An A310 with a PSU marked "WK 8728" - so that's mid 1987 - it's not the earliest design, but it's a relatively early machine. Turning the machine on, no life other than rapid "click click click" or "squeak squeak squeak", quickly identified as coming from the PSU. Unplug the PSU from the board completely and try again, now the "click click" rate is down to about 2Hz. Measure the 5V line ... it's wobbling up and down, with a 2v5 average. It's goosed. This is the first heavily failed A series PSU out of quite a considerable number I've dealt with.
Let the machine sit switched off for a day to hopefully discharge anything in the PSU and dive in. Despite how it appears that -some- A310 PSUs can be wiggled and jiggled to get the PSU out, I just couldn't find any way to do so without bending the metalwork at the front a bit. It's not a big deal, as it'll all hold itself nicely back in shape when it's bolted back in the case, but not a nice thing to be doing really - hohum. I don't like dealing with mains voltages, but needs must...
My first guess was ye olde RIFA X2 capacitors and sure enough there's two and they both looked cracked, so let's get those replaced:
I couldn't find PME271 in 15mm pitch 0.01uF so it's a PME264, which has a 600V rather than 275V rating; should be fine.
CJE
sell a PSU refurbishment kit which also includes the 470uF startup capacitor, so I changed that too, because why not.
No joy, no change in behaviour.
With some guidance from Stardotters (Thanks @techfury and others) it was decided replacing the DC-side capacitors was a reasonable next best bet.
Out come those big orange beasties, replaced with more modern ones that happen to be higher voltage rating as well as being physically smaller...
And that's sorted it. There was some apparent "brown" under one of the big capacitors, and a bit of the usual 'fish' smell when desoldering, but nothing particuarly obvious as a major capacitor leak. Only after measuring the pulled capacitors was it clear that the issue must have been the small 470uF one - wouldn't measure as any type of capacitance and had a DC resistance of 4ohms!
All back together, twiddle the 5V trimmer for an accurate 5V, -5V and check 12V (11.65V; good enough). Job sorted.
I'll feel a little less scared about servicing another in future, and would probably try and check for low resistance across output capacitors first.
