Tube Amp Tremolo Thumping

[Paul Marossy]

I have an old Gregory Mark X tube amp from 1965 which uses an LFO that messes with the power tube bias. I totally reworked the amp many years ago - fixed bad wiring (lead dress), replaced a few aging components, star grounded it, added digital reverb to it, etc. Honestly, it's my favorite tube amp. I have a few other well-known amps that are considered to have great "tone" but this one is my go to. Anyway, if I turn up the Strength knob up past a certain point I get a thumping noise. I know that it is a common problem with this kind of tremolo arrangement. I recently read that there is a trick you can do with a diode across the strength pot, but that seems to have mixed results... have read everything from it was a wonderful perfect fix to it messed up the function of their tremolo. Anyone here have experience with this? Can a leaky cap contribute to this? The schematic is here: http://www.diyguitarist.net/PDF_Files/MarkXSchemR1.pdf

One other strange problem I fixed but thought I'd mention here as well - maybe someone might benefit from it in the future. I had a noisy volume pot. It had been a long while since I've had to poke around in a tube amp. I thought maybe I had a bad capacitor passing DC but it seemed like it might actually be a bad pot. Didn't seem likely... I mean you'd have to twist that knob countless times to wear it out and I don't think the amp has been used THAT much. This is where my oscilloscope octopus curve tracer came to the rescue. Turns out it was shoddy work done by someone before I acquired it - they replaced that pot but instead of doing it right, they just cut the lugs off the old pot and soldered the clipped off lugs to the new pot. Apparently, they were too lazy/ignorant to do it the right way. Anyway, once I rewired it the proper way the scratchiness was gone! I did not expect it to be that simple or for that to be the problem. The oscilloscope actually helped me to figure out it was the pot... I could see it being noisy when I turned the knob, with the power off. After I re-worked that wiring to the pot it looked nice and smooth on the o'scope. So I guess maybe it was a semi-cold solder joint? Weird huh? I thought for sure I had a leaky capacitor. A scratchy pot doesn't necessarily mean that the pot is bad or that a cap is leaking DC, at least not in a tube amp.

[GibsonGM]

That's a good piece of info, Paul! 

[Eb7+9]

I know that it is a common problem with this kind of tremolo arrangement ...

not really,

you need equal Z's on each side of the feeder resistors for these bias modulators to work properly ... I guess it's easy to miss if you're not looking hard enough

[Paul Marossy]

I know that it is a common problem with this kind of tremolo arrangement ...

not really,

you need equal Z's on each side of the feeder resistors for these bias modulators to work properly ... I guess it's easy to miss if you're not looking hard enough

Not sure what you mean... I see mirror image each side of the phase inverter with the LFO connected to the center of the network. I was thinking that maybe that .05uF cap that connects the two areas together is possibly leaky. I tested it with curve tracer and it appeared to be OK, but I guess I ought to look for any DC getting thru with a DMM next.

[Eb7+9]

Paul,

everything looks symmetric until you look back towards the cathodyne (single triode) phase inverter ... the driving point impedance at the plate is vastly different than at the cathode ... the result, un-even LFO drive to the power tubes means it turns into perceivable signal

[Paul Marossy]

Paul,

everything looks symmetric until you look back towards the cathodyne (single triode) phase inverter ... the driving point impedance at the plate is vastly different than at the cathode ... the result, un-even LFO drive to the power tubes means it turns into perceivable signal

I thought that might be what you meant. I think the way this one is designed there is not much that you can do about it, but I still wonder if it could be improved a little bit. I don't really use the tremolo much but it would be nice to have it working as optimally as possible.

I found this which explains the pros and cons of a bias shifting tremolo. I guess this amp utilizes the least desirable way to do it?

Bias Shifting Tremolos:

How it Works-

These tremolos work by changing the bias to either the power tubes or a pre-amp tube in a amp.   By changing the operating parameters of the tube the signal can be cut off momentarily and the brought back up. This creates the tremolo effect. These tend to have a very rich and soft pulsating sound.  The signal does not pass through the tremolo circuitry so they do not effect the fidelity of the signal.

Types of bias shifting Tremolos-

There are two basic types of bias shifting tremolos: one that work on the bias of the power tubes and ones that work on the bias of a pre-amp tube.  Power tube shifting trems are generally more common and found in fixed bias amps.  Most cathode biased use pre-amp tube bias shifting trems but there is version that modulates the bias indirectly through the grid leak resistors of the power tubes

Design Concerns-

Bias shifting trem in fixed bias amps that vary power tube bias have several issues:

1. The bias must be set correctly on the power tubes for the effect to work.  The problem with this that sometimes you must bias the amp cooler than what may sound best or the trem will not function well. Another problem that occurs is beating sounds that can be very annoying.  You can minimize them but not always eliminate them.

2. Bias shifting trems that work on power tubes are hard on the power tubes and power supply which can lead to failures.  Since the trem works by affecting the bias of power tube it can affect the stability of the power section. It is not advisable to run the trem with an amp cranked up with this type of tremolo.  The strain on the tubes can lead to failure. In worst case scenario a tube failure could even damage a transformer.  Over the years I have come to conclude that power tube bias shifting tremolos are not as wise to install as the other types.  The sound is excellent, but power tube failure (which can cascade to transformer failure) is something you want to avoid.  The pre-amp bias wiggle and other tremolo types are simply better for reliability.    The power tube bias wiggle arrange works well for players who are not driving their amps too hard.

From http://carlscustomamps.com/types-of-tremolos-in-tube-amps

[Lauri]

Have you tried if it thumps less with a new set of matched power tubes? It's going to thump a lot if the power tubes aren't matched well. You could try to bypass the power tube cathode resistor with a 1000µF 50V cap and see if it helps or makes things worse. Also measure how much the power supply voltage is sagging with the tremolo. If the node feeding the preamp tubes is swinging like crazy with the tremolo turned up then the thumping might be coming from the preamp.

[Paul Marossy]

Have you tried if it thumps less with a new set of matched power tubes? It's going to thump a lot if the power tubes aren't matched well. You could try to bypass the power tube cathode resistor with a 1000µF 50V cap and see if it helps or makes things worse. Also measure how much the power supply voltage is sagging with the tremolo. If the node feeding the preamp tubes is swinging like crazy with the tremolo turned up then the thumping might be coming from the preamp.

Hmm... good points. I'll look into that. I wonder if it makes a difference whether a circuit uses a rectifier tube vs one that does not.

[Rob Strand]

Paul,

everything looks symmetric until you look back towards the cathodyne (single triode) phase inverter ... the driving point impedance at the plate is vastly different than at the cathode ... the result, un-even LFO drive to the power tubes means it turns into perceivable signal

There's definitely an illusion of symmetry there.  It doesn't take much asymmetry for the thump to bleed through.  On the positive side, the 5n caps help disconnect the LFO for the phase-splitter and that will reduce the asymmetric loading.  Nonetheless it still leaves you thinking if it's possible to lower the thump.  There will be a point where tube mismatch dominates.

[Paul Marossy]

Paul,

everything looks symmetric until you look back towards the cathodyne (single triode) phase inverter ... the driving point impedance at the plate is vastly different than at the cathode ... the result, un-even LFO drive to the power tubes means it turns into perceivable signal

There's definitely an illusion of symmetry there.  It doesn't take much asymmetry for the thump to bleed through.  On the positive side, the 5n caps help disconnect the LFO for the phase-splitter and that will reduce the asymmetric loading.  Nonetheless it still leaves you thinking if it's possible to lower the thump.  There will be a point where tube mismatch dominates.

Yeah that's thing... it seems like there ought to be a way to improve that somewhat. As I would normally use the amp with tremolo (which is not often) the thump isn't much of a problem, but if I were to use the volume pedal to kill anything going into the amp the thump is still somewhat audible thru the speaker even at lower settings. It gets very prominent past a certain point as you turn the Strength pot more clockwise. It kinda bugs me as I star grounded the amp and made it pretty remarkably quiet compared to when I first got it, but there's this one little problem left. I suppose it's just a shortcoming of the way this tremolo scheme is implemented? I had a hunch somewhere in the back of my head that the PI impedances might have something to do with it, and got confirmation on that here. I kinda had same thought that the .005uF caps ought to in theory help with that somewhat. I also changed sometime in the past the resistors on the cathodes of the EL-84s, I gave each one their own instead of a shared one like the schematic shows, and increased them just a little bit (160 vs 220). Don't know if that could be a factor or not... but probably so as that biased them a little colder. I've been modeling the whole circuit in LTSpice and there does seems to be a little asymmetry on the outputs of the PI, which probably adds to the overall chances for thumping. I guess I ought to pull the chassis and give it a good look over with the oscilloscope & DMM to see what's actually happening. It will at least be a learning opportunity.

This is where I read about placing a diode across the intensity pot: https://www.tdpri.com/threads/filtering-out-tremolo-thump-question.1026286/
I imagine that this diode thing working or not is probably highly dependent on the circuit topology. People also mentioned power tube bias - colder biasing makes it worse. It appears none of the circuits they are talking about are really anything like mine.  I guess I could try moving tubes around and see what/if affect it has on it. Could very well be that the power tubes are a little mismatched. Not looking for perfection here, just trying to see how much I can do to make it more acceptable. 

[PRR]

The 100k in the LFO is surely in series with the pot, not parallel.

We normally see a cathode resistor in this type LFO. Put a 5K trimmer in. One way won't wobble. Other way same as now. Is the a middle position where it wobbles with less enthusiasm? MmMmMm instead of waKwaKwaK?

For more time-sink, try with a 50uFd capacitor across the cathode resistor. This may or may not allow more favorable AC to DC conditions.

[Eb7+9]

Paul,

Try sticking a 68k resistor in series with the cap coming off the PI cathode ... tell us what happens to the thumping

[amptramp]

I think adding the tremolo in an earlier stage would be a better idea.  I have a transistor amp that uses a phase-shift tremolo oscillator driving a bipolar transistor right at the input stage, the equivalent of the first grid here.  Tremolo doesn't have to be a smooth sine wave modulation. This design:



shows a phase-shift oscillator in the lower left corner driving a transistor at zero bias and it works.  There is no thumping because it doesn't upset any bias conditions.  A FET would also do the job but the bias conditions for the FET would have to be select-on-test or adjustable.

[Paul Marossy]

The 100k in the LFO is surely in series with the pot, not parallel.

We normally see a cathode resistor in this type LFO. Put a 5K trimmer in. One way won't wobble. Other way same as now. Is the a middle position where it wobbles with less enthusiasm? MmMmMm instead of waKwaKwaK?

For more time-sink, try with a 50uFd capacitor across the cathode resistor. This may or may not allow more favorable AC to DC conditions.

That has been a source of confusion for me. The 100K resistor is connected to middle lug and one of the outer lugs of the 1M pot, which is connected to ground. The other lug is connected the .01uF cap at the LFO. It doesn't make much sense to me but it works OK. Maybe they were trying to change the taper of the pot?

Paul,

Try sticking a 68k resistor in series with the cap coming off the PI cathode ... tell us what happens to the thumping

I could maybe try that next. I re-worked this amp back in 2005 and I discovered yesterday that when I changed the EL-84 cathode resistors that I inadvertently did it differently than it was done originally - they had run wire from both cathodes to an 80uF section in the three section can capacitor, and then ran the 160 ohm resistor from there to to ground. So what I did yesterday is added a pair of 100uF bypass caps to my two 220 ohm resistors that replaced the single 160 ohm. That seems to help it a bit. I only get the thumping when the tone control is turned extreme CCW, when the Strength control is at max and the back is on the amp (more bass).  When it becomes kinda objectionable the tremolo is too extreme to be usable anyway. So I think I can live with it as is for the moment. I'd like to slow down the LFO a little bit but I am not sure how to do that with this LFO.

[PRR]

> I think adding the tremolo in an earlier stage would be a better idea.  I have a transistor amp that uses a phase-shift tremolo oscillator driving a bipolar transistor

Beating the power stage worked good for Fender. And the push-pull topology is very effective against oscillator bleed-through.

> When it becomes kinda objectionable the tremolo is too extreme to be usable anyway. So I think I can live with it as is for the moment. I'd like to slow down the LFO a little bit but I am not sure how to do that with this LFO.

Then the thump may be as good as it needs to be. (I would be interested in Eb7+9's 68k move though.)

Double the capacitors to halve the speed.

[Rob Strand]

I guess you have to watch out for the 'patch' fixes is they might fix what you want but they could break something else.   You would like to fix the tremolo thump without affecting the sound of the amp. You would really want to make sure the sound of the amp in unchanged (or better).    If a 'patch' fix works then at least you know you are partly on the right track, you just have to workout how to roll-out the fix.

As far as the design goes, modulating the power tubes is always going to be worst off than modulating small signal tubes.    What you will see in similar designs which modulate small signal tubes is a high-order high-pass filter *after* the tremolo modulation.   That's a way to remove some of the low-frequency thump.  The problem with modulating power tubes is there's no easy way to insert a high-pass filter between tubes and the transformer/speaker - the transformer and speaker is it.

[Paul Marossy]

One last dumb question: would a speaker impedance mismatch matter here? I discovered after measuring the OT that I apparently have an 8 ohm speaker connected to what appears to be a 4 ohm OT - primary measured 2.5K-5H/2.5K-5H and the secondary measured 3.1 ohms/8.3mH. Just wondering if an impedance mismatch could contribute to the situation. I don't intend to change the speaker as the 8 ohm OEM alnico Matchless speaker that came with the amp sounds really good and it has never been a problem, but it would be good info to know.

[Rob Strand]

One last dumb question: would a speaker impedance mismatch matter here? I discovered after measuring the OT that I apparently have an 8 ohm speaker connected to what appears to be a 4 ohm OT - primary measured 2.5K-5H/2.5K-5H and the secondary measured 3.1 ohms/8.3mH. Just wondering if an impedance mismatch could contribute to the situation. I don't intend to change the speaker as the 8 ohm OEM alnico Matchless speaker that came with the amp sounds really good and it has never been a problem, but it would be good info to know.

To the ear, what I've found is the opposite connecting an 8 ohm speaker to an amp set to 16 ohm produces a little more bass.

Your best hope is to get everything balanced as much as possible.   

You could even try a DC bias tweak by feeding DC a little DC into the grids vias large valued resistors.   Some amps have these schemes so the DC bias currents of the two output tubes can be balanced.    As a starting point you can measure the voltages across you cathode resistors - that only goes so far as the cathode resistors themselves are not exactly equal.  You can do better if you measure the resistor values and calculate the current.    Keep the amount of adjustment small and don't allow the tube currents to be adjusted with large bias differences in the bias currents.

It's important not to using the bias adjustment as a fix-all.  You want to balanced the LFO level feeding the grids first *then after that * use the bias balancing as a fine tweak.

[RGP]

Thumping in tremolo circuits is actually a common problem. The thump is most likely bleeding thru from the LFO. Roly Roper designed a very good r/c filter that tat is easy to build and wouldn't take up much space. By the way, I have an old Gregory Mark V that I love and ran into the same problem. I put that fiter in and the thumping disappeared.   

[Paul Marossy]

Thumping in tremolo circuits is actually a common problem. The thump is most likely bleeding thru from the LFO. Roly Roper designed a very good r/c filter that tat is easy to build and wouldn't take up much space. By the way, I have an old Gregory Mark V that I love and ran into the same problem. I put that fiter in and the thumping disappeared.

Interesting. Do you have a link to this circuit?