Last night I went out and did some more test shots - the first batch since I loosened the spring. All of them had visible RA trails at 60s, though some are worse than others. Siril's plot seems to show their roundness "bouncing" up and down every couple of frames, suggesting it's speeding up and slowing down. This could be due to wind, I guess, but I set up behind a barrier and all I could feel was a very light breeze.
In the longer term, stars are drifting eastward in my FOV, which seems to suggest that my mount is running slightly too fast. This was also something I noticed before my last adjustment. I'm not sure what could be causing this - some people have reported the mount running too fast when overcharged but I've used it for a few nights without recharging so that shouldn't be the problem. I also ran the calibration routine on it so it shouldn't be that.
(Aside: I'm not happy with my PA. Getting a good alignment in the polar scope is a PITA as usual, no thanks to the SGP's janky wedge - and I'm not spending $200 for the WO wedge, thank you! - but the three point alignment routines in NINA and APT seem to put me wildly off target, rather than refining the PA I get visually. Sometimes I perform the requested adjustment, wait for an update, and see no significant change in the reported error! - which makes me distrust these tools. Curious to see how well Sharpcap's routine works, supposing I can get it to run with my DSLR. In any case, the trailing I'm seeing in RA is much worse than whatever I'm getting in dec or field rotation.)
I'm beginning to wonder (a) if there is actually something legitimately wrong with my mount that can't be adjusted out, since whatever its inherent inaccuracy might be it should still be tracking the sky over the long term, not outpacing it, and (b) if I'm chasing precision that the mount simply can't deliver.
If there is a problem with the mount, it's out of warranty by now, and I doubt whether it's worth sending it in for maintenance/repair and paying a significant fraction of the mount's cost to get it fixed, knowing that it would still be a cheap, imprecise tracker afterward. If I'm going to spend that money, I think I'd better spend it on autoguiding and brute force the problem, and then if I decide to upgrade the mount later I'll still have my guiding equipment.
Still, I'd like to do more testing. Tonight I'll try different balance settings - almost even, a little more east-heavy, a little more east-heavy than that, etc - and see if I can't find a happy place. Maybe the right amount of weight will smooth out the mount's movement without making it strain overmuch. I should have done all this last summer, but at the time I was still using a ball head that I couldn't balance completely because I didn't want to spend money on rails and clamps and pan heads, and... well, live and learn.