Date: Wed, 25 Jun 2003 16:29:02 +0100 (BST)
From: Richard Saxton <email@example.com>
To: Bruno Altieri <firstname.lastname@example.org>
cc: Marcus Kirsch <email@example.com>,
Subject: Re: Some issues about EEF/PSF in CCF/UHB
the PSF issue is very confusing because of the need to keep the
CAL backwards compatible. In this way old, obsolete, algorithms are
maintained alongside the newer methods. Really we should drop the
LOW method completely and also remove the XENCIRCEN CCF which has
been totally superceded by the EXTENDED mode PSF.
> Comparing the optical PSF with the x-ray one (to investigate the effect
> of optical loading), I came across the following issues :
> - why the UHB states HEW of 13.8 and 13.0 arcsec @1.5keV at :
> while the Ghizzardi found 8.6" and 9.1" arcsec respectively ??
> This makes a big difference. My guess is that the ground values have
> been left in by error while the plots below have been updated.
The numbers in the UHB have been taken from an early Aschenbach et al.
paper. The HEW they quote is the 50% radius multiplied by two. The
equivalent numbers for the King model are 17.2 and 18.2 arcsecs but these
have been calculated assuming that the PSF is fully contained within
a radius of 5'. The Aschenbach work assumed that 100% was contained
within ~2 arcmins which is probably enough to account for the difference.
I dont think that this is very important the numbers quoted give a good
estimate of the real situation.
> - what can one trust from the XRT?_XENCIREN_NNNN.CCF ?
> Its latest issue (3) is not documented in terms of CCF release note (only
Issue 3 of the XENCIREN files was documented in XMM-CCF-REL-84. These
files are obsolete though and have been superceded by the XPSF files.
They are kept for backwards compatibility reasons.
> - the use of calview on both of these CCFs is VERY confusing,
> while I am supposed to promote calview at the SAS workshops !
> "view" --> "encircled enregy" pops you to i) "encircled energy (PSF)"
> and ii) "encircled energy (CCF)"
> - "encircled energy (PSF)" access the XPSF CCF, while its name
> tends to indicate that only the second would give info about the CCF,
> and it plots an EEF and not a PSF.
I agree that the distinction of CCF/PSF in the calview window isn't
very helpful, we should think of better names. "CCF" means the values
stored in the XENCIREN files (which are now obsolete). "PSF" means
'calculate the encircled energy function' using the mode (low, medium,
high,extended) chosen by the user in the calview screen.
> The LOW model is complietely wrong ! (dashed lines in attached plot)
True - the LOW mode is useless.
> Plots with others accuracy are sometimes odd,
> with some ripples. Is one of these plots using the CCF King's
> parameters or the model actually used by the SAS ??
> - "encircled energy (CCF)": there is no difference at at all between
> the 4 accuracy levels "LOW, MEDIUM, EXTENDED, HIGH"
> what do they mean then ?
> Why the EEF is so different above 45" ? (XPSF -> continuous line,
> XENCIREN --> dotted line)
> - why all these 512x512 PSF images in XRT?_XPSF_NNNN.CCF for ?
> that yield to sizes of 68 Mb per CCF ! just to be "calview-ed" ?
The images are the MEDIUM mode PSF. They have been produced by
SCISIM are are used directly by the CAL to derive a 1-d PSF or
encircled energy function.
Your questions point to the fact that PSF use within the CAL needs to
be rationalised. The only truly useful models are EXTENDED mode for
calculation of the encircled energy fraction in the case of
circular, annular or rectangular boxes, MEDIUM mode for 2-d PSF
use and HIGH mode for calculation of the encircled energy function for
arbitrary spatial regions.