integration times

[sarg314]

Lately I've been chasing my tail trying to find the best integration time.  It seems to have started when I bought a new guide camera; I replaced a very old QHY 5-II-M with a ZWO ASI 220 mini.  I didn't expect to see any difference with the new camera (and I may be imaging the whole thing).  I use an Off-Axis Guider on a Celestron 8" with an AM5 mount. I use a counterweight on the AM5.  I'm used to using integration times of 2.0 - 2.5 sec., with the old QHY, anyway.  G.A. always seems to tell me to use between 1.5 and 3.5 sec. I always run G.A. and always accept its recommendations.

Last night I was surprised to find that while 3.0  or 3.5 sec gave me rms > 0.9 arcsec, 1.5 sec gave me about 0.7 and sometimes even less.  The plot and scatter diagrams  definitely looked nicer at 1.5 sec. than 3.5 sec.

Some questions:

1- What can I expect from this set up? The best I ever seem to get, when seeing is good, is 0.50 or 0.55 rms.  Is this good, bad or indifferent?

2- Trying to find the best sweet spot in integration time is difficult and time consuming.  Your basically trying to match a sampling frequency to the seeing frequency without knowing the power spectrum of the seeing or how the mount is wiggling.  Is there any rule of thumb to use to select integ. time?

3- After I change integration time, how many samples (or minutes) should I wait for before deciding if an integration time is better or worse than the last one I tried?  Seeing can change a lot in a 10 minute period sometimes.

4- Is there any way to guess if the integ. time should be increased or decreased?  Or should I just always step through the whole range?