Hi Jane,
On Jan 12, 2009, at 7:20 AM, Jane Buckle wrote:
> The files can be picked up (briefly) from
>
> http://www.mrao.cam.ac.uk/~buckle/
> a20071123_00051_01_0001_test.sdf.gz --the cleaned, trimmed time
> series file
> http://www.mrao.cam.ac.uk/~buckle/
> a20071123_00051_01_reduced001.sdf.gz -- the oracdr pipeline output
It turns out the problem is in timesort. Using the first file, if you
do:
% ndfcopy in=a20071123_00051_01_0001_test'(~830,,)' out=51_trim
% timesort in=51_trim out=51_trim_ts sizelimit=! merge=true
genvar=true \
detpurge=true detectors=\'-H14,H03,H12\'
And then display 51_trim_ts_1 in GAIA, you get the line coming out
somewhere around 90km/s. Every spectrum is shifted by 98 channels, but
the WCS stays the same so you get a shift in the line.
The problem is that TIMESORT resets the origin back to 1,1,1. The
input NDF has an origin of 98,1,1, which explains the shift by 98
channels (or 97, I guess).
This would be one for David to fix. Sorry David!
Cheers,
Brad.
|