ODAC ground loop death and resurrection

Now, I know nwavguy is not liked much here, but his stuff isn't bad. For the cash, it's amazing.

It's totally prone to stupid errors, naturally (weird o2 design dependent on source,... etc).

I managed to, while experimenting with setups, destroy 2 odac's. I've never had any such problem with anything ever. Basically the ground loop just fried both boards. Once a year ago and now recently again. It's very possible I did something wrong, but I think the device just can't tolerate a stronger ground loop in a system with more gear. I have no ground loop issues, understand them well and how to defeat them, but this was not always so. I've searched online for an answer, and didn't find anything, so here goes.

The chip that got fried both times (the first time I didn't have the brains to explore and fix it myself) is u1 - a small smd linear regulator. To be expected - first in the line to get hurt. it's a mic 5205 3.3 or 3.6, not entirely sure now, would have to check in the schematic. It's super hard to find online for a reasonable price+shipping (at least here in EU), and when ordering from ebay it takes a million years to arrive - I waited 2 months.

If someone here had, has, or were to have such a problem, I have like 10 of them now, and I'll send you one for free.

I understand science poorly, though, so for all it's worth to me, you can glue a banana to it and it will work fine.