I've been doing some unguided imaging with my AP1600 w/Absolute Encoders and APCC Pro and have seen some interesting behavior with unguided imaging.
The first few nights that I run unguided after building a model of about 180 points, everything was great. I was blown away by how well it worked. The last few nights, not so much. I am seeing elongated stars and some image drift over the course of the night.
I do not believe that this is flexure. I'm imaging with my AP130GTX, and I've double checked all connections. I've double checked to make sure that the pointing model is enabled. I verified that the polar alignment is still spot on. It's a bit difficult to troubleshoot because, without guiding, there aren't any log files to examine. All I have are the subs that I can inspect.
Since we're getting into more moonlight, I've done some software updates (switched to the daily builds for NINA so that I can use the advanced scheduler). I've also set up for doing tonight's run with the guider enabled so that I can get some logs. As I was watching the session get started, I noticed something odd. Specifically, I noticed that APCC reported the temperature at over 40 degrees C, which is very wrong. I am using the Pegasus Astro Ultimate PowerBox v2 as the weather sensor.
It occurred to me that I made a change to the Pegasus software a few days ago to change from reporting the temperature in C, to reporting the temperature in F. It looks like both APCC and NINA are reporting the Fahrenheit value as Celcius. I am wondering if the significantly incorrect temperature interpretation has effected the model such that it's lost accuracy. I have reverted the Pegasus software back to reporting in C, and after tonight's run, I'm going back to unguided operation to see if I get that great result back that I was getting the first couple of nights.