Date: Wed, 25 Nov 1998 09:42:42 -0800
From: Russell E Owen
Subject: Re: 3.5-meter night log 981124
>     ...All was running smoothly and we figured out the "Horizon"
> rotation questions with help of Russ Owen and Gordon Richards. The guide
> star walks across the frame, and this shows up as guider offsets normal
> in X but huge in Y. These are taken into account and will not affect the
> guiding, though it may look strange on the guider display. I also found
> that due to this walking the guide star must be fairly bright, to allow
> short integrations, or else the star is too smeared out to be centroided
> properly.
Excellent news! I'm glad it's working after all.
One brief comment that is probably obvious: guiding during anything except
object rotation will introduce a systematic error of roughly half the
distance the star "walks" across the CCD -- hence a bigger error for longer
integrations or at points in the sky where the field is rotating more
quickly. This is just due to the affect of averaging -- the centroid is
correct for the instant of time halfway through the integration, but the
guiding correction based on that centroid is applied after the integration
is finished.
-- Russell