National Aeronautics and Space Administration

All Status Reports

Volumes 1 - 9

FOREWARD

The European Space Agency's X-ray observatory EXOSAT was operational from May 1983 to April 1986 and, during that interval, it made 1780 detailed observations of a wide variety of objects including active galactic nuclei, stellar coronae, cataclysmic variables, white dwarfs, X-ray binaries, clusters of Galaxies, and supernova remnants. Between 1983 October and 1987 September, eighteen volumes of the EXOSAT Express were published. This journal was used to rapidly disseminate information about the mission status, calibrations, analysis techniques and other useful information. It was published by the EXOSAT Observatory team, under the editorial leadership of Dave Andrews (1983-1986) and Nick White (1986-1987). The Express represents an important source of knowledge about the mission that has not been published elsewhere. As the years have gone by, many people have lost copies, or do not have access to them. At the HEASARC, we receive regular requests for particular volumes, or articles. To ensure this critical information about the EXOSAT mission is widely available and does not get lost, we have compiled all the useful articles and reprinted them here in two parts called "The Best of the EXOSAT Express." The first part begins with all the "status reports", written by Dave Andrews, which summarize the key events and happenings on the mission. These provide an historical overview of all the key events that occurred during the mission. These are followed by all the Express articles that dealt with any aspect of the spacecraft operation, calibration, or analysis techniques.

At the HEASARC, the EXOSAT database is available for online access (please contact HEASARC Feedback for further details). The database contains summary data products for each observation, plus catalogs of detected sources and an observation log. The HEASARC is currently converting these data products to FITS format. The HEASARC has also recently obtained a complete copy of the Final Observation Tapes, FOTs, from ESAs Space Information System, ESIS. These raw data will, before the end of 1993, be made available through the HEASARCs anonymous FTP account. Over the coming years, the HEASARC plans to convert these data to FITS format. Articles about these activities will be published in the HEASARCs Legacy journal (which was based upon the EXOSAT Express model). The following EXOSAT calibration articles have appeared in Legacy as follows:

Positional Accuracy of the EXOSAT CMA(LE1) -     Legacy vol 1, p 33
        J. Osborne

New EXOSAT ME Argon Calibration -               Legacy vol 1, p 53
        F. Haberl

A Recalibration of the EXOSAT Diffraction       Legacy vol 2, p 37
        Gratings - F. Paerels et al

                                                        N.E White

Volume 1: STATUS OF PROGRAMME AS OF 1ST OCTOBER 1983

EXOSAT was launched on May 26th. Switch-on of the instruments and initial checkout was completed by June 18th. Instrument calibration and performance verification was largely completed by August 15th. The observing programme proper, i.e. the undertaking of observations selected by the COPS (Committee for Observation Proposal Selection) from the response to AO1 issued in the summer of 1981, was then started.

Hardware

EXOSAT, ESA's first X-ray astronomy satellite, though conceived as early as 1969 and selected in 1973, is intended to make a major contribution in furthering this branch of astronomy following, as it does, the pioneering missions of the early '70s and the operation of the Einstein Observatory. In order to do this within various boundary conditions, such as a mass limit on the satellite of 500 kg, the spacecraft and scientific instruments are highly sophisticated and employ advanced technologies, many of them representing the ultimate in the 'state of the art'. Perhaps then it was not entirely inconceivable that there would be some teething problems in the early operational phase, despite the thorough ground testing and calibration of spacecraft and instruments. It is now possible to put these problems into perspective, and hopefully lay to rest some of the rumours and speculation that abounded during the early orbit phase.

Difficulties were experienced with operation of the attitude control system but refined operational procedures have led to a marked reduction in the anomalous behaviour. It is estimated that up to 0.75 kg of control gas out of 14 kg total was wasted through this behaviour. However, assuming that operations continue as of now, there should be no shortage of attitude control gas before orbit decay and satellite re-entry. Targets are acquired to typically 1 arcmin without trimming, and pointing is maintained to within a few arc sec.

Of the two position sensitive proportional counters (one to each telescope), one exhibited anomalous behaviour on initial switch-on and has not yet been reactivated. No clue has emerged as to why this occurred. The second was operated for a considerable period before it too started to behave anomalously. Analysis of the data has indicated a gradual build-up of low energy pulses in the detector. It is possible that the rate of this build-up can be reduced by operating at lower high voltage settings, hence lower gas gain. Reactivation and investigation of this detector at new settings is now planned for early October .

Volume 2: OBSERVATORY STATUS AS OF 31ST DECEMBER 1983

EXOSAT was launched on May 26th. Switch-on of the instruments, initial checkout, instrument calibration and performance verification was largely completed by August 15th. The observing programme proper, i.e. the undertaking of observations selected by the COPS (Committee for Observation Proposal Selection) from the response to AO1 issued in the summer of 1981, was started on August 16th and is now about two-thirds complete.

Hardware

Operational procedures, implemented as solutions to some of the early orbit problems experienced with the attitude control system, have virtually eliminated the anomalous AOCS behaviour previously reported. Two persistent problems remain however, but these have practically no impact on the operations. First, jumps in the S and Z gyro drift biases occur from time to time (always close to perigee) and result, in the case of the Z-gyro, in a spacecraft attitude drift requiring subsequent correction. There is, at present, no explanation for the phenomenon. Secondly, on board computer (OBC) memory parity interrupt errors occur regularly, with definite evidence of increased frequency as a function of time, and appear to be associated with the parity bits (no corrupted data, OBC continues to operate) of memory words physically located on the outside of the memory planes, suggesting an electromagnetic interference cause. A solution to increased frequency, should this become unacceptable, is to refresh (write to memory) from time to time the affected memory locations, presently done only when updated system software is loaded.

Analysis of flight data and ground tests of the flight spare model of the position sensitive detector (PSD) have suggested that the anomalous behaviour of PSD1 could be explained in terms of a gradual build up of a deposit on an internal electrode, eventually producing a discharge characteristic of a single electron spectrum. In orbit evacuated bake-out of PSD1, using the gas system solenoid valves to raise the temperature from typically 25 to 50 degrees C, followed by gas filling and venting have resulted in a successful reactivation of this detector at the minimum anode high voltage (2307V). This detector is now being used for specific observations as selected by COPS.

Volume 3: OBSERVATORY STATUS AS OF 29th FEBRUARY 1984

EXOSAT has been operational for just over 9 months and the programme of observations approved by the Committee on Observation Proposal Selection (COPS) in response to AO-1 is about 80% complete (see p. 11-21).

1. Hardware

1.1 Spacecraft

Z- and X-gyro drifts have been increasing monotonically for some months and are now at 0.60/hr and 0.40/hr respectively, such that the accuracy of the drift compensation applied by the attitude and orbit control electronics (AOCE) is reduced. A drift of lower magnitude is observed with the Y-gyro. Two of the four IRAS gyros (similar design) showed a high drift rate (>0.5°/hr) after one year in orbit, although 12 such gyros have been in continuous ground operation for 9 years and only one is presently outside the specification limit. If the drift rates continue to increase, there will be a gradual degradation in open-loop performance i.e. during manoeuvres or perigee passes (<50.000 km) when no update from the star tracker of the inertial reference is supplied to the gyros, resulting in larger pointing offsets after manoeuvres or on re-acquisition following perigee and, for observation periods (closed loop), a longer settling time for stable pointing. Note that, at present, stable pointing (<5 arcsec) is achieved within approximately 20 mins. of the 'closed loop' configuration and that the ±1 arcsec limit-cycling maintains inertial pointing to within ± 3 arcsec in Y and Z with a bias offset of +5 arcsec in Y (introduced to compensate for an AOCE operational anomaly). No increase is expected in fuel consumption as a result of increasing gyro drift rate.

On board computer (OBC) memory parity error interrupts and erroneous CPU overload errors continue to occur with little impact on operations. The frequency of memory parity interrupts has reduced from 1 or 2/day to 1/several days, apparently after refreshing (write to memory) the entire memory area and re-loading a new version of the in-flight software although the observed gradual reduction is not consistent with the expected behaviour.

There is no explanation for the CPU overload errors which occur about once per 4 to 6 weeks; diagnostic software has been incorporated to aid further investigation.

1.2 Payload

1.2.1 LE1

PSD1 is used routinely (Section 2) for those AO-1 observations approved as relevant PSD targets by the COPS at their meeting on 21.10.83 (Issue 2 p.4). For background limited observations the PSD (operated at its lowest gas gain) is about a factor of 2 less sensitive than the CMA for source detection (details in part 1 of the Observers Guide issued as part of AO-2).

To avoid instabilities in the CMA1 detector/HT supply combination at elevated temperatures, operational procedures have been instituted to maintain a low average detector temperature consistent with carrying out all observations. These procedures include:

• 28V power supply (analogue electronics + HT convertors) switch off for manoeuvres/perigee passes.
• angle 'preference' 130° <= <= 90°
• Observations (particularly co-ordinated) with angle > 130 or < 90 degrees (but consistent with other constraints) are scheduled and interleaved with 'preferred' angle targets.

There is no operation at present of the grating because of the partial jamming of the mechanism.

1.2.2 LE2

Reactivation of the detectors of telescope no. 2 has been attempted from time to time since the malfunctions on 28.10.83 and 11.6.83 respectively of the CMA and PSD. No explanation exists for either failure. For the CMA, ideas of temperature effects and mechanical contact problems have been discussed and the general strategy has been to maintain a cool detector (approx. 13°C) and to exchange CMA/PSD a few times prior to any test. In the case of the PSD, which exhibited anomalous behaviour 10 mins. after the first switch-on, gas flushing and in-orbit 'bake out' (temperature rise from typically 25°C to 50°C) using the gas system solenoid valves have so far proved unsuccessful.

After one abortive test on 29.12.83, a second test following the same procedure on 9.2.84 led to a 'partial reactivation' of CMA2. The detector was operated at its' minimum HT setting of 900V on plate 1 and 2830V on plate 2 for approximately 24 hours. Observations were carried out on 0851+202, OI 417, 4C55.16 (FWHM of LSF for 0851 approx. 25 arcsec) on days 40/41. Activation of CMA2 at its' nominal gain (1150V on plate 1 and 2910V on plate 2) resulted in only 30 minutes of normal operation before the malfunction reappeared. Since then no further operation of CMA2 has been achieved, although some confidence now exists that the detector has not suffered an irreversible failure and attempts to re-activate it will be continued.

1.2.3 ME

All sixteen proportional counters operate satisfactorily.

A component failure on 26.11.83 in the drive mechanism electronics for detectors E,F,G,H (lower half-experiment) such that the motor disable pulse is not automatically issued after the requisite movement, has led to a re-appraisal of the simultaneous background monitoring strategy involving no offset of these detectors. If necessary, a procedure can be established to use the OBC time-tagged co-and facility to issue the motor disable command after n seconds of movement. Further information will be given in a forthcoming issue of the Express.

1.2.4 GSPC

The GSPC continues to operate satisfactorily.

From approximately day 360, gain variations (up to 10%) of the GSPC photomultiplier have been observed. These are caused by the systematic 28V line switch-off during maneouvres instituted to maintain CMA1 as cool as possible. The gain will be measured according to the equivalent energy of light emitting diode (LED) test pulses and the CCF updated per observation. PI's with GS prime observations between day 360 and the establishment of the gain monitoring procedure will be informed separately of the estimated variation.

2. Performance and Operations

Tables 1 and 2 give the current performance parameters of the EXOSAT instruments based upon the latest analyses of PV/Cal phase data. Note refinements in the total effective area of the ME, the total effective geometric area of the GSPC and updated figures for PSD1.

PSD1 operations since re-activation total approximately 24 hours. Preliminary analysis of the data suggests that the 'LEP' (low energy pulse) behaviour differs somewhat from that observed prior to the original breakdown on 27.6.83. Initial switch-on 'LEP' count rates are reproducible at approximately 2 c/s, i.e. no increasing trend from one observation to the next, followed by relatively stable operation for 1 to 4 hours, depending apparently on 'age' of gas/whether detector 'baked-out' or not, and then a steady increase to 4/5 c/s over 1/2 hours with fluctuations superimposed. An arbitrary maximum allowable LEP of 5 c/s (below the level observed immediately prior to the breakdown on 27.6.83) has been imposed, such that PSD observations are practically limited to 300 min. at least until the phenomenon is better understood. Detector gain variations are observed consistent with the absence of automatic gain control (AGC) HT compensation for gas density fluctuations as the detector is 'topped up' for window leakage. Further details of the PSD performance will be supplied in the FOT handbook or in forthcoming issues of the EXPRESS.

During the period 1.1.84 to 29.2.84, one sun safety mode trigger on 8/1 and a giant solar 'proton' event on 16/2 disrupted normal operations. Triggering of the sun safety mode was caused by an operational error in the initiation of a fairly uncommon Y-axis only slew, and resulted in the loss of one observation. At 09.07z on 16/2 EXOSAT 'detected' the onset of a solar proton 'event' which reached a maximum particle density of 660 p cm^-2 a^-1 sr^-1 at satellite altitudes about one hour later, followed by an exponential decay over 24 hours and a continuing enhanced (approx. factor 2 higher background in GSPC/ME) activity for about 6 days. Considerable disruption to the planned observation time line between 16/2 and 22/2 occurred, particularly with respect to ME/GSPC operations (ref. AO-1 list p.21).

On 28.2.84, snow on the VILSPA antenna caused a telemetry loss for approximately three hours, (ref. p.44).

AO-1 completion is expected between the middle and end of April, slightly dependent upon re-scheduling of omitted or disrupted observations. Note that a number of A0-1 observations, approximately 55, cannot be performed during the formal AO-1 phase because of coordinated observations after April or violation of the solar constraint, and will be inserted in the appropriate place in the AO-2 time line.

3. On-board Software

A new version of the basic in-flight (BIF) software has been loaded and is running satisfactorily. BIF24 incorporates the following modifications:

• Application program slots: 5 large, 1 small
• Program ROLLER core resident/program SMC interchangeable.

BIF25 is currently under test and will ease operations by allowing all commonly used spacecraft sub-system programs i.e. ROLLER1 SMC and SAOM to be core-resident. This is achieved by deleting the program BURST, which is not useful because of background variability and solar activity, and re-configuring the program BOM as an interchangeable program. A modification has been incorporated in the ME program MHER5 (Issue 2. p.6) to allow separate accumulation of Argon and Xenon energy histograms and hence independent variation of the collection times. This feature is available on board but requires some ground support software change.

Design of a new ME application program MHER6 is complete, implementation together with the appropriate ground support software for both MHER5 and MHER6 is scheduled prior to the beginning of AO-2. MHER6 allows intensity profile accumulations for offset and on-axis experiment halves within one or two separate energy ranges, E1 < E < E2 and/or E3 < E < E4 with time resolutions < 1 SWC(32 msec). Full information will be available in Part III of the Observatory Users' Guide.

4. Observation Output

Final Observation Tape (FOT) production has been transferred from a Honeywell to a Siemens computer, increasing both speed and efficiency of production. While some delays occurred with the software conversion and scheduling organisation, current FOT production is at day 23 for LE1/GSPC and day 19 for ME. Some delays occurred in the production of GSPC/ME FOT's because of the PM gain variation monitoring (ref. Section 1.2.4) and errors in filing data from the new MHER5 mode. Despatch of tapes to PI's generally occurs within 1 week of production and we are looking at methods of reducing the gap between tape available at ESOC and in the hands of the PI.

Some observers have requested that the satellite pointing data during slews be included on the FOT to maximise the use of the ME/GSPC slew data, generated mainly for background correlation. Several bright sources have been 'scanned', some known and others unexpected. A method of obtaining the pointing history during slews is being investigated and will be implemented on the FOT if feasible. Prior to this, slew history can usually be reconstructed for specific requests from the manoeuvre history file data.

As expected, initial FOT production back-logs have now more or less migrated through the Observatory system to appear as a back-log in automatic analysis production further complicated by problems of ME energy channel conversion and resolution functions and the general organisation of the AUTO production system. Current production is at day 285 for LE1/LE2/GSPC, and day 262 for the ME with some gaps because of tape reading problems.

Despatch of the auto output to PI's generally occurs about 2-3 weeks after production. It is stressed that the raison d'etre of the automatic analysis is to provide the observer with an overview of the results of the observation (note the plans for interactive analysis) and that the output is only 'scanned' by the duty scientists for obvious faults. It is neither practical nor functionally necessary to subject the auto output to a rigorous scientific examination and observers are particularly urged to heed the various warnings in both output and documentation. Note that for the ME spectral fitting is not presently incorporated in the automatic analysis software.

In the LE automatic analysis an incorrect bin size factor for full resolution images has produced source count rate underestimations which are a few percent or zero for faint sources (20-30 photons) and up to a factor of 5 for sources with several hundred photons. Further details will be given in the next EXPRESS and PI's whose prime LE observations are affected will be informed directly.

Updating and improvement of calibration data has continued since the launch. The calibration histories will be distributed on tape during March to all institutes, and should be used instead of the calibration data on the original AO-1 FOT's. Orbit history data for AO-1 will be included on the same tape (because of errors in the orbit data on some early FOT's)

Documentation for the re-issue of the Final Observation Tape Handbook (FOTH) is in the final stage of preparation and the complete FOTH should be available during March. Included in the FOTH are notes on the LE and GSPC automatic analysis, which have been distributed separately to all AO-1 PI's (NB. one copy only to the hardware institutes). Equivalent notes on the current ME automatic analysis are reproduced on p.31.

5. Future Plans

Efforts to schedule co-ordinated observations and the rescheduling of those observations not correctly carried out during the high solar activity from 16/2 to 22/2 have increased considerably the quantity of near real time mission planning, nevertheless the time line has been published marginally 4 weeks in advance, viz. March 15th on 15/2 and March 31st on 27/2. The complete AO-1 time line should be available by about mid-March. A number of improvements have been incorporated in the mission planning software and operations, and it should be possible to maintain such an advance time line for AO-2. It must be emphasised that the mission planner or any other observatory team member concerned with the advance planning of the observations will accept inputs from the PI only (or a deputy nominated in writing by the PI). Some recent examples of multiplexing in the input stream have caused a high second derivative of the 'planning function' which can only result in delays in the issue of the time line.

Documentation for AO-2 was issued at the end of January. A response is required at the latest on 9th of March leading to evaluation and selection of the AO-2 programme by mid-April.

The review of EXOSAT scientific operations at ESOC since launch has been completed and it is intended to implement a number of changes. These changes should streamline the operations, improve the working conditions within the observatory, generate a better scientific environment and improve the service to the scientific community.

With the addition of a further HP processor, more disc storage capacity, extra terminals and by linking to the ESOC mainframe computer for CPU intensive operations, a system will be provided to enable the visiting observer to undertake interactive analysis of his data in the 48 hrs following the observation. The observer will be able to call on the support of the resident astronomers (duty Scientists) to assist him with the analysis.

The observatory team structure will be modified to provide a scientific focal point within the team. Operations in the EXOSAT Dedicated Control Room (real time EX2 system) will be undertaken by more operationally oriented personnel, leaving the resident astronomers more time for work concerned with calibration, auto analysis overview, interactive analysis support and their own research programmes. Extra manpower will be devoted to mission planning while an increase in the number of data aides should speed up the more clerically oriented tasks.

These changes will be implemented over the coming months so that by the first anniversary of routine operations the new system should be functioning fully. A key element in the change is the recruitment of Duty Scientists/Observatory Controllers and attention is drawn to the 'advertisement' on page 40. In the next issue of the EXPRESS a fuller description of the observatory computational facilities will be given.

Volume 4: OBSERVATORY STATUS AS OF 30th APRIL 1984

EXOSAT operations are approaching a triple milestone of '1 year in orbit, 1000 attitude manoeuvres and orbit number 100'. The programme of observations approved by the Committee on Observation Proposal Selection (COPS) in response to AO-1 is roughly 90% complete (about 700 observations), with a further 70 scheduled in the period up to 29/5. Subject to COPS approval the AO-2 programme should start on 30/5 and will include a small number of outstanding AO-1 observations.

1. Hardware

1.1 Spacecraft

Reference is made to the article on p.30 for information on the spacecraft hardware status. Note that the only combination of Reaction Control Equipment (RCE) and Attitude and Orbit Control Electronics (AOCE) units which has so far proved problem free is AOCE2/RCE1.

Erroneous OBC CPU overload errors, which normally have little impact on operations, have on two occasions in the period 1/3/84 - 30/4/84 resulted in the planned halt of the processor followed by a failure to re-initialize one of the spacecraft application programs and hence a re-load of the complete software system (duration approx. 90 minutes). For further operations a modification has been introduced to overcome the initialisation problem and avoid the necessity of re-loading. Additional testing and analysis is required to determine the origin of these spurious errors.

1.2 Payload

1.2.1 LE1

PSD1 operations have been suspended following a degradation in the behaviour of the 'LEP' (low energy pulses) first reported in Express No. 3 p.5. Seven PSD1 observations were carried out in the period 1/3/84 to 30/4/84, viz. NGC 5548 day 62, Her X-1 day 64, CRAB day 73, IC 443 day 78,IC 443 day 80, Zeta Pup day 81 and GKP SNR day 114, giving an integrated on-time since reactivation of about 30 hours, but showing a progressive decrease in the times to LEP instability and increase to the arbitrary safety limit (ref. section 2 p.4) of 6 c/s.

Operational procedures continue to ensure satisfactory performance of CMA1. Note that observations are scheduled preferentially for solar aspect () angles between 90° and 130° but some degree of latitude is possible to accommodate coordinated or otherwise constrained observations.

There is no operation of the grating at present because of the partial jamming of the mechanism.

1.2.3 LE2

All attempts (5) to re-activate CMA2 following the 'partial' operation reported in Express No. 3 p.4 have proved unsuccessful. Although no explanation exists for the failure, ideas of temperature effects and mechanical contact problems have led to a strategy of maintaining a cool detector (28V power supply line switched off) and operating the detector exchange mechanism from time to time, so far to no avail.

PSD2 has remained non-operational.

As a functional test, the LE2 grating was successfully moved into and out of the telescope FOV on 20/3/84.

1.2.3 ME

All sixteen proportional counters continue to operate satisfactorily.

Analysis of data from a repeat calibration observation of the CRAB NEBULA on day 85 (note that the day 73 observation was affected by a high background rate as a precursor to the solar storm on day 74) suggests that slight gain increases (<~1-2%) of the Argon detectors have occurred since the PV/Cal phase observation on day 275 (Xenon data is still in the process of analysis). Leakage of gas from the detector is a possible explanation; however this rate should have no significant detrimental effect on the instrument performance during the mission lifetime. Additional analysis is in progress and an electronic gain adjustment may be incorporated. P.I's should at present assume a linear gain change throughout the period between the two CRAB calibrations. Further information will be provided in an update to the FOT Handbook and/or in a forthcoming issue of the Express. Regular calibrations on the CRAB will be instituted.

1.2.4 GSPC

The GSPC hardware status remains stable and the detector continues to operate satisfactorily.

A note on the gain variations introduced by the procedure of switching off the 28V power supply line to LE1 in order to maintain a low detector temperature is given on p.44.

2. Performance and Operations

Tables 1 and 2 give the current performance parameters of the EXOSAT instruments based on the latest analyses of PV/Cal phase data.

PSD1 operations since re-activation total approximately 30 hours, however each successive observation has shown a decrease in the useful observing time as shown below:

Note that the first few observations (up to about 12 hours of total 'ON' time) were completed normally after the allotted observing time, whereas all recent observations were terminated when the LEP count rates increased above an arbitrarily defined 'safe' limit of 6 c/s. Similar procedures, clearly unsuccessful, of detector 'bake out' using the gas system solenoid valves to raise the temperature (typically by 25°C) together with a 'hot gas fill and drain have been carried out between each observation with the aim of maximising the period of stable LEP (2 c/s) and the time to increase to 6 c/s. From day 114 (GKP SNR) to the next PSD observation scheduled on day 138 (NGC 6964) the detector will remain drained in order to eliminate any effect of gas decomposition in the orbital radiation environment.

Observatory operations were suspended on several occasions, during the period 1.3.84 - 30.4.84. Two solar 'proton' events caused considerable disruption to the planned time line. The first occurred on 14.3.84, 28 days ± a few hours after the event on 16.2.84 reported in Express Issue 3 p.5, and reached a maximum particle density (>10 MeV) at satellite altitudes of 100 cm^-2 s^-1 sr^-1 i.e. a factor 6 on the earlier activity from the same after the maximum. However, in comparison to this event and Its predecessor, EXOSAT detected on 25/4 the onset of a huge solar 'proton' event, which reached a maximum particle density (>10MeV) of 2500 cm^-2 s^-1 sr^-1 approximately one day later and prevented normal operations for a period of 8 days. One sun safety mode trigger (see p.34) and the erroneous OBC CPU overload errors (section 1.1) caused additional loss of observing time.

A successful test of the orbit control system was carried out on 17/4 (ref. p.33). Certain restrictions will nevertheless apply for occultation planning because of AOCS operational problems with the use of Reaction Control Equipment (RCE) unit 2 which indicate that only RCE1 in conjunction with AOCE2 is problem-free thus limiting the orbit control to one thruster.

To avoid unnecessarily high integrated dosage rates in a limited number of channels of CMA1, the mission planning software now 'randomizes' the position of point sources in the FOV of LEl within a ± 30 arc sec. square of the normal pointing.

3. On-Board Software

Operational efficiency has been improved by the incorporation of a new Basic-In-Flight Software system (BIF 25) which allows all commonly used spacecraft sub-system programs to be core resident thus providing the maximum 6 slots for payload data processing programs. A further system update (BIF 26) will contain more powerful diagnostic facilities to aid the investigation of spurious CPU overload errors (section 1.1).

Difficulties have been experienced with the operation of two new application programs (MHER5/MHER6: Express Issue 3 p6), involving triggering of telemetry overload errors under certain combinations of workspace parameter definition and in conjunction with other application programs. Further modifications incorporated in the software have solved the problem, although precise limitations on time resolutions and telemetry load can only be defined after sufficient operational use.

Since the launch of EXOSAT, 6 system updates (BIF 20 - BIF 25), 10 payload application program modifications (update or new software) and 3 spacecraft application program modifications have been carried out in response to changing requirements as a result of the observational data analysis and operational experience. It is anticipated that this level of activity will shortly decrease.

4. Observation Output

Production of Final Observation Tapes (FOT's) is now routine and the goal of approximately 20 days between observation and tape availability at ESOC has been reached. Because of the iterative process of filing, analysis and checking the data generated b,y modified OBC application programs, production of certain FOT 5 containing eg. MPULS2, MHER5 (modified version, Express No. 3 p6) data does incur a delay with respect to the above 20 days.

Please note that P.I's are kindly requested to disregard the statement in AO-2 Section 2 concerning copying and return of FOT's. ESOC do not require return of the original tapes.

In response to requests from some users, the satellite attitude during slew manoeuvres will be made available both in near real time to support search of the ME/GSPC slew data for transient sources and unexpected variability of known sources and on the FOT to allow observers to reconstruct the pointing history throughout the manoeuvre. Slew auxillary data will be generated immediately after a manoeuvre and an algorithm will be provided to determine the satellite attitude using an interactive facility operated from a terminal in the DCR. This auxiliary data will be stored on the FOT and a similar algorithm will be supplied to enable users to extract the attitude data using the relevant housekeeping telemetry format. Accuracy will be <~ 5 arc minutes and the facility should be available within 2 to 3 months. Details will be provided in an update to the FOT Handbook and/or in a forthcoming issue of the Express. Note that it is not feasible to re-create the pointing history systematically for manoeuvres carried out prior to the above implementation but an analysis of archive tapes should in principle give the attitude for specific requests only.

Our general policy now is to perform the automatic analysis in three priority categories viz:

1. Backlog from day 307 of observations for non-hardware group P.I's (the hardware groups being Leiden, Leicester, Milan/Palermo, MSSL, Munich, Utrecht and SSD)
2. All current observations
3. Remaining backlog.

Note, however, that software modifications are required before data from new OBC modes such as MHER5, MHER6, GDIR, MPULS(1) and MPULS(2) or from PSD1 observations can be analysed, implying that not all category (b) analyses can proceed immediately.

5. Future Plans

AO-1 timelines have been published approximately one month in advance up to the 'final' schedule covering the period 1/5-29/5. Note that Anne Fahey has been assigned to mission planning duties as a full time task and has been practically responsible for the preparation of all recent time lines. Further changes in the division of responsibilities within the Observatory Team are expected and will be described in forthcoming issues of the Express. Considerable re-scheduling of omitted or disrupted observations (eg. solar storms on 14/3 and 25/4) plus inclusion of previously constrained AO-1 targets or coordinated observations have led to a solid timeline to the end of May. A number of AO-1 observations will be outstanding as of 29/5.

AO-2 proposal selection was undertaken by the COPS at their meeting from 25/4 to 27/4 in ESTEC. Note that the total AO-2 response represents an over-subscription for observing time of a factor of 9 for the six month observation period. Successful proposers will be informed within the next week or two.

Procurement of computer hardware and the definition of detailed requirements for the interactive analysis system to be provided to enable visiting observers to undertake a fuller analysis of their data immediately post-observation is in progress. It is planned to give a description of the present and proposed observatory computational facilities in the next issue of the Express. Our goal is to have a working analysis system available for observers' use by the end of September 1984.

Volume 5: OBSERVATORY STATUS AS OF 30TH JUNE 1984

In response to the second announcement of opportunity to participate in the observation programme of EXOSAT 560 proposals were submitted for the 6 month AO-2 observation period, a ninefold oversubscription in observation time. Of the 560 proposals submitted, 191 comprising 450 separate observations were selected as the approved A0-2 programme by the Committee on Observation Proposal Selection (COPS) at their meetings on 25-27th April and 4-5th May.

AO-2 observations have been carried out from the middle of June, with a number of AO-1 targets interleaved and it is expected that the programme will terminate around mid-December, again with outstanding or incomplete observations because of later coordination or solar constraint.

A review took place at ESTEC on 4.6.84 of the failure/performance malfunction history and current procedures for reactivation of specific elements of the LE telescopes, namely PSDl, CMA1, PSD2, CMA2 and the LEl grating. No firm conclusions were drawn regarding the reason(s) for any of the anomalies observed. Comments on the operational procedures adopted in the attempts to re-activate the various elements of the telescopes were principally in the direction of extending the existing procedures. These are noted in section 1.2.

1. Hardware

1.1 Spacecraft

All elements of the spacecraft hardware continue to operate as described in previous issues of the EXPRESS. A number of points are noteworthy.

Spurious OBC memory parity error interrupts occur once or twice per week with no impact on operations; however on one occasion during May the frequency suddenly increased to approximately one per hour, remained at the 'higher' level for 3 days and then suddenly decreased to the 'nominal' rate. Although a small solar proton event ( >10 particles cm^-2 sr^-1 s^-1 above 10 MeV) was in progress at the time, it seems unlikely that this is a real correlation since EXOSAT has experienced much larger solar storms (ref. Express Nos. 3 & 4). Note that parity errors associated with a specific memory location have been eliminated by appropriate re-arrangement of the on-board software. No CPU overload errors have occurred since the commissioning of a new version of the basic in-flight software (BIF26), although the frequency has in any case been low.

In the first year of operations the available Solar Array power has decreased by 5.5% from 288.5 watts to 272.5 watts. Note that the level at which battery discharge would occur with all elements of the satellite functioning normally is 227 watts such that if the present rate of degradation is maintained (ref Express No. 4 p.40) sufficient power should be available to ensure no interruption of operations prior to re-entry into the dense atmosphere.

Between 14th and 28th June around the summer solstice, a scan of the galactic centre (AO-1 observation, P.I.: Turner) was successfully carried out, and involved 8 pairs of long slew sequences (90°) at the slowest slew rate (43°/hr) covering the region of sky between b"= ± 5° and l"= 0 ± 90°.

1.2 Payload

1.2.1 LEl

Although the procedure of 'bake-out' of PSD1 using the gas system solenoid valves to raise the detector temperature typically from 25°C to 50°C, with a subsequent gas fill and vent at the high temperature, has not maintained the performance of PSD1 achieved immediately after re-activation (ref. Express No.3 p.5), the recommendation of the ESTEC review meeting was to continue this procedure with increased bake-out times (>5 hours) and as a parallel activity, analyse in detail the data from 30 hours of PSD1 observations. These activities will commence immediately.

Systematic operational control of the temperature of CMA1 (restricted angle, 28V power supply to the analogue electronics switched off for manoeuvres, perigee passes and during solar storms) has considerably increased the stability of the detector/HT supply combination (ref. Express No. 3 p.3). This practice will be continued.

No operation of the grating is planned because of the partial jamming of the mechanism; although the grating is clearly removed from the X-ray beam, the available evidence suggests that it Is not fully In the 'out' position and that the mechanism remains jammed with the grating at a few degrees to the optical axis.

1.2.2 LE2

There has been no operation of LE2. Recommendations of the review meeting are as follows:

PSD2

HT switch on at the lowest possible residual detector gas pressure to determine whether the discharge is internal or external to the detector. Depending on this result further procedures will be initiated.

CMA2

Regular temperature cycling (28V power supply line on/off per 1/2 days), HT switch on and variation of the second plate/anode voltage with the rear front plate voltage at nominal.

Should CMA2 prove operational at any time a number of prime grating observations will be undertaken immediately irrespective of any planned coordinated or phase dependent observations. This will be followed by calibration measurements prior to general use of the telescope.

1.2.3 ME

All sixteen proportional counters continue to operate satisfactorily.

Attention is drawn to the discussion of ME background subtraction on p.48, and the intention to re-introduce the procedure of swapping the array halves between source pointing and background monitoring on specific targets for which it is judged that slew data alone cannot provide the optimum background subtraction. Note that because of a component failure in the drive mechanism electronics of detectors E,F,G and H (Express No. 3, p.4), the OBC time tag command facility is used to initiate and stop movement of the detector quadrants E/F and G/H.

1.2.4 GSPC

The GSPC continues to operate satisfactorily.

Methods of monitoring the gain variation of the GSPC caused by the systematic switch-off of the 28V power supply line to LEl have been outlined in the previous issue of the Express (No. 4, p.44). Further analysis of the data has shown that the LED stimulation cannot easily give an absolute gain calibration (conversion from LED ch. to keV is non-deterministic) and observers are, therefore, recommended to make use of the 4.7 keV feature (and the Bi lines) for absolute gain calibration together with the LED channel for relative changes. There is, however, preliminary evidence for evolution of the Bi lines and further information on the GSPC gain calibration will be described in the next issue-of the Express.

Observatory operations have continued without interruption during the period 30/4 to 30/6. A minor solar proton event occurred on 22/5 (28 days ± few hours after the event of 25/4, Express No.4, p.5) producing however a maximum particle density insufficient to prevent EXOSAT operations. In this respect it is a pleasure to acknowledge the assistance of Solar Maximum Mission personnel who have provided us very quickly on numerous occasions with detailed information on the progress of solar activity. Indeed, SMM particle densities have proved the most appropriate criterion for re-commencement of operations during the decay phase of solar activity.

Some confusion appears to exist regarding the precise scheme of observation scheduling, liaison with the P.I. and responsibility within the Observatory Team at various stages of the procedure. The following is the typical 'flow':

• Preparation of the time line, generally 2-4 weeks in advance, and notification of the observation details to the P.I. (only) is the prime responsibility of the mission planning group (Anne Fahey/Paul Barr). Celestial constraints are checked at this stage.
• On a rota basis, duty scientists receive the time line for one orbit, together with configuration details from the proposal and notification of co-ordinated work (if available). Source confusion constraints in the ME are checked, P/L and OBC configurations determined subject to confirmation/discussion with the P.I and the configuration is considered frozen once agreement is reached. Note that P.I's should at this stage inform the duty scientist concerned, with copy to Paul Barr, of coordinated observation details which have not been discussed with the mission planning group during the time-line preparation phase.
• Observatory orbit timelines form an input to the manoeuvre support software of ESOC's orbit and attitude division. Celestial constraints are rechecked and the slew strategy between targets is determined.
• The duty scientist responsible for preparation of the orbit will normally not carry out the operations at the time of execution of the observation. Once the configuration is frozen, any 'last minute communications (which are not actively encouraged) from the P.I. on any aspect of the observation should therefore be addressed to the 'EXOSAT Duty Scientist' (copy Paul Barr) and marked 'URGENT'.
• During the observation the duty scientist has the authority to change the 'frozen' configuration, although certain modifications such as SDS changes, program re-loads, odd workspace parameters are non-standard and should be predetermined as far as possible. Note also that some OBC requirements, such as unusual sample rates which involve a new SDS, imply an interface with other Observatory Team or Spacecraft Operations Team personnel and must therefore be planned in advance. In the absence of the P.I. during the observation, telephone information, EX2 output etc. will be gladly supplied by the duty scientist in charge of the observation, but it is emphasised that this is best requested in real time.
• P.I's are notified of the cancellation of their observation or of any deviation from the planned execution by the mission planning group or the Observatory Manager.
• All information on FOT or AUTO Production and despatch is communicated to P.I's by the Data Assistants.

Note that once the Observatory Controllers assume responsibility for the real time operations, some changes will occur in the above procedures. These will be reported in a future issue of the Express.

Considerable effort in the organisation, running, checking and dispatch of automatic analyses has reduced the backlog reported in Express No.4, p.7. Equivalent production is now at day 75 (1984), although because of the priority given to the analysis of observations of non-hardware group P.I's, analyses of a number of observations prior to day 75 are outstanding. Spectral fitting, which is now included In the ME automatic analysis software is scheduled whenever the source count rate is sufficiently high to provide reasonable model fits.

Chapters 6-8 of the FOT Handbook, together with corrections to Rev.1 (chapters 1-5) will be despatched to all EXOSAT users as soon as possible. The Observatory Team is well aware of the difficulties experienced by some users because of lack of the appropriate documentation, production of which was perhaps rightly not afforded the highest priority in the early phase of the mission, and we will strive to improve this situation.

5. Future Plans

AO-1/AO-2 time lines continue to be published between 3 and 4 weeks In advance, although the fairly late availability of the approved AO-2 programme with a series of early co-ordinated observations has meant that some P.I's with observations during May and June may not have received adequate warning of the schedule. Attention is drawn to the notes on p.6 describing P.I. interaction with the Observatory Team in the steps from observation schedule to output despatch.

ESA Intends to publish the third announcement of opportunity to participate In the EXOSAT observation programme (AO-3) during August, with expected response in September and selection of the AO-3 programme in November. AO-3 will have a duration of 9 to 12 months.

Volume 6: OBSERVATORY STATUS AS OF 31.8.84

Now that the majority of the Observatory Controllers recruited to execute the real time operation of the Observatory have taken up their appointments, the Observatory Team is almost up to full strength. The new staff are gradually assuming responsibility for the operations after an initial period of training in the DCR, and it is expected that this transition will be complete by mid-October. P.I's and observers are reminded that the Duty Scientists will continue to be responsible for all liaison work to establish P/L and OBC configuration for the observations on an orbit-by-orbit basis and that the Observatory Controllers will have full responsibility to carry out the planned programme of observations. At the same time, the Duty Scientists will be available or on call on a 24 hour per day rota, to support the real time operations, principally for 'science' decisions, emergencies and general assistance to the Observatory Controllers.

Full details of the Interactive Analysis System provided by the Observatory Team to enable users to analyse their data at the Observatory will be given in a future article in the Express. However, as we intend to 'go on the air' by the early autumn, albeit in a preliminary fashion, a few general remarks are made here:

• the system will support near real time analysis of data from the current observation as well as analysis of the FOT's, if available, from prior observations.
• non-hardware group P.I's and the Observatory Team will have priority use of the system over 'hardware' group P.I's (the hardware groups being Leiden, Leicester, Milan/Palermo, MSSL, Munich, Utrecht and SSD/ESTEC).
• booking of the system will be organized within the Observatory Team by Susanne Ernst.
• the system will be available 24 hours per day; support and advice to users will be provided by the Duty Scientists on a rota basis on call or at ESOC between the hours of 08.00. and 23.00 everyday. On-call support will mean telephone contact at home or by Eurosignal bleeper, and will be requested at the discretion of the Observatory Controller. Note that in principle, the Duty Scientist supporting the Interactive Analysis System will also support the real time operations and that on-call support according to speciality cannot be provided. Such support must be arranged on an ad hoc basis with the relevant Observatory Team member.

Recent problems concerning mission planning serve to remind P.I's that ephemerides should be quoted in the standard accepted form, that is the time for phase zero, and that RA, DEC coordinates are for Epoch 1950. This will always be assumed for communications from P.I's to the Observatory Team and vice versa, unless specifically stated otherwise.

Attempts to re-activate the malfunctioning elements of the LE telescopes (PSD1, PSD2 and CMA2) have been carried out in the previous two months, all to no avail (ref. Section 1.2). These procedures will be continued on a low priority basis although any success will, of course, be reported!

1. Hardware

1.1 Spacecraft

There is no change in the status of the spacecraft hardware. Spurious OBC CPU overload errors and gyro drift jumps continue to cause minor inconvenience but measures are being adopted to overcome this (ref. Section 3).

1.2 Payload

1.2.1 LEl

CMAl continues to function as described previously; attempts to reactivate PSD1 by increased 'bake-out' durations (using the gas filling system solenoid valves to increase the detector temperature) have proved unsuccessful in that the onset of increasing LEP occurs soon after HT switch-on, and reaches the arbitrary safety limit within a few minutes. Further analysis of the existing data will be carried out with, eventually, a possible relaxation of the safety limit.

1.2.2 LE2

Temperature cycling has had no effect on the status of CMA2; PSD2 has been operated at zero residual gas pressure to determine whether the observed anomalous behaviour is caused externally or internally. Preliminary analysis suggests the latter. Further tests and analyses are in progress.

1.2.3 ME

The ME detector experiment status is unchanged; however attention is drawn to the article on p.25 describing the ME gain calibration and the observation of gain variation, particularly in Argon detectors D + G, consistent with minor gas leakage. Extrapolation of the present trends to the end of the mission suggest that continuous operation of all detectors can be maintained. Regular adjustment of the electronic amplifier gains to compensate for detector gain changes has been instituted.

1.2.4 GSPC

The GSPC status is unchanged, but note the article on p.18 describing the instrument calibration.

2. Performance and Operation.

Tables 1 and 2 on p. 8/9 give the current performance parameters of the EXOSAT instruments.

Station computer problems at VILSPA on 3/8 caused a loss of approximately 13 hours of observation time; several periods of loss of telecommanding lasting about 30 minutes have occurred since then and are believed to result from a power supply instability.

Since regular swapping of the ME array halves has been re-introduced, the logical sequence of operation has been modified such that at no time is the source not under observation, i.e. detector HT's are switched off only for those detectors currently being moved. This avoids 'missing' flaring/bursts from the source. Data produced during the period of swapping, although correctly archived, is not at present written to the FOT because the observation definition precludes periods of less than five minutes between telecommand status changes. A modification to the ME FOT production software will be incorporated as soon as possible. FOT's will be regenerated on request for the intervening period (1984 Days 185- (?) 275), if it is suspected that data from a source flare or burst may have been omitted on the original FOT.

Recent long observations with the CMA (AR LAC, LE Background Cal.) of 48 hrs have demonstrated that the improved temperature stability of the HT/detector combination has been maintained and it is clear that there is no immediately obvious degradation in performance if the Al line Is not switched off during periods of the order of one or two days. For GSPC prime observations, when gain stability is essential, all Al lines (hence GS HT) will therefore remain on during the manoeuvre to the source, provided that total LE1 Al line on time (previous observation, manoeuvre, -GSPC prime observation) does not exceed 24 hours.

Regular calibrations, with the Fe⁵⁵ filter wheel source, of the CMA low gain area and the new point source image area (ref. p.6) will be instituted.

3. On-Board Software

With the additional diagnostic facilities referred to in Express Issue 5, analysis of data from one occurrence of the spurious CPU overload error suggests a possible intermittent malfunction of the OBC instruction 'DPI' - Disable Interrupt Processing. A new application program to monitor the performance of DPI has been designed and will run continuously as a low level background activity for several weeks to aid diagnosis of the error.

Because the Basic In Flight software requires all 16k of available OBC memory and malfunction of one 8k 'module' is considered the most likely failure mode, a special 8K version of the software was designed, developed and commissioned prior to launch in order to avoid an unacceptably long hiatus in operations in the event of a failure. This is the only 'design-for-failure' situation. A major update to this 8k system to reflect some of the modifications introduced since launch (BIF22 - BIF26) is in progress. Naturally, only a subset of the facilities available in the current BIF26 can be provided in the 8K version.

Modifications will be carried out to the two ME modes, MPULS and MPULS2 to incorporate a 128 channel energy histogram with 16 phase bins mode and for MPULS2, Argon data only in the intensity samples.

Note that prior to release for general use, the revised ME Direct mode application program first reported in Express Issue 5 (p.6) requires a check using data from an observation of the CRAB which is scheduled towards the end of September. Details of this mode will be given in the next issue of the Express.

Gyro drift jumps during perigee continue to cause problems on reacquisition. An application program will be designed to monitor and correct for such jumps.

No other OBC software changes are planned and, once the above activities are complete towards the end of October, the support for OBC software will enter a maintenance phase until the end of the mission. Priority will, of course, be given to failure/malfunction - driven modifications.

4. Observation Output

Automatic analysis production is currently at day 158 (equivalent, some pre- day 158 observations outstanding) representing a delay of approximately 70 days between observation and analysis, although most non-hardware groups receive their analysis output considerably quicker because of the priority of running the analysis. Attention is drawn to the chronological list of improvements in the ME automatic analysis software (p.35).

According to the original specification of the FOT, the auxiliary data contain a set of information relating to a period of stable pointing. Recently, with the incorporation of slew attitude data into auxiliary data, this situation has changed (details in the new FOT Handbook). To ensure that slew attitude data is correctly aligned with slew observation data and at the same time, correct other outstanding anomalies, a redefinition of the procedure for incorporating auxiliary data onto the FOT is in progress.

In the meantime, a number of anomalies concerning the merging of auxiliary data onto FOT's are hereby summarised:

• the start and end times of a stable pointing (given in auxiliary data) are not always correctly correlated with the start and end times of observations on the FOT. The times themselves are correct; occasionally, however, the A4 data for the following pointing is wrongly merged with the telemetry for the current pointing.
• since June 11th 1984, FOT's contain slew attitudes for the slew following a pointing, but not for the preceeding slew.
• if an attitude trim manoeuvre is made during an observation, there will, correspondingly, be more than one set of A4 data. The exact time of the trim should be deduced from the start and end times of each auxiliary data set. It should not be deduced from the information called "guide star coordinates" contained in the observation directory, because this will lead to slightly incorrect values.
• if an attitude slew is made during an observation, there will be more than one set of A4 data, and in this case the manoeuvre times can be obtained either from the auxiliary data or from inspection of the "slew busy" flag in the observation directory.

Calibration data on FOT's for all experiments has been updated and refined during August (or early September, in some cases). and should be used for -the analysis of all FOT's. The new data sets can be identified by the "time of last update of calibration history" in record 1, which should be August or September 1984, eg. as described in the article on p.25. FOT's generated at ESOC from early September will contain updated CCF's.

5. Future Plans

As from orbit 124, the pointing direction of the spacecraft will be modified such that point sources are imaged by the LE1 CMA within a ± 30" randomization area of X = -60, Y = -18 in order to avoid high dosage in the detector 'low gain' pixels (Express No.5 p.5) . This represents a move towards the centre of the telescope FOV of ~2', whilst maintaining the maximum GSPC/ME response.

The AO-2 programme of observations should terminate during December 1984, although a number of coordinated observations cannot take place until 1985 and other targets only then emerge from the solar constraint. In addition about 19 AO-1 observations are outstanding. Regular calibrations (CMA low gain area, ME Detector gains, GSPC gains etc.) will stretch the nominal AO-2 programme further such that AO-3 observations are expected to begin during January 1985.

ESA is currently specifying the establishment of the EXOSAT data archive including the procedure for access and, once the design is implemented, retrieval will be possible for data designated as public. Details will be given in a future issue of the Express.

A number of long LE prime observations (> 6 units, but excluding deep fields) have shown no significant source detection after 3 units. Therefore In view of the real time response capability of by EXOSAT and the need to maximise the scientific return through significant results, such observations will in future be terminated after 3 units following discussion between the duty scientist/project scientist and, if possible, the P.I. This procedure will be executed sparingly.

Volume 7: OBSERVATORY STATUS AS OF 31.10.84

EXOSAT has been operational for 17 months, approximately one half of its natural lifetime as noted in the foreword, with 677 of the 717 approved AO-1 pointings and 236 of the 470 approved AO-2 pointings completed. AO-2 observations together with the small number of outstanding AO-1 pointings will continue until January/February 1985, although the solar constraint and later co-ordinations preclude several AO-2 targets from the formal observation period.

1. Hardware

There have been no changes in the status of the spacecraft or payload hardware.

Measurements of observed star separations in the star tracker field of view compared with the accurately known catalogue positions have shown that a systematic error in the star tracker calibration reference points (Local Lord Points) exists. Depending on the position of star 1 In the FOV, this can introduce a bias in the pointing direction of the Instruments. Typical values would be 2-3" with worst-case figures (star 1 near edge of FOV - 1.5 deg. from centre) of 8". Measures are in hand to determine and correct this bias, in the meantime Investigators publishing position data should verify with the Observatory Tean the likely maximum error in the result. Note that overall pointing errors are currently quoted as 10", which assumes no contribution from errors in the ST reference point coordinates.

Attempts to re-activate the malfunctioning elements of the LE telescopes have not been successful; PSD1 continues to exhibit rapidly increasing LEP (low energy pulses) very soon after HT switch on, although the initial value is always nominal at ~2 c/s. Temperature cycling and variation of the plate HT's has had no effect on the operation of CMA2. PSD2 has been operated (HT on) a number of times at 0 mb residual pressure with no indication of external breakdown, however regulation pressure operation (1100mb) immediately shows the anomalous behaviour originally observed (high count rates: Emax, Guard and qualified events).

Additional tests will be carried out during the next few weeks on these detectors, namely the activation of a 'glow discharge' In PSD2 (operation in the corona region in an attempt to clean the Internal electrodes - ? remove the ubiquitous speck of dust), continuation of the thermal cycling once per orbit on CMA2 and, somewhat dependent on PSD2 results, a relaxation of the safety criterion for PSD1 operation and/or a 'glow discharge'.

A re-calibration of all ME detectors on the CRAB (ref. p.12) has verified the gain variation first reported in Express Issue 6 (p.4). Argon detectors D+G show the strongest trend (<~12.5% per year) which can, however, at the present rate be adequately compensated throughout the mission by electronic gain adjustment.

2. Performance and Operations

Tables 1 and 2 give the current performance parameters of the EXOSAT instruments.

During the period 1/9 to 31/10, loss of observation time (10 hours) occurred on two separate but related occasions when safety mode of the AOCS was triggered, principally because of procedural errors. Certain additional checks have been introduced to reduce the likelihood of re-occurrence. Observation time (~3 hrs) was also lost when a cable short circuit caused a local fire at Vilspa, however prompt action by the Vilspa staff prevented any serious disruption. Ground system procedural problems have caused errors In the OBC software operation leading to a further loss of observation time of approximately 6 hours.

Available power from the solar array has dropped from an average 280 watts to an average 270 watts over the period July 1983 to July 1984, i.e. a rate of decrease <~ 4% per yr, which would be sufficiently low to maintain operations with an adequate margin for a 4 year mission. Preliminary estimates of fuel consumption (propane) from launch to the present, based on careful logging of a sample of slews, fine pointing and non-routine operations, suggest that the margin may not be sufficient to guarantee operation during the final months of a 4 year mission unless minor measures are taken immediately to conserve fuel (ref. p. 5 under 'Future Plans'). It is emphasised that the figures are provisional and detailed calculations are required on all slews, fine pointing etc., before a definitive statement on fuel usage can be made (an article on this subject will be included in the December issue of the Express).

Since the beginning of the mission, problems have been experienced with Y-axis slews at the highest slew rate (170.7 deg/hr). In some cases, these slews which do not proceed normally (as determined by monitoring of the Y-gyro) are terminated and restarted, leading to unnecessary fuel consumption. Marconi Space Systems (MSS) have performed several hundred slew simulations reproducing this behaviour and have defined a procedure such that interruption of the slew is unnecessary. This is currently being implemented at ESOC. Note, however, that from December 15th, the 170.7 deg/hr slew-rate will not be used (ref. p. 5).

Issue 4 of the Express (p.34) described a problem encountered with the redundant Attitude and Orbit Control Electronics (AOCE 2) which has been in use since August 1983, whereby a spurious value of the gyro drift compensation on the Y-axis can be produced when the star tracker error coordinate changes sign. Offsetting the limit cycle by 5 arcseconds on the Y-axis has 'solved' the problem, but investigations by MSS now show that this anomalous behaviour can be explained by a non-nominal component characteristic which gives rise to an erroneous "read" of the star tracker data.

Also reported in the same issue of the EXPRESS was a phenomenon encountered occasionally during perigee passage when the gyro drift biases (GDB) on the Z- and S-gyros spontaneously change level. An OBC program has been developed to monitor the GDB levels on all gyros and reset them If jumps occur. The program is activated prior to Loss of Signal (LOS) each perigee.

EXOSAT's second eclipse season, covering a period of 12 orbits with eclipses varying in duration from a few minutes to approximately 1 hour, has just ended.

3. On-Board Software

Now that all planned modifications to the on-board software are complete, maintenance of the system, including further development to cover malfunctions or new requirements, will be undertaken within the context of a software maintenance support contract of approximately 0. 2 man-year of effort. Emergency situations will be dealt with on an ad hoc basis. Reference is made to the article on p. 30 describing the revised ME.Direct mode application program.

4. Observation Output

A recent problem with the tape archiving system on the general Multi-Satellite Support System (MSSS), when approximately 7 hours data was irretrievably lost, serves to encourage P.I's to quickly scan through FOT'S on receipt and inform the Observatory promptly of any anomalies. Archiving has been carried out as follows: 1600 bpi archive tapes are produced in real time from the incoming telemetry stream, kept in a cyclic pool for a certain period and then copied to 6250 bpi master archive tapes. Until the present, this system has proved error free but a fault in the copy process discovered after recycling has illustrated an inherent weakness of the system. Steps have been taken to extend the size of the cyclic pool such that tapes have normally been delivered to the P.I. and the auto analysis carried out by the Observatory prior to recycling. Additional security checks have been implemented in the copying process from real time to master archive tapes. P.I's should nevertheless check their FOT as soon as possible after receipt for any gross structural errors, e.g. expected packets all present etc.

The FOT handbook has been completely revised and extended to include a comprehensive chapter on data processing and scientific analysis. In attempting to present as complete a picture as possible we have regrettably required considerably more time than anticipated but the final document is now ready for despatch.

Note that the majority of modifications to the real time graphics software have been completed and, as for the OBC software, maintenance support only will be available in the future.

5. Future Plans

Preliminary estimates of fuel consumption (propane) from launch to the present suggest that sane conservation may be necessary to maintain operations during the last months of a 4 year mission. This presents, of course, a difficult trade-off since the state of the hardware in mid-1987 is not easily predictable, therefore a minimal adjustment to the procedures will be instituted immediately to ensure as far as possible a slightly positive fuel margin extrapolated to May 1987. Refinements to the model and additional calibrations are in progress and may lead to further conservation measures. As from 15.12.84 the following slew strategy will be adopted, pending results of the more detailed calculations:

• no slews at 170.7 deg/hr.
• reduced number of slews at 85.3 deg/hr consistent with maintaining the observation programme.
• Use 42.7 deg/hr as the default slew rate.

On average, manoeuvre durations will be increased by a few minutes as a result of the above.

Now that EXOSAT is half-way through its natural lifetime, a detailed study is in progress to identify an optimum strategy for lifetime extension beyond April 1986 using the hydrazine originally foreseen for orbit modification for occultations. Occultation measurements are not now the prime thrust of EXOSAT although a small number may be carried out during the next year and, in preparation for this, a 'natural' occultation is planned towards the end of this year. Clearly, any use of hydrazine for occultation will reduce the possible lifetime extension and a trade-off must therefore be made. Further details will be given in a future edition of the Express.

P.I's with approved coordinated observations are urged to define as precisely as possible the times of coordination of their EXOSAT observation with other facilities such that the Observatory tean can adjust the EXOSAT schedule when necessary and still maintain the actual coordination.

Volume 8: OBSERVATORY STATUS AS OF 31.12.84

EXOSAT's third programme of observations (AO-3) will start during February/March 1985 and has a planned duration of 12 months. Outstanding AO-2 pointings (~110 of the 470 approved) will be completed in January/February or interleaved with the AO-3 programme.

1. Hardware

At 09.18Z on Day 366(84) the X-gyro motor current increased from its nominal value of ~70 ma to ~200 ma within one telemetry format (8s). The current remained high, fluctuating between ~180 ma and 276 ma (full scale on the analogue read out) and the X-gyro health monitor switched from 'GOOD' to 'FAIL'. Since gyro selection was on the basis of the health monitor, the S (Skew)-gyro was autonomously selected. De-pointing of the spacecraft occurred on all three axes as a result of the S-gyro selection. Periodic switching on of the X-gyro heater by its temperature control circuit was observed for several minutes after the gyro motor current increase, however this behaviour ceased and a short time later the temperature of the gyro box increased. On this basis, the gyro selection logic was changed from 'health monitor' to "SYZ" and the X-gyro switched off. Safety mode was triggered at this point (Y-axis pointing at the sun) with an indication that excessive rate had been detected but this was later diagnosed as spurious incremental signals generated by the X-gyro switch off.

A short test of the X-gyro was carried out while the spacecraft remained in safety mode. The gyro was switched on for approximately 7 minutes, throughout which the motor displayed a similar behaviour as earlier and the health monitor remained at 'FAIL'. During the switch on and off the Y-axis rotation rate changed, consistent with a momentum exchange between the spacecraft and X-gyro on the assumption that the X-gyro was spinning at its nominal rate.

There have been no changes In the status of the payload hardware. Additional calibration tests on the GSPC (ref. p. 25) have led to refinement of the effective areas and a re-definition of the background line feature energies.

2. Performance and Operation

Tables 1 and 2 on p. 5/6 give the current performance parameters of the EXOSAT instruments.

During the period 1.11.84 to 31.12.84 loss of observation time has resulted from ground computer system problems (3 hrs, Day 310), a power failure at ESOC (3 hrs, Day 334), solar activity (3.6 hrs, Day 355) and the X-gyro malfunction on day 366 (~16 hrs).

Regular swapping of the ME detector array halves was reintroduced (ref. EXPRESS no. 6 p. 4) to provide optimal background subtraction for weak source observations. This procedure is now the default for all observations, and P.I.'s should inform the Duty Scientist responsible for the orbit preparation if array swapping is not required.

A number of problems have arisen during source observations with respect to the requested payload configuration or OBC modes, e.g. a detection of a weak extragalactic source with the Boron filter without first using the Lexan filters. The following general configuration conditions are considered as standard:

1. Medium energy experiment array half is always offset with an array swap as above (typically once/unit).
2. Weak sources must be observed initially with a Lexan filter and detected.
3. Photometry using the low energy telescope will be performed only with a Lexan filter.

Note that responsibility for the final decision on configuration rests with the Duty Scientist/Observatory Controller who will be familiar with the science of the proposal and all payload/OBC operational details.

3. Observation Output

P.I's are reminded that the Observatory will despatch 1 copy only of the FOT per observation. There are no exceptions to this rule and P.I.'s must themselves ensure the distribution of data as appropriate to their Co.I.'s.

Chapter 8.2 of the revised (Rev. 2) FOT Handbook, dealing with ME data analysis, will be despatched shortly together with a number of single page updates to Rev. 2.

4. Future Plans

A 'natural' occultation of 4U1530-44 is scheduled for 16.01.85. This will provide a thorough test of the ground system software and operations procedures necessary to carry out occultations. Following this, a decision will be taken on the execution of a small number of approved AO occultation observations.

Specification of the concept and design of software for the generation of the EXOSAT observation log is in progress The function of this log is to provide a ready means to inform the community of the EXOSAT observation programme and to enable the Observatory Team to request information about this programme.

Software implementation and generation of the log is planned for early sumner 1985, such that public access to EXOSAT data can be supported in an efficient manner.

Data from PV/Cal and certain AO-1 observations will become public in April 1985. The observation log will identify which data is currently available and listings will be published in the EXPRESS.