Reverend Guitar 120Hz buzz

[Matador]

I have a Reverend Hellhound guitar amp that I can't seem to locate a strong 120Hz buzz signature. This one has me stumped.

Annotated schematic attached (for the Reverend Kingsnake, which is almost identical except for some power stage changes which aren't relevant here): the amp was damaged during shipping, so I had to replace the power transformer, all of the tubes, and rebuild the power supply PCB due to a crack. Then, starting at the output, I worked my way back through the circuit trying to isolate where the buzz was coming from:

1) When I plug into the return jack and inject a signal there, the amp is quiet, loud, and sounds great. So the buzz isn't coming from the phase inverter or the power section/feedback path.
2) Grounding point A on the schematic and the buzz disappears (this should be the same as using the return jack).
3) Grounding point B on the schematic and the buzz disappears.
4) Grounding point C on the schematic and the buzz disappears (the grid of V2).

So far, so good. Here's where it gets interesting.

If I pull the V4 (reverb) tube completely, the buzz remains. I also hard ground point D (the reverb level/mix) and the buzz remains. I think this eliminates the reverb send/return from the equation.

Leaving V4 out and leaving point D grounded, if I hard ground point E, the buzz remains. The only different between hard grounding point E and point C at this point is Rb, the 470K mix resistor.

I removed this resistor and tested it, and it measures 477K and appears normal. I also replaced it with a known good 470K just to be sure. However if I leave points E and D grounded, I still get a loud 120Hz buzz signature through the amp if Rb is in the PCB.

This mystifies me. Somehow Rb is developing a noise across it with zero signal flowing through it, and I can't isolate it any further. If I pull Rb completely, and inject a test signal into the 'right' terminal of the resistor, I get a clear, loud test tone with no background buzz (this is the same as driving a signal directly into V2).

At this point I must have missed something incredibly obvious....

 

[Matador]

Also, irrespective of the loud 120Hz buzz signature, if I plug a guitar into the input, I get amplification through the chain, and all of the controls work as expected: it's just unusable with the background buzz.

 

[dmp]

I don't see the test points marked on the schematic?

 

[dmp]

Sounds like the buzz is originating around the first tube.
I don't think it is isolated to Rb, although I see why it might look like that.
If you ground after Rb the buzz disappears, but if you ground before Rb the buzz remains, but often the buzz is coming from a dirty ground, and I think you are just picking that up.
First make sure you are connecting your test lead to ground in a good spot. Sometimes it can be hard to get this kind of test to work in the high impedance part of an amp.
I assume pulling V1 with everything else the same quiets the buzz (does it?).
Have you tried grounding the grid of V1a and V1b? Do these both quiet the buzz? Ground these to the ground side of the cathode resistors in a short path.
Check the ground of VR1, that is a likely spot for the problem.

 

[radardoug]

The reverb pot is wired wrong, there should not be short from top to wiper.

 

[CJ]

Do you have the reverb tank connected? It looks like they left out a 1 m grid resistor on the reverb return circuit, Marshall did the same thing on some of there models, never seen a 12ax7 get so hot with that open grid.

Weird amp with that 10k cathode r on the first stage

 

[Matador]

Ok, a few more measurements (and answers to above):

Put all tubes in, and pulled them out one by one. Pulling V4 (reverb) tube changed nothing. Pulling V1 reduced buzz by 4 or 5 db. Pulling V2 reduced it further. Pulling V3 reduced it yet some more, but it's still noticeable. Pulling phase inverter and (of course) the buzz disappears. Confirmed that injecting a sine wave directly into the return jack and the amp works quietly, at least as a power amp.

However tracing backwards from the return jack is puzzling: if I short VR2 (master) center pin, amp goes silent. Shorting between VR2 and R36 and the amp is silent. Shorting at point B makes no difference to the buzz (unlike what I said above). I think this eliminates anything before V2. I removed VR2, R36, and C13 and all are fine.

However I tried a new test which is again surprising: I tapped the signal from the 'send' jack and sent it to an external (solid state) power amp: the preamp has very little noise out of the send jack, even when V1 through V4 are all populated. There's something happening at the front end of V3, that is dependent on the source impedance leading into the V3 grid. If I hang a 1MEG resistor between the return jack and ground, then open the send jack, then I can hear the 120Hz buzz loudly.

I measured all of the PSU rail voltages: the VA supply is 424V with about 2V of ripple. The other supplies appear quiet, with no ripple measurable on VC through VE. The bias voltage looks noise free. I tried several different phase inverter tubes, all have the same noise. I also disconnected the NFB loop and there's no change either.

Finally, I placed a pedal with a buffered output (a tuner pedal) into the effects loop, and finally, the preamp and power amp work together as expected, and the amp sounds as it should (even with reverb).

So to sum up: if I break apart the preamp and power amp, I can use the preamp just fine, and it's noise free, and I can use the power amp noise free so long as the source impedance leading into C16 is low. However the noise starts when I connect them together, or have any appreciable (> 500k) source impedance leading into the power amp.

 

[radardoug]

Is the area round the 470K wired in shielded wire? Or what happens if you shield the 470K? Thats a high impedance node, and everything needs to be shielded.

 

[Matador]

Is the area round the 470K wired in shielded wire? Or what happens if you shield the 470K? Thats a high impedance node, and everything needs to be shielded.

Yes, the runs are shielded around Rb. However given the testing I did today I don't think Rb is the issue any longer.

 

[Matador]

So this got even stranger:

As I said, if I ground the input to the PI, the amp is silent (with just a faint background hum, that is a mix of 60Hz and 120Hz). If I feed the input to the PI with a low-impedance source (like a 50ohm signal generator) or a line-out, the power amp works normally: I can hear the signal clearly without background noise. The output tubes are set to 60% static dissipation (approximately 30mA @ 400V on the plate), and the negative bias on each power tube grid is about -49V. The two phases of the PI look normal and approximately equal amplitude.

However if I open the phase-invertor input, or ground it through a 1M resistor, many things change. The horrible buzz begins: the top power tube develops a grid current (about 2mA), which I can measure across its grid stopper, which drags 'up' that side's bias voltage to -37V, which causes a large imbalance between the top tube and the bottom tube (12mA static imbalance). Ultimately, this is the cause of the buzz, as the cancellation of 120Hz greatly diminishes once the current through the push-pull pair becomes so imbalanced.

However even stranger: I removed the power tubes, and installed a different matched set, and now everything works normally! The imbalance disappears, and even with an open input to the PI tube the amp is (mostly) quiet, and now works properly with the preamp. I just chalked it up to being sold a bum power tube, however even more perplexing, I installed that same tube set in a Deluxe Reverb, and it works perfectly normally as well!

So although I understand what was causing the problem, the underlying root cause still makes no sense to me. Why would the input impedance to the PI cause a static DC imbalance between the output tubes? I tried 5 or 6 different PI tubes, and all did exactly the same thing, and the circuit is a pretty standard LTP that you can find in a million other amps. The plate resistors, coupling caps, grid resistors, and bias resistors were all removed and tested and all were normal.

Can anyone explain what underlying mechanism could be at play here?

 

[Audio1Man]

Several things: First assume that someone has worked on it and made some changes.

• Remove R? 220k near “D” is the hum gone? Then hum pickup?
• Cx2/Cx3 are connected to ground and is it directly connected to the CT of 6v winding or ?
• Is the VDC 60 or 120 Hz ripple? Bad bridge?
• The LED is wired backwards
• Power tx was replaced with new sub or factory type? Hum pickup to reverb tank?

Duke

 

[radardoug]

When your output tube starts generating grid voltage it could be oscillating.

 

[murrayatuptown]

A problem with a switch contact in a jack?

 

[murrayatuptown]

There is something appealing about the efficient way of drawing dual triodes as they did...reflecting the physical properties. I'm not accustomed to seeing that...there is probably good and bad built into this, but I'm liking it at first glance...

 

[Ike Zimbel]

So this got even stranger:

As I said, if I ground the input to the PI, the amp is silent (with just a faint background hum, that is a mix of 60Hz and 120Hz). If I feed the input to the PI with a low-impedance source (like a 50ohm signal generator) or a line-out, the power amp works normally: I can hear the signal clearly without background noise. The output tubes are set to 60% static dissipation (approximately 30mA @ 400V on the plate), and the negative bias on each power tube grid is about -49V. The two phases of the PI look normal and approximately equal amplitude.

However if I open the phase-invertor input, or ground it through a 1M resistor, many things change. The horrible buzz begins: the top power tube develops a grid current (about 2mA), which I can measure across its grid stopper, which drags 'up' that side's bias voltage to -37V, which causes a large imbalance between the top tube and the bottom tube (12mA static imbalance). Ultimately, this is the cause of the buzz, as the cancellation of 120Hz greatly diminishes once the current through the push-pull pair becomes so imbalanced.

However even stranger: I removed the power tubes, and installed a different matched set, and now everything works normally! The imbalance disappears, and even with an open input to the PI tube the amp is (mostly) quiet, and now works properly with the preamp. I just chalked it up to being sold a bum power tube, however even more perplexing, I installed that same tube set in a Deluxe Reverb, and it works perfectly normally as well!

So although I understand what was causing the problem, the underlying root cause still makes no sense to me. Why would the input impedance to the PI cause a static DC imbalance between the output tubes? I tried 5 or 6 different PI tubes, and all did exactly the same thing, and the circuit is a pretty standard LTP that you can find in a million other amps. The plate resistors, coupling caps, grid resistors, and bias resistors were all removed and tested and all were normal.

Can anyone explain what underlying mechanism could be at play here?

I know this post is over 18 months old but... I have an expression WRT trouble-shooting: "When you've eliminated all of the smart things, it's going to be something stupid".
You have done an admirable amount of testing and documenting your results. However, here's my question, also raised by @murrayatuptown,
Have you checked the integrity of the Send/Return jacks, both the switch contacts, but also whether they are tight and making a firm connection to the chassis?

 

[Matador]

Have you checked the integrity of the Send/Return jacks, both the switch contacts, but also whether they are tight and making a firm connection to the chassis?

I haven't thought about it for a while, but in my last summary, I had completely bypassed the send/return loop. The experiments with the input impedance to the PI were done right at the input to the coupling cap, not involving the return jack at all.

The fact that the input impedance to the PI caused one of the tubes to enter grid current is still a mystery to me.

 

[Accelerator]

@Matador​

Starting from the very beginning of this schematic I saw something phenomenally “strange”: the R10 – 1MΩ Resistor to be parallel with the Cx101 (capacitor?...), more over on than that the Cathode Resistor of V1(a) R11 is (unreasonable) 10KΩ(?) and the Cathode Resistor of V1(b) is (unreasonable) 15KΩ(?) while they have “tried” to make “something like” the Marshall JCM800 input section but with totally wrong values…

Also the LT Capacitors Cx1 – 3300μF and the Cx2/Cx3 – 1000μF looks to me too low for the filtering of the DC heating of the pre-amp Valves (V1-V4) and I preferred something with bigger filtering Capacitors in μF values for this purpose…

I would like to know the Voltage Values of VA, VB, VC, VD, VE…

Can you measure them and post them to let me know?...

 

[Matador]

the R10 – 1MΩ Resistor to be parallel with the Cx101 (capacitor?...), more over on than that the Cathode Resistor of V1(a) R11 is (unreasonable) 10KΩ(?) and the Cathode Resistor of V1(b) is (unreasonable) 15KΩ(?) while they have “tried” to make “something like” the Marshall JCM800 input section but with totally wrong values…

The input cap across the 1M is pretty standard for RFI suppression. The high value cathode resistors with small(ish) bypass caps is found in a lot of Soldano and modded Marshall JCM800 clones (pseudo "cold clippers").

Alas this amp left my bench soon after posting this so no more voltages left to be measured.