2nd Nava isn't powering on correctly

I built my first Nava, and it is almost done, just a few voice issues to sort and it will be perfect.
I also built a second Nava, and just finished it. Before powering on, I tested both boards for any shorts between power, and found none.
I tested the power i/o board alone, and the negative rail was slightly higher than the positive, but nothing crazy, just a bit over -15V.
So far so good. I connected the boards and powered it up for the first time and smelled a bad smell. Not strongly, just a bit, so I removed power immediately.
I rechecked the power, and still no shorts. Checked the polarity of all caps and diodes against my working nava and the screen - all good.

I thought I might have imagined it, as I can be pretty paranoid about power, and after a round of soldering, maybe I was smelling a stray fume that my fan didn't get. I powered up again, all the lights were good, and I thought I was good so I started adjusting the lcd trimmer, but now there was definitely a bit of smoke.

I de-powered and took a look and in the Noise section, C53 looked a bit burnt. I thought it might be heat from Reg3 on the IO board, but they are inches apart. All of my polarized caps and diodes in that section looked good. I triple checked the power lines, and still no short.
I checked all ICs on the board, and all were correct except IC70 in the hihat section.

Any ideas? What would have made that cap go bad if not a short?

Now if you apply power, all lights go on for a moment and then go dark. They do so again when turning off - is that a bad cap?

If you let out the magic smoke and the cap looks burnt, it's done. Taps for that cap.

Yes, but the cap that looks burnt is a non-polarized cap, so it ought not to be the source of the power issue.