Traynor YCV40 - Channel Crosstalk problem

[tertiumsquid]

Hi all, I've got a Traynor YCV40 on my bench, picked it up for $140 in "broken" condition (really only needed new power tubes). The circuit is stock, except for one bias resistor change so the amp can run EL34's. I'm getting a lot of channel crosstalk through the amp. Let me explain exactly what happens:

• All knobs at zero, clean channel, strum the guitar, sound comes out of the speaker at around 80-85db. Sound coming out is quite clean and loud.
• All knobs at zero, dirty channel, strum the guitar, a tiny amount of sound comes out.
• All knobs at zero, dirty channel and boost engaged, strum the guitar, sound comes out of the speaker at around 75-80db, a little quieter than the clean channel but sounds REALLY farty and distorted.

This has led me to determine that the second gain stage V2B is the culprit, as the dirty channel with no boost doesn't use V2B.

Changes I've attempted so far: Removing V1A and V2B cathode bypass caps (C15, 22uf). Adding a 22uf 450v cap in parallel to the C+ line. Swapping preamp tubes to other known good tubes. Running the amp with no power tubes, and listening through the effects send.

Diagnostic attempts (clean channel):

1. Swapped preamp tubes with known good and/or new tubes. No change.
2. Using a signal generator, I plugged in a 1khz 1vP-P signal into the input.
3. Following the signal with my oscilloscope, I get the same 1vP-P at V1A grid, and about 22vP-P at V1A anode.
4. With EQ controls halfway, I get about 5vP-P into the volume pot.
5. With the volume turned all the way down, I get nothing out of the volume pot (as expected).
6. At the grid of V2B, I also measure nothing.
7. At the anode of V2B, I get ~800mvP-P, where I should measure nothing.
8. Following that through the chain, I get nothing at the effect send or return, but ~40mvP-P at the output of U1A (effect return opamp output), where again I should get nothing.
9. Same measurement at the grid of the phase inverter.
10. About 1vP-P at each anode of the phase inverter.
11. Same measurement at the grid of each power tube (pin 5).
12. Watching the C+ line, I get nothing (or at least nothing my scope can detect).
13. Inserting a volume control in the effect loop can effectively reduce the signal to zero. Crosstalk must be happening before the loop.

It might be down to a problem with the PCB layout and signal flow in this amp. The signal flow is V1A>V2B for clean channel and V1A>V2B>V1B>V2A for dirty channel with boost, then the signal goes back over to the master volume and channel switch relays next to V1A, back over to the opposite end of the board to the effect loop op-amps next to V3. The signal is going all the way across the board 4 times before leaving the preamp (6 times for dirty with boost). That said, I've never heard anyone complain about this problem on this amp.

Also, keep in mind right now I only have a crap oscilloscope (Jyetech DSO062, little DIY 1 meg scope, resolution only down to ~10mv and max voltage of 50v). I'm picking up an older analog scope hopefully tomorrow.

Any thoughts? I've attached the best schematic for my version of the amp.

 

[JonSnell Electronic]

Have you checked the ground connections on the controls and the input jack.

 

[tertiumsquid]

Have you checked the ground connections on the controls and the input jack.

Ah yes, forgot to mention that I did. I reflowed all the connections at the pots just in case, and often used the ground from the input as my reference for the scope.

Interesting enough, the amp didn't originally have a ground connected to the chassis; the power cable ground went to a bolt on the chassis, but not to the amp. I'd guess it was supposed to connect to ground through the input jack, but the chassis around it was all painted and didn't get a solid connection. I ran an extra wire from the chassis to the circuit ground, but that didn't make a difference to this problem.

 

[PRR]

> All knobs at zero, dirty channel and boost engaged

Is that a normal condition, all knobs zero?

I agree it is unexpected. But if it does OK at most normal knob settings, I would not get picky about the all-zero condition.

 

[rayma]

Have you checked the ground connections on the controls and the input jack.

Sometimes the pot itself will cause this problem, because of too much resistance
between the wiper and the end of the resistance element when turned to minimum.
I haven't seen this problem for a long time, though.

 

[tertiumsquid]

> All knobs at zero, dirty channel and boost engaged

Is that a normal condition, all knobs zero?

I agree it is unexpected. But if it does OK at most normal knob settings, I would not get picky about the all-zero condition.

Well, I mentioned all knobs at zero because there should be nothing coming through the amp at all. The EQ controls don't change how much signal is coming through when the volume knob is at zero.

I know most amps have some channel bleed - my Fender and Carvin both do it a little bit (maybe 60db tops), but not 80-85db. I mean, think about it, the absolute quietest this amp can possibly go is 30db louder than normal conversation level (50db), 10db louder than a vacuum cleaner (70db), 3db louder than a car driving past (77db), and about equal to a garbage disposal (80db) or a diesel train passing 100ft away (83db). That's not remotely quiet. I'll make a video on the weekend to show what I mean.

Sometimes the pot itself will cause this problem, because of too much resistance
between the wiper and the end of the resistance element when turned to minimum.
I haven't seen this problem for a long time, though.

Yeah, that's the first thing I assumed, though I haven't seen it happen for a long time either. I ruled that out by shorting the volume pot wiper to ground. Didn't change anything.

 

[tertiumsquid]

Update: I've done some more work on the amp, and have fixed the problem with the clean channel, but not the boosted gain channel. I replaced all the filter caps with Panasonic and Rubycon caps (which fixed an unrelated motorboating problem), replaced the cathode bypass electrolytics, and added a jumper to ground the input of the boost stage when the clean channel is engaged, which is what fixed the problem on the clean channel.

Unfortunately I don't have that luxury on the dirty channel, as there's no easy way of re-doing the switching in the amp, so I might be stuck with that. Suggestions are very welcome of course.

I did a quick video on the issue: Traynor YCV40 Crosstalk problem - YouTube

Hopefully that explains what I'm dealing with here. I attached a screenshot of my dB meter app; the 4 sets of spikes are, respectively, guitar with amp standby, clean channel, dirty channel, boost channel. You can see that the clean and dirty are only a little bit louder than the sound of the guitar acoustically, but the boost channel is quite a bit louder. And please believe me that I'm not just being picky about the volume, it really is loud.