Regarding grounding schemes of stereo input

[pjd3]

Hello, thank you for coming by.

You may have seen a couple of threads on some crazy person building two nearly independent 20W plexi's into one chassis and cab. Yeah, thats me.

I'd be interested in hearing what your notions would be on consderations for input grounding. I can imagine there are a few ways to go with this depending on problems that arise and well, trying different methods while examining for noise and/or buzz on my scope.

I've already comitted the two inputs to the Front center of the chassis (1.5" apart) with dedicated powers transformer in each far corner of the chassis, making what seems some adequate distance from PT to inputs. (Twin Reverb chassis 24.3" L).

So far, I can see either chancing the use of standard switched input jacks whos ground is also grounded to the chassis or, the isolated switched input jacks that would at least allow a choice to where the jacks will eventually be grounded, whether that be in close proximity or star grounded to some where else.

While I realize theres likely no surefire way of determine what the grounding scheme will finally be for the best noise perfromance of the amp, I'd very much be interested to hear what you think the potential issues could or would be, and what might try first, second, etc.

An EE where I work is a gigging bass player and also the principal engineer in the designing of what is to be the flagship defibrullator in our top company. Even he warned me that this will likely involve some experimentation to truly know what grounding scheme will provide the best performance.

Thanks for taking the time for this!
Best,
PJD3

[R.G.]

So far, I can see either chancing the use of standard switched input jacks whos ground is also grounded to the chassis or, the isolated switched input jacks that would at least allow a choice to where the jacks will eventually be grounded, whether that be in close proximity or star grounded to some where else.

That's a good place to start.

I personally prefer to use isolated jacks and run the jack "ground" wire to the local ground of the input preamp circuit it connects to. The jack is connected to the chassis very close to the jack by an RF-shunting cap, low value ceramic type.

If you use the chassis for the hardwired input ground, you may get good results, or not. This depends on the degree to which nothing else is pulling the chassis "ground" around compared to the ground in the preamp circuit. Grounding the input jacks to the chassis can give stellar results sometimes; it can also give mysterious noise and hum that no amount of tinkering with the inputs can fix because something somewhere else is causing a difference in the voltage at the preamp circuit "ground" and the chassis "ground".

All wires and metal objects are actually low-value resistors. Maybe very-very-low value, maybe noticeable. The only way to get truly zero volts across a resistor (and wire) is to make the current through it zero. If any current flows in a resistor (and necessarily, any metal/wire) then there will be a voltage across it. Whether you can then hear the voltage across it depends on the actual resistance, the size (and AC-DC-ness) of the current, and how much you amplify the voltage.

Isolating the jacks from the chassis gives you the freedom to run the jack ground to the actual input of the preamp circuit's first amplifying element. Tn only the input signal current flows in the input jack ground wire, so it can't get hum and noise from other places. And there is an asterisk, as always. This works for audio; RF follows the path of least RF impedance. that's why the ceramic cap is needed - it gives any incoming RF a low impedance path to the chassis, which acts as an RF shield, or should do so.

Using chassis-grounded jacks flips the problem on its head. Now the source of zero input volts is the chassis. But if there is current flowing through the chassis, perhaps from AC mains grounding issues, speaker return currents, power tube cathode returns, etc. then the preamp circuit ground may be wiggled around by its own ground connection, so the noise is injected from the preamp circuit ground. On some occasions, you get lucky. Chassis-grounded input jacks can get you cancellation of other noise sources. Hum on an input jack might happen to be nearly equal and opposite to tube heater hum or some such source. It's a fragile cancellation, but it's possible. Chassis-grounded jacks can also be quiet if you're very careful about how you connect other stuff, like the AC mains safety ground and the speaker return. Effectively, you make the whole chassis into the one single wire to the input jacks.

There is a test of practicality here. There will always be some hum and noise in a real system. Whether it is audible or not depends on how much you amplify it or cancel it. In general, the more gain you have, the more any noise gets amplified. So the problems get worse with high gain, rip-snorting, head-banging amplifiers.

[pjd3]

Thank you for this RG, everything you mention hits it right on the head to my concerns as much as the limits of my understanding can take me.

What's a little different about this amp is that it will truly be two full amps in one chassis including separate power transformers for each amp (amp left, amp right).

Fortunately, this amp will serve as a relatively clean pedal platform with the advantage of having the nice clean Plexi clean by itself. I've managed to get acceptable and even commendable overdrive and distortion from prudent use of pedals so, my mono amp is always clean and low volume, miced through the PA. I"m also installing Headfirst amplifiers Zero loss series FX loops on each amp for stereo sends and return. I like at least the reverb being as true to its sound as possible so, it would come back into the amp right at the phase inverter input.

It will be the same with this amp except there will be two amps to accept the stereo outputs of stereo reverbs, delays and chorus. I love the sound of a cranked Marshall but, just no venues left to make that feasible. But I'm satisfied and make the best of what I have and use.

Its interesting that occasionally there can be unforeseen benefits from a hardwired input jack. Perhaps I will just start with standard switched jacks but be ready to install isolated jacks should there be no input ground configuration that eliminates undesirable noise/buzz/etc... The boards for L and R channels are really complete and just waiting for jacks, standoffs and pots.

Thanks so much for this great synopsis on stereo input jack consideration! Very helpful

Best,

Phil Donovan

[R.G.]

You're quite welcome. I hope it helps.

I did some quick plagiarism of myself, as I've posted grounding polemics several times over the years. I looked up a few of those and picked out some of the more rational comments. It's funny - as far as I know, the term "sewer ground" may have originated in one of my longer rants. Who knew??

Two complete amps in one chassis makes for some more possibilities for hum and noise, though. The two power transformers may have different amounts of magnetic and capacitive leakage, and that can lead to hum issues as the currents from each of the two try to resolve across the chassis. They would mostly both be trying to get to the AC mains safety ground bolt position, or running in circles around the PTs. I would probably start with setting one common signal-to-chassis ground point (which should not be the chassis safety ground point) and run one wire from each of the two first filter cap negative terminals to that common point. That gives you a shot at resolving all the signal grounds from each side separately into the circuits served by that signal ground, then referencing them to one chassis point.

Star grounding isn't always going to make for zero hum and noise, as there will still be ground currents that can intrude. But it's going to be the most predictable grounding scheme, excepting for going full-differential on all signals between circuit chunks; even then, you can only get down to the common mode rejection limits of the differential scheme; nobody does that. But I've wondered how different guitar amps would be if the early ones could have afforded to use a whole 12AX7 duo-triode as a differential amplifier instead of only one half running in common cathode.

[pjd3]

Ha, interesting, and if doing DIY and not having the weight of saving pennies on the dollar, then its not terribly much more to have one more triode for a differential input. If I knew a little more about confguring tubes I may be willing to do that. Well, the first stage of this amp seems to be like most, just acting as a voltage amplifier for getting as clean a signal to the next stage. I'm sure there are scores of schematic to find that do just that but differential input. Maybe in a few of the more basic hifi tube amps - was just looking at Aikido differential circuits.

RG, if you are stll there, I'd like to make sure I'm getting your idea from your, "I would probably start with". - Are you saying to place a single lug to the chassis say right near the input jacks, connect both L and R preamp signal busses to that, then bring the ground side of both first filter caps to that common ground lug? Please feel free to straighten me out if I'm not gettng that right. And if that (or other) is the case, would you use an isolated jack as well?

Thanks for hanging in with me on this, it would be nice to have a well thought out starting point and take if from there should it need to.

It will be a great thing to play through a quiet stereo Plexi!

Appreciate your time in my behalf!

Best,
Phil D.

[R.G.]

I realized that drawing a picture would save several thousand words.
In the picture, I tried to sketch out what I would be trying to wire up if I was to be building something similar.

I would isolate both the inputs and the speaker outputs from the outer chassis. I would make exactly two connections to the outer chassis; one for the AC power safety ground, and one signal reference lug. The signal reference lug has one wire to each of the two power supply star points, and carries no significant current, only forcing the two signal grounds to be at the same potential at their respective star points, and keeping any stray RF shunted to the chassis.

RF does not play by the same rules as audio. So a small ceramic cap is needed from the isolated input jacks to the chassis, as near to the jack as possible. The RF is sucked off to the outside of the chassis by field effects, and so is siphoned away from the audio on the input.

Input jacks isolated lets you run the input ground on a wire to the input ground at the preamp circuit itself. This eliminates any possible contamination of that ground wire with currents from elsewhere. Running a ground from each input jack to its own preamp ensures that external noise voltages on one input (or the ground of one preamp circuit) do not affect the other preamp.

Isolating the speaker output jacks from chassis ensures that the speaker return currents flow back into the secondaries of the OTs. If your power amplifier circuit uses feedback from the speaker output, you will have to run both a feedback wire and a speaker-return-ground wire back to the power amp circuit. The speaker return ground carries only the feedback signal current, not anything it picked up from running across the chassis. The speaker feedback return is another "input" as far as the power amp circuit is concerned.

The preamp circuits should each have one ground wire back to the first capacitor star point for its power supply. The preamp should not grab ground from somewhere inside its power amp circuit. The sketch shows one ground wire from the power amplifier circuit to its respective first capacitor star point. As I'm writing this up, I realize that I should have shown two ground wires to the power amp star point: one for power amplifier up through the phase inverter, and a second ground wire from the output tube cathodes to the star point. This keeps output power currents off the grounds for the PI.

This all sounds complicated, but it follows a simple rule: figure out what currents have to flow, and get only "like" currents to be combined on the ground wire leading back to the first filter capacitor ground; where you want to force a reference voltage, run a ground wire that carries nearly zero current.

If this is confusing, ask questions.
[Dagnabbit, I can't get the image to insert in the post.]
https://www.dropbox.com/scl/fi/kf9to3g0 ... hkk8a&dl=0

[martin manning]

...

[R.G.]

Thanks Martin!

[pjd3]

Wow, I'm impressed. And humbled.

Thank you for giving some sight to the partially blind and giving me a great starting point, and possibly ending point.

Although my 3 amp builds have all come out great and very gig worthy, I followed classic layout to be safe thus, never ran into any hum, hiss, noise or buzz issues.

This is very unchartered territory for me and its really nice to have some experience-based guidance.

I'm placing my order today for the isolated jacks, pots, hardware to get the front panel wiring moved along. There will also be a Type-A DPDT toggle switch to allow "Channel A/Channel B/Both A and B" to make quick A/B comparisons of channels for balancing purposes. I drew out a switching diagram that would ground the grids of the input gain stages (pre-grid stop resistor) to silence each channel. Will need to make sure that doesn't corrupt any well working grounding scheme. I don't think it will.

Thanks again, I am grateful! It will be a while to finish this amp as I'm teaching myself to also build cabinets. Its coming along.

Best,
Phil D.

[Roe]

I found this paper helpful https://www.valvewizard.co.uk/Grounding.pdf

[nworbetan]

I found this paper helpful https://www.valvewizard.co.uk/Grounding.pdf

Merlin recommends using a ground reference at a point in the circuit that minimizes the power supply current flowing past/through the ground connection point. Specifically: connect the ground reference at the cathode of the input stage filter cap where there's only a few mA of current flowing through the ground wire creating extremely small noise voltages, not at the power amp or supply.

groundingversionm.png

One of the bigger points Merlin also recommends is that you don't use the amp chassis as a ground conductor; that the entire circuit exists and functions without any connection to the chassis. RG's Version 1 drawing kind of implies that too but I think it's worth mentioning in words that ideally the green lines are wires and the only connections to the chassis are the two points that are drawn as chassis connections.

Sure, lots of people use the chassis as a ground conductor and it's "fine", but who wants to make an amp that's "fine" when you can make one that's "as good as possible" instead?

[R.G.]

I'm placing my order today for the isolated jacks, pots, hardware to get the front panel wiring moved along. There will also be a Type-A DPDT toggle switch to allow "Channel A/Channel B/Both A and B" to make quick A/B comparisons of channels for balancing purposes. I drew out a switching diagram that would ground the grids of the input gain stages (pre-grid stop resistor) to silence each channel. Will need to make sure that doesn't corrupt any well working grounding scheme. I don't think it will.

One subtlety on pots and so on. It was common in the Golden Age to solder tone/volume control grounds to the shells of the pots. That often works OK, especially if the amp has carefully routed DC and feedback return paths that don't cross that part of the chassis. I never trust that. I think of the case of a pot as an extension of the shield wall of the chassis, and run any pot grounds back to the preamp from whence it came.

if you use the RF shunt capacitor on an isolated jack, it may be convenient to get a couple of ring terminals that will fit over the pot bushing to have somewhere to solder the RF bypass caps without having to drill another hole.

[quote=nworbetan post_id=469674 time=1727888673 user_id=17360
Merlin recommends using a ground reference at a point in the circuit that minimizes the power supply current flowing past/through the ground connection point. Specifically: connect the ground reference at the cathode of the input stage filter cap where there's only a few mA of current flowing through the ground wire creating extremely small noise voltages, not at the power amp or supply.
[/quote]
I'd take that a step further. I would arrange the wiring and chassis so no power supply current at all can flow through the ground wires, full stop. That's what the diagram shows, if I did the drawing correctly. The chassis has only two connections, one from the AC mains safety ground, and one from the two power supply star points. With only one connection to the DC power supply, the chassis simply can't carry any DC power supply currents.
That is, I would not connect the preamp local ground to the chassis at all. I would run it on a separate wire back to the power supply star point. I would connect the power supply star point to the chassis, and then use my ohmmeter to see if the chassis and the power star point went open circuit when I disconnected the star point to chassis reference wire. If it's done this way, the chassis gets >zero< current to/from the reference connection point. Since the reference wire has zero current flowing through it, the voltage across it is exactly zero. The ohmmeter test for single point grounding to chassis is a great way to test whether you got all the sneaky grounds connected only to the power supply star point.

One of the bigger points Merlin also recommends is that you don't use the amp chassis as a ground conductor; that the entire circuit exists and functions without any connection to the chassis. RG's Version 1 drawing kind of implies that too but I think it's worth mentioning in words that ideally the green lines are wires and the only connections to the chassis are the two points that are drawn as chassis connections.

I thought I mentioned that, but you're right - if it's not obvious, we should make it so. The green wires are the ground wires. The chassis is the gray box around the whole mess.
A chassis is a shield ground. It has two purposes other than just mechcanically holding everything in place. It (1) shields the user from electrical shocks and (2) shields the electronics from external fields, both RF and line frequency, and any electrostatic stuff.

Sure, lots of people use the chassis as a ground conductor and it's "fine", but who wants to make an amp that's "fine" when you can make one that's "as good as possible" instead?

There are actually fiver more versions of that diagram that I didn't up load yet. They gradually back off the best practices in stages. Nearly every time I post one of my grounding rants, there is a cadre of people who pop up with variations of "Well in all my [insert some long time period here] building/fixing amps, I never did that and mine are OK." And I completely believe them. Star grounding is not the only way to get absolutely minimized noise. It's just the only way you know ahead of time that will work, without tinkering. I never had time for moving grounding points around a chassis to see where the connection was quieter.

[Roe]

I found this paper helpful https://www.valvewizard.co.uk/Grounding.pdf

Merlin recommends using a ground reference at a point in the circuit that minimizes the power supply current flowing past/through the ground connection point. Specifically: connect the ground reference at the cathode of the input stage filter cap where there's only a few mA of current flowing through the ground wire creating extremely small noise voltages, not at the power amp or supply.


groundingversionm.png


One of the bigger points Merlin also recommends is that you don't use the amp chassis as a ground conductor; that the entire circuit exists and functions without any connection to the chassis. RG's Version 1 drawing kind of implies that too but I think it's worth mentioning in words that ideally the green lines are wires and the only connections to the chassis are the two points that are drawn as chassis connections.

Sure, lots of people use the chassis as a ground conductor and it's "fine", but who wants to make an amp that's "fine" when you can make one that's "as good as possible" instead?

Yes, the chassis ground connection should be only one place, not at the main ground but rather at the input stage instead.

[pjd3]

On soldering pots,

On my three amp builds (cleanish pedal platforms, all are very quiet) I either set an 18AWB solid wire on and along the preamp section of the board wiring from that to the ground lugs of the pots Or, set the 18AWG solid wire hung from the lugs of the pots say, a half inch out suspended by a short piece of solid wire from each pot lug. And then send each preamp ground over to that suspended pot wire buss. It depended on the geometry of the chassis which way I did it but I always thought soldering directly to the pots was a messy and lazy thing to do. If I was told it was unquestionably the best way to alleviate loops/hum/buzz I would do it but, I've always been happily willing to just provide a "buss wire" to either the board or suspended from the solder lugs of the pots.

To the limits of my thinking, they both seemed like equally viable ways to provide a ground to the pre and pots.

Phil D.

[pjd3]

Hi,

While the recommended grounding schemes are beginning to make electrical sense to me, I'm still not understanding the "from preamp ground to chassis" considerations - the how, where, and why it is placed on the chassis.

I'm taking it that there would be a benefit over a floating preamp ground, where the preamp ground is only connected to its associated power filter cap cathode. If the idea is to avoid ground loops, isn't the introduction of "preamp ground to chassis" adding in a potential for a ground loop? Thats the part I'm not conceptualizing very well.

Thank you, this is good stuff here, and gives me hope.

Best,
Phil D