Picked up a DDA-125 sold as not powering up, and while the first few power cycles didn't yield anything, it suddenly came to life after that, and I ran it long enough to verify the printer, some basic acquisition system stuff, the cal signal, all four inputs, etc... seemed to be working great. Hasn't powered up successfully since then.
So, suspecting the power supply immediately, I set around to take things apart. Got all the screws out of the power supply and had a lot of trouble getting it to come off - enough that it seemed like there was another screw somewhere.... but eventually the connector gave out and I saw what caused it - almost identical melting of the big power connector.
The PSU's filtering is impressive: AC input goes into an inline RFI filter, then this shielded filter module in the power supply (bottom, with the yellow tape), then into the switcher, then through a massive choke for each rail with bulk caps as well as smaller decoupling caps right on the contacts (in two sizes, no less).
What's more, I opened up the power supply and in addition to taking some measurements of the output voltage, I took some pics with the thermal camera. Most of the bits heating up quickly were power resistors, but the one part that heated up but wasn't a resistor.... was C93, the same snubber cap that failed in the OP's unit. Not only that, the only discoloration of the PCB is right under that part too - I had been suspecting it in my preliminary analysis before I found this thread. This thing may have come from a bad batch.... makes me want to look around the PSU board and just replace all of them or something.
Anyways, I have yet to pull the cap and verify that it was the main issue, but given the intermittent nature of my failure (my latest measurement of the rails actually yields pretty acceptable results, though I've measured way low previously), I wouldn't be surprised that it's gradually coming apart and replacing it will right things.
Otherwise, I plan on swapping the PSU fan with a new one of the current model (the one in their definitely is not living up to it's 29dB noise claim), and swapping the main ducted fan with something with just a hair higher CFM with lower noise. Then I've got 2x64MB SIMMs on the way to double the system memory on the CPU card, the spec sheet says it can handle up to 128MB modules and the upgrade cost $15.
Took the top shield plate off to get a good look at the acquisition board too - mostly out of curiosity, but to check for damage related to the burned connector too. 31 screws (including the 6 top screws in the plastic BNC mounts) later, it's a nice looking board with a lot of custom parts. Some high quality bodges to be found, and it looks like the input of the ADCs can be all routed to channel 2 (maybe 3?), verifying why I could get 8GS/s on a single channel when I was testing even though the manual claims you need a PP096 external connector to use all 4 ADCs at once.
Then some fancy bodges on the acquisition board - it's just not a finished product until you have to glue a resistor array to a chip.
Also been really impressed with the metalwork. You open the top and everything has a nice color and looks clean, then you get inside and it's really well designed, you can see shielding for the front end components as well as some ducting under the shield to make sure the ducted fan in back pulls all the air over the ADCs.