From mweber@head.cfa.harvard.edu Tue Jun 7 02:54:07 2005 Date: Tue, 7 Jun 2005 02:54:06 -0400 (EDT) From: Mark Weber To: Jonathan Cirtain Cc: Ed Deluca , Jason Scott , mario cosmo , Bill Podgorski Subject: E2E 20050606B Science handoff discussion. Good morning, Jonathon I did not have time to make analysis products for Best Focus I, but I will be catching up on those, so don't worry about them. The crucial info is the BF position, which we have already given to Ed. On the other hand, I'm afraid you will have to do the analysis for the on-axis x-ray alignment (Chapter 5). You can find the Image Log in our Analysis Notebook. I think you should make the contour plot first, using the Lockheed software, since that will give the Test Conductor the result he needs to proceed with testing. Ed also wants the "overlappogram". If you look in the directory for my shift, you will see a routine written by Jason for making the overlappogram. It needs several input arrays: filenames and centroid locations. Jason made some IDL save files with the variables for filenames. He didn't have time to do the xcen and ycen, so let me briefly explain. Data is 5x5 pitch/yaw positions. Note that positions in the log are in *arc-minutes*. Convert these to arc-seconds, and assume 1 pixel = 1 arc-second. Assume the (0,0) pitch/yaw position is the same as the xcen, ycen position we kept seeing during the Best Focus I test. It is something like (1288, 1028), but double-check the exact values. Jason's routine and the save files have "xroa" in their filenames, so you can recognize them. Moving on, you will probably have to do the analysis early on in the Optimized Focus sub-test. Here is the math to do the calculation: Let w_cen = the observed width of a centroid, then we consider that it has 2 components: the width of the mirror PSF w_psf, and the CCD charge-spreading width w_ccd. (w_cen)^2 = (w_psf)^2 + (w_ccd)^2 w_ccd = 0.5 pixels, according to the J-side. 1) Get w_cen from the Best Focus position in the 20050606B data. 2) Use the equation and w_ccd to solve w_psf, at BF. 3) Assume w_psf at OF = 2x w_psf at BF. 4) Use the same w_ccd value and the w_psf at OF and the equation to solve for w_centroid at OF. 5) Refer to the fine Best Focus 2nd degree polynomial plot to locate the focus step position that should give your value for w_centroid at OF. This tells you the on-axis focus position for OF. The radial pitch/yaw measurements are made relative to this OF on-axis position. Now here comes the sucky part, and I am truly sorry. I exited IDL without saving the 2nd degree polynomial plot for the Best Focus data. I was going to remake it tonight, but Chappell's archive at SAO does not have data yet, so I have nothing to work with. You will have to remake the plot unless I get there first (see below). Of course, the focus positions and gaussian widths are in our notes for the fine focus position. Here is how I used poly_fit to get the polynomial: IDL> rsswidths = sqrt((xwidths)^2 + (ywidths)^2) IDL> fsteps = findgen(41)*20 - 1150 IDL> pv = poly_fit(fsteps, rsswidths, 2, yfit) IDL> xx = findgen(401)*2 - 1150 IDL> yy = pv[0] + pv[1]*xx + pv[2]*xx^2 IDL> plot, ff, rsswidths, psym=-1 IDL> oplot, xx, yy, psym=-2 Don't worry about neatness, I will still be making a nice version of the plot for the website to atone for my sins. Lastly, I have a lot of catching up to do, so I plan to come in sometime in the morning. I will stay out of the way, but I will be working on my analysis products from 20050606B. If I get there before you guys reach the Optimized Focus sub-test, I will be able to remake the polynomial for you. Good luck! Mark