Crosstalk between gain stages of Mesa Road King II guitar amplifier

[cheater]

Hi guys,
I haven't been to this forum for a very long time so if I'm breaking any
rules please let me know!

I have a Mesa Road King II here. It's basically a very souped up Dual
Rectifier. It's not doing what I would like it to be doing.

You can find a schematic here:
http://milas.spb.ru/~kmg/files/schem...0King%20II.pdf

The amp is in pretty good condition, however the stages have a lot of
crosstalk. What I mean by this is: if I make the signal run through an
effects loop, and put empty plugs into the send/return, the amp should
be totally quiet, but a lot of sound does come through.
If you look at the schematic, the (two) loops are between V5a and V5b
which is a 12ax7. The sound that comes through is fairly loud
especially compared to normal practice-at-home levels; it's a bit
bypassed. Either way, it means to me that V6 is getting the signal
somehow through a voltage rail or some other way.

What this results in is that when my nice Decimator G-String II gate
cuts out because I've palm muted the guitar, i still get a lot of
string rattle coming through, which pretty much makes the gating
useless. The gate gets its control signal from unprocessed guitar, but
it mutes at the very end, in the last effects chain.

Here's how the leakage responds to the various controls on the amp:

chan 1/2 master: at 0, there is no output. turning up to max makes the
output loud.
chan 3/4 master: at 0, there is noticeable output. turning up to max
adds bass to the leaked signal or seems to change the resonance of
whatever "send level" is doing.
gain: at 0, no leaking. at max, noticeable output.
presence, bass, mid, treble, raw/vintage/modern switch, clean/fat/brit
or clean/fat/tweed switch: no control of leaked signal at all
output: at the lowest setting, there's noticeable output. then as you
go 5% up the output becomes duller and a bit quieter(but there's still
a lot of low signal leaking through, just the high freqs are
attenuated), then as you turn output up it becomes really, really,
really loud.

signal leaks even with tuner mute engaged, but a bit quieter. BTW,
engaging tuner mute results in a really loud popping noise.
send level (on the back) seems to not control the loudness of the leak
at all, but there is a resonance, like a 4th order resonant peaking
EQ, and as you turn the knob, the resonance sweeps frequency. So it's
like the freq knob of a really nasty peaking EQ.

taking out V4 or V5 changes nothing, the leak persists. removing
either V1, V2, V3 or V6 turns off the sound completely, including the
leak, and the usual expected hum/buzz that you get even when the
guitar is perfectly quiet.


I have several questions. I would appreciate it if anyone could help me
with this as it's really making the amp terrible!
1. Has anyone come across this kind of issue?
2. How do I fix this?
3. Where would the crosstalk be coming from? Could it be coming from
the preamp tubes' "C" voltage rail? That's the HV rail going to the preamp
tubes. Are there any other places the leak might be occuring?
4. Could this be due to dual tubes being used rather than singles? AKA
could two halves of a tube be leaking to one another?
5. Could this be fixed by re-biasing the tubes in the preamp?
Especially what I assume is a differential push-pull pair made out of
V6a and V6b.
6. Can I decouple the C rail? The idea was to do something like this:
http://tinyurl.com/j35c7dq
however I've never worked with tube circuits and don't know how the
tubes might react to this kind of thing. Naturally the values are
wrong but that's the topology.
7. Does anyone know of a good mailing list or forum that talks only
about servicing and modifying Mesa/Boogie amplifiers? Or just in
general only tube guitar amps?

Thanks everyone!

[Mick Bailey]

With the send level right down, ground the grid of V5b (clip lead across R72) and see if you still get leakage. If the power supply is getting modulated you'll still get sound. If no leakage then you're looking at something before V5b.

Are the FETs all switching?

Mesas run signals all over the place, so capacitive coupling could be an issue, but I would say only if there's another fault that causes this to become a noticable. I would avoid modifying or decoupling anything before understanding exactly what's wrong.

[pdf64]

Mick, note that when V5 tube is removed, the leakage is not affected.
My guess is that it's capacitive coupling between relays, eg RYC is carrying the full overdriven V3b cathode signal.

[cheater]

As pdf64 said, removing V5 affects nothing. I wonder how RYC could be carrying the signal from V3b - would that purely be a layout issue? I don't think RYC is part of a multi-relay box? How would one fix the relay from coupling if that's the case?

[pdf64]

p7 of the previously linked schematic shows V3b connecting to RYC2/L switch common, output feeding channel 3 or 4 tone controls.

I've resorted to screening relay cases with self adhesive copper tape, connected to 0V, to resolve such an issue.
That may or may not be feasible, depending on how densely packed the board is.
But need to confirm that's the issue first.

[Mick Bailey]

Ah, I missed the V5 scenario. The question is, does every one of these amps behave the same, or is it just a fault with this particular one?

[cheater]

Mick, I do not know if they're all the same. I don't know any other people who own a Road King II, and when I got this amp it was probably 10 years old already (just a guess). It still has the original tubes, but they don't seem to have aged, there's not a dent scratch or any wear on the amp, so I'm guessing it got rarely used if ever. Case in point, it arrived to me with the load jack connected incorrectly and would blow its fuse very quickly after being turned on, so I can't imagine it was in the hands of someone who knew what to do with it. My thoughts are the caps and tubes are still in good condition, but one of the first thing I'll do once I get my testing gear set up is check the capacitors somehow (no ESR meter here but a few 7000 series scopes, generators, and an RLC bridge).

pdf64, how does one find out whether that's the case? The only idea I can come up with is hold an RF wand connected to the scope against the relay case to pick up any EM that might be happening, and then to try and recreate similar EM radiation using a sig gen, and see if that gets picked up. But that seems like more trouble than it's worth. Maybe I should try putting a screen here and there. Dunno. I don't own any copper tape but I can probably get some. What's the best material to be doing this kind of shielding - is there anything that's "overkill"? Say, would steel sheet work better? (just for testing)

[cheater]

I also wanted to say, thanks for the great replies so far guys.

[pdf64]

Anything metallic, eg kitchen foil, will do for screening around the relay case, but self adhesive copper tape (eg Stewmac) is by far the most suitable and least hazardous in terms of personal safety and risk of damage, as it will stick in place and can be soldered to (for the 0V connection).
I think that in the 1st instance, RYB and RYC may be the most likely candidates for screening, especially if they're in close proximity on the board.

Caps that have sat uncharged for prolonged periods may deform and develop high ESR, so perhaps replacing C7,8,16 should be considered.
Testing for excessive ESR requires a specialised meter (I need to buy one of these Peak Electronic Design Limited - Atlas ESR - Equivalent Series Resistance Meter and Capacitor Analyser - Model ESR60 ) and even with a reading, then need to decide on pass/fail criteria.
The simplest way is to check whether a good new cap resolves the issue.

[cheater]

Thanks for the info.
I will get some copper tape then. C7/8/16 are on my shitlist already. We'll see if they can defend themselves. From what I understand it's enough to hook up a good cap in parallel that has low ESR.

Look at this. This isn't mine, and I assume it's a different revision (the upper left knob is silver on mine), but it shows what's ahead of me. I might have to rebuild everything between V5b and V6 but luckily that's not a huge job. ~25 passive parts and 7 active parts, I've seen worse. There seems to be enough space in the head for this kind of job.

[pdf64]

Ha, that looks quite a challenge, good luck! Is that 18 relays on there?

From what I understand it's enough to hook up a good cap in parallel that has low ESR.

For test purposes, yes.

[cheater]

Yeah, that's 18 relays. Other than relays, layout proximity, and the C rail, do you see any other ways for the leak to happen? I'm trying to solicit ideas here.

Regarding relays, I should be able to just stuff copper foil between them, ground it, and that should be enough to shield em I guess. I would go around sniffing with an RF probe looking for the largest sources of leakage.

[pdf64]

If the ribbon cables are carrying the relevant signals, they are a likely weak area for coupling.

I'm not sure that an RF probe would be useful, as it's an audio frequency signal.

[cheater]

Sorry, I just meant a radio probe, a B field probe. Just a piece of wire twisted around the plastic ring attached to a water bottle cap. That's how i debugged leaks on some other analog circuits. That goes into a mixer used as a preamp, into a loudspeaker, and possibly into a scope. It picks up normal audio.

I took the wire out of an old UTP Cat 5e cable.

[Mick Bailey]

As I think back I have come across what appeared to be crosstalk problems, though in one of the Mesas that uses vactrols instead of relays - I think there were 29 in total and many didn't fully switch out circuit segments.

With that amp though there were signs of problems with the channel controls - when one channel was active another channel's tone controls would slightly affect the sound, and with the clean channel volume on zero there was signal leakage when the gain channel volume was turned up.

Have you verified that the amp works normally without the loop?