Drawmer 1960 B+ oscillation.

[Greg Robinson]

Hi everyone, got a Drawmer 1960 compressor in for repair recently.

Owner had just bought this secondhand cheaply because it was only partially functional. It had been "repaired" before. This is an early unit (date codes point to 1988), with the potted black box modules. Appears these had failed which was apparently common, and somebody tore the whole thing to pieces "troubleshooting", taking out many, MANY pads, traces and connectors in the process, before they finally discovered the modules were at fault and installed replacements. This is my best guess because the damaged sections were seemingly at random, but the currently installed modules have a couple of long pins that have been folded over to bridge lifted pads, that wouldn't have been possible with the original factory modules which would have had pins cut short after installation.

Anyway, after tracking down all the faults, most of which had been caused by the earlier "repair" work, the B+ supply is oscillating. All parts (with the exception of the TL783 adjustable high voltage regulator - though in situ it performs fine, with phantom voltage well regulated and accurate) test within tolerance, and electrolytic caps have been replaced.
In the process of troubleshooting, I discovered that removing the 10n cap (2n2 on the schematic, silkscreen shows 10n) between the MPSA42 diff pair error amp collectors cured the oscillation, and the regulated output is solid in use. Relevant section of the schem is on page 7 of the PDF.

Now clearly, at some point in time this circuit functioned correctly as designed. I've been unable to identify why it no longer functions correctly, so I'm second-guessing myself.

Anyone have any thoughts? Thanks.

EDIT: should add, the oscillation with the cap in place is roughly 430Hz, 5vp-p sawtooth, riding on the 200V B+. Removing the cap subdues the oscillation entirely, it doesn't simply move it outside the audio band or anything like that.

[Helmholtz]

Is the 100R series resistor still there?

Did you check ESR values of the ecaps?

What is the oscillation frequency?

[Greg Robinson]

Yes, the 100Ω resistor is still on the board, could equally have removed that instead of the cap, cap was just easier to pull.
ESR on all caps is fine.
I just added that on an edit shortly before you posted, 430Hz +/-20Hz, it jumps around a little.

EDIT: should also mention, I tried the schematic value of 2n2, it also oscillated at the same frequency.

[Helmholtz]

Strange effect. Actually the RC network should suppress oscillation.

[Greg Robinson]

My thoughts exactly, hence my second guessing my results. I've left the unit powered on for a couple of hours now with the cap removed, checking periodically that it's still functioning correctly, tested with signal, including overdriving the tube stages (an advertised feature), watching B+ on my oscilloscope, all has been perfect. But as soon as I add the cap back it breaks out into oscillation.

Perhaps pertinent, with the RC network in place, switching the channels to "Mic Phantom" input causing the oscillation frequency to briefly dip to a slightly lower frequency, returning to ~430Hz after around a second. That originally had me suspecting the TL783, but with the RC network disabled it works correctly.

I'm pretty much ready to call it "good", put a note in the chassis and move on at this stage.