Thanks Bryan, Brad & John for weighing in to give me more options.  I explored a few before determining what was most suitable for me.  For future reference (and “postmortem” analysis) — here’s what I did:

First routed the mixer channels to the out bus.  Each stereo mic channel routed to an output; mono mic channels I panned hard L / hard R and routed them in pairs to an output channel.  After that, I ran the “Bounce Output Channels” export, faders on the output mixer left untouched.  Dragged the wavs into my DAW and split the paired mono tracks back into regular mono.

With no fader adjustment in the DAW, the mix sounds VERY close to the internal SD3 mix (if not the same).

I tried rendering in the DAW, which seemed like it would work fine, but the process felt a little convoluted.  Also I was getting some glitching sounds during the render, which probably was nothing, but it was enough to exasperate me.  John’s method seemed pretty useful for simplifying a bounce (hearkening back to the tape days, which I always enjoyed), but I didn’t use it just because I didn’t want to bounce via combining mics.

For anyone who’s curious: the origin of this situation was layers of stacking — a great preset from a “small room” SDX (1 kit) + the ambient mics from another “big room” SDX (2 kits).  As you can expect, things got complex with the mic routing and the bleed adjustment submenus.  Seemed like it would be much more expedient to do 2 individual exports, one for each SDX, and then dump everything into the DAW so I could have everything in a nice neat row of faders for mixing / submixing.

Anyways, I now have exactly that, and it sounds beautiful.  Thanks again for everyone’s help!