I am in the same boat as you so to speak. Bortle 8 here in my backyard. Your results match mine, more or less.
With short exposure times, you will get enough photons to work with-- it's the read noise which could be an issue. I don't think its as big of a deal with current CMOS cameras as it used to be, FWIW. YMMV.
At f/4.9 (with reducer), on m31 I'm doing 20-30 second subs using an un-modded DSLR @800 ISO. This lands my histogram at around 1/3-1/2 peaked. (Edit: I'm not using any LP filter.)
I'm getting pretty decent results with this, considering.
FWIW, even 10 seconds on M42 with these settings blows out the core.
On a positive note, at 20 seconds, no guiding is necessary and my stars are perfectly round.
This actually lines up with my plan! I ignored the tracking error part in my question but I am also unguided (Sirius EQ-G) and stars were just starting to trail at 30s. It sounds like you are getting good results at 20s so looking forward to getting a night of exposures at that. (Only had one night since I got the mount and those 2m exposures were all completely white!)
Appreciate the info...