I regularly use SGP’s focuser temperature compensation. Since the algorithm for calculating the focus position had been fixed in v3.0.2.81 in May 2018 this worked perfectly for me, including v3.0.1.169.
In two nights I now tried v3.1.0.320 and noticed that obviously in this version the old method (or something similar) based on temperature differences between frames is working again. Applying this algorithm produces small rounding errors that add up to incorrect focus positions after just a short time, see Temperature Compensation Implementation - #6 by bulrichl - Auto Focus - Main Sequence Software . The error becomes greater and greater with time.
Appended you’ll find a chart which shows temperature trend, the best focus of the underlying AF run, the focus positions applied by SGP’s temperature compensation, and the “true” focus positions that I calculated based on the result of the AF run. It is clearly recognisable that the method used by SGP in v3.1.0.320 does not react when several small temperature changes occur consecutively.
I really don’t understand why bugs that had already been successfully fixed now are introduced again into the code.
@Ken
@Jared
When will we have the correct algorithm in SGP back again?
Bernd