As I mentioned in prior posts on focus drift compensation, I have been struggling with the issue of focuser image shift with my FLI Atlas focuser. Image shifts during exposures compromise image quality and require autoguider correction.
I recently discovered a solution to this problem.
During my exposures, rather than making a focuser movement whenever temperature has changed by 0.25°C, I now make movements for every 0.05°C change in temperature.
The previous scheme with movements corresponding to a 0.25°C change occurred roughly every 10 minutes. For my setup, this change in temperature requires a focuser movement of ~11 microns or ~136 steps. With movements of this size, I observed image shifts as large as 10 microns and an RMS image shift of maybe 3 to 4 microns.
My new scheme with movements corresponding to every 0.05°C occurs roughly every 2 minutes. The focuser movement equals ~2.3 microns or ~27 steps. Image shifts are hard to measure and appear to be buried by seeing fluxuations almost always.
The accuracy and resolution of the Qti DirecTemp USB temperature probe, model DTU6024C, suffices for this application.
This solution represents a dramatic improvement for my setup, especially since autoguiding corrections can be significantly delayed due to my mount's declination backlash.