ROSAT Status Report #64:
Aug 64, 1993
Calibration note:
MPE and GSFC (in consultation with SAO) have concluded, through independent analysis of calibration and other data, that there is some temporal variation of the PSPC gain which is not included in the gain correction process currently applied to data in SASS.
This effect showed up primarily as inconsistencies between calibration spectra of N132D taken at different epochs across the ROSAT mission. Another artifact of the problem is highly significant residuals in number of spectra compared to the expected model, these appear as large negative residuals below about 0.20 keV and large positive residuals between about 0.20 to 0.4 keV.
An example of this effect is in
rosat/problems/spectral_problem.ps
(a QDP version of the file also exists there, the .pco file is the header info for the .qdp file, and both need to be copied for plotting in QDP)
An official MPE statement on this problem follows below, in addition we provide some illustrative figures on the legacy ftp account as detailed following the MPE section.
************************************************************************** ========================================================================= = = = ROSAT NEWS No. 20 --- 1-Aug-1993 = = = =-----------------------------------------------------------------------= = ROSAT Scientific Data Center at the = = Max-Planck Institut fuer Extraterrestrische Physik (MPE) = = Giessenbachstrasse, 85740 Garching, FRG = =-----------------------------------------------------------------------= = e-mail addresses (Uli Zimmermann): = = rosat_svc@mpe-garching.mpg.de (Internet) or = = MPE::ROSAT_SVC (SPAN) = = ROSAT Service Area read-access via anonymous ftp: = = ftp rosat_svc.mpe-garching.mpg.de (130.183.72.180) user: anonymous = =-----------------------------------------------------------------------= = XUV Center: 29382::GXUVDC or GXUVDC@AIT.PHYSIK.UNI-TUEBINGEN.DE = = WFC Archive access via telnet/ftp ait.physik.uni-tuebingen.de = = user: xuv (password: xuv_archive) = =========================================================================
ROSAT USERS WORKSHOP on November 3/4 at MPE
Because of the large interest shown in the ROSAT Users Workshop held last year we invite all ROSAT Users to a second one and a half day workshop at the Data Center to discuss actual ROSAT items and learn on new science gained with ROSAT. Differently to the meeting of last year we want to keep the mission aspects this time relatively short and let more room for the presentation of ROSAT results and relevant data analysis items. Also this year there will be the possibility to discuss special data analysis problems with our experts on an individual basis.
The meeting is planned to start on Wednesday, November 3, after lunch to allow for travelling in the morning and should close on Thursday in the later afternoon. Please contact us if you would like to include additional topics in the meeting agenda listed below.
We will be happy to see you at the meeting. A registration form can be found at the end of the ROSAT NEWS or can be accessed as file workshop.form from directory general in the ROSAT Service Area.
Topics planned for the workshop
- Short status of satellite, instruments and the mission future
- Status of the calibrations
- Present and future ROSAT Archives (short discussion)
- ROSAT papers (main body of the workshop)
- Discussion on data evaluation items (e.g. source detection techniques)
- EXSAS demonstration: new facilities
CALIBRATION
PSPC Systematic Gain Errors
Using a total of 11 observations of the LMC SNR N132D we were able to detect a small but systematic gain variation of the PSPC as a function of time. N132D has a pulseheight distribution which is relatively sharply peaked around PSPC channel 80 with a distribution not much broader than the energy resolution of the PSPC. Only data within off-axis angles of 15 arcmin were used in order to avoid any additional systematic errors introduced by the energy-dependent vignetting correction. The centroid energy of the 0.4-1.5 keV pulseheight distribution was calculated in "PI-channels", which are nominally corrected for PSPC gain variations using the on-board 1.49 keV aluminum calibration line. We found a slow decrease of the derived peak energy throughout the whole mission from initially 0.893 keV during the Calibration/Verification Phase observations (June 1990) to finally 0.868 keV during a maintenance observation in April 1993. The total effect amounts to about 3% throughout the mission. Since the aluminum peak position at 1.49 keV has been corrected to lie at channel 150 by the standard gain calibration procedure we are observing some nonlinear gain variation effect which might actually be more pronounced at lower energies. Recent PSPC spectra of nearby, unabsorbed stars seem to show a substantial deficiency of soft photons. We try to follow this up by calibrating on the carbon peak around 0.25 keV and on the geocoronal oxygen line at 0.54 keV. The ultimate goal is to understand the observed effect and to model it as a function of time.
It now seems understandable why the most recent response matrix (DRM_36, released December 1992), which was fudged using the 1992 observation of MKN421, had to be systematically offset with respect to the previous, nominal matrix (DRM_06) by about one channel. A similar shift is now expected between 1992 and 1993 PSPC observations. As a temporary solution we suggest observers to use the old matrix (DRM_06) for observations up to AO-1 and the new matrix (DRM_36) for observations in AO-2 and AO-3. The difference between fits with the old and the new matrix roughly accounts for the current systematic errors.
First results have also been obtained for the recent calibration of the rate-gain effect in the PANTER facility. The engineering model PSPC behind one of the X-ray mirror shells of the XMM program was exposed to 1.49 and 2.99 keV X-rays at different count rates and gain settings. Already at about 100 cts/sec in the image a decrease in gain of about 0.5% was observed. At the highest count rates of 500-600 cts/s the gain decrease was 0.7% and 1.4%, respectively. Thus this effect is substantially smaller than the above mentioned temporal gain effect and definitely only important for the brightest ROSAT PSPC sources (e.g. Crab, GX 339-4, Cyg X-2 etc.). A re-calibration of the PSPC energy scale at lower operating high voltages is still awaiting a suitable slot in the availability of the PANTER test facility.
PROCESSING
All the observations mentioned in the last ROSAT News as having too small accepted times have been reprocessed and are on the way to the PI's.
ARCHIVE
Requests for datasets from the ROSAT Data Archive can be accepted only if the relevant datasets are already contained in the archive. Please check the status by looking into the appropriate information files in the ROSAT Service Area (files rda_coord.list or rda_seq.list or us_public.list in directory archive) before sending us your request.
ROSAT PAPERS
A new list, containing all refereed ROSAT papers known to us is available in the ROSAT Service Area (file papers.refereed in directory general).
EXSAS/MIDAS
We just received the beta test release of the MIDAS November 1993 version.
==========================================================================
Anmeldung zum R O S A T Users W o r k s h o p
am 3./4. November 1993 im MPI fuer Extraterrestrische Physik in Garching
Begin: 3. November 14 Uhr End: 4. November gegen 17 Uhr
Name, Vorname: ..........................................................
Institut: ..........................................................
Ich nehme am Workshop teil ...... .......... .... (bitte ankreuzen)
sicher vielleicht nein
Ich benoetige ein Hotelzimmer vom .............. bis ..................
Thema meines Kurzbeitrags (etwa 15 min) - falls zutreffend:
...........................................................................
Bitte senden Sie diese Form so bald wie moeglich an uns zurueck:
per email: MPE::ROSAT_SVC oder 28773::ROSAT_SVC (SPAN)
ROSAT_SVC@mpe-garching.mpg.de (Internet)
per FAX: 089-3299-3569
per Brief: Wissenschaftliches Datenzentrum - ROSAT
MPI fuer Extraterrestrische Physik
Giessenbachstrasse
85740 Garching
**************************************************************************
The spectral problem can be illustrated using the following N132D datasets:
Sequence Date Gain state Many* Jun 1990 High 40' offset rp160084 May 1991 High on-axis rp141800 Dec 1991 Low on-axis rp500004 Apr 1992 Low on-axis rp141937 Apr 1993 Low on-axis
(*These data were taken at 40 arcmin off-axis and an appropriate vignetting correction was applied before comparison was made with the on-axis data).
Postscript files reside on the legacy anonymous ftp account under
rosat/problems
these are the (binned) ratios of PI observations of N132D, the denominator is the April 1993 spectrum throughout. The right hand side of each panel shows the value of chisq obtained when the ratio data were fit with a constant model:
ratio1.ps (ratio1.qdp)
The numerator is a mean spectrum of vignetting corrected data taken from 40 arcminutes off-axis. The data were taken with PSPC-C at high gain state in the PVC phase (June 1990).
ratio2.ps (ratio2.qdp)
The numerator is the high gain state May 1991 observation.
ratio3.ps (ratio3.qdp)
The numerator is the low gain state Dec 1991 observation.
ratio4.ps (ratio4.qdp)
The numerator is the low gain state Apr 1992 observation.
One can immediately see that the inconsistencies seem to be greatest when the time difference between observations is large as discussed above.
chisq.ps
Reduced chi-square obtained from fitting a simple power law with low energy neutral and uniform absorption to the spectra of 26 AGN (with more than 1000 counts each), employing the resolution matrices released on March 1992 and January 1993. Filled circles identify observations performed with PSPC B AFTER October 1991 (i.e., with the low gain setting), while open squares identify observations performed BEFORE October 1991 (i.e., with the high gain setting). ("star" symbol identifies an observation performed with PSPC C during the PV phase.) All observations which give a significantly higher chi-square value with the January 1993 matrix than with the March 92 matrix were acquired with the high gain setting.
New USRSDC Staff Members
Two new key members are being added to the USRSDC staff this month. Assuming the role of lead scientist in the Guest Observer Facility is Dr. Steve Snowden. Steve has spent the last five years at MPE as a University of Wisconsin postdoc, analyzing the soft diffuse background data from the all sky survey. He has also been a vital member of the PSPC calibration team and assisted in observation planning. He is the author of several papers concerning PSPC calibration issues.
In addition, Ms. Minhthuyen Nguyen is filling the vacancy left by the departure last month of Doug Bennett. Minhthuyen's duties will be the same as Doug's - processing the paperwork associated with ROSAT (and ASCA) Guest Observer Grants, providing technical support for the proposal evaluation process, and assisting with the upkeep and enhancements of the MIPS software. Her e-mail address is ROSAT::NGUYEN
Doug Bennett has taken a position in the EOS project office at GSFC. Doug has served the ROSAT project since 1985, longer than anyone else at GSFC. We wish him success in his new position.
Select another report
Return to the ROSAT GOF
Curator:
HEASARC Guest Observer Facility
Please use the Feedback link if you have questions on ROSAT.