need some help with nasty parasitic oscillation

[markmalin]

Hi guys!

I’m having a bugger of a time getting rid of this parasitic oscillation.  The amp otherwise works great, voltages all look normal, etc.  As you can see, nasty HF oscillation when I strike a note hard with volume above 6.  Here’s what I know so far:

-   Note in the schematic, this is where the scope probe was – so there is oscillation detected on the grid wire (500pf/grids/1M resistor).  The signal is dead clean BEFORE the 500pf cap, and has the oscillation after the cap (after meaning the side tied to the grid).  The plate side is the same but amplified (obviously).  The cathode shows no oscillation.
-   If I pull the reverb driver tube, the oscillation is gone
-   If I pull the reverb tank, I have not scoped this, but it sounds quieter.

Here’s what I’ve tried so far (other than pulling my hair out…which doesn’t seem to help)

-   I re-routed the plates-to-reverb-transformer-input wire, it used to loop around the outside of the tube socket between the two grid pins, but I ran it across the top of the tube socket.  Result: seemed a little quieter
-   I shortened the wire from the 1M and 500pf cap to the driver grids, but no noticeable difference.
-   I replaced the driver tube with a mill spec tube, and this reduced the oscillation somewhat, so I left that tube in.
-   I replaced the 500pf mica cap with a ceramic cap.  No difference
-   I replaced the 0.02 ceramic disc with an orange drop same value just for the heck of it, no difference.
-   I replaced the 1M Ohm resistor with a Dale RD65 just in case, no difference
-   If I poke around anywhere on the board near the components in this part of the circuit, or poke around the wires going to this driver tube, no noticeable difference.

I need some more ideas.  Is there a way to shunt this high frequency to ground by putting a cap across the 1M resistor??  If so, what value would I use?  Should I attempt to try a shielded wire for the 500pf to grid wire, since that seems to be the start of the oscillation?

Thanks guys!

Humbly,
Mark

[jjasilli]

I would try replacing, at lease temporarily for testing, the 1M grid leak resistor with a 1M pot ( dwell pot).  See if a lower value kills the oscillation, and still gives adequate reverb vol. 


Is there a way to shunt this high frequency to ground
Yes, and there a number of other ways to employ Local Negative Feedback at that stage.  I think one of our moderators posted a number of such circuits, google it, or re-post.  As to cap value, check those circuits or just experiment.  Can't hurt.

[shooter]

I need some more ideas.

from left field; do you have a 2nd probe?  I had a probe go south and do "that", but only >300V.

Your other ideas, shielded cable are worth trying, JJ's Idea on the R is worth trying.
I'm still learnin on reverb, but the 500pF, I think already throttles back a lot.

again from left field; can you "temp in" a 100K R in place of the reverb coil, making the driver "look" like a gain stage?

[markmalin]

JJ, I had a minute over lunch, but not to wire in a pot.  I did, however, connect a 1M resister in parallel so the resistance was halved (500k) and notice no difference (and plenty of reverb).


ASSIDE - the reverb is WAY to intense on this amp and I replaced the normal 3M3 reveb mixer (the one with the 10pf cap in parallel) with a 1.5 M to try to tame the reverb, but I've done that a few builds back.  Perhaps I should have halved the 1M at the input to the driver?


Shooter, the probe I have is new, and with our without the scope in place, I hear the oscillation riding above the note, then fading as the signal dies down a bit.


One thing I tried over lunch, guys.  I removed the reverb tank and grounded the V4a grid and the signal is quiet.  The Oscillation is not there with the reverb tank removed.  Bear in mind, the oscillation is audible with the reverb off and visible on the scope with the master volume off.  Not sure if I mentioned those two things originally in my post.


I can try a 1M pot in place of that 1M grid resistor tonight and see what happens.

[jjasilli]

There are a few ways to tame the reverb intensity including the removal of cathode bypass caps, and/or voltage dividers.  Since you have clean signal on the other side of the 500pF cap, maybe that cap is defective???


Or maybe input signal intensity is still swamping the input of that tube.  What is the signal voltage at that point?

I removed the reverb tank and grounded the V4a grid and the signal is quiet.
I don't think that tells us anything more, as you've already located the signal issue at a prior stage.

[markmalin]

There are a few ways to tame the reverb intensity including the removal of cathode bypass caps, and/or voltage dividers.  Since you have clean signal on the other side of the 500pF cap, maybe that cap is defective???

Or maybe input signal intensity is still swamping the input of that tube.  What is the signal voltage at that point?

I replaced the 500pf.  was a mica, I put a ceramic in there.  Also replaced the 0.02 fwiw (that's in the original post)


I'll measure the input signal intensity - when you say "what's the signal voltage", can I just read that on the scope (peak to peak of the signal, for example)?  Just want to be sure i"m measuring what youre asking ;)

[markmalin]

I removed the reverb tank and grounded the V4a grid and the signal is quiet.
I don't think that tells us anything more, as you've already located the signal issue at a prior stage.

JJ, I've been thinking about this and reading some other posts.  The tank I bought has input isolated and output grounded.  The jacks I bought are panel mount RCA's Fender style, so they are not isolated from the chassis.  Without the jacks isolated, could that possibly cause a ground loop?

[sluckey]

Fender never used isolated RCA jacks on the old amps.

[markmalin]

Fender never used isolated RCA jacks on the old amps.

Right. That's why I bought the ones I did.


Maybe I'm chasing a red herring, it just got me wondering...if the oscillation on the reverb driver grids stops when I remove the tank, maybe there is a ground loop.


Did I wire the reverb correctly?  Have a look at this close up.

[sluckey]

Those two RCA jacks don't reveal any problems. But, how does the return jack connect to the recovery tube? How long is that wire? How is it routed? And where is the footswitch jack? It should be connected directly to the return jack. And what is that phone jack sitting to the right of the return jack?

[markmalin]

Those two RCA jacks don't reveal any problems. But, how does the return jack connect to the recovery tube? How long is that wire? How is it routed? And where is the footswitch jack? It should be connected directly to the return jack. And what is that phone jack sitting to the right of the return jack?

The return jack connects to the recovery tube from the center of the jack to the grid (V4a).  The wire is routed down along the back of the chassis and directly to the tube socket.  Not too long, maybe 3-4"?


There is no footswitch jack, there's no footswitch for the reverb.


The jacks to the right of the return reverb jacks are line in/out (a passive effects loop.  So the FX send is next to the reverb return, and the FX return is to the right of that.  Those both have shielded wires, with their shields terminated at the 1/4" jacks.

[markmalin]

I think it was JJ who asked what voltage the input signal to the driver tube is.  Here's a picture.  The scope is set to 5v, so if I read this correctly, somewhere around 30v peak to peak?  This is with the volume close to 10.

[markmalin]

sorry for the rapid fire posts.  Here's another observation.


I disconnect the reverb send cable, the oscillation is gone.  I re-connect the send cable, disconnect the reverb return cable, it still oscillates.

[Willabe]

The tank I bought has input isolated and output grounded.  The jacks I bought are panel mount RCA's Fender style, so they are not isolated from the chassis.  Without the jacks isolated, could that possibly cause a ground loop?

No, that's why 1 of the tanks jacks is isolated. (1of the tanks jacks has to be grounded so the tanks chassis acts as a shield.)

If both of the tanks jacks were grounded and both of the chassis jacks were grounded then there would be a ground loop. 

[markmalin]

The tank I bought has input isolated and output grounded.  The jacks I bought are panel mount RCA's Fender style, so they are not isolated from the chassis.  Without the jacks isolated, could that possibly cause a ground loop?

No, that's why 1 of the tanks jacks is isolated. (1of the tanks jacks has to be grounded so the tanks chassis acts as a shield.)

If both of the tanks jacks were grounded and both of the chassis jacks were grounded then there would be a ground loop.

That makes sense.  Just to rule out the possibility, I ordered some isolated jacks from CE and am picking them up today.  This evening I'll re-wire the reverb jacks.


So back to the oscillation, I'm running out of ideas.  I may have to wait until this weekend to try JJ's idea of wiring in a pot to see if the grid resistor on the reverb driver can be dialed down to where the oscillation stops.  I also wanted to try a shielded wire on the driver grid wire.  Would also consider a cap to shunt the high frequency to ground, but I'm not sure where to start with that line of thinking.


Last night I was playing around with it and it seems now there is a crackling sound when I strum a chord really hard.      Makes me wonder if something's arching now.  That wasn't happening before.  However, I need to focus on one thing at a time. I've built this same model several times w/out these issues.  I've got to get this thing dialed in and working -- the guy I'm building it for has been really patient, which I appreciate! 


So is a 35'ish volt Peak-to-peak signal on the input of a reverb driver normal?  As it dies down to around 30 the oscillation stops.  I've never put a scope before the driver before - I do have another demo amp like this, I may open that up to see if things look similar.

[markmalin]

FWIW, I picked up 2 new isolated RCA jacks, wired them in and no difference.  I played it for a while and there is this crackling that happens when I strum a chord really hard.  After playing it for a while the crackling was getting worse, so on a whim I swapped the 12AT7 driver with a tube I bought at CE today, an Electro-harmonix 12ATWC that's supposed to be super low microphonic.  The amp's quiet now.  Will it last?...


In the course of building this amp, this is the 4th driver tube.  The first, a JJ Tesla that I figured was "bad" (DOA) because it seemed to be causing distortion and all sorts of noise.  Second was a TAD and the amp seemed quiet.  Then I turned the volume up past 6 and found the oscillation I've been chasing.  The 3rd was a NOS mill-spec 12AT7 that seemed to reduce the oscillation somewhat.  Then after the crackling started I put in the 4th, this E-H 12AT7WC.  Quiet again so far.


Granted, I didn't have a lot of time to play before the family went to bed, but I'm wondering if this thing is destroying driver tubes, or did I just need to find one that was quiet enough?

[markmalin]

More testing - seems fine now.  Here are some scope traces with the new tube in place.

[jjasilli]

Maybe run the amp with signal for a few hours into a dummyload to see if the problem  reappears. 

[markmalin]

Maybe run the amp with signal for a few hours into a dummyload to see if the problem  reappears.

I could certainly try that.  I've been playing it all weekend and it's still clean.  The traces were generated by plucking a harmonic on the 12th fret of the low E string, so not using a function generator.  just fwiw.

[jjasilli]

 

[markmalin]

Yup!  :)