DDTS item OCCcm08977


The Anomaly
===========

On Jan 25, 2010 the ACIS instrument experienced an unexpected Back End
Processor (BEP) reboot. The event resulted in a reversion to the
pre-launch version of the ACIS flight software (Ver. 11). Real-time CAP
(CAP_1144B) was run on Jan 25 to restart the ACIS software housekeeping reporting,
terminate the science run, and to dump the contents of the ACIS
software patch memory. The anomalous event had also erased the
software patches from memory. A second real-time CAP (CAP_1145B) was
run on Jan 25 to reload the patches to Version 44, and warm boot the 
ACIS BEP, and restart DEA housekeeping. A third real-time CAP
(CAP_1146) was completed on Jan 26 to reset the ACIS focal-plane
control point to -121 degree C as a final clean-up action following the anomaly. 


Anomaly Analysis
================

A switch from s/w rev 44 to 11 can occur if the BEP has had a watchdog reboot
or a cold (power-on) re-boot. There are 4 types of reboot,
distinguished by the state of various DPA H/W registers and the degree
to which they rewrite RAM from EEPROM.

All telemetry indicators point to a power-up reboot: No fatalError
packet was transmitted and the bepStartupMessage packet indicated that
this was not a watchdog reboot.   The team concluded that no
corrective action was necessary, and that one of the following two
scenarios is most likely to cause the reboot:  


1) Momentary power-drop from the PSMC to the BEP
–-----------------------------------------------
On 4 previous occasions, PSMC power to DEA or DPA dropped without
being commanded to do so.  On 3 other occasions, FEP boards had reset
without warning and the Instrument Team concluded that PSMC power had
dropped briefly. 


2) Temporary breakdown in the BEP’s power-up capacitor
------------------------------------------------------
During BEP power-up, an R-C circuit holds the Mongoose processor in a
reset state for 50msec while the remaining components power up. If 
the capacitor (22 μF tantalum) had been shorted, e.g., by ionizing
radiation, which has been observed in similar situations, it would
have reset and then rebooted the BEP.  By design, the series resistor
would prevent permanent damage to the capacitor.