Thanks for the encouragements!
I do realise my first post was not that clear...
I first bought one full DSO in good cosmetic state but not working. Seller specified that two transistors were missing on the PSU.
When the scope arrived I replaced the missing transistors and tried to power-up, it was dead. Did a little testing around for shorts on the PSU itself and on the Acq./CPU board where the PSU connects. Nothing obvious.
Looking closely at the brown Nichicons on the PSU there were dull grey solders around, the PSU has many hot spots and the smaller electrolytics were very compact models often squeezed into those hot spots.
Replacing most of them brought the PSU back to life, and the scope powered up displaying the boot splash screen with a distorted picture (monitor problems to see to later...), the traces were not on screen though, shortly after the scope restarted and appeared stuck in a boot loop, I started investigation on the CPU reset (TL7705AC), but before long there was no display at all and the PSU was hiccuping.
Checked for shorts at the PS connector again, nothing too obvious, found the rail that looked lowest and gave it a push with the lab supply. Display came back up, no traces though...
So then I went for a first peek at the frontend board on the down-underside (not far off Australia) of the scope, uncovering there was a nice spluttered burn mark on the aluminium shielding.
That turned out to have previously been a 0ohm SMD on one of the 5V rails, on the underside (getting way down south!) of the frontend PCB same scene, another fused 0ohm resistor.
This one was on the ground though (not something I'd recommend in any design...) I'll never know what the original problem was, but possibly thanks to blowing this ground link, among other-things the AD767 and the P87C52 were gone making any repair hopeless.
I had a quick look for another parts scope and the same seller was selling another in poorer cosmetic condition, missing PSU, fan, but with spare parts: An extra frontend, front panel and Acq./CPU board.
I'm guessing that these two scopes plus spares were the worst parts the seller had left after rebuilding a few scopes swapping boards around.
However, a week later I had more junk to play with and with a "new" F.E. board a scope now showing 4 traces, although 2 are often so far offset that they can't be seen.
From one 5" calibration followed by self test in the menu to the next, the results are not the same.
Things are looking like either automatic calibration failure or self test failure.
From there I checked as many supply voltages and ripples as I could find, basically at all electrolytic capacitors on Acq./CPU and F.E. boards, nothing really worrying, I did hesitate a little at 60mV spikes on a 8V supply, the caps were OK though.
Resoldered the whole worst channel, tested and swapped some relays, things did get a little better, still far from good.
I had dumped the EPROMS right at the start but being high and low data separately for a 16bit SCC68070, they needed to be grouped as a 16bit image before looking for any meaningful text strings. Anyway, a Full 45" calibration did show-up in the dump, so I then desoldered and saved the EEPROM containing the calibration data just in case...
After that I found the way into the cal menus, did a pinch of manual calibration (AD767 reference and frequency compensation on one channel) and let the 45" calibration see to the rest. It does the job by itself, although 45" takes forever....
After that all self tests passed time after time on a warm scope, a cold scope will fail offset on all channels, but I'm guessing that is why the 5" cal was left accessible to the user, to compensate the drift when needed.
(Right, being used to LeCroy this sounds awfully awkward, so I feel I must quote another manufacturer for clarification
: SPC is the acronym for Signal Path Compensation. SPC's. Running Signal Path Compensation corrects for DC inaccuracies caused by temperature variations and/or long-term drift. The benefit of running SPC is that is optimizes the oscilloscope's capability to make accurate measurements based on the ambient temperature.)
So from there I considered I had working scopes, as they were happy with their self-tests and conforming to my own checks.
Note that at the best, DC balance is poor on these OX2000 scopes.
All the monitor took was a good spraying and reajustment.
In the end I don't think I actually repaired much other than the PSU, but considering there is not even a user manual to be found on the WWW, it was a handful to actually look for a way to get at the calibration rather than go round in circles looking for a fault that wasn't there without schematics. (I did a few hours of circling all the same!)
I do like the challenges Vincent
(Sorry long post...
)
So now all I need is a PSU for the second scope, the easy way out would be using one of the spare TDS400 PSU's I have around, I/V requirements for each rail are similar enough, they also have the line trigger circuitry, all the same I'd like to see where I get with a modified ATX supply.
That will be later though, for now I have a leaky TDS620 I want to get out of the way before it corrodes through to OZ land, I just hope it won't turn out to need a sacrificial Minitel too.