| About ROSAT |
ROSAT Home Page | ROSAT Images |
|---|
[ROSAT AO-21991, NRA 91-OSSA-31991], [Hasinger1992a]
The in-orbit performance of ROSAT is closely related to the performance of its AMCS, hence the Mission Timeline ([Snowden and Schmitt1992]) is very much influenced by any problems in the attitude control system. Although all of what follows can be found in chronological order in the App. A (Mission History), the relevant events, concerning the in-orbit performance, are described in this section in more detail.
Intermittent attitude problems were experienced both during the all-sky survey and (particularly) AO-1 mission phases due to an excessive drift rate of the X-gyro. The X-gyro was therefore taken out of the attitude control loop during the sky-survey, but included again during AO-1 with a compensatory bias introduced into the control software.
On 1991 May 12, the motor current to the Y-gyro glitched to exceed
its operational limit.
While the gyro was regained for a short time,
controlling it became impossible,
and the gyro was considered to have failed beyond retrieval.
At this point, ROSAT had lost a non-redundant component of the AMCS
and normal operations became impossible.
The satellite remained in safe mode until 1991 May 18,
when attitude control was successfully regained in
reduced-pointing mode
utilizing the (unreliable) X-gyro.
The problem appeared to be electrical in nature and probably
unrelated to the drift problems experienced with the X-gyro.
However, given the problems with the X-gyro, GSOC planners had to
tightly constrain slews and acquisition of targets,
allowing them to achieve equilibrium for the gyro drift.
Slews were reduced to one direction and constrained to be short,
guide stars needed to be available when the satellite was over
Weilheim ground station
(i.e., target declinations were constrained to be > +20 
),
a reduced sun cone angle was enforced (90 
5 
)
and only one target was observed per day.
Also during this period, an increased level of solar activity led to high
levels of radiation, causing the detector to safe itself several times.
These conditions resulted in a yield of 
30,000 seconds of ROSAT
data for each day of observation.
After a long slew on 1991 July 5, ROSAT became lost due to Earth
occultation and a lack of guide stars, resulting in the S/C going
into safe-mode until recovery on July 7.
The uploading of new AMCS S/W on July 8-9 enabled a slight relaxation
in the reduced pointing phase observing constraints, allowing southern
hemisphere observations and a large sun cone of
90 
12 
.
Further modifications to the AMCS on-board software enabled a
successful resumption of normal pointing operations on 1991 November 4.
The software commanding the star trackers was improved in 1991 October
allowing the tracking of six stars in search windows of 
,
as compared to the previous limit of three stars in a 1 
window.
Magnetometer information has also been utilized since that time.
The combination of magnetometer and sun sensor
data allows a 3-axis determination on the day-side of the orbit.
The gyros are still needed for slews and velocity damping in the
control loop.
A large transient in the ROSAT Z-gyro current occurred late
on 1992 February 6, reminiscent of that
which immediately preceded the loss of the Y-gyro in 1991 May 12.
The S/C was put into safe mode and observations suspended.
The Z-gyro, although operating continuously with too high a current,
was still giving meaningful readings of angular velocities.
However, its output signal seemed to be scaled wrongly,
indicating a factor of 3-4 slower speeds
than the spacecraft was actually performing.
Further evaluation indicated the outputs of the gyro were still valid,
however the scale factor had increased by a factor of 
2.8.
The most convincing explanation of this is that 1 of 3 phases of
the gyro motor failed, and consequently the motor current had increased.
To compensate for the decreased Z-gyro output, a change of the on board
software was developed
(including a factor 2.8 decrease in the scale factor)
and successfully up-linked on February 20.
This software patch meant that the scale factors of the other gyros
can no longer be commanded,
ROSAT went back on the mission time line on 1992 February 24
and the first three slews were successfully performed.
This success indicated that the new Z-gyro scale factor was correctly
determined and appeared to be stable.
The RW-3 reaction wheel has demonstrated above nominal static and dynamic friction. This causes the wheel to stop at rotation rates close to zero.