Release Note on SWS OLP Version 8.4

In August 1999 a new version of the SWS OLP (pipeline) for automatic data 
reduction was released to be part of the new 

                    "On the Fly Reprocessing".


You can only obtain these data products by retrieving data from the IDA. By 
default the `Latest Processing' button is selected 


Improvements in the data reduction over OLP 7 are:

1.	Wavelength accuracy improved
During the mission it was discovered that the relation between the measured 
position of the grating (called LVDT) and the wavelength was shifting steadily 
at a rate of about 1 LVDT per 6 months at most. This was corrected by 
re-deriving the relation every so often. As the shift itself was steady, this 
procedure resulted in a very low amplitude sawtooth-like behaviour of the 
wavelength calibration in time. In OLP 8.4 a linear interpolation has been 
implemented to remove this saw-tooth. The wavelengths are now steady to well 
within one LVDT, which corresponds to a fraction of a resolution element. The 
main source of error in the wavelength is (as it was before) the pointing 
error, corresponding to a few LVDTs.


2.	Error information
	Two spare bytes in the AAR product (file swaa*.fits) are used to code 
the errors on the flux values, and a function DECODE_AA_ERROR will be inserted 
into the next release of OSIA.
	The two spare bytes of the AAR (SWAASPAR, see the SWS Instrument and
Data Manual section 9.4) are now used such that byte 0 contains the error
in the offset of the slope, in Jy, whereas byte 1 contains the gain error as
a percentage. These can be used to generate the 1 sigma error limits on the
spectrum by:
( spectrum - SPARE(0) ) * ( 1 - SPARE(1) )    : data - 1 sigma
( spectrum + SPARE(0) ) * ( 1 + SPARE(1) )    : data + 1 sigma


3.	SPDs and off-band data
	The SPD and AAR products of OLP 7 only contained data from a 
wavelength within one of the several SWS bands, each band defined by
a detector/aperture/order combination - for a list of these bands see
Table 3.1 of the SWS Instrument and Data Manual. The calibration of data
taken outside these bands was not guaranteed.
	In OLP 8.4, off-band data (i.e. data coming from outside these bands) 
will be present in the SPD product. This is because it is sometimes useful to 
have an overlap region between bands when comparing data. This off band data 
will not be propagated through to the AAR product.
	Note that the next release of OSIA, version 2.0, will support off band 
data.


4.	Post-Helium observations S95, S96 & S97 are now processed to SPD 
level and are present in the archive. The wavelength calibration for these 
observations, however, is not done correctly and care should be taken if they 
are retrieved from the archive. Contact helpdesk for further information.


5.	Changes to product contents
5.1	SWSPSTDV, containing the standard deviation in the SPD file (see the 
SWS Instrument and Data Manual section 9.3) now has units of microVolt/second.
The SWSPSTDV tag has not been altered at all, it is still the standard 
deviation of the slope.
5.2	SWAATINT now contains the number of seconds used to derive the AAR 
datapoint. Previously it contained the number of samples used to derive the 
AAR datapoint.


K. Leech, D. Kester, B. Vandenbussche, E. Wieprecht
31/8/99