Logged in as guest
Viewing EPICpn/10394 Full headers
Private message: yes no
Notes: Notification:
Date: Wed, 17 Mar 2004 16:28:31 GMT From: cocj@roe.ac.uk To: xmmhelp@xmm.vilspa.esa.es CC: cocj@roe.ac.uk Subject: More PSF confusion
Full_Name: Olivia Johnson Submission from: (NULL) (195.194.120.192) Hi, I'm trying to find the dependence of the encircled energy on off-axis angle, so that I can apply appropriate circular apertures for a stacking analysis. I have tried using CalView to make .fits images of the PSF at various OAA (medium accuracy model), and find that, strangely, the 80% EE radius is only very weakly dependent on OAA at 2 keV, and actually dips from OAA=0 to OAA=7 or so. I assumed I had done something wrong, and so consulted the helpdesk archive on PSF problems, where I read the following in response to message Calibration/7320, in which a user has used the King profile parameters to work out EE and is surprised by the lack of variation with OAA: >Your calculations are correct. There appears to be surprisingly little >change in the encircled energy fraction as you move off axis for all >of the cameras. Although the shape does distort and elongate beyong ~7 >arcmins, the encircled fraction stays roughly constant. I also found a plot in version 1.1 of the User's handbook which is still on-line at http://www.mssl.ucl.ac.uk/www_xmm/ukos/onlines/uhb/XMM_UHB/XMM_UHB.html which seemed to exactly reproduce my results from CalView, with very little variation with OAA and a dip at moderate OAAs. Just when I'd begun to think this was the true behavior of the EE with OAA despite the apparently obvious expansion of the PSF at large OAAs, I looked to the most recent version of the UHB (off the vilspa page) to see if anything was said about the physical reason behind this curious behavior. There I saw completely different plots of 90% EE vs. OAA, this time with significant increases in the radius as I had originally expected. Would you please explain the discrepancies, and let me know I could obtain the data used to plot Figures 9 & 10 of version 2.1 of the User's handbook. Thanks, -Olivia Johnson
From: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> To: cocj@roe.ac.uk Subject: Re: More PSF confusion (PR#10394) Date: Thu Mar 18 13:38:19 2004
Dear Olivia, I forwarded your question to our experts. I will come back to you as soon as they provide me an explanation. Best regards, Nora ..........---- Dr. Nora Loiseau XMM-Newton User Support Group
From: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> To: cocj@roe.ac.uk Subject: Re: More PSF confusion (PR#10394) Date: Mon Mar 22 11:22:39 2004
Dear Olivia, here is the reply to your question from our SAS expert. Best regards, Nora ------------------------------ the user has looked at the Medium mode PSF with calview. This is the old, now superseded method of calculating the encircled energy, which agrees with the plots in the UHB version 1.1. Using the EXTENDED PSF mode (King model), which is the current recommended method, the encircled energy has a different dependence on off-axis-angle from the medium mode although this dependence is still small. Figures 9 and 10 in Version 2.1 of the UHB are I think out of date. We are now updating them for next version. Anyway, the bottom line is that from in-orbit measurements the encircled energy seems to be a slow function of off-axis angle for all the Epic cameras. ---- Dr. Nora Loiseau XMM-Newton User Support Group
Date: Tue, 23 Mar 2004 12:19:32 +0000 (GMT) From: Olivia Johnson <cocj@roe.ac.uk> To: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> Subject: Re: More PSF confusion (PR#10394)
Dear Nora, Thanks for the quick reply. Is there a plot and/or data currently available that represent the appropriate behavior of encircled energy with off-axis angle? This is a basic instrument characteristic and should not have to be rederived by each user, and the many versions of calibration data and documentation available make this quite confusing for a novice. I need only a rough indication of radius for a given encircled energy at e.g. 2 arcminute off-axis intervals across the field of view. Thanks, Olivia On Mon, 22 Mar 2004, Nora Loiseau wrote: > Dear Olivia, > > here is the reply to your question from our SAS expert. > > Best regards, > > Nora > > ------------------------------ > > the user has looked at the Medium mode PSF with calview. > This is the old, now superseded method of calculating the encircled > energy, which agrees with the plots in the UHB version 1.1. > > Using the EXTENDED PSF mode (King model), which is the current recommended > method, the encircled energy has a different dependence on off-axis-angle from > the medium mode although this dependence is still small. > Figures 9 and 10 in Version 2.1 of the UHB are I think out of date. > We are now updating them for next version. > > Anyway, the bottom line is that from in-orbit measurements the encircled energy > seems to be a slow function of off-axis angle for all the Epic cameras. > ---- > Dr. Nora Loiseau > XMM-Newton User Support Group >
Date: Tue, 23 Mar 2004 14:36:41 +0000 (GMT) From: Olivia Johnson <cocj@roe.ac.uk> To: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> Subject: Re: More PSF confusion (PR#10394)
Hi Nora, One last follow-up on my last message. I have downloaded fits files from the extended accuracy model for both pn and MOS and performed the same encircled energy analysis I conducted on the medium accuracy model. Indeed, I now see a range in 80% encircled energy from on-axis source to those at 14' off-axis of 43.75 - 44.0 pixels for pn and 43.98 - 43.99 pixels for MOS, if the PSF is normalized at about 1'. The encircled energy curves for each off-axis position (so percentage of flux within a 50 pixel radius vs. radius in pixels) look virtually identical. I realize the sources have been radially averaged, but simply looking at the banana shape of the PSF at large off-axis angles it seems this can't be right - there must be variation in the radial distribution of light between on-axis and off-axis sources musn't there? Normalizing the PSF at just over 2', the maximum allowed by the calview images, gives a much larger 80% EE value (~80 pixels) which decreases with OAA in both instruments. Cheers, -olivia
Date: Tue, 23 Mar 2004 16:10:35 +0000 (GMT) From: Olivia Johnson <cocj@roe.ac.uk> To: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> Subject: Re: More PSF confusion (PR#10394)
Hi again Nora, Apologies, one more update. Please disregard my last email - I did not realize the image size was altered in the extended accuracy model and there was therefore a bug in my code. I have corrected this, and now for pn I see an 80% EE value for the pn of just under 25 pixels and for the MOS of about 21 pixels, both increasing slightly with off axis-angle. The 50%EE value is practically constant for MOS at just over 8 pixels, and is 10 pixels on-axis _decreasing_ to 8 for the pn. This behavior still seems strange to me. Could you confirm that this is correct and also that the pixel size for the extended accuracy model PSF images is still 1.1" Thanks, and sorry for the mix-up. -Olivia
Date: Wed, 24 Mar 2004 10:28:08 +0000 From: Nora Loiseau <nloiseau@xmm.vilspa.esa.es> To: cocj@roe.ac.uk CC: xmmhelp@xmm.vilspa.esa.es Subject: Re: More PSF confusion (PR#10394)
This is a multi-part message in MIME format. --------------010509000900090405040206 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Dear Olivia, Please find attached the 90% enclosed energy plot for MOS-1, which is different from that presented in the current UHB. The PN plot in the UHB still represents current knowledge. There are many combinations of photon energy, off-axis angle, instrument and encircled energy which could be plotted. We can not provide all of these. The user can derive the specific plots that they want using calview. In this case the general recipe is: 1. Create a cif with cifbuild and make sure that it contains the files XRT1_XPSF_0006.CCF, XRT2_XPSF_0006.CCF, XRT3_XPSF_0005.CCF 2. run calview 3. Select the required instrument select "Accuracy Level" = EXTENDED set "Energy (eV)" to required photon energy set Theta to 0 arcseconds 4. View->Encircled_energy->Encircled_energy(PSF) 5. Read from the resulting plot the radius which contains the required encircled energy fraction. 6. Set Theta to 120 arcseconds and then repeat etc.... We are sorry for the out of date documentation on the PSF, our experts are working hard to bring it up to date. Regards, Nora -- -- Dr. Nora Loiseau, XMM-Newton User Support Group phone: +34 918131298 nloiseau@xmm.vilspa.esa.es FAX: +34 918131322 --------------010509000900090405040206 Content-Type: application/postscript; name="mos1_90pcent_psf.ps" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="mos1_90pcent_psf.ps" %!PS-Adobe-3.0 EPSF-3.0 %%For: rsaxton %%Title: PGPLOT PostScript plot %%Creator: PGPLOT %%CreationDate: 22-Mar-2004 10:19 %%BoundingBox: (atend) %%DocumentFonts: (atend) %%LanguageLevel: 1 %%Orientation: Landscape %%Pages: (atend) %%EndComments %%BeginProlog /L {moveto rlineto currentpoint stroke moveto} bind def /C {rlineto currentpoint stroke moveto} bind def /D {moveto 0 0 rlineto currentpoint stroke moveto} bind def /SLW {5 mul setlinewidth} bind def /SCF /pop load def /BP {newpath moveto} bind def /LP /rlineto load def /EP {rlineto closepath eofill} bind def /MB {gsave translate MFAC dup scale 1 setlinewidth 2 setlinecap 0 setlinejoin newpath} bind def /ME /grestore load def /CC {0 360 arc stroke} bind def /FC {0 360 arc fill} bind def %%EndProlog %%Page: 1 1 %%BeginPageSetup /PGPLOT save def 0.072 0.072 scale 8149 250 translate 90 rotate 1 setlinejoin 1 setlinecap 1 SLW 1 SCF %%EndPageSetup %%PageBoundingBox: (atend) 0.000 setgray 5 SLW 8399 0 1050 780 L 0 6239 C -8399 0 C 0 -6239 C 0 108 C 0 54 1750 780 L 0 108 2450 780 L 0 54 3150 780 L 0 108 3850 780 L 0 54 4550 780 L 0 108 5250 780 L 0 54 5949 780 L 0 108 6649 780 L 0 54 7349 780 L 0 108 8049 780 L 0 54 8749 780 L 0 108 9449 780 L 0 108 1050 6911 L 0 54 1750 6965 L 0 108 2450 6911 L 0 54 3150 6965 L 0 108 3850 6911 L 0 54 4550 6965 L 0 108 5250 6911 L 0 54 5949 6965 L 0 108 6649 6911 L 0 54 7349 6965 L 0 108 8049 6911 L 0 54 8749 6965 L 0 108 9449 6911 L -22 -7 1043 650 L -14 -21 C -7 -36 C 0 -22 C 7 -36 C 14 -22 C 22 -7 C 14 0 C 22 7 C 14 22 C 7 36 C 0 22 C -7 36 C -14 21 C -22 7 C -14 0 C 0 8 2407 614 L 7 14 C 7 7 C 14 7 C 29 0 C 15 -7 C 7 -7 C 7 -14 C 0 -15 C -7 -14 C -15 -22 C -72 -72 C 101 0 C -72 -100 3871 650 L 108 0 C 0 -151 3871 650 L -8 14 5293 629 L -21 7 C -15 0 C -21 -7 C -14 -21 C -8 -36 C 0 -36 C 8 -29 C 14 -15 C 21 -7 C 8 0 C 21 7 C 15 15 C 7 21 C 0 8 C -7 21 C -15 15 C -21 7 C -8 0 C -21 -7 C -14 -15 C -8 -21 C -22 -7 6635 650 L -7 -14 C 0 -15 C 7 -14 C 15 -7 C 29 -7 C 21 -8 C 15 -14 C 7 -14 C 0 -22 C -7 -14 C -8 -8 C -21 -7 C -29 0 C -22 7 C -7 8 C -7 14 C 0 22 C 7 14 C 15 14 C 21 8 C 29 7 C 14 7 C 8 14 C 0 15 C -8 14 C -21 7 C -29 0 C 15 7 7948 622 L 21 21 C 0 -151 C -22 -7 8114 650 L -14 -21 C -7 -36 C 0 -22 C 7 -36 C 14 -22 C 22 -7 C 14 0 C 22 7 C 14 22 C 8 36 C 0 22 C -8 36 C -14 21 C -22 7 C -14 0 C 15 7 9348 622 L 21 21 C 0 -151 C 0 8 9478 614 L 7 14 C 7 7 C 15 7 C 28 0 C 15 -7 C 7 -7 C 7 -14 C 0 -15 C -7 -14 C -14 -22 C -72 -72 C 100 0 C 108 0 1050 780 L 54 0 1050 1196 L 54 0 1050 1612 L 54 0 1050 2028 L 54 0 1050 2444 L 108 0 1050 2860 L 54 0 1050 3276 L 54 0 1050 3692 L 54 0 1050 4107 L 54 0 1050 4523 L 108 0 1050 4939 L 54 0 1050 5355 L 54 0 1050 5771 L 54 0 1050 6187 L 54 0 1050 6603 L 108 0 1050 7019 L 108 0 9341 780 L 54 0 9395 1196 L 54 0 9395 1612 L 54 0 9395 2028 L 54 0 9395 2444 L 108 0 9341 2860 L 54 0 9395 3276 L 54 0 9395 3692 L 54 0 9395 4107 L 54 0 9395 4523 L 108 0 9341 4939 L 54 0 9395 5355 L 54 0 9395 5771 L 54 0 9395 6187 L 54 0 9395 6603 L 108 0 9341 7019 L 101 -72 735 730 L 0 108 C 151 0 735 730 L 7 -22 735 845 L 22 -14 C 36 -8 C 21 0 C 36 8 C 22 14 C 7 22 C 0 14 C -7 22 C -22 14 C -36 7 C -21 0 C -36 -7 C -22 -14 C -7 -22 C 0 -14 C 0 -72 735 2824 L 65 -8 C -7 8 C -8 21 C 0 22 C 8 21 C 14 15 C 22 7 C 14 0 C 22 -7 C 14 -15 C 7 -21 C 0 -22 C -7 -21 C -7 -8 C -15 -7 C 7 -21 735 2924 L 22 -15 C 36 -7 C 21 0 C 36 7 C 22 15 C 7 21 C
Date: Thu, 1 Apr 2004 19:05:46 +0100 (BST) From: Olivia Johnson <cocj@roe.ac.uk> To: Nora Loiseau <nloiseau@xmm.vilspa.esa.es> Cc: xmmhelp@xmm.vilspa.esa.es Subject: Re: More PSF confusion (PR#10394)
Hi again Nora, One last related question: Does the SAS detect chain (eboxdetect -> emldetect) use the EXTENDED or the MEDIUM accuracy model to correct the source and background counts? I'm noticing the fluxes I get from aperture photometry (corrected using the EXTENDED model) are significantly lower than those I get from the SAS detect chain, and wonder if this could be why. Cheers, -Olivia On Wed, 24 Mar 2004, Nora Loiseau wrote: > Dear Olivia, > > Please find attached the 90% enclosed energy plot for MOS-1, which is > different from that presented in the current UHB. The PN plot > in the UHB still represents current knowledge. > > There are many combinations of photon energy, off-axis angle, instrument > and encircled energy which could be plotted. We can not provide all > of these. The user can derive the specific plots that they want using > calview. > In this case the general recipe is: > > 1. Create a cif with cifbuild and make sure that it contains the files > XRT1_XPSF_0006.CCF, XRT2_XPSF_0006.CCF, XRT3_XPSF_0005.CCF > > 2. run calview > > 3. Select the required instrument > select "Accuracy Level" = EXTENDED > set "Energy (eV)" to required photon energy > set Theta to 0 arcseconds > > 4. View->Encircled_energy->Encircled_energy(PSF) > > 5. Read from the resulting plot the radius which contains the required > encircled energy fraction. > > 6. Set Theta to 120 arcseconds and then repeat etc.... > > We are sorry for the out of date documentation on the PSF, > our experts are working hard to bring it up to date. > > Regards, > > Nora > > > > > > > > -- > -- > Dr. Nora Loiseau, > XMM-Newton User Support Group phone: +34 918131298 > nloiseau@xmm.vilspa.esa.es FAX: +34 918131322 > >
From: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> To: cocj@roe.ac.uk Subject: Re: More PSF confusion (PR#10394) Date: Fri Apr 2 10:49:17 2004
Deal Olivia, medium accuracy model is used. Can you specify which method did you use for aperture photometry, and give us as many details as possible?. Cheers, Nora > One last related question: Does the SAS detect chain (eboxdetect -> > emldetect) use the EXTENDED or the MEDIUM accuracy model to correct the > source and background counts? I'm noticing the fluxes I get from > aperture photometry (corrected using the EXTENDED model) are > significantly lower than those I get from the SAS detect chain, and wonder > if this could be why. ---- Dr. Nora Loiseau XMM-Newton User Support Group
Date: Fri, 2 Apr 2004 12:43:19 +0100 (BST) From: Olivia Johnson <cocj@roe.ac.uk> To: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> Subject: Re: More PSF confusion (PR#10394)
Hi Nora, Thanks for the quick reply. Here's what I'm doing: I've run a variant of the CIAO wavdetect algorithm on the image which produces a smooth source-free image of the background, and a sourcelist which includes an estimate of source extent in pixels. Then: ap_counts = counts in a 2*4.35 pixel aperture from the counts image ap_bkg = counts " " " " " " from the background image ap_expo = counts " " " " " " from the exposure image I find the encircled energy in my aperture by interpolating the curve from the extended accuracy model, ee. Then: tot_counts = ap_counts/ee tot_background = ap_bkg * npix_source / npix_aperture where n_pix source is from the WD cell net_counts=tot_counts - tot_background exposure = ap_expo / n_pix aperture flux=counts/exposure/ecf I've attached a plot of my counts vs. the detect chain counts for the three detectors in one band. As you can see, I'm systematically low, and the SAS counts flatten out at low fluxes. I've checked that the WD background map is comparable to the splinefit one - it's alot smoother, but the level is consistent to counts equivalent to ~10^16 erg/s/cm^2. I've also just checked the discrepancy for the ee fraction between MEDIUM and EXTENDED accuracy modesl (I only checked for MOS), but this effects the net_counts at only at the 3-7% level for my aperture size. I imagine the trouble must be in the estimate of source size used to normalize the background counts. The WD algorithm defines source cells in a PSF-indepedent way, and perhaps this is larger than the assumed 100%EE radius for the SAS background count calculation? Any suggestions you have on this would be much appreciated. Thanks, -olivia On Fri, 2 Apr 2004, Nora Loiseau wrote: > Deal Olivia, > > medium accuracy model is used. Can you specify which method did you use for > aperture photometry, and give us as many details as possible?. > > Cheers, > > Nora > > > > One last related question: Does the SAS detect chain (eboxdetect -> > > emldetect) use the EXTENDED or the MEDIUM accuracy model to correct the > > source and background counts? I'm noticing the fluxes I get from > > aperture photometry (corrected using the EXTENDED model) are > > significantly lower than those I get from the SAS detect chain, and wonder > > if this could be why. > ---- > Dr. Nora Loiseau > XMM-Newton User Support Group >
From: Nora Loiseau <xmmhelp@xmm.vilspa.esa.es> To: cocj@roe.ac.uk Subject: Re: More PSF confusion (PR#10394) Date: Fri Apr 2 13:40:50 2004
Dear Olivia, this seems to be data analysis problem, not really a problem of SAS, and unhappily we do not have enough manpower for helping in that. Best regards, Nora ---- Dr. Nora Loiseau XMM-Newton User Support Group