Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/ops/swat/safing/01064/recovery.txt Дата изменения: Wed Mar 7 17:28:07 2001 Дата индексирования: Sun Mar 2 02:02:06 2014 Кодировка: Поисковые слова: п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п

From: "Rodger Doxsey"
To:
Cc:
Subject: Safemode recovery
Date: Tue, 6 Mar 2001 17:25:15 -0500

The engineering data covering the period of the safemode entry were
recovered yesterday afternoon. Analysis of the data, and review of earlier
data, led to a decision to resume science operations. The magnetic torquing
bar which was shut down by the safing action was turned on this afternoon,
and restored to the control law. It has been acting normally, with no sign
of a hardware problem. The command loads for recovering the instruments and
re-entering the science observing program are now loaded and will start
executing this evening. The first science observation will take place at
roughly 2 am local time. The net down time for the event is about 48 hours.
The scheduling team took advantage of the need to rebuild the schedule,
adjusting the timing of some GRB ToO observations to reflect slower
evolution of the object than expected. Observations lost during the safing
are being rescheduled, with the expectation that most will take place in the
next few weeks.

The situation surrounding the safemode entry is a little more complex than
originally expected, and will require further study. There is a good
possibility that this may have been a fluke incident, rather than a symptom
of any hardware problems. The flight software compares the commanded
currents in each magnetic torquer with the measured values. This check is
made every second, but data is not sent to the ground with that resolution.
If the check exceeds a specified tolerance for 120 seconds in a row, then
the test fails and safemode is entered. The test is designed to catch
failures in the hardware, such as a failed transistor, which might result in
the torquer bar not responding to the commands. It appears that there are
regular circumstances when this threshold is exceeded for periods of tens of
seconds, up to 90 seconds. This has happened dozens of times in the last
two years, but never reaching the 120 second threshold, so has never caused
a safemode entry. It has happened with all four torquer bars, in different
combinations and individually. The occurrence of these events is correlated
with the HST being in the SAA. This is significant because this is the
period when the HST sees the lowest amplitude for the Earth's magnetic
field. A current suspicion is that in these periods we can get a feedback
loop in the control law, possibly as a result of the field generated by the
torquer bars influencing strongly the sensed "earth magnetic field" from the
magnetometers. There is some evidence in the poorly sampled data that there
are large oscillations in the system at this time. To investigate this
further, the FOT will put the HST into P telemetry format for some future
SAA passages. This will get a uniform, high rate sampling of the control
law performance. There does not seem to be a single hardware component
which could fail or degrade in a way to cause these symptoms to occasionally
occur with all four torquers, which is the strongest reason why we think
that there has not been a hardware failure.

The project will be appointing an Anomaly Review Board to investigate this
situation. They are likely to be asked to recommend new settings for the
safemode test, if they can conclusively determine what happened. I am sure
they will also take a closer look to be sure that we are not ignoring
possible intermittent hardware problems.

Rodger