Problem solved! This should have occurred to me sooner, but I wasn't able to reproduce the visual noise issue on my MS8-29 chassis.
Solution:
Replace these 4 resistors with 220 ohm 0402 resistors.
Background:
MiSTercade V2 has a slightly modified video amplifier stage compared to V1.3. The RGB pots act as feedback trim resistors instead of termination trim resistors on the output. When users reported the VGA output was cleaner than the JAMMA output on their cabinet, it narrowed the issue down to the amplifiers stage themselves (Which aren't part of the VGA signal path). I sent 3 modified PCBs to
@Astro_jamma. These were marginally better but the issue persisted. A kind soul named Scott emailed me and pointed me to
this forum post, which suggests adding series resistance to the amplifier output. Thankfully, I left myself an escape when I designed the video stage by putting 0 ohm resistors between the amplifier output and the JAMMA RGBS pins. I replaced these resistors with 220 ohm resistors and sent the 4th modified board to Astro_Jamma, who tested it today and reported that it eliminated the issue.
What's different about MS8 chassis compared to MS9 or others? I don't know. This modification either limits the current of the video signals, or tunes a low pass filter on the chassis side. Either way, it works and will be rolled into MiSTercade V2.1.
This fix also explains why MiSTercade VS and V1.3 need a small tweak to the RGB potentiometers on some monitors.
I'm happy to do this repair for free for customers (just email me on my site for assistance). Not sure how to work out international shipping - hopefully there's a better solution for non-US customers who experience the issue.
For those who want to DIY, 4 of these will do the trick. Just make sure you have a steady hand, or hot air, or both:
https://www.digikey.com/en/products/detail/yageo/RC0402JR-07220RL/726442
Sorry for not catching this earlier. I didn't see it (and still can't) on my test cabinets.