Very Unstable Amp - Oscillation

[Wildcard]

Hi all,

I'm doing some work on a friends Bell Sound Systems Pacemaker PM33 PA head that he wants to use for guitar. It has a major oscillation problem and seems very unstable.

It has fresh sprague filter caps and coupling caps, virtually every resistor has been been replaced as nearly all were out of spec. (i'm using 2 good fluke meters that i know are fine). Every tube - 6SC7, 6AV6, 12AX7, 6L6GC, 6L6GC, 5U4GB are known good replacement tubes. The inputs now use shielded cable and the grids of V1 and the two 6L6's now have grid stopper resistors right on the pins. I've been through every solder joint, every ground point and they all seem fine. Voltages throughout are roughly 10% higher than on the schematic, every plate and grid are OK.

The oscillation seems better with the feedback loop removed but that then gives too much drive and some major ghost notes on the top strings...it's like having a slapback echo pedal on! (not a bad thing but he doesn't want to use it for rockabilly!). As can be expected with these things it's easier to trigger the oscillation with the bass or treble boosted on the 5-way tone selector switch.

Here's a link for the schematic:
http://i263.photobucket.com/albums/i...3schematic.jpg

Here's a short video (sorry, i didn't realise that the video was shot on it's side so you'll need to tilt you head to the left!).
http://s263.photobucket.com/albums/i...keronscope.flv

With a 400hz sinewave through the amp which is connected to an 8ohm dummy load, I get a fairly good sine wave at low volumes which squares off quite quickly as i increase the volume then with the volume at about 3/4 the oscillation starts. The amplitude of the wave grows then falls to a nearly flat line then back to a pretty troubled looking square wave about every 1-2 seconds.

When this happens there's about a 40-50vdc swing on the plates of the OP tubes. The oscillation is in every stage of the amp but gets larger as you go towards the OP.

I've checked for AC on every filter and coupling cap, there's less than 80-100mv on the two 20uf/450v caps, ditto for the 100uf/50v on the cathodes of the 6L6's, but there's around 4vAC on the 40uf/475uf main filter cap....does that sound a bit high???

Ideally I'd like to be able to re-connect the feedback loop as it's a bit wild without it but as it's so unstable I can't.

I'm running out of ideas now and really need some other input, what should I look at next? Hope you can help.

Many thanks in advance

[pdf64]

The ghost notes leads me to suspect your filter caps - just because they're new doesn't mean they're perfect. Try putting an additional 10 or 20uF in parallel with each section, except the 1st one after the 5U4, to avoid stressing it. The 4Vac doesn't sound excessive.
Has this problem just appeared, or appeared after the recent work was done?
Does the oscillation need an input signal, or does just turning up the vol start it off?
What frequency is it?
Is the 6L6 cathode bypass cap good (maybe shorting when hot?). All the heaters are referenced to it, so if it isn't decoupled adequately, problems could appear throughout the amp.
Peter.

[Robert M. Martinelli]

The oscillation seems better with the feedback loop removed

Hi,
I' m a little confused, ideally the purpose of the feedback loop would be to add stability to the design, taking back some of the signal and lowering gain, so having an improvement when opening the feedback loop sounds very strange, unless you' re having some phase problems, in this case closing the loop would result in a positive feedback, and this could be the reason why stability worsens.

As to voltages, a 10% more seems a little too much IMHO, maybe your mains voltage is closer to 120 VAC than to 110.... and 4 VAC of ripple on that filter cap seems a little too much as well, even if the cap is new, it could have spent too much time on a shelf, developing an internal ESR ( Equivalent Series Resistor ) which prevents it from working properly. If ripple is high the amp is more likely to oscillate, try replacing it - BTW, have you tried to measure the oscillations' frequency/period? If it's equal, or a multiple of the mains frequency, this indicates the ripple as the culprit.

Hope this helps

Best regards

Bob

[Bruce / Mission Amps]

I don't think 4vac on the first filter cap will create this problem.
I see lots of good sounding amps with more then that on the main filter cap.
Did you replace that cathode bypass cap on the 6L6s?

Try pulling a preamp tube out, one at a time to see if it stops.
Start with the very first one to see if it is simple capacitive coupled feedback from a stage closer to the power amp leaking signal through wiring back to the front.
These things probably weren't made to accept a huge signal on the front end and maybe just a simple mod to reduce gain will lead you down the right path.

[Wildcard]

Thanks Peter,
Hard to say if it did it before, think it came from ebay, my friend switched it on and it went kaboom! The cap can exploded...jesus...that took about a week to clean up! The cathode bypass is a brand new sprague also. I think i can get it to oscillate without a signal if i turn it up to full quickly with either the bass or treble boosted. I'll report back about the frequency and will try to add some extra caps...

Cheers,

Martin

[Wildcard]

Thanks Bob,
I'm 240v UK through a step down transformer. Yep, it seems like it could be a phase problem, whats the best way to test for this?
The motorboating is worse with the loop connected, strangely it seems a little more stable without it but will freak out when pushed....plus there's the ghost notes too....
Thanks, Martin

[Bruce / Mission Amps]

Thanks Peter,
Hard to say if it did it before, think it came from ebay, my friend switched it on and it went kaboom! The cap can exploded...jesus...that took about a week to clean up! The cathode bypass is a brand new sprague also. I think i can get it to oscillate without a signal if i turn it up to full quickly with either the bass or treble boosted. I'll report back about the frequency and will try to add some extra caps...

Cheers,

Martin

The first stages are contact bias... short their grids to ground with a big 100nF (.1uF) cap and that will tell you if are getting amplified signal back into the first stages.

[PRNDL]

The last oscillation problem I had was solved by changing the lead dress and component locations.

With the "chopstick method", you listen to the amp in oscillation while moving and tapping on things with a wood stick until you do something that makes it a lot worse or a lot better.

There were two wires in a Fender Vibroverb that turned out to be problematic. One wire caused a "pan flute" whistling type of sound when moved a few millimeters. Moving the other wire increased or decreased the noise.

[Wildcard]

The first stages are contact bias... short their grids to ground with a big 100nF (.1uF) cap and that will tell you if are getting amplified signal back into the first stages.

Thank you Bruce, before I go ahead and short the grids, I just need to clarify what i'm looking for...should I get no signal at all if all is OK???

FYI the cathode bypass cap is a new one too...infact it's the second new one...I changed it a couple of days back just make sure.

Cheers

Martin

[Wildcard]

I've been through it with the chopstick, moved anything I can, makes no difference at all, been over the solder joints in great detail, re-flowed or replace just about everything I can....

[PRNDL]

This is a shot in the dark, but you might want to disconnect the hum balance from the cathode RC and ground the center lug of the pot with it in the middle. I'm wondering if the issue is the lamps are off and putting an AC voltage on the bias, which may cause the waveform you describe. (Or you could scope the bias and see if it's steady DC). Either way, I don't think it's a great way to get a DC voltage on the heaters.

[Wildcard]

Thanks PRNDL, I had wondered about grounding that centre lug, i'm actually getting 6.8-9vac on the heaters...bit high really. I've got a few things to go on nowe, i'll try and do the work this evening and get back to you guys in the morning. Thanks again!

[Wildcard]

points to the second stage...

OK, I've added extra filter caps in parallel with everything except the tube rectifier cap and that made no difference at all. I've disconnected the hum balance wiper from the cathodes of the 6L6's and grounded the wiper....no difference. I pulled tubes starting from V1 and the oscillation stops when I pull V2 (6AV6), i've tried 2 other good tubes but that makes no difference.

The frequency of the oscillation looks to be 120hz.

I've tried shorting the grids of the input tubes to ground and have run a signal through the amp and nothing is getting through.

I can definitely make the oscillation happen with nothing plugged into the amp by simply turning the amp up to full quickly.

If it looks to be connected with V2 what can I try next???

Thanks to everyone so far for your input! You guys are fantastic.

Cheers
Martin

[Robert M. Martinelli]

Hi Martin,
the 6AV6 is a double diode/triode, according to the schematic only the triode is used, the two diodes' anodes ( pins 5 and 6 ) should be grounded, can you check is they actually are at GND potential in your amp? Maybe I' m only shooting in the dark, but if they' re not grounded some signal could get in and leak through to the triode due to the tube's inner capacitance between the electrodes and cause the amp to oscillate...Any electrode left "open" in a twin/triple tube could be a source for disturbances.
Hope this helps
Best regards
Bob

[Enzo]

Are you sure it is oscillation and not loud ripple? 120Hz is suspicious.

Pull V1 and it still oscillates? Pull V2 and it stops?


Then ground the grid of V2. That stop it? We want to determine if V2 is just passing along something from before it.

Is there any AC or DC voltage on that grid? Any chance something is shorted to that grid? Look at the V2 socket, any chance two pins are touching each other?