Not only bad caps!

[Elitist]

Also in the recent batch were a bunch of HEC-350 (some appended AR-PTF). Very nice, very high quality - but dead. All had the same fault - blown fuse and varistors across the reservoirs. Knowing the history is half the battle. Apparently all the occupants of the building had been advised to switch everything off whilst their HV transformer was replaced and pulse tested. Need I say more?! The varistors did their job magnificently, absorbing the first 39J, spraying everything in the vicinity with silver, carbon and zinc oxide; ignorance delivered some welcome kit to yours truly. Another case of "No serviceable parts inside" is a lie to dissuade the uninitiated......

[PeteS in CA]

Re: Not only bad caps!

Sounds like that transformer replacement may have been the root cause of the failures of a lot of those P/Ss. On the other P/Ss, it sounds like enough of the voltage surge got through after the MOVs fried and before the fuse blew to destroy the I/P lytics and the switch devices. Or so it seems from what I read in that thread last PM.

[Elitist]

Re: Not only bad caps!

Probably correct analysis, Pete. As you know, the cheaper end of the market may be using the same designs, perhaps even the same pcb s, but they save $$$ by leaving the varistor slots empty (amongst other components!), thus allowing the MOSFETs and NPNs to blow before the wire fuses catch up and explode, closely followed by the house tripping out. I've also seen the spaces for the thermal element (often labelled TH1) shorted across with a wire link, even on quite well engineered units.

[Elitist]

Re: Not only bad caps!

There may be some basic guidelines for attempting PSU repairs. This is for NON-EXPERTS without professional test gear.

Of course, flash over the underside of the pcb with eg a big old insulated screwdriver before touching ANYTHING, including the heatsinks. Marigolds?!

If several caps have gone bad with visible bulges, leaks, etc, it may be sufficient just to replace them, possibly also the fuse.

If the fuse and the varistors across the big reservoirs are blown, it may be sufficient just to replace them, although the varistors can be omitted with additional risk of total destruction in the event of a future major surge.
Combinations of these faults may, however, be symptomatic of other problems.

Where there is no obvious damage, the NPNs/MOSFETs (out-of-circuit) and associated diodes+resistors should be the first suspects. If faulty, they are usually too expensive to replace with single-purchase devices unless the PSU is of a special design, for example. Sometimes near-equivs can be purloined from similar units with other faults, although it is important to heed the specs fairly carefully.

In all cases, the big NPN fast switch on the input side should be removed for testing - only a few are likely to be found to have failed, though.

Other than that, it's just a question of plugging in, switching on after standing well back. If it goes bang again, you've got a nice bag of spares for the cost of a fuse.

Doubtless, our resident professional gurus will be able to add some better-informed wisdom on this topic.

[Elitist]

Re: Not only bad caps!

Epilogue:
Probably the final blast relating to the batch described above, most of which have now been repaired or stripped for parts. It transpires that surges can also blow one (!) or two of the rectifying diodes in the input bridge, as well as causing the more obvious damage. So far none of the single-package bridges seem to have sustained any failures, but only individual members of the four discrete diodes. These generally fail to short circuit so are readily identified in circuit.
Regarding repairs, most of the carbon deposits can removed with IPA and a cotton bud. However stubborn burnt deposits can be further cleaned with scouring powder (Vim) and an old toothbrush, or, in extremis, by scraping the pcb with an old scalpel or Stanley.