I don’t think we can be of much help. Our code for filter wheels has been in place and operational for more than 2 years now. Every known filter wheel seems to operate just fine (except possibly the Moravian). I don’t say this lightly, but the data supports that SGPro is likely not an issue in this equation. I cannot speak for APT’s coding practices so it’s not really a valid comparison.
You should not need to do this, but if your filter wheel requires it, you can add sort of a settling time before SGPro will start imaging after a filter change:
If you want to post logs when this happens we can see if we are able to help you track this down.