DEC VT 320 with Terminal Disease - help?

All, yep, it's me again. The 4000 VLC started and ran perfectly. I left it on for about 8 hours as a burn-in test. When I came back, it was (apparently) still running - but the VT-320 it was talking through had died. Arrgggh. I've done a little investigating, so hopefully I won't need quite so many exchanges to get this one going. The external symptoms of the VT320 were: no display; when power switch is on, the LK201 lights flash about once a second. No other sign of life. I turned it off, unplugged it, and opened it up. The power supply board at the side has an output indicator LED which also flashes about once a second. I turned it off, unplugged the power supply board from the mainboard, and turned it on again. In this condition (which I'll call "unloaded"), the output indicator LED comes on and stays on. In addition, the pins on the output connector go to more or less their correct voltage, as determined by labels on the mainboard at the other end of that connector harness: Labelled Ground 5V 12V 18V Actual Ground 6.5V 12.5V 22V The power supply has a big (10-pin) transformer inside a cage (with some other components) on it. The transformer seems to divide the circuitry in two, inasmuch as only the "Ground" traces seem to go to both sides. On the input (upstream, plugs into wall) side, there is a network connected to a UC3842N IC. Google led me to a spec sheet for that, which calls it a "Current-mode PWM Controller". The spec sheet also has a "typical application" schematic for an "Off-line Flyback Regulator" which looks to be at least somewhat similar to the upstream side of the power supply, though I have not traced the latter out well enough to be sure yet. There are also several large electrolytic capacitors on the input side of the power supply. Possible signs of trouble are there, in the form of scorch marks on the PWB around two power resistors (? They have what looks like a powder blue ceramic exterior, are marked 120 Ohms +/- 5% and ?? (maybe 47 Ohms), and are much bigger than most of the resistors on the board). In the "loaded" condition, the resistor with the worst scorch marks around it (R523, as marked on the circuit board) has 0V across it, with a small spike upward (as shown by the needle jerking on my analog VOM) each time the output LED flashes. In the "unloaded" condition, my VOM shows 15V across that resistor, and it gets pretty warm pretty fast. Questions for the group: 1) Have I got good evidence that the fault this time is actually on the power supply board, and not a pull-down on the mainboard, as it was on the VAX 4000 VLC? 2) Is there a good way (without use of an ESR meter, which I haven't got) to isolate which component is bad, or should I just replace the capacitors on suspicion? The power supply, at least, looks very easy to work on. Nice big components, circuit board is only 2 layers (front and back) and not very many traces on the component side, and the traces are exposed and easy to get to in its operating configuration. I'm being careful around the 120VAC (rectified to 160 DC in the unloaded condition). Thanks in advance for any help you can give me! - Mark