[Date Prev][Date Next][Thread Prev][Thread Next][Search]
[Main Index]
[Thread Index]
[HEASARC Archives]
The effect on SIS data of RDD and RBM_CONT
Dear ASCA PIs and Archival Researchers:
This message contains important information about:
1. The effect of RDD (Residual Dark Distribution) on SIS data
2. The dependence of SIS background on RBM_CONT (count rate in the
Radiation Belt Monitor)
These two items can also be found on the ASCA GOF WWW homepage, the URL
of which is:
http://heasarc.gsfc.nasa.gov/docs/asca/ascagof.html
Please look under "SIS News".
Koji Mukai & Charles Day, ASCA GOF
Comments and questions, please, to ascahelp@athena.gsfc.nasa.gov
------------------------------------------------------------------------
1. THE EFFECT OF RDD (RESIDUAL DARK DISTRIBUTION) ON SIS DATA
Recent investigations by the SIS team (Rasmussen, as well as Yamashita
and Dotani) have revealed the overriding importance of RDD.
1.1 What is RDD?
CCDs always have read-out noise: pixels without photon or particle
events will have a mean "dark level" and some distribution around this
mean level. This is usually treated as a Gaussian distribution. In the
case of ASCA, the on-board processor maintains a coarse map of the dark
levels (each chip is divided into 16 by 16 regions by default) and
subtracts these from individual pixel values. The dark frame map is
calculated assuming a symmetric noise around the mean dark level.
Unfortunately, the cumulative effects of radiation damage have created a
population of active pixels which skews this distribution, leading to
imperfect dark level subtraction on-board. This effect is called the
Residual Dark Distribution, or RDD.
Charges created by X-ray photons may be recorded in a single pixel or in
two or more neighboring pixels. When a significant charge is detected in
a pixel, its neighbors are examined for significant charges. If only the
central pixel has a significant charge, its value is recorded and the
event is assigned a "grade" of 0. If some of the neighbors have
significant charges, then these are added to that of the central pixel,
and a higher grade is assigned to the event. The calculation of the
final PHA value and the grade is done on-board for Bright mode data, and
on the ground for Faint mode data. The latter gives a finer control over
the process but cannot remove all the effects of RDD.
The effect of RDD is to impair the ability of the on-board processor to
assign the correct grade to an event. For example, a grade 0 event may
be recognized as a higher grade event because a neighboring pixel is
very active. For higher grades, this effect is worse and can even result
in a true photon event being classified as grade 7, i.e., as a rejected
particle event. Thus the effect of RDD is a complex interplay of the
analog electronics, on-board digital processing and off-line processing,
_and_ involves the intrinsic randomness of these active pixels.
RDD is clocking-mode dependent: the longer you integrate, the bigger RDD
becomes. Therefore, it is most pronounced in 4-CCD mode (16 s
integration), and less important in 1-CCD mode (4 s integration).
1.2 The Bottom Line:
* Grade branching ratios are changing (i.e., the proportions of events
in various grades).
* The Quantum Detection Efficiency (QDE) of SIS is degrading (because
real events are being rejected).
* The spectral resolution is degrading (because of higher noise and
because there are apparently fewer events at the lower grades).
* The energy scale is changing.
* 4-CCD mode is affected by RDD much more than 1-CCD mode.
1.3 Quantitative effects of RDD
All these effects are strong in 4-CCD mode, while 1-CCD mode remains
little affected. Here are some numbers, using the current understanding
of the RDD effect as parameterized by the SIS team.
---------------------------------------------------------
| QDE (as approximate percentage of nominal) |
|--------------------------------------------------------
| 2 years after launch 3 years after launch |
| |
| 4-CCD 55-70 25-40 |
| 2-CCD 90-95 70-80 |
| 1-CCD 100 95 |
---------------------------------------------------------
---------------------------------------------------------
| SPECTRAL RESOLUTION (at 6.4 keV in eV) |
|--------------------------------------------------------
| 2 years after launch 3 years after launch |
| |
| 4-CCD 150 --- |
| 2-CCD 60 80 |
| 1-CCD --- 20 |
---------------------------------------------------------
---------------------------------------------------------
| ENERGY SCALE OFFSET* (in eV) |
|--------------------------------------------------------
| 2 years after launch 3 years after launch |
| |
| 4-CCD 100 200 |
| 2-CCD 15 40 |
| 1-CCD --- < 5 |
---------------------------------------------------------
* The RDD-induced energy scale offset is independent of energy
1.4 Recommendations for dealing with RDD
FOR UPCOMING OBSERVATIONS: The GOF recommends that the use of 4-CCD mode
be discontinued and that 2-CCD mode be used with caution.
FOR EXISTING DATA: The calibration effort continues. SIS team members
are experimenting with various approximations to generate better and
time-dependent response matrices. Watch this space for new releases.
FOR AO-4 PROPOSALS: Those proposing to observe extended sources should
perhaps consider focusing on the GIS - its performance has shown almost
no detectable degradation.
------------------------------------------------------------------------
2. THE DEPENDENCE OF SIS BACKGROUND ON RBM_CONT
2.1 What is RBM_CONT?
In his study of dark Earth data, Keith Gendreau of the SIS team has
found that the SIS internal background depends on the mkf parameter
RBM_CONT (the radiation belt monitor count rate).
Currently, we use COR (Cut-Off Rigidity) in our default screening. This
is a geometric parameter which predicts the actual particle background
well. RBM_CONT, on the other hand, is an in-situ measurement: during SAA
passages, RBM_CONT values can be above 10000 cps. Outside SAA, RBM_CONT
averages near 1 cps at COR=14 GeV/c and 10 cps at COR=4 GeV/c, with
frequent episodes of RBM_CONT in the 100-5000 cps range at COR<6 GeV/c.
Even excluding the SAA passages, however, there are persistent
occurrences of moderately high (~several hundred cps) RBM_CONT when COR
is in the 7-11 GeV/c range.
SIS dark Earth data collected with the conditions COR>6 (default) and
RBM_CONT>50 (currently not used in default selection) show enhanced
detector background, particularly in the Al K-alpha line at 1.49 keV (NB
normally this line is undetectable due to cosmic X-ray background). We
intend to carry out further studies after which we will determine
whether to keep the default screening as is, replace with an RBM_CONT
criterion, or use a combination. Such a combination may yield the
optimum data selection.
2.2 RBM_CONT and your data
In the meantime, GOs can investigate how the count rate of their data
depends on RBM_CONT and COR by using the "mkfbin" and "plot mkf"
commands in XSELECT (see example below). This is especially recommended
for GOs interested in weak features around 1.5 keV. GOs can screen their
data based on RBM_CONT by using the "select mkf" command in XSELECT
manually or by using the "-q" option of ascascreen. With this option,
ascascreen will produce three files: <name>.xco, <name>_obscat.sel and
<name>_mkf.sel (where <name> is the output file name you entered). Edit
the <name>_mkf.sel file and add something like:
&& RBM_CONT<100
(and optionally remove "&& COR>6"). Then type "xselect @<name>".
2.3 An example plotting RBM_CONT, COR and SAA in XSELECT
test-ASCA-SIS0-BRIGHT > mkfbin
> Enter list of mkf parameters to bin >[ELV BR_EARTH S0_PIXL1] SAA COR RBM_CONT
> Give print-out time interval >[32] 32
test-ASCA-SIS0-BRIGHT > plot mkf
The available parameters are:
----------------------------
0 - TIME
1 - SAA
2 - COR
3 - RBM_CONT
> Enter independent variable ( 0 for TIME ) >[0] 0
> Enter dependent variables (e.g. 1-6) >[1-3] 1-3
PLT> win 4
PLT> log y
PLT> r y 0.1 1e5
[In plot mkf, n-th quantity is plotted in "window n+1" (in this case,
RBM_CONT is plotted in "window 4"). RBM_CONT panel is useless unless you
change the y scale to logarithmic; 0.1 and 100000 are good lower and
upper limits.]