Up to top level
AO15   AO16   AO17   AO18   AO19   Backgrounds   Calibration   Conference   Data   Docs   EPICMOS   EPICpn   Feedback   Gallery   Misc   OM   Pending   PhD_Theses   Publications   RGS   RadMonitor   SAS_Hardware   SAS_WS   SASv16.0   SASv16.0_Installation   SASv16.1   SASv16.1_Installation   SASv17.0   SASv17.0_Installation   SASv18.0   SASv18.0_Installation   SciSim   Simulators_other   Suggestions   Trash   Visibility   XMM-bouncing   XMM-news   XRPS   XSA   esas   incoming  

Logged in as guest

Viewing Backgrounds/8446
Full headers

From: gfossati@rice.edu
Subject: epic background level and sensitivity
Compose reply
Download message
Move To:
2 replies: 1 2
0 followups:

Private message: yes  no

Notes:

Notification:


Date: Tue, 22 Apr 2003 00:23:16 GMT
From: gfossati@rice.edu
To: xmmhelp@xmm.vilspa.esa.es
CC: gfossati@rice.edu
Subject: epic background level and sensitivity
Full_Name: Giovanni Fossati
Submission from: (NULL) (128.42.10.124)


Hi, I was trying to make some back of the envelope estimates about the 
EPIC sensitivities, based on the numbers of the Watson et al paper referenced
in the UHB, but I can't reproduce their numbers, reported in their fig.3 (also
reported in the UHB).
I have tried different scenarios for the "integration" area, namely just one PSF
FWHM, or the 50% encircled energy radius, or the 90% encircled energy radius.
For source rates I am using flux-rate conversion for pn and mos thin from the
latest pimms.
Well, even in the best case (one PSF FWHM, but retaining all pimms counts...
which is
unlikely that will fall within on PSF FWHM) I get exposure times for a 5-sigma
detection that are at least a factor of 2 larger than Watson's plot.
For more reasonable "extraction radius" (e.g. the 50% radius) the difference
gets even larger.  They end up way above their curves.

Assuming pimms rates are correct (and I don't have a reason to doubt it) is it
just that Watson et al used more generous estimates of the flux-count rate
conversion? (after all the paper is from 2001).

Are there updated values for the background rates, to use instead of Watson's?
I have read the paper by Read and Ponman, for instance, where they give
estimates
for the photon and particle bkg rates: for MOS thin their rates are comparable,
or better than Watson's, but for PN-thin they are worse by a factor or 2-3.

Anyway, aside from recommending to use SciSim, or whatever advanced tool, since
that would be an overkill... I would appreciate if you could comment on this
"problem", that I think may be interesting for many people "out there".
Moreover, since feasibility is a pretty important piece of the proposal writing,
and selection, it would quite important to ensure up-front a certain uniformity
in what numbers people get and report.

It would be great if you could provide some guidelines as to what numbers to use
for this simple estimates (including for instance what "extraction radius" would
be the most sensible choice).

Thank you very much.
Take care,

  Giovanni

------------------------------------------------------------------------------
 Dr. Giovanni Fossati                            +---------------------------+
 Dept. of Physics and Astronomy, MS 108          | Ph    : (713)-348-3425    |
 Rice University                                 | Fax   : (713)-348-5143    |
 6100 Main street, Houston, TX 77005, USA        | e-mail: gfossati@rice.edu |
-------------------------------------------------+---------------------------+



Reply 1

Resend
From: Matteo Guainazzi <xmmhelp@xmm.vilspa.esa.es>
To: gfossati@rice.edu
Subject: Re: epic background level and sensitivity (PR#8446)
Date: Tue Apr 22 19:47:33 2003
Dear Giovanni,

 I have spent some time with your question and your detailed analysis.
I apologize in advance if my answer is not complete, but I would not
like to leave you, waiting for the pieces of information, that I can
provide you with in the meanwhile.

 I think that your questions can be summarized as follows:

1. is PIMMS reliable?

 The answer is, as far as we know, yes. PIMMS uses effective area
files which are admittedly a bit old (~18 months). However, our
estimate of the effective area in the EPIC cameras has not changed
by a factor 2-3 or larger in the last 1.5 years! Therefore, PIMMS
cannot be considered fully reliable, for the level of accuracy that
such a tool requires

2. is the background analysis in the Read & Ponman (2003) paper
"more updated" than in the Watson et al. (2001) and in the UHB?

 The answer is, obviously, yes. Their results - please note that
the paper is still "submitted" and was uploaded to astro-ph earlier
this month - were not available for inclusion in the version of the
User Hanbook published for the AO3. Does a user make something
wrong by using/quoting their results? This is up to the OTAC
to decide of course, but as an individual scientist I do not
see anything wrong with it, as their results are now public.
Still, the information reported in the User Handbook are
considered the official reference for the AO3.

 Just as side note, there are other detailed considerations that
one should take into account when dealing with background
issues. For instance, the dependence of the soft X-ray cosmic
diffuse background on the sky position - as shown by the ROSAT map by
Snowden et al. (1997). A look at the paper by Lumb et al.
(2002; A&A, 389, 93) may be useful in this respect as well.
Other caveats are reported at Page 47 of the User Handbook,
as you surely know

3. is there any general guideline a user should follow to estimate
which extraction region one should consider etc. etc.?

 The details of the analysis cannot be pre-determined in advance
for the most general user, as they depend on the specific scientific
problem to be addressed. Even limiting ourselves to the case of
point-like sources, an extraction region encompassing a large
fraction (i.e. 90%) of the encicled energy fraction is advisable.
As you probably know, in SAS v5.4.1 the implementation of the tool
"egetspec" within "xmmselect" contains an optimization algorithm,
which calculates the optimal extraction region on the basis of the
best signal-to-noise ratio. On the other hand, "eboxdetect", for
instance, uses by default pretty small boxes. Again, I believe that
no users makes an error is s/he uses and applies the criteria s/he
deems better for the specific scientific problem s/he is dealing with.
If s/he prefers to use "reference" criteria, these criteria are to
be found in the SAS documentation (i.e. the User Manual) and in the
User Handbook

4. why do the exposure times derived from Fig.37 of the UHB and
using an estimate of the source count rate from PIMMS differ so
much?

 This is still under investigation. I will come back to you as
soon as I have a useful answer for you.

 Please, do not hesitate to ask further if my answers to the point
1. to 3. above need any further clarifications/integrations.

 Regards, Matteo


Reply 2

Resend
From: Matteo Guainazzi <xmmhelp@xmm.vilspa.esa.es>
To: gfossati@rice.edu
Subject: Re: epic background level and sensitivity (PR#8446)
Date: Wed Apr 23 07:18:23 2003
Dear Giovanni,

 concerning the 4th point of you message, I have to ask you for a
clarification, as I am not sure I am exactely reproducing what you
are doing. Please, do not hesitate to tell me what I am doing
differently from you, and therefore to contradict my conclusions:

1. you choose a reference flux (i.e., 10^{-14} cgs in the, e.g.,
2-10 keV band) and a source count extration radius (e.g. 50",
corresponding to an encircled energy fraction ~90%). Let's
call A the corresponsing area

2. you run PIMMS - using a power-law model with \Gamma=1.7 and
N_H=3x10^{20} cm$^{-2} - and get the total source count rates
for a source with that spectral shape.

3. you correct these counts for the encircled energy fraction
(90%)

4. you calculate the expected background count rate in 50",
from the background flux (B) in the UH

5. you estimate the time (T) necessary for a 5-sigma detection
under the assumption of "purely Poissonian statistics", i.e.:
T~25(BxA)/S^2

 If you are doing the above, this implies that you can tune
the extraction area, in order to match your estimation of T
with the results of the Watson et al. (2001) paper. If you
try this exercise, e.g., for MOS and pn in the 2-10 keV band,
the extraction radius you get is ~15-17". This is in agreement
with the qualitative statement that the sensitivity just for
detection is maximised at some small value of PSF enclosed.

 Please, let me know what I have misunderstood of your whole
procedure.

 Regards, Matteo

Up to top level
AO15   AO16   AO17   AO18   AO19   Backgrounds   Calibration   Conference   Data   Docs   EPICMOS   EPICpn   Feedback   Gallery   Misc   OM   Pending   PhD_Theses   Publications   RGS   RadMonitor   SAS_Hardware   SAS_WS   SASv16.0   SASv16.0_Installation   SASv16.1   SASv16.1_Installation   SASv17.0   SASv17.0_Installation   SASv18.0   SASv18.0_Installation   SciSim   Simulators_other   Suggestions   Trash   Visibility   XMM-bouncing   XMM-news   XRPS   XSA   esas   incoming  

Logged in as guest


Please make your (short) question the subject of your request!


Web interface using JitterBug ... back to the XMM home page