While analyzing older LETG/HZ43 observations to compare against
the HV test data I noticed the latest (ObsID 13025) had a distorted
0th order (see below). Next to it is 0th order in CHIP coords.
The dither covers crsv=98 (left) to 100 (right). You'll notice the
dither trace on crsv=100 seems smeared out. Below I plot the
PSFs in tg_d (cross-disp) and tg_r (dispersion axis, along crsv),
filtering on different CRSV values (splitting 99 in half). You can see
that the PSF for crsv=100 is offset and quite broad.
Here's the sky image of 0th order using only events with crsv=98:99 (left)
and then using only crsv=100 (right).
The green circle is centered on the same pixel in each.
This problem started to show up around 3 years ago when, because of
aimpoint drift, the dither pattern started to move onto crsv=100.
Here are the PSFs of all longish LETG/HRC-S HZ43 calibration observations
to date, along with CHIPY distributions for the most recent four:
So far (though I have yet to look more closely) this problem has significantly affected one GO observation--he was about to publish a paper claiming a hitherto unknown companion star, but we caught it in time (thanks to Margarita K). I would assume several obserations have this problem but the GOs usually don't care about 0th order. We need to study this more but it seems like a good idea to change the default Yoffset by around 20" and avoid this tap.