Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/ops/swat/safing/97099/safe.txt Дата изменения: Thu Apr 10 02:15:26 1997 Дата индексирования: Sun Mar 2 00:58:32 2014 Кодировка: Поисковые слова: cygnus loop

From: DOXSEY@avion.stsci.edu
To: jay@stsci.edu, mcclure@stsci.edu, mackenty@stsci.edu, miller@stsci.edu,
NORMAN@stsci.edu
Subject: HST entered Zero Gyro safemode this afternoon, Gyro 4 probably failed
Date: Wed, 09 Apr 1997 18:05:34 -0500 (EST)



Safemode entry on April 9

The HST entered zero gyro safemode this afternoon at about 17:14 UT,
or 1:15 local time. At this point, it looks like gyro 4 has failed
and will no longer be usable. The scenario is believed to be as follows,
but followup engineering analysis will be taking place for verification.
There are telemetry indicators that gyro 4 failed about an hour and a half
before the safemode entry. It failed into a state where the rate output
of the gyro was zero counts. At the time, we were guiding with FGSs and
observing with the WFPC-II. In this situation, zero is a perfectly legal,
and common, output from the gyro. The flight software checks on the sanity
of the gyro data, done by comparing the outputs of the 4 gyros with expected
rates, did not exceed the error thresholds so the gyro was left in the
control law. At about 17:14 the HST started a large vehicle slew to
the next target. At this point, the zero counts from the gyro differed from
the expected values by a large amount, and the gyro was automatically
taken out of the loop by the DF224. However, by this time the erroneous
counts from the gyro, used in the control algorithm, had resulted in a
divergence of the computed from actual slew path, so eventually the other
gyros failed the sanity test as well. When the second gyro was dropped
off-line, that left too few to control the HST on gyros, so the zero
gyro safemode was entered. This general failure path had been discussed
back in the days before SM-1, when we were short of gyros, and it was
understood that if a gyro fails to a legal and valid output, then it
might take some time until the system detected that fact. During previous
gyro failures, they have gone to a saturated, full scale output which
is easy to detect.

The zero gyro safemode has been used only once before, when it
was tested after the control algorithm was installed (I think sometime
in 1992, but my memory is fading). In this mode the control law uses
the sun sensors during the day portion of the orbit to control the
pointing. In the night portion it uses rates derived from the
magnetometers to keep the drift low. As it re-emerges and sees the sun again
the sun sensors once again are used for pointing control. There is
typically 10 - 20 degrees of drift during each night period, so it
takes a few minutes to get the pointing back to nominal and full
sunlight on the arrays. The entry into this mode appeared to be nominal
and performance so far has been as expected.

The current plan is to use one of the ground simulators to reproduce
the postulated sequence of events. The gyro 4 in the simulator will be
forced to "fail" to zero output, and then the same slew will be executed.
This should result in the same sequence of calculations in the DF, with the
other gyros being pulled off-line in the same sequence. The data from the
other three gyros (3,5, and 6) have looked nominally after safemode
entry. The measured rates are what we would expect in the zero gyro
mode, so there is no reason at this time to think that we have a problem
with any of the others. Overnight, the HST will be recovered to the
normal inertial hold software sunpoint, using these gyros for control.
Tomorrow morning the project will review the data and determine which
of the spare gyros to activate (gyros 1 and 2 have been in "storage",
turned off, for some time now). Once the HST is in the inertial hold
software sunpoint, the MOSES staff will use FHSTs to get star maps
and determine the precise attitude. This will allow the PRESTO staff
to start on the health and safety SMS, which probably will begin
executing at roughly 101:00:00:00 (Thursday evening). If things go
smoothly, then we will pick up the science timeline roughly tweleve
hours later.

Unlike the previous HST safing, this event will cost us some
important SMOV proposals, including the NIMCOS absolute photometric
measurements and some STIS and NICMOS EROs. It will gain us another
serendipitous annealling of the STIS CCD. The PRESTO staff will be
circulating a list of the observations we expect to loose given the
current plan. We will not be rebuilding next week's SMS, it is too
late for that at this point.


From: David Lychenheim
To: dtaylor@stsci.edu, reinhart@stsci.edu
Cc: chapman@stsci.edu
Subject: Safemode Status
Date: Wed, 09 Apr 1997 14:27:35 -0400

SI Configuration prior to safemode

STIS Operate
NICMOS SAAOPR
WFPCII Operate
FOC Hold (side A)
COSTAR Hold

099:17:17:13 enter software sunpoint safemode
SI Configuration after Safemode

STIS Safe
NICMOS SAAOPR
WFPCII Operate
FOC Safe to (side B)
COSTAR Hold

Per SI SE Colleen Townsley