Carlsbro Showman parametric/1 oscillating amp , 1990

[N_Cook]

1 channel good the other one even with a 27 ohm load oscillates at 65
percent of mains and above, bursts peaking at 40 KHz dropping back to much
lower f, extinguishing and restarting over 5 seconds or so.
power amp probably much like Carlsbro S600M
Anyone have any leads to schematic?
uses 4x MJ15024 , 2x SM2177A, 2x SM2178A each ch
predriver DC levels (prior to oscillation) certainly different to the good
ch but no obvious reason other than rain got into the amp
A lamp will go in each of the +/- rails before I fiddle again

ignore any apparent email address appearing with this posting

 

[N_Cook]

Removed the triac speaker prot cct just in case a reaction with that, same
result. Checked ESR of all cap types.
Upto +/-40V rails , below oscillation, all DC agree near enough on both
channels all TO3,TO220 and TO92, zeners.
Go up to about +/-45V and the 6V, 5W bulbs start glowing that I added into
the supply rails , alternately swapping dim and bright pairs very odd. Next
I'll disconnect 1 of each complementary pairing of MJ15024 and see what
happens. Any other suggestions? Next stage is sketching out the cct

 

[David Nebenzahl]

Are you posting from a cellphone? Do you have an internet service provider
charging you by the byte?

Nah, he just uses this public forum to post badly-formatted, cryptic
notes to himself for future reference, apparently.

 

[N_Cook]

Still oscillatated with one pair and also with the other pair. Checking over
the board again there was no continuity between line level i/p ground and
the pa ground. An innocuous 1/4W resistor the same body colour as all the
others but overlooked at 2.2 ohm bands was o/c linking the grounds. No
discolouration except I managed to originally see the third band as yellow ,
not gold. No overheating on the adjascent pcb , so presumably a fusible
resistor.
In hindsight the slow "hunting" effect of the osc giving alternate loading
to main + and - rails should have suggested a ground or zero problem.

 

[N_Cook]

The "good channel" also had an o/c 2.2R but presumably because full o/c no
oscillation there.
The failed one in the oscillating channel had 3.7K across the ends so
presumably not fusible resistors. Must have been a very high current surge
to knock out normal MO resistors without any sign of overheating.