Viewing Calibration/4719
Full headers

Private message: yes  no

Notes:

Notification:

Date: Tue, 26 Feb 2002 16:31:00 GMT
From: nevalaj@email.uah.edu
To: xmmhelp@xvsoc01.vilspa.esa.es
CC: nevalaj@email.uah.edu
Subject: CCD gap and bad pixel corrections

Full_Name: Jukka Nevalainen
Submission from: (NULL) (146.229.9.129)


What is the logic behind reducing the arfgen produced effective area by the
ratio of
the dead area within the source region to the good area? I mean, the effective
area 
for given energy is not proportional to the geometric area, but is a weighted
mean of
the effective area values of the pixels in the area. In the public data sets the
bad 
pixels are set to give 0 counts and thus in arfgen these pixels give no
contribution 
to the effective area. So I think the correct way is not to correct for the bad

pixels, as suggested and implemented in SAS. 

Reply 1

From: Matthias Ehle <xmmhelp@xmm.vilspa.esa.es>
To: nevalaj@email.uah.edu
Subject: Re: CCD gap and bad pixel corrections (PR#4719)
Date: Wed Feb 27 09:06:59 2002

Dear Jukka Nevalainen,

thanks for your question and remarks. I have contacted our arfgen SAS expert
and will let you know his anwer asap.

Cheers,
Matthias.

Matthias Ehle 
XMM-Newton SOC 
User Support Group

Reply 2

From: Matthias Ehle <xmmhelp@xmm.vilspa.esa.es>
To: nevalaj@email.uah.edu
Subject: Re: CCD gap and bad pixel corrections (PR#4719)
Date: Wed Feb 27 09:26:11 2002

Dear Jukka,

please, find below the reply from the arfgen s/w developer:
---
   The answer to this depends on how you want to use the resulting ARF.
If you have a point source you would like to know how much flux was
emitted by the source. To find this you have to factor in the point
spread function which tells you how much of the flux you have lost by
taking a source area of finite size. The chip gaps and bad pixels can
be thought of as part of the point spread function.

For an extended source, if you are interested in the total flux
emitted by the source then the above logic applies. If you want the
surface brightness, i.e. the flux per unit area, then it is not correct
to take account of the chip gaps or bad pixels because these are
already handled in the calculation which gives the area on the sky,
i.e. the BACKSCAL keyword. In this case withbadpixcorr=no should be
set on the arfgen command line.

This is a touch philosophical but if you think of it from the PSF
perspective it makes sense.
---
I hope this answers your question...

Best regards,
Matthias.