Orange Rockerverb reverb driver oscillation

[htrd]

Hello, I'm back!

This time I have a problem with Orange Rockerverb 50 Mk2. It has some annoying high frequency fizz on high volume, at the edge of breakup of single channel, where PA starts to distort. Both channels affected.Long story short I was able to track it down to reverb driver stage:
- if I pull out driver tube, fizz is gone;
- if I plug out reverb tank, fizz is gone;
- if I pull out the reverb recovery stage, fizz is present;
- if I plug out only the tank output, fizz is present.

So it looks like this oscillation is going to EL34 screen grids through the HT node. According to the unofficial schematic from the Tube-Town forum, the reverb transformer uses the screen supply without the additional filtering.I measured R21, R22, R23, R66 and C12 without desoldering, looks ok (R21 is more like 186K). Also I've tried to scope the oscillation, it appears right after the R23 and goes to the grids of both driver stages. Can it be bad R23 or the problem is more complex?

Schematic contains few errors, I didn't trace them all, but R22 is 2K2, reverb driver and recovery stages are 12AT7, lead channel gain is 1M and so on.
I attached an audio sample, with and without a reverb tank.
reverb_tank_distortion.mp3
Orange_RockerverbMKII_PS.pdf
Orange_RockerverbMKII_Pre.pdf
Orange_RockerverbMKII_PWR.pdf

[Jon Snell]

So you unplug the tank and the fizz has gone.
You unplug the driver, (which valve do you mean?) and fizz has gone.
Pull out the recovery stage and fizz still present. Do you mean V11?
Disconnect the tank feed and fizz is present.

Is that correct?

[htrd]

Reverb driver is V6, reverb recovery stage is V11, both are 12AT7. I forgot to mention that tubes are good, replaced both, no effect.

Disconnect the tank feed and fizz is present.

Tank output. Input stays connected to the reverb transformer.
Everything else is correct. So the fizz is gone if I pull out V6 or disconnect tank input.

[Helmholtz]

Make sure R23 is actually 220k.

I don't think this is oscillation/instability and I don't think the distortion gets to the screens, assuming the 100µ screen filter cap is good and well connected.

You might have grid conduction in the reverb driver or more likely something like this:
https://music-electronics-forum.com/...ibrolux-reverb

Does lifting the cathode bypass cap cure the problem?

[htrd]

Make sure R23 is actually 220k.

I don't think this is oscillation/instability and I don't think the distortion gets to the screens, assuming the 100µ screen filter cap is good and well connected.

You might have grid conduction in the reverb driver or more likely something like this:
https://music-electronics-forum.com/...ibrolux-reverb

Does lifting the cathode bypass cap cure the problem?

R23 is 220K, I've measured it. I haven't done any soldering experiments yet - big PCB with direct mounted potentiometers, but problem depends on overall volume and amount of low frequency in case of guitar signal. I was thinking about screens because the fizz is present with pulled out recovery stage and disengaged reverb via footswitch or reverb potentiometer. Looks like the only way to the signal path is through the B+.

[Helmholtz]

Looks like the only way to the signal path is through the B+.

Not saying it's impossible but the path from recovery tube grid back into the dry signal is much more likely.

Can you post scope pics of the recovery grid signal and the signal between R9 and C3?

[htrd]

Recovery tube was pulled out.

Can you post scope pics of the recovery grid signal and the signal between R9 and C3?

Sure, will do tomorrow. Point between C6 and R9 was clean, distortion appears between R23 and R66, on both grids.

[htrd]

Ok, I've got some new results. Looks like you are right and this is transducer saturation.
First pic - tank connected, driver installed, there is slight kink between R9 and C3. Reverb is disabled via footswitch, RL2 on the schematic.



At the reverb recovery grid everything is ok. Reverb enabled, of course.



Afterwards I loaded the reverb transformer with 10R across the RCA jack and the fizz was gone. First pic is between R23 and R66, the second - between R9 and C3.

[Helmholtz]

Afterwards I loaded the reverb transformer with 10R across the RCA jack and the fizz was gone.

Nice test!

You might even consider permanently wiring a resistor of a few ohms in SERIES with the drive transducer to somewhat lower the drive.

Please let me see the signal at the reverb driver grids.

[htrd]

Third picture, between R23 and R66 is the first driver grid. Second one was also clean.

I was thinking about swapping 12AT7 with 12AU7 or even lifting a bypass capacitor to lower the driver gain. Anyway reverb is extremely spacious and almost unusable at higher mixer settings.
Reverb tank is Ruby RRVS3AB2C1B, looks like the direct replacement of Accutronics 9AB2C1B. You wrote that tank swap didn't help in your case, right?

[Helmholtz]

Third picture, between R23 and R66 is the first driver grid.

I meant with tank connected to see the grid distortion.

... or even lifting a bypass capacitor to lower the driver gain.

I already suggested that.

But do try the 10R in series. It will attenuate some undesirable bass content in the drive signal.
It's mainly the lower frequencies that cause saturation.
You might have to experiment with different resistor values though.

You wrote that tank swap didn't help in your case, right?

Yes, tried 3 Fender type Accutronics tanks.

[htrd]

I meant with tank connected to see the grid distortion.

It's mainly the lower frequencies that cause saturation.

Yes, I noticed that lowering bass in the tonestack or using single coils instead of humbuckers helps a lot.

I think I have Accutronics 4ab3c1b in my Twin, I should try it next time just to satisfy the interest.

[Helmholtz]

So no grid conduction but transducer saturation + Miller effect, just like what I had.

I tried 4AB and 9AB type Accutronics tanks. No difference regarding saturation, but I generally prefer the 2-string (4AB) in my Fender amps.

[htrd]

Well, interest is satisfied - 4AB gave the same result.

[htrd]

But do try the 10R in series. It will attenuate some undesirable bass content in the drive signal.
It's mainly the lower frequencies that cause saturation.
You might have to experiment with different resistor values though.

Yes, that works, thank you.
We made a few temporary RCA adapters with different resistors in series to quickly test a few options, starting with 8R and went up to 51R when the ringing became negligible. It is still there, but only at the loudest peaks when playing guitar, not the sinewave. On the other hand, the reverberation volume became very low, so now I need to raise the gain in the recovery stage. Maybe just swap at7 with ax7, a low effort solution?
Also it looks like the driver stage has a rather low headroom, grids were clipping with just 1Vpp sine wave into the return.

Test adapter


0.7Vpp sine into the return, reverb driver grid. 51R in series with transducer.


Same but 1Vpp.