Rainbow Electronics MAX34440 - Manual (Page 21)

PMBus 6-Channel Power-Supply Manager
MAX34440
21
Fault Management and Reporting
For reporting faults/warnings to the host on a real-time
basis, the device asserts the open-drain ALERT pin (if
enabled in MFR_MODE) and sets the appropriate bit in
the various status registers. On recognition of the ALERT
assertion, the host or system manager is expected to poll
the I
2 C bus to determine the device asserting ALERT.
The host sends the SMBus ARA (0001 100). The device
ACKs the SMBus ARA, transmits its slave address, and
deasserts ALERT. The system controller then communi-
cates with PMBus commands to retrieve the fault/warn-
ing status information from the device.
See the individual command sections for more details.
Faults and warnings that are latched in the status regis-
ters are cleared when any one of the following conditions
occurs:
• A CLEAR_FAULTS command is received.
• The RST pin is toggled.
• Bias power to the device is removed and then reapplied.
One or more latched-off power supplies is only restarted
when one of the following occurs:
• The output is commanded through the CONTROL
pin, the OPERATION command, or the combined
action of the CONTROL pin and OPERATION com-
mand to turn off and then turn back on.
• The RST pin is toggled.
• Bias power to the device is removed and then reapplied.
A power supply is not allowed to turn on if any faults the
supply responds to are detected. Only after the faults
clear is the power supply allowed to turn on. When global
supplies are being sequenced on, a fault on any of the
supplies keeps all supplies from being turned on.
A system-wide power-up (OPERATION command is
received to turn the supplies on when PAGE is 255 or the
CONTROL pin is toggled to turn on the supplies) allows
all enabled power supplies to power-up. If any faults are
detected once the supplies start to turn on, the response
of MFR_FAULT_RESPONSE is performed.
The device responds to fault conditions accord-
ing to the manufacturer fault response command
(MFR_FAULT_RESPONSE). This command byte deter-
mines how the device should respond to each particular
fault. Table 5 illustrates the required conditions and fault
actions for specific parameters.
System Watchdog Timer
The device uses an internal watchdog timer that is inter-
nally reset every 5ms. In the event that the device is
locked up and this watchdog reset does not occur after
500ms, the device automatically resets. After the reset
occurs, the device reloads all configuration values that
were stored to flash and begins normal operation. After
the reset, the device also does the following:
1) Sets the NONE OF THE ABOVE bit in STATUS_BYTE.
2) Sets the NONE OF THE ABOVE and MFR bits in
STATUS_WORD.
3) Sets the WATCHDOG bit in STATUS_MFR_SPECIFIC.
4) Notifies the host through ALERT assertion (if enabled
in MFR_MODE).
Table 5. Device Parametric Monitoring States
PARAMETER REQUIRED CONDITIONS FOR ACTIVE MONITORING ACTION DURING A FAULT
Overvoltage Power Supply Enabled (TON_MAX_FAULT_LIMIT ≠ 0000h) Continue Monitoring
Undervoltage
• Power Supply Enabled (TON_MAX_FAULT_LIMIT ≠ 0000h)
• PSEN Output is Active
• Channel’s VOUT Must Have Exceeded VOUT_UV_FAULT During
Channel Power-Up
Stop Monitoring While the Power
Supply is Off
Overcurrent
• Power Supply Enabled (TON_MAX_FAULT_LIMIT ≠ 0000h)
• Current Monitoring Enabled (IOUT_OC_FAULT_LIMIT ≠ 0000h)
Continue Monitoring
Power-Up Time Power Supply Enabled (TON_MAX_FAULT_LIMIT ≠ 0000h) Monitor Only During Power-On
Overtemperature Temp Sensor Enabled (ENABLE in MFR_TEMP_SENSOR_CONFIG = 1) Continue Monitoring