Mike,
I hope you don't mind my butting in here, I'm sure you have some useful input which you can add. I have a bit of time right now so I thought I'd assist straight away.
Simon,
From my review of the data I ended up using 24 Ha, 35 O3 and 27 s2. A few of the Ha subs I used contain a bright artefact near the top right, but appeared to be eliminated during integration. It appeared in 3 frames before moving out of the frame. I'm used to seeing the effect of geosynchronous satellites, but whatever it was is pretty large.
Some frames suffered from the gain not being set correctly and I eliminated those just in case they caused problems. Nine of the S2 frames were shifted, the mount slewed to a slightly different position, but I included those.
Sorry for that wordy explanation, which is not really relevant I know.
Date and time stamps of the ones I included.
Ha
20200815: 021557 - 033924
20200816 thru 20200817: 231551 - 010948
I selected 20200815 030126 for the alignment reference frame.
O3
20200812 thru 20200813 : 231753 - 003400
20200818 thru 20200819 : 234231 - 005823
20200819 thru 20200820 : 233915 - 003443
S2
20200815: 035209 - 043205
20200817: 014157 - 031332
20200818: 023819 - 035437
I hope that helps. If you decide to use those, there should be 24, 35 and 27 subs respectively.
I did not use any flat frames, they appeared to make matters worse, and I used the -5 degrees darks as the older ones suffer from major light leak problems. That may well have been your issue. I ignored the warnings about incompatibility between the light frames and the master dark when running the script due to the temperature difference.
Cheers, Ray
Mike I see you changed the colour scheme since I last viewed it. Very nice.