Subject: Bad events

Dear All,

another AI closed: AI29_6 on JC and CAO to make spectra of bad events to see
just how bad they are. Here are the results:

http://www.dsri.dk/~oxborrow/sdast/CAO.longStaring102.scw35.ps.tar.gz

These results show the JEM-X1 events from 7 sub-SCWs in SCW 35 of revolution
102, which was a long staring during the crab calibration. All events have been
corrected both for position and gain by the ISSW, and the spectra are given in
PI channels, hence the sudden jumps where channel width increases. Events have
been sorted into  6 catagories: Hotspot events (STATUS=8192); Events from
outside the active detector area (STATUS=4096); spillover events from the
calibration sources (STATUS= 1024); events from neighbours of bad anodes 
(STATUS=32) and events from bad anodes (STATUS=16). There were no other kinds
of bad events present in these science windows (which is only to be expected),
and only events that we flagged as one of the above have been included. 
That is to say, events that we both from calibration sources and a bad anode
etc. have been left out of the spectra. `Shadowgrams' or detector images have
been included in the figures to demonstrate the origin of each sort of bad
event.

HOTSPOT EVENTS: There is no genuine hotspot in these science windows, but the
program nicely flags calibration events that even fall outside the area marked
for calibration events in -DETE-MOD. This problem was particularly bad for the
two Fe-55 sources that were mounted somewhat hastily and overflowed quite
extensively until the positions of the calibration areas were adjusted after
rev 102. The spectrum of the flagged hotspot events is dominated by 6Kev
photons from the Fe-55 sources with no sign of the Xe peak. These events can be
caught by using the STATUS<256 selection, and this should be done for all
science analysis.

OUTSIDE EVENTS: Though many of these events originate from the outer rim of the
detector, and could conceivably be considered real events (unlike the little
cluster in the the upper left hand corner), this spectrum deviates considerably
at middle energies from that of the good events, and barely shows any Xe peak.
Evidently this rim is very noisy, and would generally be cut out using
radiumLimit anyway, but will also be removed using STATUS<256.

CALIBRATION OVERSPILL EVENTS: As expected this spectrum is dominated by Fe-55
(6KeV from sources 1 and 3) and Cd-109 photons (22KeV from sources 2 and 4)
in proportion to the amount of overspill produced by each type of source. 
Significantly, the small Xe peak appears at higher energy than for the good
events, suggesting that the gain in these highly irradiated areas is greater
than over the rest of the plate, as expected.
These should also be removed using STATUS<256 selection.

NEIGHBOURS TO BAD ANODES AND BAD ANODES: These events show no significant
deviation from the good events spectrum, suggesting that though counting
activity may be reduced in these areas,  gain is not systematically affected.
This confirms the observation that the second SPAG layer that includes
anode-by-anode corrections beyond those introduced by the primary SPAG table
do not noticeably improve results. Also, the gain changes introduced for bad
anodes and their neighbours follow no discernible pattern.  Hence we are correct
to keep these events for imaging and spectral analysis (they are not caught by
the STATUS<256 selection), but where flux is being calculated it is perhaps
best to remove these events and simply consider these areas inactive since we 
cannot for a given science window automatically estimate how many events are
misssing from a dead or unstable area.

CONCLUSION: Keep the current STATUS<256 selection for imaging and spectral
extraction; and keep the STATUS==0 selection for flux determination. In other
words, we're already doing the right thing by these bad events. The only 
area we don't have covered entirely is flux determination in the presence of
transient hotspots, and the means to do that ( the -DETE-STA table) will be
available for OSA 5.0.

Best wishes,

Carol Anne




.DANISH.SPACE.RESEARCH.INSTITUTE.DANISH.SPACE.RESEARCH.INSTITUTE.DANISH.
Dr. Carol Anne Oxborrow
Email: oxborrow@dsri.dk	Homepage: http://www.dsri.dk/~oxborrow
Telephone (direct): +45 35 32 57 33
Telephone (secretary): +45 35 32 57 01
Fax: +45 35 36 24 75