LFC data reduction using NOAO mscred
(the "lfcred" page, dedicated to René Magritte)
![[René Magritte, The Betrayl of Images]](pipe.jpg)
LFC data reduction using NOAO mscred(the "lfcred" page, dedicated to René Magritte) |
|
The lfcred suite is a collection of IRAF routines, calibration data, and other tidbits meant to facilitate the reduction of Palomar LFC data using the NOAO mscred reduction package.
You can download the current or old revisions of lfcred, or you can examine the CVS repository for lfcred to see what changes have been made to the components.
Once you have obtained lfcred, consult the included GUIDE file for instructions on installing and using the software.
This software and its instructions are currently in quite a rough state, as described in GUIDE. Contributions of code, calibrations, or suggestions are welcome!
NOTE: If you are preparing for an observing run, you should plan to take an astrometry calibration field which is rich in USNO-A2 stars (low Galactic latitude). You should be in good shape with 15 seconds in any filter but u'. To get enough stars in u' (about 200-300 per chip) would require much more than the 300 second integration I've tried. One filter (typically Rs/r') is probably OK, but taking an image in all of your filters won't hurt. You will need to construct your own WCS (see link below) if the focal plane has been worked on since my last solution or if you're using the instrument in the E-W orientation. With the LFC being worked on now and then these days, and with LFC astrometry in its early stages, I must recommend that everyone take astrometry frames on each run!
NEWS (11 March 2002): According to Paul Price, you should just use the same, unbinned astrometric solutions for binned mode. There is some magic in IRAF that takes care of the binning transparently. (This is actually much older news that I forgot to write here.) (16 December 2001):: We all know that it can be hard to find LFC guide stars in u'. I've written a program that will query the GSC and list likely guide stars at your LFC pointings. (12 October 2001): Thanks to Paul Price, we now have changes to lfcassemble to support 2x2 binning. I don't do 2x2, so I have not personally tested this. Also, subsets and ccdtype now work, but you have to set mscred/ccdred instrument parameters--see GUIDE. (2 October 2001): There's a new release that has an updated crosstalk file that works with the new lfcassemble (no 32k clipping). Recent versions of mscred seem to have a bug in xtalkcor that precludes crosstalk corrections of zero. I have used a very small correction instead and will contact the mscred people about this. Also, please note that I have not addressed the header translation that mscred needs to differentiate bias vs. flat vs. object exposures, so you should leave ccdtype="" in ccdproc and other tasks like flatcombine. (29 September 2001): There's a new release that preserved the whole dynamic range of the chips (using the BZERO header). Also some minor changes to the documentation. Note in particular that lfcassemble does not support any binning besides 1x1 without serious effort on your part. BUT the xtalk.dat file is now incorrect. If someone can measure for each chip the crosstalk ADU (above sky) that's present divided by the ADU of the saturated source, that would be a big help. (4 September 2001): There are new astrometric solutions and bad pixel masks from our 2001 Aug 23 UT observations. I had been using order=6 polynomials for the fit, but have found that order=4 fits the USNO-A2 stars just as well, and is better behaved at the edges of the focal plane and in regions where the stars are more sparse. (We used this solution for LRIS slitmasks with excellent results.)
Useful links: