Our efforts of late have been directed toward calibrations and improving the efficiency and reliabilty of our data-processing software. The calibrations fall into two areas. One is the physical orientation of the detectors relative to the spacecraft axes. To aid in this, special observations of Sco X-1 were carried out early in April and we now know the camera orientations of a given dwell to about one arc minute. The other required calibration is the determination of the model for the expected SSC image of a point source. The model is a function of angular position and with spectral shape. Errors in the model produce reduced signals, increased errors, and increased source confusion. The model is now known well enough to produce fairly reliable "quick-look" results (see below).
In the process of these calibration efforts, the entire database of observations has been processed several times during the past month with improved quality each time. Based on the appearance of the light curve from the Crab nebula (which should be a constant intensity), we have now removed or understand most (but not all) systematic effects that produce spurious data points. We therefore now have the tools in place to produce a "quick-look" set of light curves for 103 cataloged sources (including upper limits). We will deliver these to the Guest Observer Facility (GOF) at GSFC about 10 days from now. They will be delivered with a cautionary note that any single data point should be treated with caution. The quality of the data points depends upon the source intensity, source spectrum, the location of the source in the field of view, and the degree of source confusion. The Crab is being detected with an rms of about 4% (2.2% statistical and an estimated 3% systematic) in each dwell. The threshold for detection is well below this for an isolated source, but it is not yet well determined.
Searches for flaring sources (Targets of Opportunity; TOO's) have been carried out on the data base. The only "new" source found was the Granat-discovered GRS1739-278. Heretofore, we have not been able to carry out TOO searches on a daily basis, but we are now implementing such a system, It is just now being tested on the real-time data. A TOO alarm would be indicated by several sightings of a source in an unusual state. After testing, it should be a short step to implement a dwell-by-dwell TOO-search system, the advertised goal of this experiment.
Dramatic variability from the galactic binary ("microquasar") 1915+10 was noted in the ASM data. This was the first ASM TOO which led to PCU observations that are rich in temporal variability. Results will be reported at the HEAD meeting by Greiner, Morgan, and Remillard.
In the meantime, the MIT EDS (Experiment Data System) has been reprogrammed to detect anomalies in the data from a PCA detector and to send a turn-off command (via the spacecraft). This is in place and working well.
Fnally, we are pleased to be working on useful data (rather than hardware
problems) and are striving to get the ASM data into the hands of the community
in a reliable form on the shortest timescale possible.
ASM Status Report for 30 August, 1996
Analysis Status (enlarged source catalog and improved light curves)
The light curves posted on the MIT WEB site have been updated
weekly since our last status report in April. The light curves
have guided the observing program of the PCA/HEXTE on a number of
occasions, e.g. the entry of Cyg X-1 into a high state, the
renewed activity in the microquasars GRS1915+105 and GRO 1655-40,
in Cir X-1, etc. The ASM data have been included in a number of
papers, published and in press, by authors within and outside our
group. The ASM results have been particularly fruitful when used
in conjunction with BATSE (which features a higher
non-overlapping energy band) and with the PCA (ASM provides
context of PCA observations).
Improvement of the analysis software and calibrations have continued and have provided light curves of increasingly higher quality. On several occasions, the entire light curves from launch were replaced to incorporate these improvements. This was accomplished through filtering the solutions tables or by a reanalysis of the entire database. The most notable improvement was an algorithm to reject solutions wherein two catalog sources lie in the same "confusion cell" (0.1 deg x 6 deg in SSC-FOV coordinates) because we found that such occurrences led to spurious intensities.
In early August, the entire ASM database since launch was reprocessed when it was decided to enlarge the catalog from 112 to 158 sources many of which are known to be below the ASM threshold for detection. Their inclusion in the analysis was implemented for two reasons: (1) the detection of transient behavior of these sources is facilitated, and (2) it provides a substantial number of null source detections which allow a better determination of the 'background' and detection threshold.
Unfortunately, the increase of catalog size led to fewer light-curve data points for sources in crowded regions because two modeled sources were more likely to fall in the same confusion cell and thereby to be rejected. Accordingly a second-pass analysis of the entire database was carried out wherein the sources that were demonstrably below threshold for the previous 25 weeks were temporarily removed from the catalog, and the entire database was then reanalyzed.
This "pass 2" processing for the entire database was performed after removing 63 "dead" sources from the catalog. The remaining 95 "possibly active" sources were initially re-analyzed. This effort (through Week 28) netted 36770 ADDITIONAL data points as a result of removing source confusion involving a "dead" source. This is an overall increase of 13% in the ASM light curves, but the big winners were the SMC, LMC, and galactic center. In the extreme case, the yield for SMC X-1 ( and its nice 60 day period) increased from 453 to 1860 data points..... LMC X-3 rose from 1176 to 2092....galactic center sources (GCX-1, GRS1739-278, etc) rose from ~450 to ~850.
The light curves currently posted "through Week 29; 8/22/96" include the results of this global (since launch) Pass-2 analysis for the "possibly active" sources. The Pass-1 light curves for the "dead" sources are also posted. We expect to continue this two-pass system on our weekly updates of the light curves.
The software for TOO searches is in place as follows: Cataloged sources are flagged with alarm levels tied to TOO-proposal criteria and these now operate at the SOF and at MIT on the realtime data. Three alarms on the same source trigger an alarm, and these could occur within an hour or two for a real effect. The software for real-time search for new UNcataloged sources is also in place at the SOC and at MIT. Lists of candidate source positions from the crosscorrelation analyses of the residuals are generated automatically, and maps of lines of position can be generated for likely positions. This process still takes considerable personal judgment, however, and anslyses down to ~50-100 mCrab are currently undertaken once or twice per week. One of our major goals is to refine and further automate this process to increase its reliability and efficiency.
ASM INSTRUMENT STATUS (gain measurements and anode failure)
The ASM has continued to operate with 21/24 of its aperture over the past 5 months since all three detectors were brought on line following the January failures. One additional anode of SSC2 has exhibited a dip in its position response since launch. This is indicative of a high-resistance spot, but this dip has not worsened. This anode is not yet used in the analysis because of this increased level of non linearity, so the light curves to date represent 20/24 of the aperture. The high productivity of the ASM is evident in the released light curves of 158 sources of which 95 are considered "active". These light curves are continuous since Feb 20. Until the incident described below, no further breakdown events were noted. SSC1 continued to operate with all 8 anodes, SSC2 with 7 anodes (including the one with the dip), and SSC3 with 6 anodes. The rotation drives and readouts, the rotation-control software, and safety turn-off fixes have all worked reliably with no obvious anomalies.
SSC1 was known to have a slow leak before launch. It has been so small that it has heretofore been undetectable in space. Gain measurements with the weak onboard Fe55 sources may now be detecting a gain increase of ~6%/year in this detector, averaged over 8 months. The values obtained by Don Smith with 90% confidence errors (which may be underestimates) are:
SSC1 : 5.6 +- 4.2 % change per year
SSC2 : 2.2 +- 4.3
SSC3 : -1.3 +- 1.1
If this is the true gain rate of change in SSC1, it appears that SSC1 should have 5 years or more of life ahead of it, and the others much more.
Anode failure in SSC3
We have just encountered another failure of an anode in SSC3. Anode 6 (scale 0-7) showed no apparent counts in the realtime position histogram displays on 26 August. This is indicative of a spot of high resistance so the counts "appear" to be located at the ends of the anode. Accumulations of diffuse background data for previous weeks showed (beginning at the end of June; Week 22), a slowly growing dip in the position histograms for that anode. This is indicative of such a high-resistance spot presumably caused by low-level breakdown. From the January events, it was expected that serious breakdown would start soon after the Aug 26 observation. In fact, it began on the next day, Aug 27.
We are now following the procedure that was successfully used by us previously, namely to let the breakdown proceed until it eventually erodes all the carbon off the affected quartz anode. On Aug. 28 and 29, the breakdown was proceeding with the measurement chains of all but two anodes squelched (to prevent the pickup counts from causing a high-voltage turnoff). If experience holds, the breakdown will cease in about 3 days and the unaffected anodes will return to normal operation.
If this proceeds as expected, SSC3 will resume operation with five of its 8 anodes operative. This is the SSC which views the "polar" region and hence covers less sky with more exposure per orbit. We note that this is the first failure since the Jan 6 breakdown of SSC3 and the Jan 12 breakdown of SSC2. These detectors were "fixed" in March and have operated continuously since March 12 and 19 respectively, or about 5 months. We are optimistic that we will shortly have all three detectors operational again and the situation will remain stable thereafter.
We do not expect that the deteriorating anode introduced
significant errors into the light curves until possibly Week 29,
which began 16 August. The data for SSC3 since Week 22 will be
reanalyzed without the affected anode and will be posted with the
Week 30 data about Sept. 4.
ASM Status Report - 5 September 1996
The breakdown of Anode 6 in SSC 3 has ceased. Since about 22:00 (UT)
on 4 Sept 1996, the counter has been in normal operation with five of
its eight anodes (0, 1, 2, 5, and 7 on a 0-7 scale) functioning properly.
The week 30 data distribution includes a re-processing of weeks 22-30 to avoid the rapid evolution in anode 6 of SSC3 that resulted in its recent failure.
Analysis Update: 6 February 1997
We must note that this update constitutes a complete revision of the
ASM light curves based on revised geometric models in which the
SSCs' position responses are modelled in a time-dependent manner.
For bright sources, the statistical scatter has been reduced significantly.
We have also been able to relax several data filters, e.g. the size of
our 'confusion cell' (now 0.06 x 6 degrees with long axis aligned along
the SSC collimator), and the calibrated region in our field of view
(+- 5 deg. alonmg short axis and from -41.5 to +46. degrees along long axis).
Finally, the number of monitored sources has risen to 171, of which 68
appear to be entirely 'off' during the ASM Mission so far.
- The ASM team (Bradt, Chakrabarty, Cui, Levine, Morgan, Remillard,
Jernigan, Shirey, Smith)
A compilation of old status reports, from January 16 through March 28, 1996 is also available, should you wish to read about the history of our instrument's performance.