|
|
Suggested methods for dealing with data contaminated by solar X-rays
If the angle between the LAC pointing direction and the Sun is
less than 90° then the data may be contaminated with solar
X-rays. In this case there are two possible ways to proceed with the
analysis:
- Sort the raw data as normal, using sortac. Then run
select to extract only the data with the SUNSHINE flag set to
zero (the lower and upper limit on SUNSHINE are set to cover zero but
not one). This will restrict the observational intervals to those in
the Earth's shadow. Whilst this may lead to a large reduction in the
useful length of an observation, it does ensure that the end result is
free of solar contamination. An example of this approach is
illustrated below:
qcl>select
** SELECT Version 2.4 Tue Jun 17 15:08:21 1997 **
SEL_DATIN Input data filename =mkn335_91a.cube
SEL_OPT Extract a TIME / PHAD / CUBE file? = PHAD =cube
SEL_DATOUT Output data filename = mkn335_91a.cube =mkn335_91a.cube_ex
SEL_NPHA How many PHA ranges? = 48 =
SEL_NUM_SEL Number of HK parameters to select on (-1 = default list)
= 1 =
SEL_HK_NAME Name of HK parameter =sunshine
SEL_WINDOW Normal or inverse window = Norm =
SEL_PAR_LO Lower limit on SUNSHINE = 0.0 =-0.2
SEL_PAR_HI Upper limit on SUNSHINE = 0.0 =0.02
SEL_NUM_PER Number of time windows = 1 =
SEL_ST_PER001 Start time of window - secs from begining = 0.0 =
SEL_END_PER001 End time of window - secs from begining = 50304.0 =
DATIN mkn335_91a.cube_cl
Shape : 393 x 48 x 2 = 37728
TIME PHA IDENT
Time binning is 128.0000 secs
... searching 393 time bins from range of 393
ID_MASK( 2)
0101010101010101
1010101010101010
OUTPUT mkn335_91a.cube_ex
Shape : 393 x 48 x 2 = 37728
TIME PHA IDENT
11 time bins are occupied
qcl>
- An alternative approach, that works for MPC1 data, makes use of the fact that certain LAC detectors may be shielded from the Sun by the body of the spacecraft and will therefore not be sensitive to solar X-rays. To see if this is the case, the user may proceed as follows:
- Run sortac with the parameter GNG_SQUISH set to
"NO". This will keep the data from the eight different
detectors separate.
- Run select as above to restrict the observational
intervals to those in (a) the Earth's shadow (SUNSHINE = 0) and (b)
the Sun (SUNSHINE = 1). This produces two spectral files with full
detector information. Setting SEL_WINDOW = Inv selects data when the
sunshine flag is one rather than zero.
- Subtract the two spectral
files using subcb:
qcl>subcb
** SUBCB Version 1.0 Tue Jun 17 15:33:37 1997 **
SCB_DATIN1 First input data file = =mkn335_91a_sun.phad
SCB_DATIN2 Second input data file = =mkn335_91a_nosun.phad
SCB_DATOUT File containing FILE1-FILE2 = mkn335_91a.phad
SCB_DATIN1: mkn335_91a_sun.phad
Shape : 1 x 48 x 16 = 768
TIME PHA IDENT
IDENT masks
1 1000000000000000
2 0100000000000000
3 0010000000000000
4 0001000000000000
5 0000100000000000
6 0000010000000000
7 0000001000000000
8 0000000100000000
9 0000000010000000
10 0000000001000000
11 0000000000100000
12 0000000000010000
13 0000000000001000
14 0000000000000100
15 0000000000000010
16 0000000000000001
SCB_DATIN2: mkn335_91a_nosun.phad
Shape : 1 x 48 x 16 = 768
TIME PHA IDENT
IDENT masks
1 1000000000000000
2 0100000000000000
3 0010000000000000
4 0001000000000000
5 0000100000000000
6 0000010000000000
7 0000001000000000
8 0000000100000000
9 0000000010000000
10 0000000001000000
11 0000000000100000
12 0000000000010000
13 0000000000001000
14 0000000000000100
15 0000000000000010
16 0000000000000001
DATIN2 is a PHAD. Subtracting DATIN2 from each PHAD in DATIN1.
qcl>
- Plot the Sun minus Earth shadow spectrum (using the qcl
procedure pl_phad.qcl; see section
6.2). If solar contamination is significant in a particular
detector then there will be excess counts in the subtracted spectrum
below 5 keV. Identify detectors with no excess. Note that only
even detectors need be plotted as they refer to the top-layer; the
mid-layer is insensitive to genuine X-rays (and hence solar X-rays)
below ~5 keV.
- Run hyper on the original, uncompressed cube (i.e.
the cube with no sunshine selection) to select only the detectors with
no soft excess in the sunlit data:
qcl>hyper
** HYPER Version 1.2 Tue Jun 17 15:49:51 1997 **
HYP_DATIN data file name =mkn335_91a.cube
HYP_DATOUT new data file name = mkn335_91a.cube_sq =
Shape : 393 x 48 x 16 = 301824
TIME PHA IDENT
IDENT masks
1 1000000000000000
2 0100000000000000
3 0010000000000000
4 0001000000000000
5 0000100000000000
6 0000010000000000
7 0000001000000000
8 0000000100000000
9 0000000010000000
10 0000000001000000
11 0000000000100000
12 0000000000010000
13 0000000000001000
14 0000000000000100
15 0000000000000010
16 0000000000000001
HYP_OPTION TIME, PHA, or IDENT dim. =ident
HYP_NCHOUT no.of output ident chans = 1 =2
HYP_COMB001 chan combination = 1-16 =2,4,6,8,10,12,14,16
HYP_COMB002 chan combination =1,3,5,7,9,11,13,15
New ID channel 1 : ID chans 2,4,6,8,10,12,14,16
New ID channel 2 : ID chans 1,3,5,7,9,11,13,15
Squished file
Shape : 393 x 48 x 2 = 37728
TIME PHA IDENT
qcl>
HYP_NCHOUT and HYP_COMB are the number of output detector idents
and ident combination respectively. Even and odd idents refer to the
top- and mid-layers respectively.
It will not usually be possible to perform an universal
background subtraction (see section
3) on the resulting cube as the background model coefficient files
generally assume that all 8 LAC detectors are present (the LAC
background is not identical in every detector). It will be necessary
to create a local background cube using the same detector combination
selected for the source cube.
Further details regarding the programs select and
hyper are given in section
5.
[Top of page]
[Next: HK Data]
[Previous: Telemetry]
[Back to Data Analysis]
|
