UVCS/SOHO Post-Sept.'97 LYA Flat Field
UVCS/SOHO post-Sept.'97 LYA flat field
Peter L. Smith, Nigel Atkins & Raid Suleiman
with help from L. Gardner, J. Kohl, R. Frazin, and others
The flat field data for the LYA detector "today" (i.e., after the voltage
reduction in Sept.97) that were displayed and discussed at the UVCS
Science meeting in Maine (September 1998) are available:
To download the file hold the shift key and click here
Let rsuleiman@cfa.harvard.edu know if you have any trouble retrieving
them.
The image below shows the whole detector.
The figures below show the relative response of columns 350:950.
Note that the dark current is not uniform, but is lower where
the detector efficiency is lower.
Some caveats:
- The "flat-fielding" in the spatial direction is based:
- on an observation of zeta Tau on part of the OVI detector,
- that is transferred to a broader part of that detector using
scattered disk light measurements,
- which are used to determine the spatial distribution of a
second set of scattered disk light measurements (the "source
function") on the OVI detector,
- that is the source function for scattered disk light
measurements that were made simultaneously on the the LYA detector.
This process could lead to small (5% perhaps) systematic errors
in the macroscale (i.e., over large areas) flat field in the
spatial direction.
- The flat-fielding in the spectral direction was done by observing
scattered disk light on three occasions:
columns (approx.) | binning | source function |
760:870 | 3 x 1
| measured
by the method discussed in (1) above |
610:780 | 3 x 1
|
measured, but the portion of the OVI detector
that was used seems to have deteriorated
by 8% between the two sets of measurements.
Therefore, this region was flat-fielded by
comparison to 760:870 in the region of
overlap. |
415:640 | 6 x 1
| not measured;
flat-fielded by comparison to
610:780 in the region of overlap. |
Again, there is a possibility of small (5% perhaps) systematic errors
in the macroscale flat field. Estimation of these is problematic.
- The edge effects (columns ~420 and 875; row ~35) are actifacts. They
do not indicate actual performance of the detector at these locations.
statistical thoughts
We sampled two "undamaged" portions of the detector.
rows | columns | bin size
(rows x columns) | bin-to-bin standard deviation |
100:270 | 459:510 | 6 x 1 | 11% |
100:270 | 798:854 | 3 x 1 | 8% |
The statistical uncertainty of the data is estimated to be +/-3%.
last update 3 February, 1999, by R. M. Suleiman