Last time I tried the latest Jan 9 version. It seems, both old problems are gone. No 'discarding start request'. Linear algorithm reverts focus back to where it started.
Yet, the focus position drifted all the way from 16500 to 28000 in 6 hours. See the attached logs.
Only a few first focus attempts were somewhat successful. After that clouds rolled in, so star detection errors are legit.
From what I read in the log, failing focus session reverts the focus asynchronously. Next focus session starts quickly before focus motor reached target position and reads current position, not the intended one. So, either the reverting operation should wait until the motor actually completes the move. Or reading 'current position' should yield the target position if there is move command in flight. My diagnosis may be wrong, of course.

Read More...