Update: it took several days to get this post approved, me being new here and all that. The festival started last night (Friday) and goes until Sunday. I had a chance to listen to the recordings I made last night. Viewing the waveforms (I captured the M and S mics, and the sound board feed separately) and can see that I need to remove about 15ms of delay from the SB's feed to time-align it with what the M mic is hearing. This is likely due to built-in latency in the digital SB.
I used an omni for the M mic, and that was a mistake: it picked up the bass rumble of nearby generators, trucks, other stages the next block over, and general low-end hubub of a city. I didn't use much of that mic in the mix.
I'm setting up a card for today's broadcast and recording.
Because the station is limited to broadcasting in mono only, whatever micing I use has to be compatible with a mono mixdown. The sound board feed is pure mono (the sound board guy cares about the PA, not the recording) so adding in some S mic (figure-8, matrixed) gives a nice spaciousness to the sound. If I turn off the S mic, the soundfield collapses into plain mono. The M-S micing also picks up crowd noise, applause, cheering, which the onstage mics don't. So I think this is going to work out real good.