up to six system voltages simultaneously. The MAX16067 integrates an analog-to-digital
converter (ADC) and configurable outputs for sequencing power supplies. Device configuration information,
including overvoltage and undervoltage limits, time
delay settings, and the sequencing order is stored in
nonvolatile flash memory. During a fault condition, fault
flags and channel voltages can be automatically stored
in the nonvolatile flash memory for later readback.
The internal 1% accurate, 10-bit ADC measures each
input and compares the result to one overvoltage and
one undervoltage limit. A fault signal asserts when a
monitored voltage falls outside the set limits.
The MAX16067 supports a power-supply voltage of up to
14V and can be powered directly from the 12V intermediate bus in many systems.
The integrated sequencer provides precise control over
the power-up and power-down order of up to six power
supplies. Three outputs (EN_OUT1 to EN_OUT3) are
configurable with charge-pump outputs to directly drive
external n-channel MOSFETs.
The MAX16067 includes six programmable generalpurpose inputs/outputs (GPIOs). GPIOs are flash configurable as a fault output, as a watchdog input or output,
or as a manual reset.
Features
S Operates from 2.8V to 14V
S 1% Accurate, 10-Bit ADC Monitors 6 Voltage
Inputs
S Analog EN Monitoring Input
S 6 Monitored Inputs with Overvoltage and
Undervoltage Limits
S Nonvolatile Fault Event Logger
S Power-Up and Power-Down Sequencing
Capability
S 6 Outputs for Sequencing/Power-Good Indicators
S 3 Configurable Charge-Pump Outputs
S Six General-Purpose Inputs/Outputs Configurable
as:
Dedicated Fault Output
Watchdog Timer Function
Manual Reset
SMBus Alert
Fault Propagation Input/Output
S SMBus and JTAG Interface
S Supports Cascading with MAX16065/MAX16066
S Flash-Configurable Time Delays and Thresholds
S -40NC to +85NC Extended Operating Temperature
Range
Applications
The MAX16067 features nonvolatile fault memory for
recording information during system shutdown events.
The fault logger records a failure in the internal flash
and sets a lock bit protecting the stored fault data from
accidental erasure.
An SMBus™ or a JTAG serial interface configures the
MAX16067. The MAX16067 is available in a 32-pin, 5mm
x 5mm, TQFN package and is fully specified over the
-40NC to +85NC extended temperature range.
Networking Equipment
Telecom Equipment (Base Stations, Access)
Storage/Raid Systems
Servers
Typical Operating Circuit appears at end of data sheet.
Ordering Information/Selector Guide
PART
PIN-PACKAGE
VOLTAGEDETECTOR INPUTS
GENERAL-PURPOSE
INPUTS/OUTPUTS
MAX16067ETJ+
32 TQFN-EP*
6
Note: This device is specified over the -40NC to +85NC extended temperature range.
+Denotes a lead(Pb)-free/RoHS-compliant package.
*EP = Exposed pad.
6
SEQUENCING
OUTPUTS
6
SMBus is a trademark of Intel Corp.
________________________________________________________________ Maxim Integrated Products 1
For pricing, delivery, and ordering information, please contact Maxim Direct at 1-888-629-4642,
or visit Maxim’s website at www.maxim-ic.com.
MAX16067
General Description
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
ABSOLUTE MAXIMUM RATINGS
VCC to GND................……………………………….-0.3V to +15V
MON_, SCL, SDA, A0 to GND.................................-0.3V to +6V
EN, TCK, TMS, TDI to GND.....................................-0.3V to +4V
TDO to GND.............................................-0.3V to (VDBP + 0.3V)
EN_OUT1, EN_OUT2, EN_OUT3
(configured as open-drain) to GND...................-0.3V to +15V
EN_OUT1, EN_OUT2, EN_OUT3
(configured as charge pump) to GND...............-0.3V to +15V
EN_OUT4, EN_OUT5, EN_OUT6, RESET, GPIO_
(configured as open-drain) to GND......................-0.3V to +6V
EN_OUT_, RESET, GPIO_ (configured as push-pull)
to GND..................................................-0.3V to (VDBP + 0.3V)
DBP, ABP to GND........................................-0.3V to minimum of
(4V and (VCC + 0.3V))
Continuous Current (all other pins).................................. Q20mA
Continuous Current (GND, pin 5)..................................... Q30mA
Continuous Power Dissipation (TA = +70NC)
32-Pin TQFN (derate 34.5mW/NC above +70NC)...... 2759mW*
Thermal Resistance (Note 1)
BJA. ...............................................................................29NC/W
BJC..................................................................................2NC/W
Operating Temperature Range........................... -40NC to +85NC
Junction Temperature......................................................+150NC
Storage Temperature Range............................. -65NC to +150NC
Lead Temperature (soldering, 10s).................................+300NC
Soldering Temperature (reflow).......................................+260NC
*As per JEDEC 51 Standard, Multilayer Board (PCB).
Note 1: Package thermal resistances were obtained using the method described in JEDEC specification JESD51-7, using a fourlayer board. For detailed information on package thermal considerations, refer to www.maxim-ic.com/thermal-tutorial.
Stresses beyond those listed under “Absolute Maximum Ratings” may cause permanent damage to the device. These are stress ratings only, and functional
operation of the device at these or any other conditions beyond those indicated in the operational sections of the specifications is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect device reliability.
ELECTRICAL CHARACTERISTICS
(VCC = 2.8V to 14V, TA = TJ = -40NC to +85NC, unless otherwise specified. Typical values are at VABP = VDBP = VCC = 3.3V,
TA = +25NC.) (Note 2)
PARAMETER
Operating Voltage Range
Undervoltage Lockout
Undervoltage Lockout Hysteresis
Minimum Flash Operating Voltage
SYMBOL
VCC
VUVLO
CONDITIONS
RESET output asserted low
MIN
2.8
MAX
14
Minimum voltage on VCC to ensure the
device is flash configurable
2.7
UVLOHYS
VFLASH
TYP
1.2
55
Minimum voltage on VCC to ensure flash
erase and write operations
UNITS
V
V
mV
2.7
V
ICC1
No load on any output
2.8
4
Supply Current
ICC2
No load on any output, during flash writing
cycle
7.7
14
DBP Regulator Voltage
VDBP
VCC = 5V, CDBP = 1FF, no load
2.8
3
3.2
ABP Regulator Voltage
VABP
VCC = 5V, CABP = 1FF, no load
2.85
3
3.15
V
Boot Time
tBOOT
VCC > VUVLO
100
200
Fs
VCC = VABP = VDBP = 3.6V (Note 3)
Flash Writing Time
8-byte word
Internal Timing Accuracy
(Note 4)
mA
5
122
-10
V
ms
+10
%
ADC
Resolution
10
Bits
TA = +25NC
0.35
TA = -40NC to +85NC
0.75
Gain Error
ADCGAIN
Offset Error
ADCOFF
1.50
LSB
Integral Nonlinearity
ADCINL
1
LSB
2 _______________________________________________________________________________________
%
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
(VCC = 2.8V to 14V, TA = TJ = -40NC to +85NC, unless otherwise specified. Typical values are at VABP = VDBP = VCC = 3.3V,
TA = +25NC.) (Note 2)
PARAMETER
Differential Nonlinearity
ADC Total Monitoring Cycle Time
ADC MON_ Ranges
ADC LSB Step Size
SYMBOL
CONDITIONS
MIN
TYP
ADCDNL
tCYCLE
ADCRNG
ADCLSB
Monitoring all 6 inputs, no MON_ fault
detected
24
MON_ range set to ‘00’
5.552
MON_ range set to ‘01’
2.776
MON_ range set to ‘10’
1.388
MON_ range set to ‘00’
5.42
MON_ range set to ‘01’
2.71
MON_ range set to ‘10’
1.35
ADC Input Leakage Current
MAX
UNITS
1
LSB
30
Fs
V
mV
1
FA
ENABLE INPUT (EN)
EN Input-Voltage Threshold
EN Input Current
VTH_EN_R
EN voltage rising
VTH_EN_F
EN voltage falling
IEN
EN Input-Voltage Range
1.24
1.195
1.215
1.235
V
-0.5
+0.5
FA
0
3.6
V
OUTPUTS (EN_OUT_, RESET, GPIO_)
Output Voltage Low
VOL
Maximum Output Sink Current
Output-Voltage High (Push-Pull)
VOH
Output-Voltage High (EN_OUT1,
EN_OUT2, EN_OUT3 Configured
as Charge Pumps)
ISINK = 2mA
0.4
ISINK = 10mA, GPIO_ only
0.7
VCC = 1.2V, ISINK = 100FA (RESET only)
0.3
Total current into EN_OUT_, RESET, GPIO_,
VCC = 3.3V
30
ISOURCE =100FA
2.4
VOH_CP
IEN_OUT_= 1FA
11
11.7
EN_OUT_ Pullup Current
(Charge Pump)
ICH_UP
VEN_OUT_ = 1V
5.4
7.9
Output Leakage Current
(Open Drain)
IOUT_LKG
V
mA
V
13
FA
1
EN_OUT1, EN_OUT2, EN_OUT3 > 11.8V
V
5
FA
INPUTS (A0, GPIO_)
Input Logic-Low
VIL
Input Logic-High
VIH
2.0
0.8
V
V
WDI Pulse Width
tWDI
100
ns
MR Pulse Width
tMR
2
Fs
SMBus INTERFACE
Logic-Input Low Voltage
VIL
Input voltage falling
Logic-Input High Voltage
VIH
Input voltage rising
2.0
VCC shorted to GND, VMON_ = 0 or 6V
-1
Input Leakage Current
Output Sink Current
VOL
Input Capacitance
CIN
0.8
V
+1
FA
V
ISINK = 3mA
0.4
5
V
pF
_______________________________________________________________________________________ 3
MAX16067
ELECTRICAL CHARACTERISTICS (continued)
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
ELECTRICAL CHARACTERISTICS (continued)
(VCC = 2.8V to 14V, TA = TJ = -40NC to +85NC, unless otherwise specified. Typical values are at VABP = VDBP = VCC = 3.3V,
TA = +25NC.) (Note 2)
PARAMETER
SYMBOL
CONDITIONS
MIN
TYP
MAX
UNITS
400
kHz
SMBus TIMING
Serial-Clock Frequency
fSCL
Bus Free Time Between STOP
and START Condition
tBUF
1.3
Fs
START Condition Setup Time
tSU:STA
0.6
Fs
START Condition Hold Time
tHD:STA
0.6
Fs
STOP Condition Setup Time
tSU:STO
0.6
Fs
Clock Low Period
tLOW
1.3
Fs
Clock High Period
tHIGH
0.6
Fs
Data Setup Time
tSU:DAT
Output Fall Time
tOF
Data Hold Time
Pulse Width of Spike Suppressed
SMBus Timeout
tHD:DAT
100
From 50% SCL falling to SDA
change
Receive
0.15
Transmit
0.3
tSP
tTIMEOUT
ns
250
10pF P CBUS P 400pF
0.9
250
SMBCLK time low for reset
22
ns
Fs
ns
35
ms
0.8
V
JTAG INTERFACE
TDI, TMS, TCK Logic-Low Input
Voltage
VIL
Input voltage falling
TDI, TMS, TCK Logic-High Input
Voltage
VIH
Input voltage rising
TDO Logic-Output Low Voltage
VOL_TDO
TDO Logic-Output High Voltage
VOH_TDO
TDI, TMS Pullup Resistors
RJPU
I/O Capacitance
CI/O
TCK Clock Period
TCK High/Low Time
2.0
V
ISINK = 3mA
0.4
ISOURCE = 200FA
2.4
Pullup to DBP
30
V
V
50
65
5
t1
kI
pF
1000
500
ns
t2, t3
50
ns
TCK to TMS, TDI Setup Time
t4
15
TCK to TMS, TDI Hold Time
t5
15
TCK to TDO Delay
t6
500
ns
TCK to TDO High-Z Delay
t7
500
ns
ns
ns
Note 2: Specifications are guaranteed for the stated global conditions, unless otherwise noted. 100% production tested at TA =
+25NC and TA = +85NC. Specifications at TA = -40NC are guaranteed by design.
Note 3: For VCC of 3.6V or lower, connect VCC, DBP, and ABP together. For higher supply applications, connect only VCC to the
supply rail.
Note 4: Applies to RESET (except for reset timeout period of 25Fs), fault, autoretry, sequence delays, and watchdog timeout.
4 _______________________________________________________________________________________
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
MAX16067
SDA
tBUF
tSU:DAT
tSU:STA
tHD:DAT
tLOW
tHD:STA
tSU:STO
SCL
tHIGH
tHD:STA
tR
tF
START
CONDITION
STOP
CONDITION
REPEATED START
CONDITION
START
CONDITION
Figure 1. SMBus Timing Diagram
t1
t2
t3
TCK
t4
t5
TDI, TMS
t6
t7
TDO
Figure 2. JTAG Timing Diagram
_______________________________________________________________________________________ 5
Typical Operating Characteristics
(Typical values are at VCC = 3.3V, TA = +25NC.)
VCC SUPPLY CURRENT
vs. VCC SUPPLY VOLTAGE
NORMALIZED MON_THRESHOLD
vs. TEMPERATURE
TA = +25°C
TA = -40°C
1.5
ABP AND DBP
REGULATORS ACTIVE
1.0
0.5
0
0
2
4
6
8
VCC (V)
10
12
1.025
1.010
0.995
0.980
5.6V RANGE
HALF-SCALE
PUV THRESHOLD
0.965
25
20
15
10
0
20
40
60
1
MAX16067 toc05
0.995
0.980
MAX16067 toc03
60
50
40
30
20
0
-20
-40
0
20
40
60
80
2
4
TEMPERATURE (°C)
0.40
MAX16067 toc07
0.35
600
MIN
EN_OUT1–
EN_OUT3
0.30
VOUT (V)
700
0.25
0.20
0.15
400
RESET, GPIO_,
AND EN_OUT4–
EN_OUT6
0.10
300
8
DEGLITCH VALUE
OUTPUT VOLTAGE vs. SINK CURRENT
(OUT = LOW)
800
DELAY (ns)
70
10
MAX
0.05
200
0
100
-40
-20
0
20
40
TEMPERATURE (°C)
60
80
80
1.010
1100
500
60
90
1.025
MR TO RESET PROPAGATION DELAY
vs. TEMPERATURE
900
40
TRANSIENT DURATION
vs. MON_ DEGLITCH
EN OVERDRIVE (mV)
1000
20
NORMALIZED TIMING ACCURACY
vs. TEMPERATURE
1.040
100
10
0
TEMPERATURE (°C)
0.950
0
-20
-40
TEMPERATURE (°C)
0.965
5
0.980
80
1.055
NORMALIZED SLOT DELAY
MAX16067 toc04
30
0.995
0.950
-20
-40
TRANSIENT DURATION
vs. THRESHOLD OVERDRIVE (EN)
35
1.010
0.965
0.950
14
1.025
MAX16067 toc08
2.0
1.040
MAX16067 toc06
2.5
1.040
NORMALIZED EN THRESHOLD
ICC (mA)
3.0
TRANSIENT DURATION (us)
TA = +85°C
1.055
MAX16067 toc02
3.5
NORMALIZED EN THRESHOLD
vs. TEMPERATURE
1.055
NORMALIZED MON_ THRESHOLD
ABP AND DBP CONNECTED
TO VCC
MAX16067 toc01
4.0
TRANSIENT DURATION (µs)
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
80
0
5
10
15
ISINK (mA)
6 _______________________________________________________________________________________
20
16
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
OUTPUT-VOLTAGE HIGH vs. SOURCE
CURRENT (CHARGE-PUMP OUTPUT)
3.3
3.2
0.8
0.6
3.1
6
0.4
GPIO_
3.0
2.9
2.8
2.7
4
2.5
0
1
2
3
4
5
6
7
8
200
0
ISOURCE (µA)
0
-0.2
-0.6
-0.8
2.4
400
600
800
1000
-1.0
1200
0
128 256 384 512 640 768 896 1024
ISOURCE (µA)
CODE (LSB)
FET TURN_ON WITH CHARGE PUMP
DIFFERENTIAL NONLINEARITY vs. CODE
1.0
MAX16067 toc13
MAX16067 toc12
0.8
0.6
VEN_OUT1
5V/div
0.4
0.2
0
IOUT
1A/div
-0.2
-0.4
-0.6
VOUT
5V/div
-0.8
-1.0
0
100ms/div
128 256 384 512 640 768 896 1024
CODE (LSB)
RESET OUTPUT CURRENT
vs. VCC SUPPLY VOLTAGE
SEQUENCING
MAX16067 toc14
MAX16067 toc15
25
ABP AND DBP CONNECTED TO VCC
20
VEN_OUT1
2V/div
VEN_OUT2
2V/div
VEN_OUT3
2V/div
VEN_OUT4
2V/div
OUTPUT CURRENT (mA)
0
0.2
-0.4
EN_OUT1,
EN_OUT2,
EN_OUT3
2.6
2
INL (LSB)
VOUT (V)
8
DNL (LSB)
VOUT (V)
10
RESET
MAX16067 toc11
12
INTEGRAL NONLINEARITY vs. CODE
1.0
MAX16067 toc10
3.4
MAX16067 toc09
14
OUTPUT-VOLTAGE HIGH vs.
SOURCE CURRENT (PUSH-PULL OUTPUT)
15
ABP AND DBP
REGULATORS ACTIVE
10
5
VRESET = 0.3V
200µs/div
0
0
2
4
6
8
10
12
14
VCC (V)
_______________________________________________________________________________________ 7
MAX16067
Typical Operating Characteristics (continued)
(Typical values are at VCC = 3.3V, TA = +25NC.)
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
GND
ABP
VCC
DBP
EN
EN_OUT1
EN_OUT2
TOP VIEW
MON1
MAX16067
Pin Configuration
24
23
22
21
20
19
18
17
MON2 25
16
EN_OUT3
MON3 26
15
EN_OUT4
MON4 27
14
EN_OUT5
MON5 28
13
EN_OUT6
12
TMS
11
TCK
10
TDI
9
TDO
MAX16067
MON6 29
RESET 30
GPIO1 31
*EP
+
6
7
8
SDA
GPIO5
5
SCL
GPIO4
4
AO
3
GND
2
GPIO6
1
GPIO3
GPIO2 32
TQFN
*CONNECT EXPOSED PAD TO GND.
Pin Description
PIN
NAME
FUNCTION
1–4, 31, 32
GPIO3–GPIO6,
GPIO1, GPIO2
General-Purpose Inputs/Outputs. Each GPIO_ can be configured to act as an input, a push-pull
output, an open-drain output, or a special function.
5, 23
GND
Ground. Connect all GNDs together.
6
A0
7
SCL
SMBus Serial-Clock Input
8
SDA
SMBus Serial-Data Open-Drain Input/Output
9
TDO
JTAG Test Data Output
10
TDI
JTAG Test Data Input
11
TCK
JTAG Test Clock
12
TMS
JTAG Test Mode Select
13–18
EN_OUT6–
EN_OUT1
19
EN
Four-State SMBus Address. Address is sampled upon POR.
Outputs. Set EN_OUT_ with an active-high/active-low logic and with push-pull or open-drain
configuration. EN_OUT_ can be asserted by a combination of MON_ voltages configurable
through the flash. EN_OUT1–EN_OUT3 can be configured with a charge-pump output (+12V
above GND) that can drive an external n-channel MOSFET. All EN_OUT_ can be configured as
GPIOs.
Analog Enable Input. All outputs deassert when VEN is below the enable threshold.
8 _______________________________________________________________________________________
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
PIN
NAME
FUNCTION
20
DBP
Digital Bypass. All push-pull outputs are referenced to DBP. Bypass DBP with a 1FF capacitor
to GND.
21
VCC
Power-Supply Input. Bypass VCC to GND with a 10FF ceramic capacitor.
22
ABP
Analog Bypass. Bypass ABP to GND with a 1FF ceramic capacitor.
24–29
MON1–MON6
30
RESET
—
EP
Monitor Voltage Inputs. Set the monitor voltage range through the configuration registers.
Measured values are written to the ADC registers and can be read back through the SMBus or
JTAG interface.
Configurable Reset Output
Exposed Pad. Internally connected to GND. Connect to ground, but do not use EP as the main
ground connection.
Functional Diagram
VCC ABP DBP
REF
REG
EN
DECODE
LOGIC
VTH_EN
ALERT
GPIO1
FAULT
GPIO2
EXTFAULT
MON2
MON4
MON5
MON6
GPIO
CONTROL
MAX16067
10-BIT ADC
(SAR)
ADC
REGISTERS
GPIO5
EN_OUT1–
EN_OUT6
SEQUENCER
STATE
MACHINE
VOLTAGE
AND
SCALING
MUX
DIGITAL
COMPARATORS
GPIO6
GPIO4
WDI
WATCHDOG
WDO
TIMER
MON1
MON3
GPIO3
MR
RESET
OUTPUT
LOGIC
RESET
RAM
REGISTERS
SMBus
INTERFACE
AO SCL SDA
FLASH
REGISTERS
GND
JTAG
INTERFACE
TDO TDI TCK TMS
_______________________________________________________________________________________ 9
MAX16067
Pin Description (continued)
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Detailed Description
The MAX16067 manages up to six system power supplies. After boot-up, if EN is high and the software-enable
bit is set to ‘1,’ a power-up sequence begins based on
the configuration stored in flash and the EN_OUT_s are
controlled accordingly. When the power-up sequence is
successfully completed, the monitoring phase begins.
An internal multiplexer cycles through each MON_ input.
At each multiplexer stop, the 10-bit ADC converts the
monitored analog voltage to a digital result and stores
the result in a register. Each time a conversion cycle
(5Fs, max) completes, internal logic circuitry compares
the conversion results to the overvoltage and undervoltage thresholds stored in memory. When a result violates
a programmed threshold, the conversion can be configured to generate a fault. GPIO_ can be programmed
to assert on combinations of faults. Additionally, faults
can be configured to shut off the system and trigger the
nonvolatile fault logger, which writes all fault information
automatically to the flash and write-protects the data to
prevent accidental erasure.
The MAX16067 contains both SMBus and JTAG serial
interfaces for accessing registers and flash. Use only
one interface at any given time. For more information
on how to access the internal memory through these
interfaces, see the SMBus-Compatible Serial Interface
and JTAG Serial Interface sections. The memory map
is divided into three pages with access controlled by
special SMBus and JTAG commands.
The factory-default values at POR (power-on reset) for all
RAM registers are ‘0’s. POR occurs when VCC reaches
the undervoltage-lockout threshold (UVLO) of 2.7V
(max). At POR, the device begins a boot-up sequence.
During the boot-up sequence, all monitored inputs are
masked from initiating faults and flash contents are
copied to the respective register locations. During bootup, the MAX16067 is not accessible through the serial
interface. The boot-up sequence takes up to 150Fs, after
which the device is ready for normal operation. RESET
is asserted low up to the boot-up phase after which it
assumes its programmed active state. RESET remains
active for its programmed timeout period once sequencing is completed and all monitored channels are within
their respective thresholds. Up to the boot-up phase, the
GPIO_s and EN_OUT_s are high impedance.
Power
Apply 2.8V to 14V to VCC to power the MAX16067.
Bypass VCC to ground with a 10FF capacitor. Two internal voltage regulators, ABP and DBP, supply power to
the analog and digital circuitry within the device. For
operation at 3.6V or lower, disable the regulators by connecting ABP and DBP to VCC.
ABP is a 3.0V (typ) voltage regulator that powers the internal analog circuitry. Bypass ABP to GND with a 1FF ceramic capacitor installed as close as possible to the device.
DBP is an internal 3.0V (typ) voltage regulator. DBP
powers flash and digital circuitry. All push-pull outputs
refer to DBP. DBP supplies the input voltage to the internal charge pump when the programmable outputs are
configured as charge-pump outputs. Bypass the DBP
output to GND with a 1FF ceramic capacitor installed as
close as possible to the device.
Do not power external circuitry from ABP or DBP.
Sequencing
To sequence a system of power supplies safely, the output voltage of a power supply must be good before the
next power supply may turn on. Connect EN_OUT_ outputs to the enable input of the external power supplies
and connect MON_ inputs to the output of the power
supplies for voltage monitoring. More than one MON_
can be used if the power supply has multiple outputs.
Sequence Order
The MAX16067 provides a system of ordered slots to
sequence multiple power supplies. To determine the
sequence order, assign each EN_OUT_ to a slot ranging
from Slot 1 to Slot 6 (Table 6b). EN_OUT_(s) assigned to
Slot 1 are turned on first, followed by outputs assigned
to Slot 2 through Slot 6. Multiple EN_OUT_s assigned to
the same slot turn on at the same time.
Each slot includes a built-in configurable sequence
delay (registers r77h to r7Dh) ranging from 80Fs to
5.079s. During a reverse sequence, slots are turned off
in reverse order starting from Slot 6. The MAX16067 can
be configured to power down in simultaneous mode or
in reverse-sequence mode as set in r75h[0]. Set r75h[0]
to ‘1’ for reverse sequence power-down.
See Tables 5 and 6 for the MON_ and EN_OUT_ slot assignment bits, and Tables 2 and 3 for the sequence delays.
During power-up or power-down sequencing, the current sequencer state can be found in r21h[3:0].
10 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
REGISTER
ADDRESS
BIT RANGE
21h
DESCRIPTION
[3:0]
Current-sequencer state
0000 = Slot0
0001 = Slot1
0010 = Slot2
0011 = Slot3
0100 = Slot4
0101 = Slot5
0110 = Slot6
0111 = Power-on mode
1000 = Fault state
1001 to 1111 = Unused
[7:4]
Reserved
A sequencing delay occurs between each slot and is
configured in registers 77h–7Dh as shown in Table 2.
Each sequencing delay is stored as an 8-bit value and
is calculated as follows:
(
)
t SEQ = 5 × 10 −6 × 2 a × (16 + b)
where tSEQ is in seconds, a is the decimal value of the
4 MSBs and b is the decimal value of the 4 LSBs. See
Table 3 for example calculations.
Enable Input (EN)
To initiate sequencing and enable monitoring, the voltage at EN must be above 1.24V (typ) and the software
enable bit in r73h[0] must be set to ‘1.’ To power down
and disable monitoring, either pull EN below 1.215V
(typ) or set the software enable bit to ‘0.’ See Table 4 for
the software enable bit configurations. Connect EN to
ABP if not used.
If a fault condition occurs during the power-up cycle,
the EN_OUT_ outputs are powered down immediately,
regardless of the state of EN. In the monitoring state,
if EN falls below the threshold, the sequencing state
machine begins the power-down sequence. If EN rises
above the threshold during the power-down sequence,
the sequence state machine continues the power-down
sequence until all the channels are powered off and then
the device immediately begins the power-up sequence.
When in the monitoring state, and when EN falls below
the undervoltage threshold, a register bit, ENRESET
(r20h[2]), is set to a ‘1.’ This register bit latches and must
be cleared through software. This bit indicates if RESET
asserted low due to EN going under the threshold. The
POR state of ENRESET is ‘0’. The bit is only set on a falling edge of the EN comparator output or the software
enable bit. If operating in latch-on fault mode, toggle
EN or toggle the software enable bit to clear the latch
condition and restart the device once the fault condition
has been removed.
Table 2. Slot Delay Register
REGISTER
ADDRESS
FLASH
ADDRESS
BIT RANGE
77h
277h
[7:0]
Sequence slot 0 to slot 1 delay
78h
278h
[7:0]
Sequence slot 1 to slot 2 delay
79h
279h
[7:0]
Sequence slot 2 to slot 3 delay
7Ah
27Ah
[7:0]
Sequence slot 3 to slot 4 delay
7Bh
27Bh
[7:0]
Sequence slot 4 to slot 5 delay
7Ch
27Ch
[7:0]
Sequence slot 5 to slot 6 delay
7Dh
27Dh
[7:0]
Sequence slot 6 to power-on state delay
DESCRIPTION
______________________________________________________________________________________ 11
MAX16067
Table 1. Current Sequencer Slot
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Table 3. Power-Up/Power-Down Slot Delays
Code
(
)
(
Value
)
t SEQ = 5 × 10 −6 × 2 a × (16 + b) = 5 × 10 −6 × 2 0 × (16 + 0) = 80µs
0000 0000
•
•
•
(
)
•
•
•
(
)
t SEQ = 5 × 10 −6 × 2 a × (16 + b) = 5 × 10 −6 × 215 × (16 + 15) = 5.079s
1111 1111
Table 4. Software Enable Configurations
REGISTER
ADDRESS
73h
FLASH
ADDRESS
273h
BIT RANGE
DESCRIPTION
[0]
Software enable
1 = Sequencing enabled
0 = Power-down
[1]
Reserved
[2]
1 = Margin mode enabled
[3]
Reserved
[4]
Independent watchdog mode enable
1 = Watchdog timer is independent of sequencer
0 = Watchdog timer boots after sequence completes
Monitoring Inputs While Sequencing
An enabled MON_ input can be assigned to a slot
ranging from Slot 1 to Slot 6. EN_OUT_s are always
asserted at the beginning of a slot. The supply voltages
connected to the MON_ inputs must exceed the undervoltage threshold before the programmed fault timeout
period expires, otherwise, a fault condition occurs. Once
a MON_ input crosses the undervoltage threshold, the
monitoring for overvoltage begins. The undervoltage
and overvoltage threshold checking cannot be disabled
during power-up and power-down. See Tables 5 and
6 for the MON_ slot assignment bits. The programmed
sequence delay is then counted before moving to the
next slot.
Slot 0 does not monitor any MON_ input and does not
control any EN_OUT_. Slot 0 waits for the software
enable bit r73h[0] to be a logic-high and for the voltage
on EN to rise above 1.24V (typ) before initiating the power-up sequence and counting its own sequence delay.
Any MON_ input that suffers a fault during power-up
sequencing causes all the EN_OUT_s to turn off and
the sequencer to shut down regardless of the state of
the critical fault enables (see the Faults section). If a
MON_ input is less critical to system operation, it can
be configured as “monitoring only” (see Table 6a) for
sequencing. Monitoring for MON_ inputs assigned as
“monitoring only” begins after sequencing is complete,
and can trigger a critical fault only if specifically configured to do so using the critical fault enables.
12 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
SLOT 1
SLOT 2
MAX16067
SLOT 0
SLOT 6
tFAULT
SLOT1-SLOT2
DELAY
EN_OUT1
SLOT1-SLOT2
DELAY
OV
BOTH ARE
ASSIGNED
TO SLOT 1
MON4
UV
UV/OV
MONITORING BEGINS
WHEN MON4
REACHES UV
THRESHOLD
MON4 MUST
REACH UV
THRESHOLD BY THIS
TIME
EN_OUT2
BOTH ARE
ASSIGNED
TO SLOT 2
RESET
MON3
TIMEOUT
MON5
RESET
EN
Figure 3. Delay and Reset Timing
______________________________________________________________________________________ 13
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Table 5. MON_ and EN_OUT_ Assignment Registers
REGISTER
ADDRESS
FLASH
ADDRESS
BIT RANGE
[2:0]
7Eh
27Eh
[3]
[6:4]
[7]
[2:0]
7Fh
27Fh
[3]
[6:4]
[7]
[2:0]
80h
280h
81h–83h
281h–283h
84h
284h
85h
285h
86h
286h
[3]
[6:4]
DESCRIPTION
MON1
Not used
MON2
Not used
MON3
Not used
MON4
Not used
MON5
Not used
MON6
[7]
Not used
—
Not used
[3:0]
EN_OUT1
[7:4]
EN_OUT2
[3:0]
EN_OUT3
[7:4]
EN_OUT4
[3:0]
EN_OUT5
[7:4]
EN_OUT6
Table 6a. MON_ Slot Assignment Codes
SLOT ASSIGNMENT
CODE
Table 6b. EN_OUT_ Slot Assignment
Codes
SLOT ASSIGNMENT
MON_ DESCRIPTION
000
Not assigned
001
Slot 1
010
Slot 2
011
Slot 3
100
Slot 4
101
Slot 5
110
Slot 6
111
Monitoring-only state
CODE
EN_OUT_ DESCRIPTION
0000
0001
0010
0011
0100
0101
0110
1101
1110
—
Not assigned
Slot 1
Slot 2
Slot 3
Slot 4
Slot 5
Slot 6
General-purpose input
General-purpose output
All other unspecified codes are not assigned.
14 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
When the tFAULT counter expires before all MON_ inputs
assigned to the slot are above the fault UV threshold, a
fault asserts. EN_OUT_ outputs are disabled and the
MAX16067 returns to the fault state. Register r75h[4:1]
sets the tFAULT delay. See Table 7 for details.
After the voltages on all MON_ inputs assigned to the
last slot exceed the UV fault threshold and the slot delay
expires, the MAX16067 starts the reset timeout counter.
After the reset timeout, RESET deasserts. See Table 22
for more information on setting the reset timeout.
Table 7. tFAULT Delay Settings
r75h[4:1]
FAULT DELAY
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
120Fs
150Fs
250Fs
380Fs
600Fs
1ms
1.5ms
2.5ms
4ms
6ms
10ms
15ms
25ms
40ms
60ms
100ms
Power-Down
Power-down starts when EN is pulled low or the software
enable bit is set to ‘0.’ Power down EN_OUT_s simultaneously or in reverse sequence mode by setting the
reverse sequence bit (r75h[0]) appropriately. Set r75h[0]
to ‘1’ to power down in reverse sequence.
is programmable in registers r43h–r44h (see Table
8). When MON_ configuration registers are set to ’11,’
MON_ voltages are not monitored and the multiplexer
does not stop at these inputs, decreasing the total cycle
time. These inputs cannot be configured to trigger fault
conditions.
Reverse Sequence Mode
When the MAX16067 is fully powered up and EN is
pulled low or the software enable bit is set to ‘0’, the
EN_OUT_s assigned to Slot 6 deassert, the MAX16067
waits for the Slot 6 sequence delay and then proceeds to
the previous slot (Slot 5), and so on until the EN_OUT_s
assigned to Slot 1 turn off. When simultaneous powerdown is selected (r75h[0] is set to ‘0’), all EN_OUT_s turn
off at the same time.
The two programmable thresholds for each monitored
voltage include an overvoltage and an undervoltage
threshold. See the Faults section for more information
on setting overvoltage and undervoltage thresholds. All
voltage thresholds are 8 bits wide. The 8 MSBs of the
10-bit ADC conversion result are compared to these
overvoltage and undervoltage thresholds.
Voltage Monitoring
The MAX16067 features an internal 10-bit ADC that monitors the MON_ voltage inputs. An internal multiplexer
cycles through each of the enabled inputs, taking less
than 24Fs for a complete monitoring cycle. Each acquisition takes approximately 4Fs. At each multiplexer stop,
the 10-bit ADC converts the analog input to a digital
result and stores the result in a register. ADC conversion
results are stored in registers r00h–r0Bh (see Table 9).
Use the SMBus or JTAG serial interface to read ADC
conversion results.
The MAX16067 provides six inputs, MON1–MON6,
for voltage monitoring. Each input-voltage range
For any undervoltage or overvoltage condition to be
monitored and any faults detected, the MON_ input must
be assigned to a sequence order or set to monitoring
mode as described in the Sequencing section. Inputs
that are not enabled are not converted by the ADC; they
contain the last value acquired before that channel was
disabled. The ADC conversion result registers are reset
to 00h at boot-up. These registers are not reset when a
reboot command is executed.
To temporarily disable voltage monitoring during voltage
margining conditions, set r73h[2] to ‘1’ to enable margining mode functionality. Faults, except for faults triggered
by EXTFAULT pulled low externally, are not recorded
when the device is in margining mode but the ADC
continues to run and conversion results continue to be
available. Set r73h[2] back to ‘0’ for normal functionality.
______________________________________________________________________________________ 15
MAX16067
Power-Up
On power-up, when EN is high and the software enable
bit is ‘1’, the MAX16067 begins sequencing with Slot
0. After the sequencing delay for Slot 0 expires, the
sequencer advances to Slot 1, and all EN_OUT_s
assigned to the slot assert. All MON_ inputs assigned to
Slot 1 are monitored and when the voltage rises above
the undervoltage (UV) fault threshold, the sequence
delay counter is started. When the sequence delay
expires, the MAX16067 proceeds to the next slot.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Table 8. ADC Configuration Registers
REGISTER ADDRESS
43h
44h
FLASH ADDRESS
BIT RANGE
DESCRIPTION
[1:0]
MON1 full-scale range
00 = 5.6V
01 = 2.8V
10 = 1.4V
11 = channel not converted
[3:2]
MON2 full-scale range
00 = 5.6V
01 = 2.8V
10 = 1.4V
11 = channel not converted
[5:4]
MON3 full-scale range
00 = 5.6V
01 = 2.8V
10 = 1.4V
11 = channel not converted
[7:6]
MON4 full-scale range
00 = 5.6V
01 = 2.8V
10 = 1.4V
11 = channel not converted
[1:0]
MON5 full-scale range
00 = 5.6V
01 = 2.8V
10 = 1.4V
11 = channel not converted
[3:2]
MON6 full-scale range
00 = 5.6V
01 = 2.8V
10 = 1.4V
11 = channel not converted
[7:4]
Not used
243h
244h
Table 9. ADC Conversion Results (Read Only)
REGISTER ADDRESS
BIT RANGE
00h
[7:0]
MON1 result (MSB)
DESCRIPTION
01h
[7:6]
MON1 result (LSB)
02h
[7:0]
MON2 result (MSB)
03h
[7:6]
MON2 result (LSB)
04h
[7:0]
MON3 result (MSB)
05h
[7:6]
MON3 result (LSB)
06h
[7:0]
MON4 result (MSB)
07h
[7:6]
MON4 result (LSB)
08h
[7:0]
MON5 result (MSB)
09h
[7:6]
MON5 result (LSB)
0Ah
[7:0]
MON6 result (MSB)
0Bh
[7:6]
MON6 result (LSB)
16 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
When GPIO1–GPIO6 are configured as general-purpose
inputs/outputs, read values from the GPIO ports through
r1Eh and write values to GPIOs through r3Eh. Note that
r3Eh has a corresponding flash register, which programs the default state of a general purpose output. See
Table 12 for more information on reading and writing to
the GPIO.
Table 10. GPIO_ Configuration Registers
REGISTER ADDRESS
FLASH ADDRESS
3Fh
BIT RANGE
23Fh
40h
240h
DESCRIPTION
[1:0]
GPIO1 configuration
[3:2]
GPIO2 configuration
[5:4]
GPIO3 configuration
[7:6]
GPIO4 configuration
[1:0]
GPIO5 configuration
[3:2]
GPIO6 configuration
[4]
ARAEN bit
[7:5]
Not used
Table 11. GPIO_ Function Configuration Bits
GPIO1
GPIO2
GPIO3
GPIO4
GPIO5
GPIO6
00
Logic input
Logic input
Logic input
Logic input
Logic input
Logic input
01
Logic output
(push-pull)
Logic output
(push-pull)
Logic output
(push-pull)
Logic output
(push-pull)
Logic output
(push-pull)
Logic output
(push-pull)
10
Logic output
(open drain)
Logic output
(open drain)
Logic output
(open drain)
Logic output
(open drain)
Logic output
(open drain)
Logic output
(open drain)
11
ALERT (open drain)
FAULT (open drain)
MR input
WDI
WDO
(open drain)
EXTFAULT
(open drain)
Table 12. GPIO_ State Registers
REGISTER
ADDRESS
1Eh
FLASH
ADDRESS
—
BIT RANGE
[0]
GPIO1 input state
[1]
GPIO2 input state
[2]
GPIO3 input state
[3]
GPIO4 input state
[4]
GPIO5 input state
[5]
GPIO6 input state
[7:6]
3Eh
23Eh
DESCRIPTION
Not used
[0]
GPIO1 output state
[1]
GPIO2 output state
[2]
GPIO3 output state
[3]
GPIO4 output state
[4]
GPIO5 output state
[5]
GPIO6 output state
[7:6]
Not used
______________________________________________________________________________________ 17
MAX16067
General-Purpose Inputs/Outputs
GPIO1–GPIO6 are programmable general-purpose
inputs/outputs. GPIO1–GPIO6 are configurable as a
manual reset input, a watchdog timer input and output,
logic inputs/outputs, and fault-dependent outputs. When
programmed as outputs, GPIOs are open-drain or pushpull. See Tables 10 and 11 for more detailed information
on configuring GPIO1–GPIO6.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
ALERT
GPIO1 is configurable as the SMBus alert signal, ALERT.
ALERT asserts when any fault condition occurs. When the
SMBus host sends the ARA (Alert Response Address),
the MAX16067 responds with its slave address and
deasserts ALERT. ALERT is an open-drain output.
Set the ARAEN bit in r40h[4] to ‘1’ to disable the ARA
feature. Under these conditions, the device does not
respond to an ARA on the SMBus line.
FAULT
GPIO2 is configurable as a dedicated fault output,
FAULT. FAULT asserts when an overvoltage or undervoltage condition occurs on the selected inputs. FAULT
dependencies are set using registers r36h and r37h
(see Table 13). When FAULT depends on more than one
MON_, the fault output asserts when one or more MON_
exceeds a programmed threshold voltage. FAULT acts
independently of the critical fault system, described in
the Critical Faults section. Use r37h[7] to set the polarity
of FAULT.
Manual Reset (MR)
GPIO3 is configurable to act as an active-low manual reset
input, MR. Drive MR low to assert RESET. RESET remains
asserted for the selected reset timeout period after MR
transitions from low to high. When connecting MR to a pushbutton, use a pullup resistor. See the Reset Output section
for more information on selecting a reset timeout period.
Watchdog Input (WDI) and Output (WDO)
GPIO4 and GPIO5 are configurable as the watchdog
timer input (WDI) and output, WDO, respectively. See
Table 23 for configuration details. WDO is an open-drain,
active-low output. See the Watchdog Timer section for
more information about the operation of the watchdog
timer.
External Fault (EXTFAULT)
GPIO6 is configurable as the external fault input/output,
EXTFAULT. EXTFAULT asserts if any monitored voltage exceeds an overvoltage or undervoltage threshold.
EXTFAULT also asserts if a power-up or power-down
sequencing fault occurs. This signal can be used to
cascade multiple MAX16067s.
Pull EXTFAULT low externally to force the sequencer to
enter a fault state. Under these conditions, all outputs
deassert.
Two configuration bits determine the behavior of the
MAX16067 when EXTFAULT is pulled low by an external device. Register bit r72h[5], if set to a ‘1’, causes
the sequencer state machine to enter the fault state,
deasserting all the outputs when EXTFAULT is pulled
low. When this happens, the flag bit r1Ch[6] is set to
indicate the cause of the fault. If register bit r6Dh[2] is
set in addition to r72h[5], EXTFAULT going low triggers
a nonvolatile fault log operation.
Table 13. FAULT Dependencies
REGISTER
ADDRESS
36h
FLASH
ADDRESS
236h
BIT RANGE
[0]
FAULT depends on MON1 undervoltage threshold
[1]
FAULT depends on MON2 undervoltage threshold
[2]
FAULT depends on MON3 undervoltage threshold
[3]
FAULT depends on MON4 undervoltage threshold
[4]
FAULT depends on MON5 undervoltage threshold
[5]
FAULT depends on MON6 undervoltage threshold
[7:6]
37h
237h
DESCRIPTION
Not used
[0]
FAULT depends on MON1 overvoltage threshold
[1]
FAULT depends on MON2 overvoltage threshold
[2]
FAULT depends on MON3 overvoltage threshold
[3]
FAULT depends on MON4 overvoltage threshold
[4]
FAULT depends on MON5 overvoltage threshold
[5]
FAULT depends on MON6 overvoltage threshold
[6]
Not used
[7]
0 = FAULT is an active-low digital output
1 = FAULT is an active-high digital output
18 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
The MAX16067 is capable of measuring overvoltage
and undervoltage fault events. Fault conditions are
detected at the end of each ADC conversion. An overvoltage event occurs when the voltage at a monitored
input exceeds the overvoltage threshold for that input.
An undervoltage event occurs when the voltage at a
monitored input falls below the undervoltage threshold.
Fault thresholds are set in registers r49h–r59h as shown
in Table 14. Disabled inputs are not monitored for fault
conditions and are skipped over by the input multiplexer.
Only the upper 8 bits of a conversion result are compared with the programmed fault thresholds.
Table 14. Fault Threshold Registers
REGISTER
ADDRESS
FLASH
ADDRESS
BIT RANGE
48h
248h
[7:0]
Not used
49h
249h
[7:0]
MON1 overvoltage threshold
4Ah
24Ah
[7:0]
MON1 undervoltage threshold
DESCRIPTION
4Bh
24Bh
[7:0]
Not used
4Ch
24Ch
[7:0]
MON2 overvoltage threshold
4Dh
24Dh
[7:0]
MON2 undervoltage threshold
4Eh
24Eh
[7:0]
Not used
4Fh
24Fh
[7:0]
MON3 overvoltage threshold
50h
250h
[7:0]
MON3 undervoltage threshold
51h
251h
[7:0]
Not used
52h
252h
[7:0]
MON4 overvoltage threshold
53h
253h
[7:0]
MON4 undervoltage threshold
54h
254h
[7:0]
Not used
55h
255h
[7:0]
MON5 overvoltage threshold
56h
256h
[7:0]
MON5 undervoltage threshold
57h
257h
[7:0]
Not used
58h
258h
[7:0]
MON6 overvoltage threshold
59h
259h
[7:0]
MON6 undervoltage threshold
______________________________________________________________________________________ 19
MAX16067
Faults
The MAX16067 monitors the input (MON_) channels and
compares the results with an overvoltage threshold and
an undervoltage threshold. Based on these conditions,
the MAX16067 asserts various fault outputs and save
specific information about the channel conditions and
voltages into the nonvolatile flash. Once a critical fault
event occurs, the failing channel condition, ADC conversions at the time of the fault, or both can be saved by
configuring the event logger. The event logger records
a single failure in the internal flash and sets a lock bit
which protects the stored fault data from accidental erasure on a subsequent power-up.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Deglitch
Fault conditions are detected at the end of each conversion. When the voltage on an input falls outside a monitored threshold for one acquisition, the input multiplexer
remains on that channel and performs several successive conversions. To trigger a fault, the input must stay
outside the threshold for a certain number of acquisitions
as determined by the deglitch setting in r74h[6:5] (see
Table 15).
in Table 16. Clear a fault flag by writing a ‘1’ to the appropriate bit in the flag register. Unlike the fault signals sent
to the fault outputs, these bits are masked by the critical
fault enable bits (see Table 17). The fault flag is only set
when the matching enable bit in the critical fault enable
register is also set.
Fault Flags
Fault flags indicate the fault status of a particular input.
The fault flag of any monitored input in the device can be
read at any time from registers r1Bh and r1Ch, as shown
The SMB Alert (ALERT) bit is set if the MAX16067 has
asserted the SMBus Alert output. Clear by writing a ‘1’.
See the SMBALERT (ALERT) section for more details.
If GPIO6 is configured as the EXTFAULT input/output
and EXTFAULT is pulled low by an external circuit, bit
r1Ch[6] is set.
Table 15. Deglitch Configuration
REGISTER
ADDRESS
74h
FLASH
ADDRESS
274h
BIT RANGE
[6:5]
DESCRIPTION
Voltage comparator deglitch configuration
00 = 2 cycles
01 = 4 cycles
10 = 8 cycles
11 = 16 cycles
Table 16. Fault Flags
REGISTER
ADDRESS
1Bh
BIT RANGE
[0]
MON1 undervoltage threshold
[1]
MON2 undervoltage threshold
[2]
MON3 undervoltage threshold
[3]
MON4 undervoltage threshold
[4]
MON5 undervoltage threshold
[5]
MON6 undervoltage threshold
[7:6]
1Ch
DESCRIPTION
Reserved
[0]
MON1 overvoltage threshold
[1]
MON2 overvoltage threshold
[2]
MON3 overvoltage threshold
[3]
MON4 overvoltage threshold
[4]
MON5 overvoltage threshold
[5]
MON6 overvoltage threshold
[6]
External fault (EXTFAULT)
SMB alert
[7]
20 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Logged fault information is stored in flash registers
r200h–r208h (see Table 18). After fault information is
logged, the flash is locked and must be unlocked to
enable a new fault log to be stored. Write a ‘0’ to r8Ch[1]
to unlock the configuration flash. Fault information can
be configured to store ADC conversion results and/or
fault flags in registers. Select the critical fault configuration in r6Dh[1:0]. Set r6Dh[1:0] to ‘11’ to turn off the fault
logger. All stored ADC results are 8 bits wide (MSBs of
the conversion).
Power-Up/Power-Down Faults
All EN_OUT_s deassert when an overvoltage or undervoltage fault is detected during power-up/power-down
and the MAX16067 enters to the fault state. Fault information can be stored to flash depending on r6D[1:0]
(see Table 17).
Table 17. Critical Fault Configuration
REGISTER
ADDRESS
FLASH
ADDRESS
BIT
RANGE
[1:0]
6Dh
26Dh
[2]
[7:3]
6Eh
6Fh
70h
71h
72h
26Eh
26Fh
270h
271h
272h
DESCRIPTION
Fault Information to Log
00 = Save failed line flags and ADC values in flash
01 = Save only failed line flags in flash
10 = Save only ADC values in flash
11 = Do not save anything
1 = Fault log triggered when EXTFAULT is pulled low externally
Not used
[0]
1 = Fault log triggered when MON1 is below its undervoltage threshold
[1]
1 = Fault log triggered when MON2 is below its undervoltage threshold
[2]
1 = Fault log triggered when MON3 is below its undervoltage threshold
[3]
1 = Fault log triggered when MON4 is below its undervoltage threshold
[4]
1 = Fault log triggered when MON5 is below its undervoltage threshold
[5]
1 = Fault log triggered when MON6 is below its undervoltage threshold
[7:6]
Not used
[3:0]
Not used
[4]
1 = Fault log triggered when MON1 is above its overvoltage threshold
[5]
1 = Fault log triggered when MON2 is above its overvoltage threshold
[6]
1 = Fault log triggered when MON3 is above its overvoltage threshold
[7]
1 = Fault log triggered when MON4 is above its overvoltage threshold
[0]
1 = Fault log triggered when MON5 is above its overvoltage threshold
[1]
1 = Fault log triggered when MON6 is above its overvoltage threshold
[7:2]
Not used
[7:0]
Not used
[4:0]
Not used
[5]
[7:6]
1 = EXTFAULT pulled low externally causes sequencer to enter fault state, turning
off all EN_OUT_s
0 = EXTFAULT pulled low externally does not cause sequencer to enter fault state
Not used
______________________________________________________________________________________ 21
MAX16067
Critical Faults
During normal operation, a fault condition can be configured to shut down all the EN_OUT_s and store fault
information in the flash memory by setting the appropriate critical fault enable bits. During power-up and
power-down, all sequenced MON inputs are considered
critical. Faults during power-up and power-down always
cause the EN_OUT_s to turn off and can store fault information in the flash memory, depending on the contents
of r6Dh[1:0]. Set the appropriate critical fault enable bits
in registers r6Eh–r72h (see Table 17) for a fault condition
to trigger a critical fault.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Table 18. Nonvolatile Fault Log Registers
FLASH
ADDRESS
200h
201h
BIT RANGE
[3:0]
Sequencer state where the fault has happened (see Table 1 for state codes).
Fault has happened during power-up if bit [3] = 0 and during power-down if [3] = 1.
Bits [2:0] indicate the slot number.
[7:4]
Not used
[0]
Fault log triggered on MON1 falling below its undervoltage threshold
[1]
Fault log triggered on MON2 falling below its undervoltage threshold
[2]
Fault log triggered on MON3 falling below its undervoltage threshold
[3]
Fault log triggered on MON4 falling below its undervoltage threshold
[4]
Fault log triggered on MON5 falling below its undervoltage threshold
[5]
Fault log triggered on MON6 falling below its undervoltage threshold
[7:6]
202h
DESCRIPTION
Not used
[0]
Fault log triggered on MON1 exceeding its overvoltage threshold
[1]
Fault log triggered on MON2 exceeding its overvoltage threshold
[2]
Fault log triggered on MON3 exceeding its overvoltage threshold
[3]
Fault log triggered on MON4 exceeding its overvoltage threshold
[4]
Fault log triggered on MON5 exceeding its overvoltage threshold
[5]
Fault log triggered on MON6 exceeding its overvoltage threshold
[6]
Fault log triggered on EXTFAULT
Not used
[7]
203h
[7:0]
MON1 ADC output (8 MSBs)
204h
[7:0]
MON2 ADC output (8 MSBs)
205h
[7:0]
MON3 ADC output (8 MSBs)
206h
[7:0]
MON4 ADC output (8 MSBs)
207h
[7:0]
MON5 ADC output (8 MSBs)
208h
[7:0]
MON6 ADC output (8 MSBs)
Autoretry/Latch Mode
The MAX16067 can be configured for one of two fault
management methods: autoretry or latch-on-fault. Set
r74h[4:3] to ‘00’ to select the latch-on-fault mode. In
this configuration, EN_OUT_s deassert after a critical
fault event. The device does not reinitiate the power-up
sequence until EN is toggled or the software enable bit
is toggled. See the Enable Input (EN) section for more
information on setting the software enable bit.
Set r74h[4:3] to a value other than ‘00’ to select autoretry
mode (see Table 19). In this configuration, the device
shuts down after a critical fault event then restarts following a configurable delay. Use r74h[2:0] to select an
autoretry delay from 20ms to 1.6s. See Table 19 for more
information on setting the autoretry delay.
When fault information is stored in flash (see the Critical
Faults section) and autoretry mode is selected, set an
autoretry delay greater than the time required for the
storing operation. When fault information is stored in
flash and latch-on-fault mode is chosen, toggle EN or
reset the software enable bit only after the completion of
the storing operation. When saving information about the
failed lines only, ensure a delay of at least 12ms before
the restart procedure. Otherwise, ensure a minimum
153ms timeout, to ensure that ADC conversions are
completed and values are stored correctly in flash.
22 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
REGISTER
ADDRESS
74h
FLASH
ADDRESS
BIT
RANGE
DESCRIPTION
[2:0]
Retry Delay
000 = 20ms
001 = 40ms
010 = 80ms
011 = 150ms
100 = 280ms
101 = 540ms
110 = 1s
111 = 2s
[4:3]
Autoretry/Latch Mode
00 = Latch
01 = Reserved
10 = Reserved
11 = Always retry
274h
Programmable Outputs
(EN_OUT1–EN_OUT6)
The MAX16067 includes six programmable outputs.
These outputs are capable of connecting to either the
enable (EN) inputs of a DC-DC or LDO power supply, or to drive the gate of an n-channel MOSFET in
charge-pump mode. Selectable output configurations
include: active-low or active-high, open-drain or pushpull. EN_OUT1–EN_OUT3 can act as charge-pump
outputs, EN_OUT1–EN_OUT6 can be configured as
general-purpose inputs or general-purpose outputs. Use
the registers r30h–r33h to configure outputs. See Table
20 for detailed information on configuring EN_OUT1–
EN_OUT6.
In charge-pump configuration: EN_OUT1, EN_OUT2,
and EN_OUT3 act as high-voltage charge-pump outputs to drive up to three external n-channel MOSFETs.
During sequencing, an EN_OUT_ output is configured
as a charge-pump output 11V above GND. See the
Sequencing section for more detailed information on
power-supply sequencing.
In open-drain output configuration: Connect an external pullup resistor from the output to an external voltage up to 5.5V (EN_OUT4, EN_OUT5, EN_OUT6) or
14V (EN_OUT1, EN_OUT2, EN_OUT3) when configured
as an open-drain output. Choose the pullup resistor
depending on the number of devices connected to the
open-drain output and the allowable current consumption. The open-drain output configuration allows wiredOR connection.
In push-pull configuration: The MAX16067’s programmable outputs are referenced to VDBP.
______________________________________________________________________________________ 23
MAX16067
Table 19. Autoretry Configuration
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Table 20. EN_OUT1–EN_OUT6 Configuration
REGISTER
ADDRESS
30h
31h
33h
FLASH
ADDRESS
BIT RANGE
[1:0]
EN_OUT1 Configuration
00 = Active-low, open drain
01 = Active-high, open drain
10 = Active-low, push-pull
11 = Active-high, push-pull
[3:2]
EN_OUT2 Configuration
00 = Active-low, open drain
01 = Active-high, open drain
10 = Active-low, push-pull
11 = Active-high, push-pull
[5:4]
EN_OUT3 Configuration
00 = Active-low, open drain
01 = Active-high, open drain
10 = Active-low, push-pull
11 = Active-high, push-pull
[7:6]
EN_OUT4 Configuration
00 = Active-low, open drain
01 = Active-high, open drain
10 = Active-low, push-pull
11 = Active-high, push-pull
[1:0]
EN_OUT5 Configuration
00 = Active-low, open drain
01 = Active-high, open drain
10 = Active-low, push-pull
11 = Active-high, push-pull
[3:2]
EN_OUT6 Configuration
00 = Active-low, open drain
01 = Active-high, open drain
10 = Active-low, push-pull
11 = Active-high, push-pull
[7:4]
Not used
230h
231h
233h
DESCRIPTION
[0]
EN_OUT1 Charge-Pump Output Configuration
0 = Charge-pump output disabled
1 = Charge-pump output enabled (active-high)
[1]
EN_OUT2 Charge-Pump Output Configuration
0 = Charge-pump output disabled
1 = Charge-pump output enabled (active-high)
[2]
EN_OUT3 Charge-Pump Output Configuration
0 = Charge-pump output disabled
1 = Charge-pump output enabled (active-high)
[7:3]
Not used
24 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
EN_OUT_ State During Power-Up
When VCC is ramped from 0V to the operating supply voltage, the EN_OUT_ output is high impedance
until VCC reaches UVLO and then EN_OUT_ goes into
the configured deasserted state. See Figures 4 and 5.
Configure RESET as an active-low push-pull or opendrain output pulled up to VCC through a 10kI resistor
for Figures 4 and 5.
Table 21. EN_OUT_ GPIO State Registers
REGISTER
ADDRESS
1Fh
FLASH
ADDRESS
—
BIT RANGE
[0]
EN_OUT1 input state
[1]
EN_OUT2 input state
[2]
EN_OUT3 input state
[3]
EN_OUT4 input state
[4]
EN_OUT5 input state
[5]
EN_OUT6 input state
[7:6]
34h
234h
DESCRIPTION
Not used
[0]
1 = Assert EN_OUT1
[1]
1 = Assert EN_OUT2
[2]
1 = Assert EN_OUT3
[3]
1 = Assert EN_OUT4
[4]
1 = Assert EN_OUT5
[5]
1 = Assert EN_OUT6
[7:6]
Not used
MAX16067 fig04
MAX16067 fig05
VCC
2V/div
0V
VCC
2V/div
UVLO
0V
RESET
2V/div
0V
EN_OUT_
2V/div
0V
20ms/div
Figure 4. RESET and EN_OUT_ During Power-Up, EN_OUT_ is
in Open-Drain Active-Low Configuration
RESET
2V/div
0V
ASSERTED
LOW
HIGH-Z
EN_OUT_
2V/div
0V
10ms/div
Figure 5. RESET and EN_OUT_ During Power-Up, EN_OUT_ is
in Push-Pull Active-High Configuration
______________________________________________________________________________________ 25
MAX16067
EN_OUT_s as GPIO
EN_OUT1–EN_OUT6 can be configured as general-purpose inputs by setting the sequencing slot assignments
in r84h–r86h to ‘1101’ or as general-purpose outputs by
setting the slot assignments to ‘1110’. See Tables 5 and
6. If an EN_OUT_ is configured as a general-purpose
input, the state of the GPIO can be read from r1Fh (see
Table 21). If an EN_OUT_ is configured as a generalpurpose output, it is controlled by r34h.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Reset Output
The reset output, RESET, indicates the status of the
sequencer and the monitored inputs. It asserts during power-up/power-down and deasserts following the reset timeout period once the power-up sequence is complete. The
power-up sequence is complete when any MON_ inputs
assigned to slot 6 exceed the undervoltage thresholds and
the slot 6 sequence delay expires. When no MON_ inputs
are assigned to slot 6, the power-up sequence is complete
after the slot sequence delay expires.
During normal monitoring, RESET can be configured to
assert when any combination of MON_ inputs violates
configurable combinations of undervoltage or overvoltage
thresholds. Select the combination of MON_ inputs using
r3Ch[5:0] and r3Dh[5:0]. Note that MON_ inputs configured as critical faults always cause RESET to assert
regardless of these configuration bits.
RESET can be configured as push-pull or open drain
using r3Bh[3], and active high or active low using
r3Bh[2]. Select the reset timeout by loading a value from
Table 22 into r3Bh[7:4].
To generate a one-shot pulse on RESET, write a ‘1’ into
r3Bh[0]. The pulse width is the configured reset timeout.
Register bit r3Bh[0] clears automatically (see Table 22).
The current state of RESET can be checked by reading
r20h[0].
Table 22. Reset Output Configuration
REGISTER
ADDRESS
3Bh
FLASH
ADDRESS
BIT RANGE
DESCRIPTION
[0]
RESET Soft Trigger
0 = Normal RESET behavior
1 = Force RESET to assert
[1]
Not used
[2]
0 = Active low
1 = Active high
[3]
0 = Open drain
1 = Push-pull
23Bh
[7:4]
Reset Timeout Period
0000 = 25Fs
0001 = 1.5ms
0010 = 2.5ms
0011 = 4ms
0100 = 6ms
0101 = 10ms
0110 = 15ms
0111 = 25ms
1000 = 40ms
1001 = 60ms
1010 = 100ms
1011 = 150ms
1100 = 250ms
1101 = 400ms
1110 = 600ms
1111 = 1s
26 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
REGISTER
ADDRESS
3Ch
FLASH
ADDRESS
23Ch
BIT RANGE
[0]
1 = RESET depends on MON1 undervoltage
[1]
1 = RESET depends on MON2 undervoltage
[2]
1 = RESET depends on MON3 undervoltage
[3]
1 = RESET depends on MON4 undervoltage
[4]
1 = RESET depends on MON5 undervoltage
[5]
1 = RESET depends on MON6 undervoltage
[7:6]
3Dh
23Dh
DESCRIPTION
Not used
[0]
1 = RESET depends on MON1 overvoltage
[1]
1 = RESET depends on MON2 overvoltage
[2]
1 = RESET depends on MON3 overvoltage
[3]
1 = RESET depends on MON4 overvoltage
[4]
1 = RESET depends on MON5 overvoltage
[5]
1 = RESET depends on MON6 overvoltage
[7:6]
Not used
Watchdog Timer
The watchdog timer operates together with or independently of the MAX16067. When operating in dependent
mode, the watchdog is not activated until the sequencing is complete and RESET is deasserted. When operating in independent mode, the watchdog timer is
independent of the sequencing operation and activates
immediately after VCC exceeds the UVLO threshold
and the boot phase is complete. Set r73h[4] to ‘0’ to
configure the watchdog in dependent mode. Set r73h[4]
to ‘1’ to configure the watchdog in independent mode.
See Table 23 for more information on configuring the
watchdog timer in dependent or independent mode. The
watchdog timer can be reset by toggling the WDI input
(GPIO4) or by writing a ‘1’ to r75h[5].
Dependent Watchdog Timer Operation
Use the watchdog timer to monitor FP activity in two
modes. Flexible timeout architecture provides an adjustable watchdog startup delay of up to 300s, allowing
complicated systems to complete lengthy boot-up routines. An adjustable watchdog timeout allows the supervisor to provide quick alerts when the processor activity
fails. After each reset event (VCC drops below UVLO
then returns above UVLO, software reboot, manual reset
(MR), EN input going low then high, or watchdog reset)
and once sequencing is complete, the watchdog startup
delay provides an extended time for the system to power
up and fully initialize all FP and system components
before assuming responsibility for routine watchdog
updates. Set r76h[6:4] to a value other than ‘000’ to
enable the watchdog startup delay. Set r76h[6:4] to ‘000’
to disable the watchdog startup delay.
The normal watchdog timeout period, tWDI, begins after
the first transition on WDI before the conclusion of the
long startup watchdog period, tWDI_STARTUP (Figures 6
and 7). During the normal operating mode, WDO asserts
if the FP does not toggle WDI with a valid transition (highto-low or low-to-high) within the standard timeout period,
tWDI. WDO remains asserted until WDI is toggled or
RESET is asserted (Figure 7).
While EN is low, the watchdog timer is in reset. The
watchdog timer does not begin counting until the poweron mode is reached and RESET is deasserted. The
watchdog timer is reset and WDO deasserts any time
RESET is asserted (Figure 8). The watchdog timer is
held in reset while RESET is asserted.
The watchdog can be configured to control the RESET
output as well as the WDO output. RESET asserts for
the reset timeout, tRP, when the watchdog timer expires
and the Watchdog Reset Output Enable bit (r76h[7]) is
set to ‘1’. When RESET is asserted, the watchdog timer
is cleared and WDO is deasserted, therefore, WDO
pulses low for a short time (approximately 1Fs) when
______________________________________________________________________________________ 27
MAX16067
Table 22. Reset Output Configuration (continued)
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
the watchdog timer expires. RESET is not affected by
the watchdog timer when the Watchdog Reset Output
Enable bit (r76h[7]) is set to ‘0’. If a RESET is asserted
by the watchdog timeout, the WDRESET bit is set to ‘1’. A
connected processor can check this bit to see the reset
was due to a watchdog timeout.
exceeding UVLO and once the boot-up sequence is finished. When RESET is asserted by the sequencer state
machine, the watchdog timer and WDO are not affected.
See Table 23 for more information on configuring watchdog functionality.
In independent mode, if the Watchdog Reset Output
Enable bit r76h[7] is set to ‘1,’ when the watchdog timer
expires, WDO asserts then RESET asserts. WDO is then
deasserts. WDO is low for approximately 1Fs. If the
Watchdog Reset Output Enable bit (r76h[7]) is set to
‘0,’ when the watchdog timer expires, WDO asserts but
RESET is not affected.
There is a startup delay if r76h[6:4] is set to a value different than ‘000’. If r76h[6:4] is set to ‘000’, there is not a
startup delay. See Table 23 for delay times.
Independent Watchdog Timer Operation
When r73h[3] is ‘1,’ the watchdog timer operates in
the independent mode. In the independent mode, the
watchdog timer operates as if it were a separate device.
The watchdog timer is activated immediately upon VCC
VTH
LAST MON_
< tWDI
tWDI_STARTUP
WDI
< tWDI
tRP
RESET
Figure 6. Normal Watchdog Startup Sequence
VCC
WDI
< tWDI
< tWDI
> tWDI
< tWDI
< tWDI
< tWDI
< tWDI
0V
tWDI
VCC
WDO
0V
Figure 7. Watchdog Timer Operation
28 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
< tWDI
tWDI
tRP
< tWDI_STARTUP
< tWDI
0V
VCC
RESET
0V
VCC
WDO
0V
1µs
Figure 8. Watchdog Startup Sequence with Watchdog Reset Enable Bit Set to ‘1’
Table 23. Watchdog Configuration
REGISTER
ADDRESS
FLASH
ADDRESS
BIT
RANGE
73h
273h
[4]
1 = Independent mode
0 = Dependent mode
[7]
1 = Watchdog reset output enabled
0 = Watchdog reset output disabled
76h
DESCRIPTION
[6:4]
Watchdog Startup Delay
000 = No initial timeout
001 = 30s
010 = 40s
011 = 80s
100 = 120s
101 = 160s
110 = 220s
111 = 300s
[3:0]
Watchdog Timeout
0000 = Watchdog disabled
0001 = 1ms
0010 = 2ms
0011 = 4ms
0100 = 8ms
0101 = 14ms
0110 = 27ms
0111 = 50ms
1000 = 100ms
1001 = 200ms
1010 = 400ms
1011 = 750ms
1100 = 1.4s
1101 = 2.7s
1110 = 5s
1111 = 10s
276h
______________________________________________________________________________________ 29
MAX16067
VCC
WDI
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
User-Defined Register
Register r8Ah provides storage space for a user-defined
configuration or firmware version number. Note that this
register controls the contents of the JTAG USERCODE
register bits 7-0. The user-defined register is stored at
r28Ah in the flash memory.
Memory Lock Bits
Register r8Ch contains the lock bits for the configuration
registers, configuration flash, user flash, and fault register lock. See Table 24 for details.
SMBus-Compatible Interface
The MAX16067 features an SMBus-compatible, 2-wire
serial interface consisting of a serial-data line (SDA)
and a serial-clock line (SCL). SDA and SCL facilitate
bidirectional communication between the MAX16067
and the master device at clock rates up to 400kHz.
Figure 1 shows the 2-wire interface timing diagram. The
MAX16067 is a transmit/receive, slave-only device, relying upon a master device to generate a clock signal.
The master device (typically a microcontroller) initiates
a data transfer on the bus and generates SCL to permit
that transfer.
A master device communicates to the MAX16067 by
transmitting the proper address followed by command
and/or data words. The slave address input, A0, is
capable of detecting four different states, allowing multiple identical devices to share the same serial bus. The
slave address is described further in the Slave Address
section. Each transmit sequence is framed by a START
(S) or REPEATED START (Sr) condition and a STOP (P)
condition. Each word transmitted over the bus is 8 bits
long and is always followed by an acknowledge pulse.
SCL is a logic input, while SDA is an open-drain input/
output. SCL and SDA both require external pullup resistors to generate the logic-high voltage. Use 4.7kI for
most applications.
Bit Transfer
Each clock pulse transfers one data bit. The data on
SDA must remain stable while SCL is high (Figure 9);
otherwise, the MAX16067 registers a START or STOP
condition (Figure 10) from the master. SDA and SCL idle
high when the bus is not busy.
START and STOP Conditions
Both SCL and SDA idle high when the bus is not busy.
A master device signals the beginning of a transmission
with a START condition by transitioning SDA from high to
low while SCL is high. The master device issues a STOP
condition by transitioning SDA from low to high while
SCL is high. A STOP condition frees the bus for another
transmission. The bus remains active if a REPEATED
START condition is generated, such as in the block read
protocol (see Figure 1, SMBus Timing Diagram).
Early STOP Conditions
The MAX16067 recognizes a STOP condition at any point
during transmission except if a STOP condition occurs in
the same high pulse as a START condition. This condition
is not a legal SMBus format; at least one clock pulse must
separate any START and STOP condition.
Table 24. Memory Lock Bits
REGISTER
ADDRESS
8Ch
FLASH
ADDRESS
BIT RANGE
DESCRIPTION
[0]
Configuration Register Lock
1 = Locked
0 = Unlocked
[1]
Flash Fault Register Lock
1 = Locked
0 = Unlocked
[2]
Flash Configuration Lock
1 = Locked
0 = Unlocked
[3]
User Flash Lock
1 = Locked
0 = Unlocked
28Ch
[7.4]
Not used
30 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
SDA
SCL
SCL
DATA LINE STABLE, CHANGE OF
DATA ALLOWED
DATA VALID
Figure 9. Bit Transfer
S
P
START
CONDITION
STOP
CONDITION
Figure 10. START and STOP Conditions
CLOCK PULSE FOR ACKNOWLEDGE
1
SCL
2
8
9
SDA BY
TRANSMITTER
S
NACK
SDA BY
RECEIVER
ACK
Figure 11. Acknowledge
REPEATED START Conditions
A REPEATED START can be sent instead of a STOP
condition to maintain control of the bus during a read
operation. The START and REPEATED START conditions
are functionally identical.
Acknowledge
The acknowledge bit (ACK) is the 9th bit attached to
any 8-bit data word. The receiving device always generates an ACK. The MAX16067 generates an ACK when
receiving an address or data by pulling SDA low during
the 9th clock period (Figure 11). When transmitting data,
such as when the master device reads data back from
the MAX16067, the device waits for the master device to
generate an ACK. Monitoring ACK allows for detection
of unsuccessful data transfers. An unsuccessful data
transfer occurs if the receiving device is busy or if a system fault has occurred. In the event of an unsuccessful
data transfer, the bus master can reattempt communication at a later time. The MAX16067 generates a NACK
after the command byte received during a software
reboot, while writing to the flash, or when receiving an
illegal memory address.
______________________________________________________________________________________ 31
MAX16067
SDA
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
calculation does not include ACK, NACK, START, STOP,
or REPEATED START.
Slave Address
Use the slave address input, A0, to allow multiple identical devices to share the same serial bus. Connect A0 to
GND, DBP (or an external supply voltage greater than
2V), SCL, or SDA to set the device address on the bus.
See Table 26 for a listing of all possible 7-bit addresses.
Command Codes
The MAX16067 uses eight command codes for block
read, block write, and other commands. See Table 27
for a list of command codes.
The slave address can also be set to a custom value by
loading the address into register r8Bh[6:0]. See Table
25. If r8Bh[6:0] is loaded with 00h, the address is set by
input A0. Do not set the address to 09h or 7Fh to avoid
address conflicts. The slave address setting takes effect
immediately after writing to the register.
To initiate a software reboot, send A7h using the send
byte format. A software-initiated reboot is functionally the same as a hardware-initiated power-on reset.
During boot-up, flash configuration data in the range
of 230h–28Ch is copied to r30h–r8Ch registers in the
default page.
Packet Error Checking (PEC)
The MAX16067 features a packet-error checking (PEC)
mode that is useful to improve the reliability of the communication bus by detecting bit errors. By enabling PEC,
an extra CRC-8 error check byte is added in the data
string during each read and/or write sequence. Enable
PEC by writing a ‘1’ to r8Bh[7].
Send command code A8h to trigger a fault store to flash.
Configure the Critical Fault Log Control register (6Dh) to
store ADC conversion results and/or fault flags.
While in the flash page, send command code A9h to
access the flash page (addresses from 200h–2FFh).
Once command code A9h has been sent, all addresses
are recognized as flash addresses only. Send command
code AAh to return to the default page (addresses from
000h–0FFh). Send command code ABh to access the
user flash-page (addresses from 300h–3FFh), and send
command code ACh to return to the flash page.
The CRC-8 byte is calculated using the polynomial
C = X8 + X2 + X + 1
The PEC calculation includes all bytes in the transmission, including address, command and data. The PEC
Table 25. SMBus Settings Register
REGISTER
ADDRESS
FLASH
ADDRESS
8Bh
28Bh
BIT RANGE
DESCRIPTION
[6:0]
SMBus Slave Address Register. Set to 00h to use A0 pin address
setting.
[7]
1= Enable PEC (Packet Error Check).
Table 26. Setting the SMBus Slave Address
SLAVE ADDRESSES
A0
SLAVE ADDRESS
0
1010 100R
1
1010 101R
SCL
1010 110R
SDA
1010 111R
R = Read/write select bit.
32 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
COMMAND CODE
MAX16067
Table 27. Command Codes
ACTION
A5h
Block write
A6h
Block read
A7h
Reboot flash in register file
A8h
Trigger emergency save to flash
A9h
Flash page access ON
AAh
Flash page access OFF
ABh
User flash access ON (must be in flash page already)
ACh
User flash access OFF (return to flash page)
Restrictions When Writing to Flash
Flash must be written to 8 bytes at a time. The initial
address must be aligned to 8-byte boundaries—the 3
LSBs of the initial address must be ‘000’. Write the 8
bytes using a single block write command or using eight
successive Write Byte commands. A write operation
requires 122ms for each 8-byte block. After programming a block, check r20h[1] (see Table 31) to make sure
the write operation is complete before attempting to write
the next block.
Send Byte
The send byte protocol allows the master device to
send one byte of data to the slave device (see Figure
12). The send byte presets a register pointer address
for a subsequent read or write. The slave sends a NACK
instead of an ACK if the master tries to send a memory
address or command code that is not allowed. If the
master sends A5h or A6h, the data is ACK, because this
could be the start of the write block or read block. If the
master sends a STOP condition before the slave asserts
an ACK, the internal address pointer does not change.
If the master sends A7h, this signifies a software reboot.
The send byte procedure is as follows:
Receive Byte
The receive byte protocol allows the master device to
read the register content of the MAX16067 (see Figure
12). The flash or register address must be preset with
a send byte or write word protocol first. Once the read
is complete, the internal pointer increases by one.
Repeating the receive byte protocol reads the contents
of the next address. The receive byte procedure follows:
1) The master sends a START condition.
2) The master sends the 7-bit slave address and a read
bit (high).
3) The addressed slave asserts an ACK on SDA.
4) The slave sends 8 data bits.
5) The master asserts a NACK on SDA.
6) The master generates a STOP condition.
Write Byte
The write byte protocol (see Figure 12) allows the master device to write a single byte in the default page,
extended page, or flash page, depending on which
page is currently selected. The write byte procedure is
as follows:
1)The master sends a START condition.
1)The master sends a START condition.
2)The master sends the 7-bit slave address and a write
bit (low).
2)The master sends the 7-bit slave address and a write
bit (low).
3)The addressed slave asserts an ACK on SDA.
3)The addressed slave asserts an ACK on SDA.
4)The master sends an 8-bit memory address or command code.
4)The master sends an 8-bit memory address.
5)The addressed slave asserts an ACK (or NACK) on
SDA.
6)The master sends an 8-bit data byte.
6)The master sends a STOP condition.
5)The addressed slave asserts an ACK on SDA.
7)The addressed slave asserts an ACK on SDA.
8)The master sends a STOP condition.
______________________________________________________________________________________ 33
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
To write a single byte, only the 8-bit memory address
and a single 8-bit data byte are sent. The data byte is
written to the addressed location if the memory address
is valid. The slave asserts a NACK at step 5 if the memory address is not valid.
When PEC is enabled, the write byte protocol becomes:
1) The master sends a START condition.
2) The master sends the 7-bit slave ID plus a write bit
(low).
3) The addressed slave asserts an ACK on the data
line.
4) The master sends an 8-bit command code.
5) The active slave asserts an ACK on the data line.
6) The master sends an 8-bit data byte.
7) The slave asserts an ACK on the data line.
8) The master sends an 8-bit PEC byte.
9) The slave asserts an ACK on the data line (if PEC is
good, otherwise NACK).
When PEC is enabled, the read byte protocol becomes:
1) The master sends a START condition.
2) The master sends the 7-bit slave ID plus a write bit
(low).
3) The addressed slave asserts an ACK on the data
line.
4) The master sends 8 data bits.
5) The active slave asserts an ACK on the data line.
6) The master sends a REPEATED START condition.
7) The master sends the 7-bit slave ID plus a read bit
(high).
8) The addressed slave asserts an ACK on the data
line.
9) The slave sends 8 data bits.
10) The master asserts an ACK on the data line.
11) The slave sends an 8-bit PEC byte.
12) The master asserts a NACK on the data line.
10) The master generates a STOP condition.
13) The master generates a STOP condition.
Read Byte
The read byte protocol (see Figure 12) allows the master
device to read a single byte located in the default page,
extended page, or flash page depending on which page is
currently selected. The read byte procedure is as follows:
Block Write
The block write protocol (see Figure 12) allows the
master device to write a block of data (1–16 bytes) to
memory. Preload the destination address by a previous
send byte command; otherwise the block write command begins to write at the current address pointer.
After the last byte is written, the address pointer remains
preset to the next valid address. If the number of bytes
to be written causes the address pointer to exceed 8Fh
for configuration registers or configuration flash or FFh
for user flash, the address pointer stays at 8Fh or FFh,
respectively, overwriting this memory address with the
remaining bytes of data. The slave generates a NACK at
step 5 if the command code is invalid or if the device is
busy, and the address pointer is not altered.
1) The master sends a START condition.
2) The master sends the 7-bit slave address and a write
bit (low).
3) The addressed slave asserts an ACK on SDA.
4) The master sends an 8-bit memory address.
5) The addressed slave asserts an ACK on SDA.
6) The master sends a REPEATED START condition.
7) The master sends the 7-bit slave address and a read
bit (high).
The block write procedure is as follows:
8) The addressed slave asserts an ACK on SDA.
1
The master sends a START condition.
9) The slave sends an 8-bit data byte.
2
The master sends the 7-bit slave address and a
write bit (low).
10)The master asserts a NACK on SDA.
11)The master sends a STOP condition.
If the memory address is not valid, it is NACKed by the
slave at step 5 and the address pointer is not modified.
3) The addressed slave asserts an ACK on SDA.
4) The master sends the 8-bit command code for block
write (94h).
5) The addressed slave asserts an ACK on SDA.
34 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
3) The addressed slave asserts an ACK on SDA.
7) The addressed slave asserts an ACK on SDA.
4) The master sends 8 bits of the block read command
(95h).
8) The master sends 8 bits of data.
5) The slave asserts an ACK on SDA, unless busy.
9) The addressed slave asserts an ACK on SDA.
6) The master generates a REPEATED START condition.
10) Repeat steps 8 and 9 n - 1 times.
11) The master sends a STOP condition.
When PEC is enabled, the block write protocol becomes:
1) The master sends a START condition.
2) The master sends the 7-bit slave ID plus a write bit
(low).
7) The master sends the 7-bit slave address and a
read bit (high).
8) The slave asserts an ACK on SDA.
9) The slave sends the 8-bit byte count (16).
10) The master asserts an ACK on SDA.
3) The addressed slave asserts an ACK on the data
line.
11) The slave sends 8 bits of data.
4) The master sends 8 bits of the block write command
code.
13) Repeat steps 11 and 12 up to fifteen times.
5) The slave asserts an ACK on the data line.
15) The master sends a STOP condition.
6) The master sends 8 bits byte count (min 1, max 16) n.
When PEC is enabled, the block read protocol becomes:
7) The slave asserts an ACK on the data line.
1) The master sends a START condition.
8) The master sends 8 bits of data.
2) The master sends the 7-bit slave ID plus a write bit
(low).
9) The slave asserts an ACK on the data line.
10)Repeat 8 and 9 n - 1 times.
12) The master asserts an ACK on SDA.
14) The master asserts a NACK on SDA.
11)The master sends an 8-bit PEC byte.
3) The addressed slave asserts an ACK on the data
line.
12)The slave asserts an ACK on the data line (if PEC is
good, otherwise NACK).
4 The master sends 8 bits of the block read command
code.
13)The master generates a STOP condition.
5) The slave asserts an ACK on the data line unless
busy.
Block Read
The block read protocol (see Figure 12) allows the
master device to read a block of up to 16 bytes from
memory. Read fewer than 16 bytes of data by issuing
an early STOP condition from the master, or by generating a NACK with the master. The destination address
should be preloaded by a previous send byte command;
otherwise, the block read command begins to read at
the current address pointer. If the number of bytes to
be read causes the address pointer to exceed 8Fh for
the configuration register or configuration flash or FFh
in user flash, the address pointer stays at 8Fh or FFh,
respectively.
6) The master sends a REPEATED START condition.
7) The master sends the 7-bit slave ID plus a read bit
(high).
8) The slave asserts an ACK on the data line.
9) The slave sends 8-bit byte count (16).
10)The master asserts an ACK on the data line.
11)The slave sends 8 bits of data.
12)The master asserts an ACK on the data line.
13)Repeat 11 and 12 up to 15 times.
14)The slave sends an 8-bit PEC byte.
The block read procedure is the following:
15)The master asserts a NACK on the data line.
1) The master sends a START condition.
16)The master generates a STOP condition.
2) The master sends the 7-bit slave address and a
write bit (low).
______________________________________________________________________________________ 35
MAX16067
6) The master sends the 8-bit byte count (1 byte to 16
bytes), n.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
SEND BYTE FORMAT
S
ADDRESS
RECEIVE BYTE FORMAT
R/W ACK
7 BITS
0
COMMAND
ACK
8 BITS
0
0
P
ADDRESS
S
7 BITS
SLAVE ADDRESS: Address DATA BYTE: Presets the internal
of the slave on the serial
address pointer or represents
interface bus.
a command.
ADDRESS
0
COMMAND
ACK
DATA
ACK
8 BITS
0
8 BITS
0
0
SLAVE ADDRESS: Address COMMAND BYTE:
of the slave on the serial
Sets the internal
interface bus.
address pointer.
READ BYTE FORMAT
S
SLAVE
ADDRESS
R/W ACK
7 BITS
0
NACK
8 BITS
1
P
SMBALERT#
R/W ACK
7 BITS
1
DATA
SLAVE ADDRESS: Address DATA BYTE: Data is read from
of the slave on the serial
the location pointed to by the
interface bus.
internal address pointer.
WRITE BYTE FORMAT
S
R/W ACK
0
COMMAND
ACK
8 BITS
0
0
P
S
SLAVE
ADDRESS
R/W ACK
7 BITS
1
R/W ACK
0001100
D.C.
0
DATA
NACK
8 BITS
1
P
ALERT RESPONSE ADDRESS: SLAVE ADDRESS: Slave places
Only the device that
its own address on the
interrupted the master
serial bus.
responds to this address.
DATA BYTE: Data is written to
the locations set by the
internal address pointer.
SR
ADDRESS
0
DATA BYTE NACK
8 BITS
P
1
DATA BYTE: Data is written to
the locations set by the
internal address pointer.
SLAVE ADDRESS: Address COMMAND BYTE:
of the slave on the serial
Sets the internal
interface bus.
address pointer.
BLOCK WRITE FORMAT
S
ADDRESS
R/W ACK
7 BITS
0
COMMAND
ACK
BYTE
COUNT = N
8 BITS
0
8 BITS
0
ACK DATA BYTE 1 ACK DATA BYTE … ACK DATA BYTE N ACK
8 BITS
0
SLAVE ADDRESS: Address COMMAND BYTE:
of the slave on the
FAh
serial interface bus.
8 BITS
0
0
8 BITS
SLAVE TO MASTER
P
0
MASTER TO SLAVE
DATA BYTE: Data is written to the locations
set by the internal address pointer.
BLOCK READ FORMAT
S
ADDRESS
7 BITS
R/W ACK
0
COMMAND
ACK
8 BITS
0
0
SR
SLAVE ADDRESS: Address COMMAND BYTE:
of the slave on the
FBh
serial interface bus.
ADDRESS
R/W ACK
7 BITS
1
0
BYTE
COUNT = N
ACK DATA BYTE N ACK DATA BYTE … ACK DATA BYTE N NACK
8 BITS
0
8 BITS
0
8 BITS
0
8 BITS
P
1
SLAVE ADDRESS: Address DATA BYTE: Data is read from the locations
of the slave on the
set by the internal address pointer.
serial interface bus.
WRITE BYTE FORMAT WITH PEC
S
ADDRESS
7 BITS
R/W ACK
0
COMMAND
ACK
DATA
ACK
PEC
ACK
8 BITS
0
8 BITS
0
8 BITS
0
0
P
READ BYTE FORMAT WITH PEC
S
ADDRESS
7 BITS
R/W ACK
0
COMMAND
0
8 BITS
ACK SR
0
ADDRESS
R/W ACK
1
7 BITS
0
DATA
ACK
PEC
8 BITS
0
8 BITS
NACK P
1
BLOCK WRITE WITH PEC
S
ADDRESS
7 BITS
R/W ACK
0
COMMAND
8 BITS
0
ACK BYTE COUNT N ACK DATA BYTE 1 ACK
0
8 BITS
0
8 BITS
0
DATA BYTE ACK
8 BITS
0
DATA N
ACK
PEC
ACK
8 BITS
0
8 BITS
0
DATA BYTE
ACK
DATA N
ACK
PEC
NACK
8 BITS
0
8 BITS
0
8 BITS
1
P
BLOCK READ WITH PEC
S
ADDRESS
7 BITS
R/W ACK
0
COMMAND
0
S = START CONDITION
P = STOP CONDITION
Sr = REPEATED START CONDITION
D.C. = DON’T CARE
8 BITS
ACK SR
0
ADDRESS
7 BITS
R/W ACK BYTE COUNT N ACK DATA BYTE 1 ACK
1
0
8 BITS
0
8 BITS
ACK = ACKNOWLEDGE, SDA PULLED LOW DURING RISING EDGE OF SCL.
NACK = NOT ACKNOWLEDGE, SDA LEFT HIGH DURING RISING EDGE OF SCL.
ALL DATA IS CLOCKED IN/OUT OF THE DEVICE ON RISING EDGES OF SCL.
0
= SDA TRANSITIONS FROM HIGH TO LOW DURING PERIOD OF SCL.
= SDA TRANSITIONS FROM LOW TO HIGH DURING PERIOD OF SCL.
Figure 12. SMBus Protocols
36 �������������������������������������������������������������������������������������
P
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
REGISTERS
AND EEPROM
gered the SMBus alert. The master must send the ARA
before clearing r1Ch[7]. Clear r1Ch[7] by writing a ‘1’. If
GPIO1 is configured as the SMBus alert output but the
SMBus alert feature is disabled (r40h[4] is set to ‘0’),
GPIO1 acts as an additional fault output.
JTAG Serial Interface
The MAX16067 features a JTAG port that complies with
a subset of the IEEE 1149.1 specification. Either the
SMBus or the JTAG interface can be used to access
internal memory; however, only one interface is allowed
to run at a time. The MAX16067 contains extra JTAG
instructions and registers not included in the JTAG
specification that provide access to internal memory.
The extra instructions include LOAD ADDRESS, WRITE,
READ, REBOOT, and SAVE.
01100
01011
01010
01001
01000
00111
MEMORY WRITE REGISTER
[LENGTH = 8 BITS]
00110
MEMORY READ REGISTER
[LENGTH = 8 BITS]
00101
MEMORY ADDRESS REGISTER
[LENGTH = 8 BITS]
00100
USER CODE REGISTER
[LENGTH = 32 BITS]
00011
MUX 1
COMMAND
DECODER
01100
IDENTIFICATION REGISTER
[LENGTH = 32 BITS]
00000
BYPASS REGISTER
[LENGTH = 1 BIT]
11111
SETFLSHADD
01011
RSTFLSHADD
01010
RSTUSRFLSH
01001
SETUSRFLSH
01000
SAVE
00111
REBOOT
VDB
INSTRUCTION REGISTER
[LENGTH = 5 BITS]
RPU
MUX 2
TDO
TDI
TMS
TEST ACCESS PORT
(TAP) CONTROLLER
TCK
Figure 13. JTAG Block Diagram
______________________________________________________________________________________ 37
MAX16067
SMBALERT (ALERT)
The MAX16067 supports the SMBus alert protocol. To
enable the SMBus alert output, set r40h[4] to ‘1’, then
configure GPIO1 to act as the SMBus alert (ALERT)
according to Table 11. This output is open drain and
uses the wired-OR configuration with other devices
on the SMBus. During a fault, the MAX16067 asserts
ALERT low, signaling the master that an interrupt has
occurred. The master responds by sending the ARA
(Alert Response Address) protocol on the SMBus. This
protocol is a read byte with 09h as the slave address.
The slave acknowledges the ARA (09h) address and
sends its own SMBus address to the master. The slave
then deasserts ALERT. The master can then query the
slave and determine the cause of the fault. By checking
r1C[7], the master can confirm that the MAX16067 trig-
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Test Access Port (TAP)
instruction does not call for a parallel load or the selected
test data register does not allow parallel loads, the test
data register remains at its current value. On the rising
edge of TCK, the controller goes to the shift-DR state if
TMS is low or it goes to the exit1-DR state if TMS is high.
Controller State Machine
The TAP controller is a finite state machine that responds
to the logic level at TMS on the rising edge of TCK. See
Figure 14 for a diagram of the finite state machine. The
possible states are described as follows:
Shift-DR: The test data register selected by the current
instruction connects between TDI and TDO and shifts
data one stage toward its serial output on each rising
edge of TCK while TMS is low. On the rising edge of TCK,
the controller goes to the exit1-DR state if TMS is high.
Test-Logic-Reset: At power-up, the TAP controller
is in the test-logic-reset state. The instruction register
contains the IDCODE instruction. All system logic of the
device operates normally. This state can be reached
from any state by driving TMS high for five clock cycles.
Exit1-DR: While in this state, a rising edge on TCK puts the
controller in the update-DR state. A rising edge on TCK with
TMS low puts the controller in the pause-DR state.
Run-Test/Idle: The run-test/idle state is used between
scan operations or during specific tests. The instruction
register and test data registers remain idle.
Pause-DR: Shifting of the test data registers halts while
in this state. All test data registers retain their previous
state. The controller remains in this state while TMS is
low. A rising edge on TCK with TMS high puts the controller in the exit2-DR state.
Select-DR-Scan: All test data registers retain their previous state. With TMS low, a rising edge of TCK moves the
controller into the capture-DR state and initiates a scan
sequence. TMS high during a rising edge on TCK moves
the controller to the select-IR-scan state.
Exit2-DR: A rising edge on TCK with TMS high while in
this state puts the controller in the update-DR state. A rising edge on TCK with TMS low enters the shift-DR state.
Capture-DR: Data can be parallel-loaded into the test
data registers selected by the current instruction. If the
1
TEST-LOGIC-RESET
0
0
RUN-TEST/IDLE
1
SELECT-DR-SCAN
1
SELECT-IR-SCAN
0
1
0
1
CAPTURE-DR
CAPTURE-IR
0
0
SHIFT-DR
1
1
EXIT1-DR
1
EXIT1-IR
0
0
PAUSE-DR
PAUSE-IR
0
1
0
1
0
EXIT2-DR
EXIT2-IR
1
1
UPDATE-DR
1
0
SHIFT-IR
0
1
0
1
UPDATE-IR
0
1
Figure 14. Tap Controller State Diagram
38 �������������������������������������������������������������������������������������
0
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Select-IR-Scan: All test data registers retain the previous states. The instruction register remains unchanged
during this state. With TMS low, a rising edge on TCK
moves the controller into the capture-IR state. TMS high
during a rising edge on TCK puts the controller back into
the test-logic-reset state.
Capture-IR: Use the capture-IR state to load the shift
register in the instruction register with a fixed value.
This value is loaded on the rising edge of TCK. If TMS is
high on the rising edge of TCK, the controller enters the
exit1-IR state. If TMS is low on the rising edge of TCK,
the controller enters the shift-IR state.
Shift-IR: In this state, the shift register in the instruction
register connects between TDI and TDO and shifts data
one stage for every rising edge of TCK toward the TDO
serial output while TMS is low. The parallel outputs of
the instruction register as well as all test data registers
remain at the previous states. A rising edge on TCK with
TMS high moves the controller to the exit1-IR state. A
rising edge on TCK with TMS low keeps the controller in
the shift-IR state while moving data one stage through
the instruction shift register.
Exit1-IR: A rising edge on TCK with TMS low puts the
controller in the pause-IR state. If TMS is high on the rising edge of TCK, the controller enters the update-IR state.
Pause-IR: Shifting of the instruction shift register halts
temporarily. With TMS high, a rising edge on TCK puts
the controller in the exit2-IR state. The controller remains
in the pause-IR state if TMS is low during a rising edge
on TCK.
Exit2-IR: A rising edge on TCK with TMS high puts the
controller in the update-IR state. The controller loops
back to shift-IR if TMS is low during a rising edge of TCK
in this state.
Update-IR: The instruction code that has been shifted
into the instruction shift register latches to the parallel
outputs of the instruction register on the falling edge of
TCK as the controller enters this state. Once latched,
this instruction becomes the current instruction. A rising
edge on TCK with TMS low puts the controller in the runtest/idle state. With TMS high, the controller enters the
select-DR-scan state.
Instruction Register
The instruction register contains a shift register as well
as a latched 5-bit wide parallel output. When the TAP
controller enters the shift-IR state, the instruction shift
register connects between TDI and TDO. While in the
shift-IR state, a rising edge on TCK with TMS low shifts
the data one stage toward the serial output at TDO. A
rising edge on TCK in the exit1-IR state or the exit2-IR
state with TMS high moves the controller to the updateIR state. The falling edge of that same TCK latches the
data in the instruction shift register to the instruction
register parallel output. Table 28 shows the instructions
supported by the MAX16067 and the respective operational binary codes.
Table 28. JTAG Instruction Set
INSTRUCTION
CODE
NOTES
BYPASS
0x1F
Mandatory instruction code
IDCODE
0x00
Load manufacturer ID code/part number
USERCODE
0x03
Load user code
LOAD ADDRESS
0x04
Load address register content
READ DATA
0x05
Read data pointed by current address
WRITE DATA
0x06
Write data pointed by current address
REBOOT
0x07
Reboot FLASH data content into register file
SAVE
0x08
Trigger emergency save to flash
SETFLSHADD
0x09
Flash page access ON
RSTFLSHADD
0x0A
Flash page access OFF
SETUSRFLSH
0x0B
User flash access ON (must be in flash page already)
RSTUSRFLSH
0x0C
User flash access OFF (return to flash page)
______________________________________________________________________________________ 39
MAX16067
Update-DR: A falling edge on TCK while in the updateDR state latches the data from the shift register path of
the test data registers into a set of output latches. This
prevents changes at the parallel output because of
changes in the shift register. On the rising edge of TCK,
the controller goes to the run-test/idle state if TMS is low
or goes to the select-DR-scan state if TMS is high.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
BYPASS: When the BYPASS instruction is latched into
the instruction register, TDI connects to TDO through the
1-bit bypass test data register. This allows data to pass
from TDI to TDO without affecting the device’s operation.
memory in the MAX16067. When the LOAD ADDRESS
instruction latches into the instruction register, TDI connects to TDO through the 8-bit memory address test
data register during the shift-DR state.
IDCODE: When the IDCODE instruction is latched into
the parallel instruction register, the identification data
register is selected. The device identification code is
loaded into the identification data register on the rising
edge of TCK following entry into the capture-DR state.
Shift-DR can be used to shift the identification code
out serially through TDO. During test-logic-reset, the
IDCODE instruction is forced into the instruction register.
The identification code always has a ‘1’ in the LSB position. The next 11 bits identify the manufacturer’s JEDEC
number and number of continuation bytes followed by
16 bits for the device and 4 bits for the version. See
Table 29.
READ DATA: This is an extension to the standard
IEEE 1149.1 instruction set to support access to the
memory in the MAX16067. When the READ instruction
latches into the instruction register, TDI connects to TDO
through the 8-bit memory read test data register during
the shift-DR state.
USERCODE: When the USERCODE instruction latches
into the parallel instruction register, the user-code data
register is selected. The device user-code loads into the
user-code data register on the rising edge of TCK following entry into the capture-DR state. Shift-DR can be
used to shift the user-code out serially through TDO. See
Table 30. This instruction can be used to help identify
multiple MAX16067 devices connected in a JTAG chain.
LOAD ADDRESS: This is an extension to the standard
IEEE 1149.1 instruction set to support access to the
WRITE DATA: This is an extension to the standard
IEEE 1149.1 instruction set to support access to the
memory in the MAX16067. When the WRITE instruction
latches into the instruction register, TDI connects to TDO
through the 8-bit memory write test data register during
the shift-DR state.
REBOOT: This is an extension to the standard IEEE
1149.1 instruction set to initiate a software controlled
reset to the MAX16067. When the REBOOT instruction latches into the instruction register, the MAX16067
resets and immediately begins the boot-up sequence.
SAVE: This is an extension to the standard IEEE 1149.1
instruction set that triggers a fault log. The current ADC
conversion results along with fault information are saved
to flash depending on the configuration of the Critical
Fault Log Control register (r6Dh).
Table 29. 32-Bit Identification Code
MSB LSB
VERSION (4 BITS)
PART NUMBER (16 BITS)
MANUFACTURER (11 BITS)
FIXED VALUE (1 BIT)
0001
1000000000000001
00011001011
1
Table 30. 32-Bit User-Code Data
MSB
LSB
DON’T CARE
00000000000000000
SMBUS SLAVE ID
USER ID (r8A[7:0])
See Table 26
40 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
RSTFLSHADD: This is an extension to the standard
IEEE 1149.1 instruction set. Use RSTFLSHADD to return
to the default page and disable access to the flash page.
SETUSRFLSH: This is an extension to the standard
IEEE 1149.1 instruction set that allows access to the user
flash page. When on the configuration flash page, send
the SETUSRFLSH command, all addresses are recognized as flash addresses only. Use this page to access
registers 300h–3FFh.
RSTUSRFLSH: This is an extension to the standard
IEEE 1149.1 instruction set. Use RSTUSRFLSH to return
to the configuration flash page and disable access to
the user flash.
Restrictions When Writing to Flash
Flash must be written to 8 bytes at a time. The initial
address must be aligned to 8-byte boundaries—the
3 LSBs of the initial address must be ‘000’. Write the
8 bytes using 8 successive Write Data commands. A
write operation requires 122ms for each 8-byte block.
After programming a block, check r20h[1] (see Table
31) to make sure the write operation is complete before
attempting to write the next block.
Applications Information
Unprogrammed Device Behavior
When the flash has not been programmed using the
JTAG or SMBus interface, the default configuration of
the EN_OUT_ outputs is open drain active low. This
means that the EN_OUT_ outputs are high impedance.
When it is necessary to hold an EN_OUT_ high or low to
prevent premature startup of a power supply before the
flash is programmed, connect a resistor from EN_OUT_
to ground or the supply voltage. Avoid connecting a
resistor to ground when the output is to be configured as
open drain with a separate pullup resistor.
Device Behavior at Power-Up
When VCC is ramped from 0V, the RESET output is
high impedance until VCC reaches 1.4V, at which point
RESET goes low. All other outputs are high impedance
until VCC reaches 2.7V, then the flash contents are copied into register memory. This takes 150Fs (max) after
which the outputs assume their programmed states.
Programming the MAX16067 in Circuit
The MAX16067 can be programmed in the application
circuit by taking into account the following points during
circuit design:
UThe MAX16067 needs to be powered from an intermediate voltage bus or auxiliary voltage supply so
programming can occur even when the board’s power
supplies are off. This could also be achieved by using
ORing diodes so that power can be provided through
the programming connector.
UThe SMBus or JTAG bus lines should not connect
through a bus multiplexer powered from a voltage rail
controlled by the MAX16067. If the device needs to be
controlled by an on-board FP, consider connecting the
FP to one bus (such as SMBus) and use the other bus
for in-circuit programming.
UAn unprogrammed MAX16067’s EN_OUT_s go high
impedance. Ensure that this does not cause undesired
circuit behavior. If necessary, connect pulldown resistors to prevent power supplies from turning on.
Table 31. RESET State, Flash State, and Reset Reason
REGISTER ADDRESS
r20h
BIT RANGE
DESCRIPTION
[0]
Reset Output State
0 = RESET is low
1 = RESET is high
[1]
1 = Flash memory is busy
[2]
1 = Last reset asserted due to EN going low
[3]
1 = Last reset asserted due to watchdog timeout
[7:4]
Not used
______________________________________________________________________________________ 41
MAX16067
SETFLSHADD: This is an extension to the standard
IEEE 1149.1 instruction set that allows access to the
flash page. Flash registers include ADC conversion
results, DACOUT enables, and GPIO input/output data.
Use this page to access registers 200h-2FFh.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Maintaining Power During
a Fault Condition
Power to the MAX16067 must be maintained for a specific period of time to ensure a successful flash fault log
operation during a fault that removes power to the circuit.
Table 32 shows the amount of time required depends on
the settings in the fault control register (r6Dh[1:0]).
Maintain power for shutdown during fault conditions in
applications where the always-on power supply cannot
be relied upon by placing a diode and a large capacitor
between the voltage source, VIN, and VCC (Figure 15).
The capacitor value depends on VIN and the time delay
required, tFAULT_SAVE. Use the following formula to calculate the capacitor size:
C = (tFAULT_SAVE x ICC(MAX))/(VIN - VDIODE - VUVLO)
where the capacitance is in Farads and tFAULT_SAVE is in
seconds, ICC(MAX) is 14mA, VDIODE is the voltage drop
across the diode, and VUVLO is 2.7V. For example, with
a VIN of 14V, a diode drop of 0.7V, and a tFAULT_SAVE
of 153ms, the minimum required capacitance is 202FF.
Driving High-Side MOSFET Switches
Up to three of the programmable outputs (EN_OUT1,
EN_OUT2, EN_OUT3) of the MAX16067 can be configured as charge-pump outputs to drive the gates
of series-pass n-channel MOSFETS. When driving
MOSFETs, these outputs act as simple power switches
to turn on the voltage supply rails. Approximate the slew
rate, SR, using the following formula:
SR = ICP/(CGATE + CEXT)
where ICP is the 6FA (typ) charge-pump source current, CGATE is the gate capacitance of the MOSFET,
and CEXT is the capacitance connected from the gate
to ground. If more than three series-pass MOSFETs
are required for an application, additional series-pass
p-channel MOSFETs can be connected to outputs configured as active-low open drain (Figure 16). Connect
a pullup resistor from the gate to the source of the
MOSFET, and ensure the absolute maximum ratings of
the MAX16067 are not exceeded.
Configuring the Device
An evaluation kit and a graphical user interface (GUI) are
available to create a custom configuration for the device.
Refer to the MAX16067 Evaluation Kit for configuration.
Cascading Multiple MAX16067s
Multiple MAX16067s can be cascaded to increase the
number of rails controlled for sequencing and monitoring. There are many ways to cascade the devices
depending on the desired behavior. In general, there are
several techniques as follows:
U Configure a GPIO on each device to be EXTFAULT
(open-drain). Externally wire them together with a single
pullup resistor. Set register bits r72h[5] and r6Dh[2] to ‘1’
Table 32. Maximum Write Time
r6Dh[1:0] VALUE
DESCRIPTION
MAXIMUM WRITE TIME (ms)
00
Save flags and ADC readings
153
01
Save flags
102
10
Save ADC readings
153
11
Do not save anything
—
VIN
R
VCC
VIN
VOUT
C
MAX16067
GND
Figure 15. Power Circuit for Shutdown During Fault Conditions
MON_
EN_OUT_
MAX16067
Figure 16. Connection for a p-Channel Series-Pass MOSFET
42 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
U Connect open-drain RESET outputs together to obtain
a master system reset signal.
U Connect all EN inputs together for a master enable signal. Since the internal timings of each cascaded device
are not synchronized, EN_OUT_s placed in the same
slot on different devices do not come up in the desired
order even if the sequence delays are identical.
U Consider using an external FP to control the EN inputs
or the software enable bits of cascaded devices, monitoring the RESET outputs as a power-good signal.
U For a large number of voltage rails, the MAX16067
can be cascaded hierarchically by using one master
device’s EN_OUT_s to control the EN inputs of several
slave devices.
Controlling Power Supplies
Without Using the Sequencer
A FP can control power supplies manually without involving the sequencing slot system by controlling EN_OUT_s
configured as GPIO. The output of a power supply controlled this way can be monitored using a MON_ input
configured as “monitoring only” (see the Monitoring
Inputs While Sequencing section). To monitor the supply for critical faults, the FP will need to manually set the
critical fault enable bit in r6Eh–r72h after turning on the
EN_OUT_, and manually clearing the critical fault enable
bit before turning off the EN_OUT_.
Layout and Bypassing
Bypass DBP and ABP each with a 1FF ceramic capacitor
to GND. Bypass VCC with a 10FF capacitor to ground.
Avoid routing digital return currents through a sensitive
analog area, such as an analog supply input return path
or ABP’s bypass capacitor ground connection. Use
dedicated analog and digital ground planes. Connect
the capacitors as close as possible to the device.
Figure 17. Graphical User Interface Screenshot
______________________________________________________________________________________ 43
MAX16067
so that all faults propagate between devices. If a critical
fault occurs on one device, EXTFAULT asserts, triggering
the nonvolatile fault logger in all cascaded devices and
recording a snapshot of all system voltages.
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Register Map
FLASH
ADDRESS
REGISTER
ADDRESS
READ/
WRITE
DESCRIPTION
ADC VALUES, FAULT REGISTERS, GPIOs AS INPUT PORTS—NOT IN FLASH
—
000
R
MON1 ADC output, MSBs
—
001
R
MON1 ADC output, LSBs
—
002
R
MON2 ADC output, MSBs
—
003
R
MON2 ADC output, LSBs
—
004
R
MON3 ADC output, MSBs
—
005
R
MON3 ADC output, LSBs
—
006
R
MON4 ADC output, MSBs
—
007
R
MON4 ADC output, LSBs
—
008
R
MON5 ADC output, MSBs
—
009
R
MON5 ADC output, LSBs
—
00A
R
MON6 ADC output, MSBs
—
00B
R
MON6 ADC output, LSBs
—
00C–01A
—
Reserved
—
01B
R/W
Fault register—failed line flags
—
01C
R/W
Fault register—failed line flags
—
01D
—
Reserved
—
01E
R
GPIO data in (read only)
—
01F
R
EN_OUT_ as GPIO data in (read only)
—
020
R/W
—
021
R
Flash status/reset output monitor
Current sequencer slot
GPIO AND OUTPUT DEPENDENCIES/CONFIGURATIONS
230
030
R/W
EN_OUT_ configuration
231
031
R/W
EN_OUT_ configuration
232
032
—
233
033
R/W
Charge-pump configuration bits
234
034
R/W
EN_OUT_ as GPIO data out
235
035
—
236
036
R/W
FAULT dependencies
237
037
R/W
238–23A
038–03A
—
FAULT dependencies
Reserved
Reserved
Reserved
23B
03B
R/W
RESET output configuration
23C
03C
R/W
RESET output dependencies
23D
03D
R/W
RESET output dependencies
23E
03E
R/W
GPIO data out
23F
03F
R/W
GPIO configuration
240
040
R/W
GPIO configuration, ARANEN (ARA Enable)
241–242
041–042
—
Reserved
44 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
FLASH
ADDRESS
REGISTER
ADDRESS
READ/
WRITE
DESCRIPTION
ADC—CONVERSIONS
243
043
R/W
ADCs voltage ranges for MON_ monitoring
244
044
R/W
ADCs voltage ranges for MON_ monitoring
245–247
045–047
—
Reserved
Reserved
INPUT THRESHOLDS
248
048
—
249
049
R/W
MON1 OV threshold
24A
04A
R/W
MON1 UV threshold
24B
04B
—
24C
04C
R/W
MON2 OV threshold
24D
04D
R/W
MON2 UV threshold
24E
04E
—
24F
04F
R/W
MON3 OV threshold
250
050
R/W
MON3 UV threshold
251
051
—
252
052
R/W
MON4 OV threshold
253
053
R/W
MON4 UV threshold
254
054
—
255
055
R/W
MON5 OV threshold
256
056
R/W
MON5 UV threshold
257
057
—
258
058
R/W
MON6 OV threshold
259
059
R/W
MON6 UV threshold
25A–26C
05A–06C
—
26D
06D
R/W
26E
06E
R/W
Save after EXTFAULT fault control
Faults causing store in flash
26F
06F
R/W
Faults causing store in flash
270
070
R/W
Faults causing store in flash
271
071
—
272
072
R/W
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
FAULT SETUP
Reserved
EXTFAULT enable
______________________________________________________________________________________ 45
MAX16067
Register Map (continued)
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Register Map (continued)
FLASH
ADDRESS
REGISTER
ADDRESS
READ/
WRITE
273
073
R/W
Watchdog independent mode, margin enabled, soft RESET functionality
274
074
R/W
ADC fault deglitch/autoretry configuration
275
075
R/W
WDI toggle/fault timeout, reverse sequencing bit
276
076
R/W
WDRESET, WD timers
277
077
R/W
Sequence delay for Slot 0
278
078
R/W
Sequence delay for Slot 1
279
079
R/W
Sequence delay for Slot 2
27A
07A
R/W
Sequence delay for Slot 3
DESCRIPTION
TIMEOUTS
27B
07B
R/W
Sequence delay for Slot 4
27C
07C
R/W
Sequence delay for Slot 5
27D
07D
R/W
Sequence delay for Slot 6
27E
07E
R/W
Assign MON_ input from Slot 1 to Slot 6 or for monitoring
27F
07F
R/W
Assign MON_ input from Slot 1 to Slot 6 or for monitoring
280
080
R/W
Assign MON_ input from Slot 1 to Slot 6 or for monitoring
281–283
081–083
—
284
084
R/W
Assign EN_OUT_ from Slot 1 to Slot 6
285
085
R/W
Assign EN_OUT_ from Slot 1 to Slot 6
286
086
R/W
Assign EN_OUT_ from Slot 1 to Slot 6
287–289
087–089
—
28A
08A
R/W
Customer use (version)
MISCELLANEOUS
Reserved
Reserved
28B
08B
R/W
PEC enable/SMBus address
28C
08C
R/W
Lock bits
28D
08D
R
Revision code
46 �������������������������������������������������������������������������������������
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
+48V IN
OUT
IN
DC-DC
EN
LATCH
CIRCUIT
IN
OUT
IN
OUT
IN
OUT
DC-DC
DC-DC
DC-DC
EN
EN
EN
EN_OUT1
MON1
EN_OUT2
MON2
EN_OUT3–
EN_OUT6
VCC
3.3V
MON3–
MON6 FAULT
RESET
MAX16067
SCL
EN
ABP
DBP
GND
AO
Package Information
Chip Information
PROCESS: BiCMOS
SDA
For the latest package outline information and land patterns, go to www.maxim-ic.com/packages. Note that
a “+”, “#”, or “-” in the package code indicates RoHS
status only. Package drawings may show a different suffix character, but the drawing pertains to the package
regardless of RoHS status.
PACKAGE TYPE
PACKAGE CODE
DOCUMENT NO.
32 TQFN-EP
T3255+4
21-0140
______________________________________________________________________________________ 47
MAX16067
Typical Operating Circuit
MAX16067
6-Channel, Flash-Configurable System Manager
with Nonvolatile Fault Registers
Revision History
REVISION
NUMBER
REVISION
DATE
0
10/09
Initial release
1
2/10
Updated Absolute Maximum Ratings and Table 19
DESCRIPTION
PAGES
CHANGED
—
2, 23
Maxim cannot assume responsibility for use of any circuitry other than circuitry entirely embodied in a Maxim product. No circuit patent licenses are implied.
Maxim reserves the right to change the circuitry and specifications without notice at any time.
48
© 2010
Maxim Integrated Products, 120 San Gabriel Drive, Sunnyvale, CA 94086 408-737-7600
Maxim Integrated Products Maxim is a registered trademark of Maxim Integrated Products, Inc.