This is an interesting thread. Though I've gotten in late on the discussion, a few points come to mind:

1) Although a guide camera may be able to take exposures as short as 20mS, this will result in the system trying to 'chase the seeing' as others have pointed out. Long exposures of 1 second or more will average out the perceived position of the guide star due to seeing conditions. The problem with longer exposures is that you will tend to saturate the image sensor when using bright guide stars. A saturated guide star image will cause problems with the guider software routines that compute the star centroid when measuring star position. One way around this is to use a shorter exposure but to average several pictures together before computing the guide star position. That way you don't saturate on the bright guide stars and also average out the seeing induced variations. Not many guider software programs provide for this type of operation, however.

2) Auto-guiding is a closed loop system in that a measurement of the guide star's position is made and then a correction is sent to the mount to correct for any error. The process repeats indefinitely or at least until the sun rises. The key premise here is that you cannot make a guide correction faster than you can see the result. Otherwise you have an unstable system. So, if you send a guide correction to your mount, you better wait for the telescope movement to complete before taking another picture. If you try to take a picture of your guide star before the previous guide correction has finished, you will erroneously conclude that the last correction didn't correct for the error and you will probably do the wrong thing. Most telescopes have a lot of mass and can't respond very quickly. That is what AO systems promise. They can respond faster since they are moving an optical element which has far less mass than your telescope mount.

3) Auto-guiding attempts to correct pointing problems regardless of their cause. These can be any or all of problems with polar alignment, periodic error, flexure, etc. So, if for instance you have poor polar alignment, this will be detected by the auto-guider and a large correction will be sent you your mount. Better polar alignment will result in less drift between guide star pictures and smaller corrections being made. Your stars will get progressively rounder with smaller and smaller guide corrections. Likewise, auto- guiding will correct for periodic error in the mount's drive train but it would be better to have the mount's drive electronics perform the periodic error correction. It is, after all, a very predictable error and is best handled by the drive electronics. The auto-guider would just try to make a correction after the error has occurred whereas the drive electronics would be pro-active and not let the error occur in the first place. The moral of the story is that there is no substitute for addressing the sources of error where they lay. Your auto-guider will have an easier job and your pictures will show better results. I always say... ' the best auto-guider correction is one that doesn't have to be made'.

4) I'm the designer of the Starfish Guide camera and, yes, it uses the same image sensor as the Orion Starshoot Autoguider. Apart from using the same image sensor the two cameras are distinctly different as the following references point out:

http://www.fishcamp.com/starfish_shootout.html

http://www.astrophotoinsight.com/node/162

Regards,

Bob Piatek

fishcamp engineering
105 W. Clark Ave.
Orcutt, CA 93455

http://www.fishcamp.com
TEL: 805-937-6365
FAX: 805-937-6252