Changes in the upcoming LWS OLP Version 8
Improvements to glitch checking. SPL currently checks the detector
ramps for glitches. Before the glitch checking is performed, "bad" points
are discarded from the ramps. However, many of these points, while not
suitable for processing into SPD data, are good enough to use for checking
for glitches. In fact, in some cases they are required in order to detect
the glitch in the first place. For example, a large glitch may cause the
detector voltage to "hit the rails" for the remainder of the ramp. If these
points are discarded then the ramps appears normal and no glitch is detected.
The solution implemented allows the set of points checked for glitches
to be different from the set of points used for the determination of the
photocurrent for the ramp. The set of points used for glitch checking will
be the raw data points with only the portion of the ramp affected by the
reset removed. This update should result in more glitches being correctly
Addition of data following a glitch. SPL currently discards a fixed
number of ramps following each glitch, since these ramps are often affected
by the glitch. However, it would be useful to have access to this discarded
data for "expert" users to be able to examine.
The solution implemented is to include the discarded data in the
LSPD file, but to flag the data as "invalid" within the status word. An
additional flag within the status word will indicate that the data has
been discarded because it is following a glitch. These data will NOT be
copied into the LSAN file to avoid any observer using them as though they
are valid data. SCREW-396
Addition of mini-scan counter. For L03 observations the line number
fields in the LSPD and LSAN files do not contain useful information. For
OLP 8 the "mini-scan" number is written in place of the line number. The
mini-scan number is a counter which is incremented for each different grating
position used. This is effectively a "line" number, so it is logical to
write this into the line number field. SCREW-396
Fit to problem with drift correction. The correction for the drift
in responsivity is currently being performed at the wrong stage in AAL.
The drift correction is being performed at the end of AAL, while the theory
behind the drift correction requires it to be done at the start. In practice
this will only make a difference for observations with very low signal
levels, comparable to the dark current.
As part of the solution to this problem, the LSNR file will no longer
be generated by AAL. This file contained intermediate results, before the
application of the responsivity drift correction and absolute responsivity
correction. With the drift correction being performed at the start of the
pipeline there is little point in the generation of this file. SPR-3037.
Use of standard dark currents for FP observations. For FP observations
it has been found more reliable to used standard dark current values, rather
than the values measured during the surrounding illuminator flashes. The
solution involves the creation of a new calibration file named LCDK. The
LCDK file will be a binary extension FITS file containing the dark current
values for each detector. In case of future need, the file can contain
different dark currents for different ranges of revolution numbers. SCREW-396
Updated flux uncertainty calculation. The uncertainty value inserted
into the LSANFLX field of the LSAN file has been updated to make it a more
meaningful value. The uncertainty value is now calculated using a new algorithm,
as specified in the note "Error calculation in LWS offline processing"
by Bruce Swinyard. SCREW-396
Improved FP processing. In previous OLP versions the LWS Fabry-Perot
data were converted from photocurrent to flux by using the grating position
to identify a particular calibration curve in a look up table. Owing to
the inaccuracy of the grating position sometimes the wrong FP position
was used and no calibration curve was found. This problem has proved intractable
and errors continued to occur in OLP V7. An interactive analysis routine
has been written to cure the problem but this is not sufficiently robust
to be reliably automated. It has therefore been decided to stop attempting
to remove the signature of the grating resolution element from the pipeline
products and leave this aspect of the processing to IA.
In addition to this change, a new method of calibrating the total
throughput of the FP's has been devised using Mars as the calibration standard.
This method uses a fitted polynomial between the wavelength and the FP
throughput and allows the FP LSAN products to be put into W cm-2 um-1 for
the first time. All L04 and L03 AA products will be in these flux units.
The spectra will still have the relative change in flux over the grating
resolution element present for the reasons given above - this can be removed
by the use of LIA.
This update uses a new calibration file name LCTP. This consists of
a FITS header only. The following calibration files will no longer be used:
LCLRZ, LCSRZ, LCLR_n (n=0..9), LCSR_n (n=0..9).
SCREW-396. This update also fixes SPRs 2853 and 2855, which were caused
by the problems described above.
Update to GR spectral responsivity warning flag. The status word
in the LSAN file contains a flag which warns that the grating wavelength
is outside of the nominal calibration range in the LCGR file. L03 and L04
observations may use such grating positions. For L03 observations and for
the active detector on L04 observations this warning flag is not appropriate
and will no longer be set.
SCREW-396. This update also fixes SPR 2856, where all of the data
for detector 8 had the warning flag set.
Updated AAL calibration files. The LCIR and LCGR files have been
updated using an improved model of Uranus. SCREW-396
This information was kindly provided by Andrew Harwood
of the LWS Instrument
Team at the Rutherford Appleton Labs, UK