[geeks] power supply trouble?

[der Mouse]

I have some hardware that's acting up and I suspect the power supply.
What's more, I suspect a particular cause.  But before I go to the
lengths necessary to address it, I thought I'd see if anyone could
sanity-check my suspicions....

The device is a flatscreen display.  (There's actually a slight sunhelp
tie-in - beyond its just being something of mine, that is - in that
it's the main screen on my principal house workstation - a SS20.)  It's
worked fine for some time (some small number of years).  Recently, it
started being reluctant to come on, either when switched on with the
power button or resuming from a no-sync-signal power-save.  The symptom
was that the green light on the power button would blink on briefly
(100-200 msec?), then go off for a second or two, then blink again,
repeat.  At first this would repeat for some four or five cycles, then
it would power up apparently normally and run fine until next
power-down or power-save.  (Even then, if I resumed it promptly, it
would come back immediately.)  It looked reminiscent of a power supply
with its overcurrent trip activating: power long enough to realize it's
overloaded, then shut down and retry in a second or two.

It gradually (over a month or so) started taking more and more cycles
before starting to work.  Today, I tried to wake it up when I got up
and it was still blinking when I left for work an hour or two later.
When I got home, hours later, it was _still_ trying.

So I opened it up.  It's got two pc boards, and looking at the
interfaces, it's pretty clear the division is "power supply" and
"powered device": the interface between them is five pins, marked
"+12v", "GND", "+5v", "ADJ", and "On/Off".  (Actually, eight pins
mechanically, but only five pins electrically; the first three lines
each have two pins.  The etch runs clearly tie those three pin pairs
together.)

The only thing that's apparent as a possible cause is that of the dozen
electrolytic caps on the board, some six or seven have visibly bulged
tops; one of them even has something that could be a small amount of
some internal compound leaking out.  (Mercifully, the main power filter
capacitor is not among them.  Based on a little eyeball circuit
tracing, I think the bulging caps are between +5 and ground.)

So, the suspicion I'm looking for a sanity-check on: I suspect that
these capacitors are the problem and that if I replace them with good
caps there's a reasonably good chance it'll go back to working fine.

I considered just running it off an ordinary peecee power supply, since
it appears to need only +5 and +12.  But (a) I don't know whether the
ADJ line is actually important from the powered-device point of view,
(b) I don't know whether the on/off line is interface-compatible with
an ATX power supply's soft-power line, and (c) the power supply has
four other connectors that feed (if the markings are to be believed)
6.5mA of 720VAC to the backlight lamps.  It looks as though the
backlight supply is well isolated from the rest and probably just eats
+5 or +12, but I'm not sure which and would rather not experiment if I
don't have to.  Besides, I'd rather not use an external power supply if
I don't need to.  (And as a side note on (b), it's also possible the
powered-device part depends on the on/off line's interface in some
way.)

Thoughts?

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse at rodents-montreal.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B