For the most part, it is fairly easy to remove and replace the (well-known) poor quality TEAPO capacitors (and any other suspect, swollen, domed or leaking caps) once you get the hang of it. The smaller, lime-green ones on the AcBel “sub-board” (mentioned above by ssp3) might not be quite as easy - BUT if the computer has exhibited the “failure” symptoms as described above… those smaller caps on that board might as well be replaced while everything is disassembled.
The capacitors themselves cannot be tested in-circuit (while on board) without an ESR tester / meter [or other suitable device(s) / LCR bridge?] and this would represent an added cost that would likely exceed Andy’s cost of repair, coupled with the cost of all the needed replacement capacitors. SO, unless you’ve many PSU’s to repair or replace / or you intend to offer this as a questionably lucrative service to others… the purchase of an ESR meter (etc.) adds an increased cost when you could just remove and replace the suspected BAD capacitors… which has worked for me previously. No guarantees implied. May not always work but worth a shot for minor cost.
Again, Andy’s fees are
very reasonable.
I’m certainly no expert and I continue to learn as I go, but that’s just me. Besides, I’ve enough G4’s here to keep me busy until my own
MTBefore
F and I probably should invest in an ESR meter. Others with a similar curiosity and the will to try… and to possibly learn as they go - might attempt the shotgun (scatter-gun) approach of simply removing and replacing most of the caps (and especially the TEAPOs). Replacement capacitors themselves are very inexpensive - all things considered.
The AcBel PSU that we've been discussing here did not power its’ fan when I got it. Instead of repairing or re-capping the PSU then, I powered the fan with an external power source wired directly to the fan because the PSU still powered everything else just fine (for well over a year)… until that too failed. I then simply installed another working PSU (thanks to darthnVader) and put the failed one “on the shelf” until GorfTheChosen began his quest here and I thought it might be time to attempt a tandem repair effort.
So, should Mr. Gorf choose not to proceed and instead opt to send his PSU’s to Andy (who has offered great prices for his services btw) this AcBel PSU will probably just go back onto the shelf until later. BUT, we’ve already learned a bit about those other tiny caps on that sub-board in this little adventure, thanks to ssp3. As for “not powering up a switched mode PSU without a load attached”… again I am definitely no expert - but the MDD’s PSU didn’t seem to mind this previously, nor did this QS AcBel PSU. So, do or don’t… it’s up to you. YMMV / Proceed entirely at your own risk.
*Note: The MDD’s PSU still had its’ fans attached in-circuit so there was a “load” present during jumpering. The AcBel’s PSU fan was not - as that circuit was incomplete or non-functional. Power testing at that fan connector on that sub-board (with the PSU power jumpered) provided a voltage reading that simply was there at first and then quickly bled off to zero volts. It still powered the QS (after that jumpering) for a well over a year before eventual total failure of the PSU.
"No user serviceable parts inside."Might be willing to friendly wager that I could remove the AcBel sub-board and then replace it without damaging the “main PSU”.
(The location of C216 might actually necessitate this.) Here’s the sub-board for the Delta PSU.
Much more open and easier access than the AcBel - but much smaller caps.
And it has a hard-post-soldered config to the main board using 19 base solder points.
The AcBel sub-board has a pin-in-socket config, held in place only by the “white goop” on the backside.
(Maybe a thorough cleaning of the pins and the sockets might be in order?)
Addendum:
I lied.
AcBel sub-board has 15 legs soldered to the main board (actually only 12 of them are soldered). Used ChipQuik SMD 291 spread with a toothpick on each solder point and then several 1/4” cuts of expanded solder wick - clamped in hemostats for better pinpoint control for each point to remove solder. Took some time but no problems. White goop removal / carefully using X-Acto #17 (flat squared blade) and a #11 (normal angled blade) to final, careful cut under bottom seal of sub-board. I did not heat these blades.
AND I had noticed what appeared to be burn or heat-related darkening here (see yellow dot). Might also note burned-look on top of C216. Shirley (don’t call me…) the loose leg and that possible intermittent connection would have definitely caused some possible damage over time? AND now, might be a bit easier to access & replace the five, domed TEAPO caps on the main board… as well as removing and testing the much smaller caps off of the sub-board… before ordering new ones.
Positive (+) legs for the caps on the sub-board are clearly marked on that board… & now have this pic for future reference if necessary. Should be much easier to solder sub-board back than it was to remove it.
IF I continue with this effort, I’ll post images of the solder points for the cap replacements on both the main board and the sub-board. (Unless someone beats me to this.)
