Now for the Buzz/hum chase - 2 channel tube amp

[pjd3]

Hello,

I hoped that this thread may serve as somewhat of an ongoing discussion on ideas for chasing the all feared "tube amp buzz".

So, 2 nights ago I tubed up one side of my latest project stereo 2-channel tube amp build with known good tubes. It really is two of Slucky's 20- watt Plexi's in one chassis/cabinet. I realize that by the very thought of undergoing something like this that a host of "invisible" grounding issues could (or would) occur, and my work would be cut out for me. But hey, it's a great time for amp building with websites like this so, I decided to go forth and make this a voyage of discovery. and it is already.

Just to mention, aside from the buzzes that are occurring, my guitar sounded really good, solid, rich and full, even better perhaps than my first mono 20 watt plexi that I have gigged with for a couple of years and receive many compliments on tone.

Per recommendation, I have the preamp grounds connected with alligator clip wires to the first filter cap negative. With no preamp tubes, no hum or buzz, completely quiet. With a 5751 in the phase inverter, still, no hum or buzz - completely quiet, even with pre-phase inverter master volume all the way up.

Enter a 12ax7 in V2, there is a bad hum, or better, set of hums. As a first experiment, I attached a clip lead to the preamp ground buss and shorted that to the chassis. The worst part of the hum, which I believe was a loud 60 cycle hum went completely away. But, I was left still with hum that was an octave higher than the one eradicated with the short to the chassis. I could only assume that was a 120 hz hum still there, not anywhere near as obtrusive but still very apparent, show stopping apparent. The volumes were set for a low bedroom volume. No amount of skating the alligator clip around the chassis would change either the timbre or the volume of the 120 hz hum, steady. FYI, I currently have the earth safety grounds of each amp in the corners of the chassis right near the PT and other power section wiring. I managed to keep all power components and wiring at the far corner and very end of the chassis, while inputs are in the dead center of the front panel, at least 13-14 inches apart from one another. I suspected that was an adequate distancing of inputs-to-power section.

So far, what this is telling me is that the power section is not the source of the hum in itself, and the hum does not rear its head until V2 is placed in the circuit. I should also mention that each amp in the chassis has its own PT and each PT has its own private AC cord. So, the only power and earth ground is coming soley from the left channel amp being active at this point, one earth ground in that corner. Also, all the inputs and speaker output jacks are floating per suggestion from another thread with RG Green being the first to voice that recommendation. This could very well be some iffy wiring on my part concerning the connections associated with V2, and it would probably behoove me to get in there and go over it, thoroughly. Now that I think of it, I could probably experiment with "how" I am bringing the preamp ground to first filter cap, making sure I'm not presenting ground loops with the two clip leads involved. I have the "potentiometer" grounds clip leaded to the board ground buss, then the board ground buss clip headed over to the first filter cap. That "seems" Ok in theory but, perhaps I could improve that. I moved around the clip lead wires to see if there were any changes in either the quality or magnitude of the hum but, it didn't change anything.

Well, if you made it this far, thank you very much for your time in my behalf. I understand grounding in a new layout can be a very precarious thing, and its all my fault! But, I'm determined to discover the source of the hum, one way or the other - too much effort and cash invested already!

Any insights or recommendation are welcome and will be implemented in experiment.

Thanks, and best,
Phil D. (pjd3)

[R.G.]

Good on you for thinking your way through. You're completely right - there are many sources of hum in a tube guitar amp. I tried to write down the ones I could think of in Guitar Amp Wiring Notes, link below. If you haven't yet, you might want to give it a read.
https://www.dropbox.com/scl/fi/vpijj3al ... k7985&dl=0

Also good on you for distinguishing between line frequency (60 hz for you) and 2x line (120Hz). If it's 120Hz, it has to be coming from the rectification or ripple voltage. If it's line frequency (60Hz) then it is not rectifier ripple, it's from AC line coupling, either conductive, capacitive, or magnetic. EXCEPT for one rare, degenerative case; a defective diode that does not conduct in the power supply can make power supply ripple be line frequency, not 2x line frequency. This is hard to find unless you know to look. I've seen this exactly once, as a green engineerling, when I was working for TIC (Three Initial Corporation), but it stuck with me.

Enter a 12ax7 in V2, there is a bad hum, or better, set of hums. As a first experiment, I attached a clip lead to the preamp ground buss and shorted that to the chassis. The worst part of the hum, which I believe was a loud 60 cycle hum went completely away.
Best practice for chassis grounding is to connect one and only one wire between chassis and signal ground. There are find points of exactly where to do this, but only one wire. Otherwise, the chassis can produce hum by currents conducted across it. There are variations of chassis grounding that can be less bad, or even with unnoticeable hum, but it's hunt-and-peck, as you're doing. Single wire chassis ground is the only way to know ahead of time that the chassis won't be involved in hum.

But, I was left still with hum that was an octave higher than the one eradicated with the short to the chassis. I could only assume that was a 120 hz hum still there, not anywhere near as obtrusive but still very apparent, show stopping apparent. The volumes were set for a low bedroom volume. No amount of skating the alligator clip around the chassis would change either the timbre or the volume of the 120 hz hum, steady.
That says that the big 120Hz hum is coming from power supply ripple, not chassis coupling, so no, changing chassis ground point will make no difference. This is power supply ripple to the stage you just added. Maybe a not-connected bypass cap, maybe even a grid wire near a wire carrying lots of rectifier current, something like that.

FYI, I currently have the earth safety grounds of each amp in the corners of the chassis right near the PT and other power section wiring. I managed to keep all power components and wiring at the far corner and very end of the chassis, while inputs are in the dead center of the front panel, at least 13-14 inches apart from one another. I suspected that was an adequate distancing of inputs-to-power section. [...] each amp in the chassis has its own PT and each PT has its own private AC cord. So, the only power and earth ground is coming soley from the left channel amp being active at this point, one earth ground in that corner.

Hmmm. With two chassis, two line cords, you're setting yourself up for hum issues just like two separate amps. Are you finding this hum with only one amp powered and the other line cord disconnected?

Also, all the inputs and speaker output jacks are floating per suggestion from another thread with RG Keen being the first to voice that recommendation.
Good choice. Is there a single wire from signal ground somewhere (and first filter cap negative is a good choice) to the chassis? Mr. Ohmmeter will be helpful here.

Now that I think of it, I could probably experiment with "how" I am bringing the preamp ground to first filter cap, making sure I'm not presenting ground loops with the two clip leads involved. I have the "potentiometer" grounds clip leaded to the board ground buss, then the board ground buss clip headed over to the first filter cap. That "seems" Ok in theory but, perhaps I could improve that. I moved around the clip lead wires to see if there were any changes in either the quality or magnitude of the hum but, it didn't change anything.

I would not expect exactly where the grounding leads connect to first filter cap to affect 120Hz hum. There are cases where it is possible, but these would be unusual. Check for where your bypass caps in the dropping chain get grounded. Ideally, they should be to the first filter cap negative ( the "One True Ground"), but if you have done local star ground groups, all with a private wire to the One True Ground, they can also be local to the stage.

Do you have an oscilloscope? Scoping the ripple voltage on B+ at each stage in the dropper chain can help a lot. VTVM (rarer today than scopes ) or even an audio probe can tell you about the magnitude of the ripple at each stage. Have you checked that the PT center tap goes ONLY to the first filter cap negative? If it goes anywhere else along the way, or if any signal ground touches that wire anywhere except at the One True Ground, you get 120Hz ripple that can't be cured.

[sluckey]

all the inputs and speaker output jacks are floating

The input jacks and speaker jacks must be connected to signal ground by some means. IOW, you must measure zero ohms between the sleeve of those jacks and the HT center tap.

[R.G.]

all the inputs and speaker output jacks are floating

The input jacks and speaker jacks must be connected to signal ground by some means. IOW, you must measure zero ohms between the sleeve of those jacks and the HT center tap.

I think he meant he carried ground from the jack to the actual circuit. Right pdj3?
Yeah, the input jacks have to connect the sleeves to signal ground.

[sluckey]

I'm terrible at "thinking what was meant" versus "what was said", so I don't even try anymore.

Since this amp has a typical Marshall NFB loop, it's necessary to ground the sleeve of the speaker jack also. Overlooking this will result in a nasty buzz.

[R.G.]

Yep, dead right.

The issue with the need to ground one side of the speaker output in amps with feedback is covered in Guitar Amp Wiring Notes. Isolating the speaker output jack and running a ground cable and why it's needed is in there.

To be fair, not isolating the speaker output jack is only a minor exposure, an issue that only crops up in truly hard cases, but isolating it makes certain that the chassis is not introducing hum from magnetic field loops and/or oscillation from ground-carried feedback. Making sure the chassis is only connected to signal ground with one wire completely eliminates these issues. But we all generally get away with it.

[pjd3]

Hello,

Thanks again for all your generous insights and suggestions.

I should have used the work "isolated" instead of "floating". RG had recommended the option to isolate input and output ground from the chassis to avoid any possible (or certain) conditions of chassis currents that would be difficult to resolve. Isolating the inp/outputs would allow the option to bring their grounds to the best position for reducing noise and interference.

Probably tonight, I will rewire some of the power wiring as I did find I did not bring a dedicated connection from PT centertap to the first filter cap. What I did was bring the PT centertap to a solder lug halfway to the filter cap where things like the power tube cathodes and possibly bias grounds met, and maybe even something else. So, whether that could be the cause of the remnant 120Hz buzz, I will rewire that tonight, see what happens or doesn't happen.

To clarify a bit more on earth ground(s), I have each amp's earth ground, at least for now, connected at its associated corner of the chassis. Right now, I am only trying to get the left side amp working quietly, with only its AC cord plugged in. So for all practical purposes, it should be like the right side amp isnt even there (as far as I can tell, its cord is not plugged in and there is nothing in common with one amp to the other except for the chassis, and, the channel selector switch - A DPDT mini-toggle that selectively grounds the inputs/V1 grids. Center position doesn't ground anything). I believe the internals of those switches are isolated from the chassis)). Of course, this means that there are not yet any guarantees for how interaction between the 2 amps will play out, even when I get both of them quietly grounded by themselves. I have nightmares of having to futily bandsaw the chassis right down the middle and connect them with some non-ferrous/non-conductive plate material. The funny thing is, that is a doable thing, because I designed the layout in a completely mirror image/symmetrical manner. When this is done, you could bandsaw the entire amp right down the middle and have two fully featured working amps. I just thought that would be cool. Ha. We will one day see just how cool that actually turned out to be!

I will also need to recheck where I did ground the speaker jack sleeves. It was to a main power ground. I did follow "version 1" on the diagrams that were given to me in the other post on this amp. It sounds like from what you said that it is worth paying attention to, and that will be one of the things I experiment with should other attempts at alleviating the 120Hz buzz fail.

So, for now I will just concentrate on rewiring the PT centertap to go uninterrupted to the first filter cap, and make sure all the other grounds have their own connection where they best go. There are surely some things to try and of course, I"m hopeful it will land a quiet amp.

Thanks again, all.

I'll be back to report on the next wave of modification.
Best,
Phil D. (pjd3)

[R.G.]

Probably tonight, I will rewire some of the power wiring as I did find I did not bring a dedicated connection from PT centertap to the first filter cap. What I did was bring the PT centertap to a solder lug halfway to the filter cap where things like the power tube cathodes and possibly bias grounds met, and maybe even something else. So, whether that could be the cause of the remnant 120Hz buzz, I will rewire that tonight, see what happens or doesn't happen.

That particular wiring path does open you up to 120Hz buzz. To the extent it's possible, solder/heatshrink an extension on the PT CT and move it to the first filter cap negative. Bias ground going there also lets the PT CT and the power tube return currents wiggle around the bias 0V reference.

To clarify a bit more on earth ground(s), I have each amp's earth ground, at least for now, connected at its associated corner of the chassis. Right now, I am only trying to get the left side amp working quietly, with only its AC cord plugged in. So for all practical purposes, it should be like the right side amp isnt even there (as far as I can tell, its cord is not plugged in and there is nothing in common with one amp to the other except for the chassis, and, the channel selector switch - A DPDT mini-toggle that selectively grounds the inputs/V1 grids. Center position doesn't ground anything). I believe the internals of those switches are isolated from the chassis)). Of course, this means that there are not yet any guarantees for how interaction between the 2 amps will play out, even when I get both of them quietly grounded by themselves. I have nightmares of having to futily bandsaw the chassis right down the middle and connect them with some non-ferrous/non-conductive plate material. The funny thing is, that is a doable thing, because I designed the layout in a completely mirror image/symmetrical manner. When this is done, you could bandsaw the entire amp right down the middle and have two fully featured working amps. I just thought that would be cool. Ha. We will one day see just how cool that actually turned out to be!

The nice thing about having only one wire from signal ground to chassis is that it guarantees that two circuits cannot interact through the chassis. If the amp circuits are common only at:
- AC Line
- AC Neutral
- Safety ground
- and connected signal grounds,
you're minimizing the chance for interaction.
As a practical matter, if the two amps have separate line power switches, I would not hesitate to use only one line cord; I'd prefer it. At best, both amp line cords will be plugged into two halves of a duplex outlet. At worst, having two line cords makes it possible to connect them to two different outlets, possibly with different Line Neutral and safety ground offsets.

You're kind of stuck with connecting the signal grounds if the setup involves one guitar chain. The input sleeve/ground is going to get connected through cables if nothing else. That's where to start checking for interference when you get both going. As Guitar Wiring Notes mentions, input ground is a little ambiguous. Lowest input noise is when the first tube circuit gets the sleeve ground local to the tube. Lowest conducted RF happens when the input jack is either connected to the chassis, or shunted by a ceramic cap right at the jack. Two amps open up the possibility that different AC leakages will wiggle the input tubes' signal grounds around. The real answer is to make input stages be differential amplifiers, but that ship sailed a century ago.

I will also need to recheck where I did ground the speaker jack sleeves. It was to a main power ground. I did follow "version 1" on the diagrams that were given to me in the other post on this amp. It sounds like from what you said that it is worth paying attention to, and that will be one of the things I experiment with should other attempts at alleviating the 120Hz buzz fail.

So, for now I will just concentrate on rewiring the PT centertap to go uninterrupted to the first filter cap, and make sure all the other grounds have their own connection where they best go. There are surely some things to try and of course, I"m hopeful it will land a quiet amp.

Good. You have some things to try. The whole trick in (low voltage, low frequency...) grounding is summed up by:
> ground is a point, not a wire or bus
> all wires are really resistors that make voltage by V I*R
> think in terms of where the currents flow for which wire/resistors cause noise voltages.

[pjd3]

I kiss the ground you guys walk on.... well, maybe not that far, BUT!

I created a dedicated wire from PT centertap to first filter cap ground and behold. At my personal gig volume, you cant hear a mouse chewing on old cheese. I am happy as ten pigs in shit. Thank you. Plus, a bonus, the guitar tone is remarkable, its notably better than the first Slucky 20W plexi I've been gigging with for 2 years, and that amp sounds great. I really wasn't that confident that this move would do it, it just seemed too easy. But wahlah, thar it is matey.

I still need to read your last post, I was too excited to let you know that I had a pretty great victory here. But, as usual, I'm sure it is packed with good information to have and hold.

Now, I'm eager to get the "other amp" sounding working on par with the one I'm doing now, get that quiet, and then see what happens with both amps on. Thats where the rubber meets the road, (oh oh, I'm hearing an intermittent scratching sound coming from my idleing amp, probably a cord or tube). What I will do however, is while the volume is moderatly high, experiment with the grounding locations of the potentiometer grounds and the preamp to power ground to see exactly what paths lead to the quietest amp. I guess thats the thing to do since at this moment, those connections are being made with cheap alligator-to-alligator clips (probably whats causing the farting right now).

I will read your last post and then get to work over the weekend and see about birthing the sound of the "other amp". I'd consider it a moderate success if I can get both of them even working good individually.. Wish me luck (I may need a good dose of it)..

Have a great safe happy Fourth of July everyone, and come back with all your fingers intact.
Best,
Happy amp camper Phil, Thank you all,

[pjd3]

Back to the saftey earth ground topic.

I also imagined that one common earth ground point, starting with the middle rear of the chassis would be a right place to start, and very well be a finish.

I let myself get influence by one of the top engineers in my company (Zoll Medical). This guy is very autistic but, they say the smartest guy in the company, a genious. Ha, he's one of the few they let use the $230,000 oscilloscope as he gets the real esoteric baffling problems and projects to resolve that no one else can. (They don't let me use that scope). I asked him If I could outline amp to see what he thought I should do about placing earth grounds. After listening very carefully and asking lots of questions (isolated inp/outputs were one of them) he said he thought I should create a common ground point in the middle of the chassis, which matched my initial instinct. But then about 10 minutes later he came looking for me and said No, instead do your saftey grounds separately in the far corners of the amp in the vicinity of the associated power transformer/AC section. I decided to go with that first, with the caveat that I would be ready at an instant to go common ground point. One other engineer thought I should bring one earth saftey across the amp to meet the other safety ground in one of the corners. Interesting. I personally don't have much idea of what will be the best but, surely really to try em all if it calls for that.

OK, its really nice to start getting things and make progress but, still a long way to go!

Thank you again, such great guidance over here!
Best,
Phil Donovan

[pjd3]

Basteerd.

I call my son in to hear how good my Left channel amp sounds, and how quiet it is. So he comes in, I pick up the guitar, and the second I touch the guitar, there is a "Pop" and then no output (unless you turn the volumes way up and then you barely hear it), Big damn.

I'm hoping this is just a bad cable or connector, they are getting finicky and old. All the voltages and bias voltages were present and all the PT and OT coils measure normal expected ohm. It was getting late, didn't want to be playing around high voltages. Ahhhhh, I'll be back.

Best,
Phil D (pjd3)

[R.G.]

Yeee-haaah!! Congratulations!

I've run into the rectifier-return-wire issue in my own amps before. It is amazing how much that one-wire move changes things.

On the safety-ground topic:
There are formal rules on exactly how to connect safety ground to chassis. The easy way to satisfy this is to use an IEC power-entry module, perhaps with an internal fuse, and take a wire to chassis right from the inlet module. The location of the safety ground wire to chassis is dominated by safety issues, not signal quality issues. It needs to be "near" the line cord entry, and not used for any other purpose; transformer mounting bolts for instance, are not allowed for this. Much more on this in Guitar Amp Wiring Notes.

If it were me, for your case of single chassis, two independent amplifiers, I would have one AC line cord entering the chassis, and one AC mains fuse. I would split the AC into two paths, one to each power switch. This eliminates complexity on where the AC safety ground runs.

The two signal grounds will have to be tied together somewhere. There is a discussion to be had on whether the two signal grounds should be tied and one wire from one of them to the chassis, or whether both first filter cap negatives should each have a wire that connects to the single chassis tie point. In either case, it should only be one connection from signal ground to chassis. Input jacks will need ceramic caps from sleeve to chassis at the jack to shunt RF off to the chassis.

There should never be any significant current from one amp signal ground to the other, so there will be no voltage drop between them. With chassis-isolated input jacks, each input circuit sees only the incoming sleeve voltage from the cable to amplify. If you plug different signal sources into the input jack, that ground connection wire will carry a current, and you could still get hum, but it will be obvious that it's the input sources doing it.

The purpose of tying the chassis to the AC mains safety ground is - well, safety. With only one signal ground connection to the chassis, you can't have any chassis voltage drop issues, so exactly where the signal ground ties to the chassis is not related to where the AC mains safety ground chassis point. They should not be the same point.

Electrical safety rules say that every bit of metal that a user can touch must be tied to AC safety ground so that if a wire breaks inside or a transformer primary shorts to chassis, that kind of thing, that the touchable metal will be held at substantially earth-rod voltage until the mains breaker pops or the AC fuse opens. There are formal requirements for any external user-accessible metal to conduct - um, IIRC - 25A of current with less than a couple of volts of drop.

In normal operation, there is zero mains current in the AC safety ground wire and in the chassis, so nothing happens there. The chassis also ties switch and pot bushings, etc. to safety ground, so the user is protected there, too.

In normal operation, the chassis also acts as a Faraday cage, a metal enclosure that keeps RF and external magnetic fields from going inside the chassis. This is a separate action from the safety-ground nature of the chassis.

But I'm just running on about grounding. You're well on your way to a great setup there. Keep up the status reports!