6V6 Champ style strange distortion

[tubesinside]

Hope all you experts can help. I'm an experienced tube amp service engineer and have not come across this before. I have built several PP amps but my most recent project is a 5 Watt Champ style combo with digital reverb and VVR. All is working fine except when I really drive the amp O/P stage I get a pronounced spike at the trailing edge of the waveform. This sound horrible and needs to be sorted.
The O/P tube is a JJ 6V6 and the traffo is a Hammond 125BSE connected for 5k and 8 Ohm output.
No schematic yet but all the usual precautions.
Cathode bias resistor is 680 R and the HT is 370 V.
Bias current is 34 mA. Clean sound is sweet...mild overdrive is smooth. More overdrive is nasty.
I have a 68k grid block resistor and a 390R Screen Grid resistor.

I'm lucky to have all the necessary test equipment to test the amp. Strangely the output waveform is perfect when connected to an 8 Ohm test load (not the real world, I know !).
I have tried the following to no avail:
Different speaker, different 6V6, increased/decreased loading on O/P traffo, increased/decreased bias current, 200 Ohm resistor across speaker jack, protection diodes from anode of tube to earth, 330pF bypass cap between grid and cathode of 6V6.

Help

[Andy Le Blanc]

Is that the 125B universal SE/P-P?

Square wave puts stress on SE opt. current induction.

I once melted the wax out of a champ opt with a marshal high gain front end.
Everything was 5/5 with the circuit.

Hammond 125B can only handle 5w, max DC bias is 45ma., you might need to go
with a more robust opt...

[JD0x0]

well im no expert but to me it sounds like it could be two things.

1. You may be clipping the digital reverb which would get a nasty tone. Although im not sure how you built your amp so it's hard to say.

2. It could also be what is described as "hash" which is a nasty overtone that usually clashes with the bass frequencies. The tone is compared to someone clearing phlem out of their throat.

[tubesinside]

Thanks for the suggestions so far.....

Andy, It is the special SE traffo with airgap designed for this application. I have tried the amp with lower bias to no avail...don't think its stressing the traffo. I get 4 watts clean and about 5 watts before obvious clipping.

JD0x0, I have eliminated the circuitry before the O/P tube driver. Nothing before the driver tube affects the sound.
I think it has something to do with the high impedance of the speaker at certain frequencies causing the transformer to 'ring' causing the spike.
the spike is visible at the speaker output and at the anode of the tube when seriously overdriven.

I will post the schematic of the amp tomorrow for reference.

C

[jjman]

68k or 6.8k grid stopper? Is it there to avoid grid current when turning the VVR down? Could 68k cause too much miller and manifest in this manner?

[tubesinside]

The grid stoppers are there so that I can drive the tubes hard without the blocking effect. I could change the values. I cannot see any evidence of the spiked distortion at the grid of the 6V6. I have uploaded a rough sketch of the circuit. The Belton reverb module is really good for the price.

[Andy Le Blanc]

Do you know the signal amplitude applied to the 6v6 that creates the transient?

You need to go backwards for a bit, remove the 220r, the feedback resistor,
the vvr and the extra diodes in the PS. Test the 6v6 to the amplitude that
caused the transient. first all by itself, then include the preceding stage,
then the feedback, then with the vvr. find out where in to game plan the
distortion is being introduced add one element at a time.

Its a simple system, if your not getting the distortion from the pre it'll be the power side.

Usually its the added "features" that really mess things up.

[chmuzz]

Hi,

+1 with JD0x0
According to your schematics, you don't seem to have a voltage divider at Belton module intrance (At full tilt, belton module accepts 1.5V Peak maximum at this point).

Otherwise, I'm trying to drive this module too in a EL84 SE amp, but with an ECC832. your recovery/mixer stage is quite simple. Does it works well ?

[tubesinside]

Hi Chmuzz/JD0x0
The cathode resistors are arranged as a 3:1 attenuator on the 12AU7 cathode follower. The maximum signal into the Belton is within the 1.5V peak...works very well....no sign of distortion at the output of the module. I have also eliminated the module from the equation by completely removing it...(plugable leads). I will post signal levels in a couple of days for anybody that wants to use that module. The recovery/mixer circuit is simple and works well.
I have already tried this circuit without the VVR etc. But will follow your advice and revert.

[chmuzz]

Hi Chmuzz/JD0x0
The cathode resistors are arranged as a 3:1 attenuator on the 12AU7 cathode follower.

OK Sorry, my old eyes didn't noticed it !:shock:

[tubesinside]

Update:
I have disconnected all of the preamp stages to the 6V6 and now drive the grid directly from my test set (through a capacitor) but with the 220K bias resistor and the 68k Grid stopper still connected. With a sine wave I can drive the tube to full output with 5.5 V RMS but not enough signal output to be able to be able to drive it into that horrible distortion. However, when I drive it with a square wave into a test load it's not too bad as Picture 1 below. Then I drive the same square wave into the speaker and I get severe ringing / peaking on the square wave. (Picture 2).
I'm thinking that I should be looking for my distortion elsewhere...maybe drop in a 12AX7 into the reverb recovery/mixer position and pad down the output?
I did manage to get rid of the overshoot by adding a .022 capacitor between the anode of the 6V6 and ground. Also got rid of my high frequency response. Smaller value caps do not work.

[Andy Le Blanc]

You might need a conjunctive filter in series with the opt or a different opt.

Ive seen that before, operating a SE with square wave to a point where the
problems are created. If thats the issue your gonna have to reduce the input.
to where its manageable with the SE design. Check out the Garnet "Herzog".

It might be loading issue?

[jjman]

I've never used the square wave mode for testing my amps. Only sine. Looks like overshoot with no ringing. My only real knowledge would be from the below.