Date: 10 Sep 1997 10:03:19 -0400 From: David Huenemoerder To: melvis@head-cfa.harvard.edu, kim@head-cfa.harvard.edu Subject: mtgs, expmaps cc: adam@head-cfa.harvard.edu,jhk,vancura@cfa.harvard.edu , wise, houck , dd, davis Exposure maps: Mtg 1: Monday Adam, Joel, and I met to go over some instrument specific details relevant to exposure maps. - HRC has a time-dependent live-time factor. This has to be factored into the aspect histogram. It comes as a L1 HRC product. [AD - called what?] - ACIS has a global live-time factor, which is the ratio of the frametime (which is in the header) to the "nominal" frametime. This factor gets applied as a scalar multiplication on the entire exposure map. [JHK- what is the frametime keyword? - where does "nominal" frametime come from?] - ACIS can drop frames if there are too many events to process in the nominal frame time. This can be expressed as a Good-Time-Interval per chip, which have to be accounted for as the aspect histogram is built. Hence, if there are frame dropouts, then there will be multiple aspect histograms (or deltas for each chip on the basic histogram). - At full resolution, the aspect histogram is about 200 x 200 x a_few (0.2 arcsec x 0.2 arcsec over 40 arcsec dither, a few is for the roll-range, and is probably from one to ten). This is not a strain. Detector QE maps are on the order of 30x30 arrays - also no problem. High-spatial resolution maps can be made on small regions. IF a final map is needed near an edge or bad pixel, then low and high resolution maps need to be merged into a high-resolution map (making a high-resolution map of an entire detector may be a problem, due to size). Mtg 2: Yesterday, I met w/ John Davis, Mike Wise, and John Houck to discuss exposure maps. John D. has a formal description of the instrument response - starting w/ the source, and folding through to the detector - in which he has included terms for the aspect motion (as well as mirror PSF and detector RMF redistributions). While we agreed that this is a correct description, it is abstract enough that it is difficult to see what assumptions and approximations lead it to my description (if any!). My current job is to relate John's description to calibration products, aspect solution, and investigate assumptions such as PSF or RMF invariance over the scale of dither. In particular, his "aspect integral term" needs to be compared to the traditional "aspect histogram" to see if they are indeed the same. John noticed that formally, one cannot decouple the aspect from the detector RMF or mirror PSF (a couple terms don't commute). Physically, I think this means that IF the PSF changes significantly over the scale of the dither (unlikely, but unverified vs off-axis angle), and IF the RMF changes substantially over the scale of the dither (guaranteed, for small regions of ACIS), then the traditional exposure map (as currently defined), is wrong. The current definition only averages QE, and ignores RMF. The advantage of adopting John's formalism is that: - it is general - we can include as many affects as we want. We can make approximations which reduce the integrals to XSPEC compatible response matrices plus the traditional exposure map, for instance. (and will because 1st order analysis needs require such) - subtle affects can be accommodated up front when necessary, without having ad hoc back-end fixes (we could patch the RMF detail that way with an additional map of "chip-type-fraction"). In this way for edge regions, we have a clear definition of a *consistent* exposure map and RMF. 2- it will work for slew data (I was taking vignetting out of a sum as a necessary simplification) 3- it provides a framework for correct description of extended sources (though there is an integral John says he doesn't know how to do (and if HE doesn't know...); it can, however be treated in the forward-folding direction). 4- The formalism is general regarding grating/imaging modes (the grating is treated as another redistribution). Relevant Documentation: My description of exposure maps is at: http://space.mit.edu/~dph/Maps/Exp_maps.ps.gz My description in the tools had a different implementation which specified a convolution of exposure map w/ detector map: (exp tools in the current tool book - 2000 series, prefaced w/ "exp_"). MIT-local copy of specifications is: file:/nfs/wiwaxia/h1/dph/h1/ASC/TG/Flight/Development/Expmap/Specs.ps.gz SAO-local copy is: file:/proj/asc/dph/Expmap/Specs.ps.gz John's draft document on responses in general: (see section 4): MIT-local: file:/nfs/wiwaxia/h1/davis/tex/grsp.ps SAO-local: [John - I took the liberty...] file:/proj/asc/dph/Expmap/grsp.ps.gz -- Dave David Huenemoerder (617-253-4283; fax: -0861) Center for Space Research/AXAF Science Center MIT NE80-6023, Cambridge, MA 02139 http://space.mit.edu/~dph