So, this has been a month of frustration, and therein, learning, both in terms of technique and potential upgrades to my current equipment. Thanks to all those who have pitched in their comments and ideas.
Below is the result of 4 nights combined imaging, each with a different emphasis in trying to get a better set of subs, and each limited by either or both weather (poor seeing and transparency) and technical issues (flakey intervalometer). Normally, one wouldn't mash such data together by simply putting each night into a tab in DSS and pushing "Go", but I neither have the skills nor probably the software to process them separately for subsequent merging. A challenge for myself next month, assuming the new imaging target won't be so, um, time challenging, is to try APP instead of DSS+ST. The "problem" this month has been that, as messy as this has been, adding each additional night's data has improved the resulting image. So, this is surely not the best that can be done with the data, but still better than what I'd get otherwise.
OTA: Celestron C8N 8" f/5 Newtonian with a Paracorr-2 coma corrector. Net 1,150mm focal length
Mount: Celestron AVX
Guiding: Celestron 80mm f/7.5 guide scope, ASI174mm Mini camera. PHD2 software running on a Raspberry Pi 3B
Camera: Nikon D3200, unmodified. 3 nights at ISO 12,800; one night at ISO 3200.
Sub Exposures varied from 30 seconds to 2 minutes; three nights without and one night with an Astronomik UHC filter. One night was at a really dark site (until the Moon came up), somewhat sabotaged by a flakey intervalometer which shortened the maximum sub exposure times to 30 seconds each, and a nasty breeze that pushed the AVX and PHD2's ability to keep stars round. Total image integration time 4 hours 34 minutes, 43 seconds, which is a personal record.
Stacked in DSS, processed with StarTools. The image here is about a 50% crop, due to the wildly varied alignment from one night to the next. I'm rather impressed that DSS was actually able to make sense out of the lot. On the final night I finally got Plate Solving to work, so at least that issue has been more or less solved for future imaging.