Peavey PV2000 slave amp

dead output transistors

I now see it is DDT (TM) for speaker protection , there was no DC on the output of either channel so presumably false protect cct and the crowbar triac is ok. Still can't find what DDT stands for though

A quick Googling shows that DDT stands for "distortion detection technique", a system that prevents (or reduces) clipping.

Why should I have had to do this for you?

Reseat the socketed devices. Temperature isn't a factor, humidity more likely the cause of slight corrosion of the pinned devices.

--
Live Fast Die Young, Leave A Pretty Corpse

Don't Open This.

-- Jeff Liebermann snipped-for-privacy@cruzio.com

150 Felker St #D Santa Cruz CA 95060 Skype: JeffLiebermann AE6KS 831-336-2558

Looks as though a TO92 has decided to take in condensation, go ohmic and move the bias a bit negative. TO92 failures the same, with a load of Crest audio slave amps left over winter ,almost outdoors , by someone else.

At least these amps have separate LV DC supplies and the biasing part can be probed out of casing , powered from +/-15V bench ps without +/-92V around

TO92 60V, 0.6A , 9V on one end of the 47K base resistor and 8mV on the other. Checks out DVM diode test and no obvious DVM resistance between pins. Now removed can see ,under x30, slight build up of corrossion crud where the legs enter the plastic, marginally more than the layer of corrossion over the rest of the legs. I will let the owner see the result under pocket magnifier , that and bill might change his storage arrangements.

Some people just don't realize that electronics of this type really need to be stored in a climate controlled storage unit.

--
Live Fast Die Young, Leave A Pretty Corpse

The other owner , with the Crest audio amps thought covering with a plastic sheet , include a bag of silical gel, and leave in a Dutch barn was sufficient. He never even recharged the silica gel - having no knowledge of that process.

I wonder why TO92 and not TO220 or TO3 fail in these circumstances. I suppose the leads corrode and swell and then like concrete spalling due to rusting of rebar, the plastic splits enough for capilliary water condensate to migrate up into the die. Something to do with small diameter leads for TO92? I will try to remember to split open this TO92 and see if it cleaves along the corrossion plane. Different maker to the Crest TO92 so not a batch problem.

I think the corrosion infiltrated the die. Like ice crystals in concrete as you said. Repeated heat thaw cycles exacerbate the issue. Furthermore, not having a fresh/sufficient desiccant furthers the process. Why the TO92 suffers more would only be guessing. May be a batch issue. I know there are plenty that don't suffer from extreme conditions. But they may be saved from frequent use.

--
Live Fast Die Young, Leave A Pretty Corpse

Different manuafacturer and series of TO92 transistors rather than different amp maker , so less likely a batch problem. In TO92, TO126 and TO220 the encapsulation all goes down to the die IIRC. Perhaps corrossion on small diameter leads has a better forcing/splitting effect into the plastic, then capilliary. Continue the bad environment long enough and the other classes of encapsulation would start failing also is my guess. Further back in thread 80mV not 8mV on the base

Assuming it is B-E junction failure going ohmic. It will be low current transistors showing the effect first, it may be inside other higher current ones but not manifesting as a problem yet. For fault finding I suppose one route is to check TO92 first and of those the lightest current ones first eg highest base resistor values first , if no other knowledge. A trouble with these sort of biasing circuit areas is an almost total lack of grounded nodes so everything is floating up or down together.

looks as though the precipitating problem was failing 15V zener. For some odd reason there are LV +/- 15V LV supplies and +/-15 supplies from the +/-82V rails linked together via diodes at the pa. The LV +15V supply is from 7815 but the -15V is from 1W zener. The -92V supplied 15V zener stable but the LV one starts at -15.6V (diode offset from the other one of 15.2V presumably) and drops through 15.8V . When the difference between nominal +/-15V rails is about 1V it goes outside the capture range of the half U100 dual op amp that feeds back correction into the pa bias area (why there is no bias adjust on these amps)