Why terminate circuit ground(s) to chassis, rather than directly to mains ground point?

[chas.wahl]

In order to avoid further digression of @pfarrell's 5F6-A build thread, I'm starting a new one, and including most recent relevant discussion that's transpired there so far. After reviewing that discussion, and mulling this over, it occurs to me to ask the question that's this thread's title: Why terminate circuit ground(s) to chassis, rather than directly to the mains grounding point (with a bona fide conductor)? By this I mean, as an alternative to the single or split ground points terminated to the (non-optimally-conductive*) chassis, why not wire this (or these) termination directly to the ground reference where the mains ground is fastened to the chassis? It seems to me that would take the chassis out of the picture, except as a mains-ground-protected element for user safety, and provide a more conductive and direct ground source reference for preamp and power sections. Yet nobody seems to do this, so far as I can tell.
*See this recent post for relative electrical conductivity of copper (1.0), aluminum (0.63), carbon steel (0.12) and stainless steel (0.024).

Is there something obvious (but not to me, yet) that I am missing? Sure, there are lots of circuit components that are metallic and physically attached to the chassis base material (pot bushings, for instance) that Leo Fender, in fact, used as part of the circuit grounding; but more recent theory and practice has seemingly evolved away from extensive use of the chassis for conduction of circuiting. Why not pretty much eliminate that entirely?**

Anyone care to opine about which is better (more reliable), as a ground reference point: 1) the sleeve lug of a single jack fastened to its mounting hole in the chassis, or 2) a dedicated screw/nut/ring terminal next to said single jack (itself isolated from chassis)?

I would think dedicated lug on chassis with appropriate thread locking, because of the forces exerted on the input jack.

I wonder how important the jack isolation really is to the single-point grounding scheme? @2L man ? It's so physically close to the single-point ground. Even if not using isolation, I would still run a wire from jack to chassis ground, so as not to rely on the jack's connection to the chassis.

Looking at the "theoretically ideal" situation, having a "single dedicated ground at inputs" for a 4-input amp with isolated jacks means what looks like a lot of wiring. If each of the sleeve lugs, and any input wire shields employed, are all soldered to the "bus" (not "buss", which is a kiss), that's up to 8 joints, unless one daisy-chains them. And whether one daisy-chains, or not, in what order should the 1) bright high, 2) bright low, 3) normal high and 4) normal low be attached to the bus, in order of their proximity to the ground terminal? I imagine that the likelihood of jumpering jacks in use should be considered in all this, including which inputs one is most likely to use as the primary one for an instrument.

Last year I spent a while searching for and reading every single post I could find in which @2L man talked about grounding, because I wanted it crystal clear in my head. Here are some of his quotes:

"I think it does not matter if you isolate input jacks or not if all return current flow to power supply thru Common bus and no current leak to chassis at amp input."

"Your amp has lots of steel input jacks and I think you can still keep them connect to chassis because there does not flow operative current when a wire is used for that."

"You can use four non isolated steel input jacks which tie secondary zero volt (common ground) to Safety Earth. When this is only place "ground" and Chassis meet no current leak can happen and jacks often come with star washers and contuinity comes good enough."

"You can use steel input jacks or use ground lug when you use insulated input jacks. Output jack must always be insulated."

I don't want to speak on his behalf, but my understanding is that his opinion always comes from the perspective that the amp is wired to eliminate all return current from flowing through the chassis. I honestly think that the answer to your question is that it doesn't matter which way you do it in his system. Can there be a ground loop issue in a chassis in which there is no operative current flowing, whatsoever?

If you follow his methods, any electron that moves has only one straight-line path back to the power transformer, and it's never the chassis. Really cool.

Good point. Heh, I'm gonna change my story. Maybe mechanically reliable is a bigger deal than eliminating a small ground loop by isolation. My usual grounds guru, Merlin, is a bit ambiguous on the exact location; his grounds chapter, sec. 15.9, says, "the ground-to-chassis connection should always be made right at the input jack of an amplifier circuit", but since he also advocates isolated input jacks, I take this as "very near the input jack."



Yes, that was my point about isolating the jacks in a 4-holer, but it's not *that* bad. I drew a few options when I was considering one:

View attachment 1075670

Once the jacks were isolated, would order matter? or jumpers or which jack you use? Didn't do it; in a crowded tweed chassis it seemed like a bit much, but I was equally put off by drilling out the four jack holes, especially in SS.

Interesting. Blencowe advises isolating input jacks and 2L man doesn't seem to believe that it matters. When jacks are not isolated, and yet have sleeves wired to a dedicated ground terminal, there are, arguably, at least two paths between each and the "safety ground" reference to mains service: directly through chassis to mains ground, and indirectly from sleeve to "circuit ground" terminal and thence to mains ground. If the ground wires for non-isolated jacks are daisy-chained, that adds another path per jack to ground (from input jack to its partner-by-ground-wiring, and thence through chassis to mains ground. It gets dizzying.

Of course, there's always a less/least resistive path among those possible; but does that prevent ground loop action from being a problem? But I'm inclined to agree with Merlin Blencowe's approach, and attempt to limit a) number of paths to ultimate ground and b) sharing of ground path with that of other jacks, excepting that necessary right at the circuit ground terminal to chassis, by virtue of physical-mechanical constraints (how to make a 4-pronged star ground there).

How does "current flow" come into play in an input circuit? Is the instrument cable plug's tip the source of current, with respect to the input?

My apologies if I've taken too much of a detour into the weeds, for what's intended to be a 5F6-A build thread. I'll cease and desist.

On typical Tweed steel chassis I use non insulated steel jacks and I connect thicker Common ground to one of their lugs. Lately I have began to solder thin wires between their lugs in case the nut loosen. This make short parallel current paths there but when there flow very little current I think it does not matter.***

When I use insulated jacks I install soldering lug where the secondary Common begin and then solder wires to each jack where amp signal path begin. The drawing King Fan posted above the top right is what I might have used most often.

I think all these function equally good for audio frequencies.

*** Earlier today on another forum was a case where the use of guitar volume potentiometer did produce scratc sounds to amp output and reply said that bad input tube grid can leak DC current and make the guitar pickups a circuit thru instroment cable so now I must question my trought of this This potential current show as a voltage which possible to measure. Obviously voltage is higher when no guitar is plugged?

So to be clear, the optimization here (if deemed necessary) in case of 5E3 etc. would be to isolate all 4x jacks and run separate ground to chassis lug for all?

Obviously that kind tube grid leak current is separate and possibly serious issue and there nothing in grounding help.

Grid leak biasing is based to control grid internal leak current and then there is a coupling capacitor on amp input.

Isn't it operative current flow that's the issue, and not guitar signal flow? If the HT center tap is connected to the chassis, then current will flow through the chassis back to your HT center tap. Because your input jacks are also grounded to the chassis (whether instulated or non-insulated), some of this current flow can pass through your input jack grounds and induce noise because it's so close to the hot connection of the guitar signal at the input jacks.

I must confess that I do not fully understand what @gabasa is talking about ^^^here^^^. It seems to have something to do with this thread, which I missed until just now:
https://www.tdpri.com/threads/another-amp-re-grounded.1113580/
but I see in his photo that he's got grounds collected at the PT center tap, using that for ground reference rather than the mains ground connection to chassis.
**This thread seems to be an approach to answering my question above "Why not pretty much eliminate that entirely?" -- but it's not clear to me how this "solves" the issue. As an solution, it certainly seems to be an "outlier", not one I've seen before, and the discussion there regarding being shocked in a club venue is somewhat alarming.

Those interested might also refer to a very recent thread by @gabasa
https://www.tdpri.com/threads/tracing-operative-current-flow-in-a-5e3.1125894/#post-11807179
which may have been intended, at least in part, as an answer to my question posed yesterday (and quoted above): "How does "current flow" come into play in an input circuit? Is the instrument cable plug's tip the source of current, with respect to the input?". Unfortunately, I don't see an answer to that question; maybe I'm just not smart enough to understand how current flow works. I'm also confused by much of what @2L man has said; that could be a language problem, or again, my lack of understanding. Possibly current has nothing to do with input signal and the "termination of circuit ground" issue, but I'm left wondering, still.

 

[King Fan]

Heh, I understand it a bit more every time I read it; seems like Merlin's logic on circuit ground location rests on just a few concepts:

First, his single anchor is a ground *reference* only -- a proper star or single bus will carry no current: "...having dutifully obeyed a sensible ground scheme, we must finally connect circuit ground to the chassis (which is bonded to earth) at one point only. This connection is purely for safety / shielding purposes; no circuit current flows in this connection under normal circumstances. It is merely a voltage reference, not a return for current."

Then, he puts it near the input "...because, inevitably, we will want to plug other powered devices into the amp, such as mains-powered effects pedals or external preamps. These appliances will have their own earthed chassis so connecting two together will create a ground loop via the audio interconnect and the mains earth. This loop will have a huge area so even fairly innocuous magnetic fields may produce significant noise currents in it, and power supply ground currents can also flow from one device into the other, which introduces another source of hum."

Finally, his Fig 15.15 and related discussion builds on this 'device chain' idea to show why the ground reference "should always be made right at the input jack of an amplifier circuit."

 

[sds1]

No current flows to earth so there should be no penalty for having single-point ground and earth on opposite ends of the chassis, and so conversely there should be no benefit from running the single point ground directly to earth.

That said...

Merlin says the single point should be on the input end of the amp but as far as I can see doesn't say why that location is preferred. Anyone know?

EDIT: Oh gosh thanks @King Fan there's a whole section on it.

 

[Scottcurry13]

There is a long article by Aiken that discusses grounding. He doesn’t say anything against using the safety ground point as the single ground point.

Grounding

Guitar amplifiers

www.aikenamps.com

 

[chas.wahl]

In case it's unclear (much of the discussion about this is done without graphic aids at all, and I always prefer a schematic to a layout), here is an illustration of what I've proposed, with the grounding circuit dashed, and the leg from input end of circuit ground to mains chassis ground point shown in heavier line weight:

I used a 5F2-A because it's a simpler example.

If tying the circuit ground to mains chassis ground point would cause ground loop problems (as in @King Fan's quotation from Merlin Blencowe above) that could be avoided by terminating the circuit ground to chassis at input jacks, then I'd like to hear (or see) how that happens.

It seems to me that the ground reference is, ultimately, coming from the mains chassis ground point (no?) and (if necessary) through the chassis to wherever the electronic circuit ground is terminated to the chassis. Is that wrong? If that's correct, then why not make the connection thereto as direct and as conductive (and isolated from the chassis itself) as possible?

 

[Peegoo]

Based on what I know about RF and gain circuitry, I think direct grounding to the chassis reduces the opportunity to induce RF into the signal because any length of wire acts as an antenna. It's the same reason why it's good practice to keep all wires--signal and ground in a guitar circuit--as short as possible.

In my amp builds I carry all grounds from the board to a common point to cancel any differential in ground potential from various points on the board. But for pots, jacks and tubes, I direct ground to the chassis.

My builds always have very quiet operation.

 

[King Fan]

There is a long article by Aiken that discusses grounding. He doesn’t say anything against using the safety ground point as the single ground point.

Grounding

Guitar amplifiers

www.aikenamps.com

I'd forgotten that. Long and complicated. And he misspells 'bus' (I'm joking... mostly), and (less joking) suggests soldering the safety earth to the chassis, which in my limited legal understanding wouldn't meet modern safety codes. But NVM.

In the complexity, he does insist if you ground at the input jacks that 1) they must be isolated 2) you must run "a 0.01uF capacitor... from the shield terminal of the input jack directly to the chassis with the shortest possible wires," and 3) "you *must* solder the input side of the buss to the chassis right at the input jack."

He doesn't discuss the mains-powered devices upstream of the amp that Merlin considers in anchoring the circuit ground reference near / at the input jack. To be fair, maybe most guitarists don't use mains-powered devices upstream of the amp (I assume 9V pedals etc don't count?), and if they do (6G15?) they may adopt a ground-lift plan. Dunno, we don't play that way in my sandlot league...

Now I should admit my first amp (5F2a) tho it had a single bus and isolated input jacks, had the bus anchored (insert sensitivity warning here) with the power amp grounds near the PT. Didn't know I was following Aiken. Worked fine. But I did later revise it to anchor at the input end...

 

[chas.wahl]

Now I should admit my first amp (5F2a) initially had a single bus, isolated input jacks, with the bus anchored (insert sensitivity warning here) with the power amp grounds near the PT. Didn't know I was following Aiken. Worked fine. But I did later revise it to anchor at the input end...

And did the revision result in a noticeable improvement, or not?

In the complexity, he does insist if you ground at the input jacks that 1) they must be isolated 2) you must run "a 0.01uF capacitor... from the shield terminal of the input jack directly to the chassis with the shortest possible wires," and 3) "you *must* solder the input side of the buss to the chassis right at the input jack."

I'm going to guess that the "shield terminal of the input jack" is the sleeve; correct? And might this capacitor serve as the connection between the (isolated) sleeve lug and the circuit ground point, or is that a different connection (parallel to the cap? doesn't make sense to me). I don't see anybody doing the cap thing -- have I just missed it?

In my amp builds I carry all grounds from the board to a common point to cancel any differential in ground potential from various points on the board. But for pots, jacks and tubes, I direct ground to the chassis.

Does this mean that (referencing my schematic above) all the points labeled "to B" or "to C" (jacks, cathodes and pots) would go direct to the chassis and not to the ground bus?

 

[Snfoilhat]

I thought that the ideal safety ground was supposed to be a dedicated piece of hardware connecting your power cord or iec inlet to chassis. If so, then it’s not necessary and maybe less safe to piggyback other circuit (e.g. ground reference) or mechanical (e.g. transformer mount) functions onto that.

I dont think what happens on the secondary side of the power transformer cares about any ‘true’ reference to earth at all. Ground potential is just whatever the guitar and strings are at relative to what the other components held above or below them are

 

[Phrygian77]

First, the chassis is metal and must be earth grounded for safety. Second, the chassis importantly serves as an EMI/RF shield. Using it for both requires compromises or a more complicated approach. Directly connecting the circuit ground to the chassis is a compromise.

 

[JPKmusicman]

I thought that the ideal safety ground was supposed to be a dedicated piece of hardware connecting your power cord or iec inlet to chassis. If so, then it’s not necessary and maybe less safe to piggyback other circuit (e.g. ground reference) or mechanical (e.g. transformer mount) functions onto that.

That's basically what UL stated in an appliance code I read many years ago. You don't want anything else physically attached to the chassis safety ground point. Just the ground for the mains. It's for safety reasons. That way it's not tampered with during the rest of the appliance assembly or future service (unless a new cord set is installed for example).

 

[chas.wahl]

First, the chassis is metal and must be earth grounded for safety. Second, the chassis importantly serves as an EMI/RF shield. Using it for both requires compromises or a more complicated approach. Directly connecting the circuit ground to the chassis is a compromise.

This doesn't make sense to me. Does "using it for both" mean "grounded for safety and EMI/RF shield" -- that's a compromise already, right? And direct connection of circuit ground there is yet another compromise? I've asked where is the source for the ground reference for the circuit, and I'm not understanding where that source is, if not the mains ground point indirectly through the chassis (which seems like a compromise to me). If electrical code insists on that compromise, then OK, I see some logic there -- but that's a different issue (prevention of tampering, as opposed to electrical efficiency).

 

[gabasa]

Hi Chas, I hope your ok with me posting two of my illustrations from my other conversation over to here, because I think that in essence, both of our topics are the same. In the drawings below are the current path from the PT secondaries, through V1, and back to the PT center tap. They both illustrate the exact same thing, but the first is in schematic form and the second is in layout form. In the second one, the only current path I could find back to the power transformer is through the input jacks and chassis. The electrons could take another path, if there were one, but there isn't a path that exists elsewhere that I can see.

You can see the compromises once you see that PT current has no choice but to flow through the input jacks and chassis in the second drawing. Unless, of course, I'm missing something, which was the point of my other thread.

 

[Phrygian77]

This doesn't make sense to me. Does "using it for both" mean "grounded for safety and EMI/RF shield" -- that's a compromise already, right? And direct connection of circuit ground there is yet another compromise? I've asked where is the source for the ground reference for the circuit, and I'm not understanding where that source is, if not the mains ground point indirectly through the chassis (which seems like a compromise to me). If electrical code insists on that compromise, then OK, I see some logic there -- but that's a different issue (prevention of tampering, as opposed to electrical efficiency).

I think the standards organizations only insist on not compromising the earth safety ground, so that means you have to be more crafty about RF shielding and avoiding ground loops.

Ground (common) and references concern differences in potential, different impedances, between one part of a circuit and another. You're never dealing with perfect conductors, so there is always some impedance between here and there.

 

[Peegoo]

Does this mean that (referencing my schematic above) all the points labeled "to B" or "to C" (jacks, cathodes and pots) would go direct to the chassis and not to the ground bus?

Yes, that is how I like to do it. If it's a jack that has common ground continuity with the chassis per the schematic, I ground the jack at its location.

 

[chas.wahl]

You can see the compromises once you see that PT current has no choice but to flow through the input jacks and chassis in the second drawing. Unless, of course, I'm missing something, which was the point of my other thread.

Let's stick with the schematic. I don't see any return current flowing through the jacks. And if you follow the "current flow" diagram in Merlin Blencowe's "Grounding" article, you see that as the circuit (shown with ground return) progresses further toward the input, the magnitude of current lessens by degrees, apparently -- to the point that it's presumably next to nothing at the input, which is the whole point of an amplifier. So what is the problem being solved by grounding everything to the PT HT center tap? Seems to me like taking the reference back to the source of a lot of noise and AC activity.

If the answer to my question is that direct connection to mains ground point is off limits due to appliance code, and (as @Phrygian77 notes) we simply have to put up with a less efficient connection to ground reference, and it doesn't matter that much, then I think that's a conclusion I will accept. My understanding of Merlin Blencowe's discussion about ground loops and attached devices (at an input) is still very murky; but I will follow that up and try to improve. Thank you all, for your patience and forebearance.

 

[Phrygian77]

Yes, that is how I like to do it. If it's a jack that has common ground continuity with the chassis per the schematic, I ground the jack at its location.

The alternative, if this is an input or output for sensitive signals, is to isolate the jack and then RF reference the shield to the chassis at that location via a capacitor. The capacitor blocks AC at lower frequencies to prevent ground current loops.

 

[tele_savales]

I've been reading and re-reading Blencowe and Aiken on this. Hoping that eventually it'll come to me. Iirc, Aiken doesn't seem to think isolating the input jacks is essential? I may have read that wrong or be partially mis-remembering. Honestly it seems to me there are best practice concepts and then you kind of wing it for your particular circuit and hope for the best.

Up until now I've been using split ground, I'll probably try a unified buss and insolated jacks (including output) next build. Also I've been remiss in grounding my output tube cathodes to the power section ground.

I have a mental block in understanding running the buss wire to the input jack vs running the bus wire to a solid ground lug and the input jack ground lug to the buss, other than stability of said ground buss and the ability to run a capacitor from the jack to the buss. Does it really make a difference?

EDIT: I have a number of mental blocks.

 

[sds1]

Let's stick with the schematic. I don't see any return current flowing through the jacks. And if you follow the "current flow" diagram in Merlin Blencowe's "Grounding" article, you see that as the circuit (shown with ground return) progresses further toward the input, the magnitude of current lessens by degrees, apparently -- to the point that it's presumably next to nothing at the input, which is the whole point of an amplifier. So what is the problem being solved by grounding everything to the PT HT center tap? Seems to me like taking the reference back to the source of a lot of noise and AC activity.

The current is returning to the PT HT no matter what. We cannot control that, we only have control over the physical path that the return current follows.

You have to look at the layout he provided. The PT HT is connected to a power amp ground bus. The entire preamp is grounded to a separate bus that connects to the chassis at the input jack. For preamp DC currents to return to the PT HT, the only path available is via the chassis via the input jacks (on the blue line he drew).

He is citing this as an example of what NOT to do.

 

[2L man]

The concept "ground" does not exist anymore!!!

On Mains (primary side) it has become a Safety Earth.

On secondary the operative "ground" is where current return back to power supply. English is not my language but re-turn is what can be seen to happen to feed current when it turn 180 decrees back to PS. Same time this turm make feed voltage drop to zero!

When I began study electronics it was named as "zero volt". So when I see upside turned T-symbol I always think zero Never chassis! And I managed to do electronics work almost 42 years

When I see chassis I think Safety Earth and Shield.

When power supply has positive and negative feeds to balanced system, like for example solid state AB class amplifiers, there flow only fraction of current on 0V and it become more as a reference.

Signal use 0V as a reference often. Some vintage effects use positive as reference inside the circuit when PNP transistors are used.

Often tube amps have negative bias supply but it feed so little current and most return thru 0V.