GNS seems to often report an error due to the fact that guiding has not settled after a dither (at least this is what appears to be happening). Since it gets its settling data from SGP I assume that if SGP fails to report settling quickly enough, the error is triggered. Someone can correct me if I am wrong.
What I have noticed about SGP when settling from a dither is that it can take maybe a 3-10 cycles to report that it has settled to the user set level where an exposure can start (in my case typically .4 or .5 pixels RMS). My question is where is this RMS being calculated, PhD or SGP? Wherever it is, I would suggest that it be done over either smaller set of guide corrections or a user settable number. What happens now is that it is often as many as 3-10 cycles (or as long as an extra minute or more) before SGP will allow an exposure to start, despite it being very obvious from watching PhD that the tracking has long since gotten below the required error level.
For settling, the method you describe is far more complex than what is happening. SGPro will start the next exposure when the guider reports that its error in pixels is below your defined threshold continuously for the defined number of seconds. We receive error data from PHD2 for as often as you expose. That’s all settling is.
I wonder if others are seeing what I am. It does not seem to be doing what you describe. It is very common for me to watch the PhD graph and the corrections are well under what SGP is showing at the same time. It is easy to know since on one of my systems the sampling is very close to 1 arcsec/pixel. Is there a lag?
I will have another look at this next time I image but I am pretty sure that is what it is doing.