TDS694C doesn't boot when acquisition is powered

[SuperSVGA]

I have a TDS694 which only starts up when either the +25V (3.3V) to the acquisition board is disconnected, or the data bus for the acquisition board is disconnected.

Here's the brief history I have with this scope.
When I first powered it up, it started up fine and reported Acquisition as failed.
I tried to trigger on the probe compensation signal, but it didn't look like triggering was working.
I left it running for a few minutes and looked away, and when I looked back the display was off. It was still running though, I could turn channels on and off and heard the relays click.
At that point I thought I might have heard an angry hissing, so I quickly unplugged it and began tearing it apart.
I did a lot of searching of the low voltage power supply and display boards, but the only thing I found appeared to the electrolytic capacitor on the CRT neck board, which either was burnt or dirty, and may have had a hole in it, but I'm not certain. It still seemed to test alright, but I replaced it anyway.
I also dumped the Dallas DS1250Y and socketed it.

After putting everything back together, it powers up but doesn't go through the normal boot sequence. Garbage flashes on the screen briefly as it does normally, but then some relays click and all the front panel lights go off.

Eventually I discovered that if I disconnect the +25V cable going to the acquisition board, it fully boots up and somehow all the tests pass. But no waveforms are displayed at all.

That +25V line labeled "J124" is unhelpfully referred to as "TDS 694C" in the service manual. From what I can tell it's only for powering the DC converter that generates 3.3V.
I've taken the 3.3V converter out and tested it, but couldn't find any problems there. I also tried directly powering it with 25V from a bench supply, but it made no difference.

The other odd thing is, with the +25V connected but the D1 Bus disconnected, the scope starts up, with several tests failing.

Any ideas what might be wrong? I'm not sure the unit I have is a regular production unit either...

[m k]

If you heard hissing the electrolytic must have been a bit bigger.
They can also be quite a failure ninjas.

Have you tried external and current limited 3.3V supply?

[SuperSVGA]

Have you tried external and current limited 3.3V supply?

I'll have to try that next time I have room to crack it open. Hopefully it doesn't need the full current of the converter, I think it was rated at 10A but the most I can supply with the bench supply is probably around 6A.

If that doesn't work I'll probably just have to make myself crazy by spending hours probing at the board. Something is signaling to the system to shut off when that 3.3v is supplied, and that signal must go through the D1 bus somehow.

This thing is just a massive pain to take out and put in repeatedly. I hope I don't have to get to the analog boards again...