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

[Phrygian77]

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.

Ground loops are only going to be a significant problem when there's significant ground return current. The real potential problem with not isolating the jacks (and ultimately the whole circuit ground) from chassis/earth is a shared signal return path and earth ground path. This happens when there is another device that is connected to both the common ground and the earth ground through two different connections, for example, when you run a guitar to two separate amps, the amp chassis are connected together through the instrument cable shield and through the mains electrical ground. The electrons will do what they do. Even though you provided them with a nice signal cable to get from point A to B, they will also inevitably make they're way through the ground wires in the power cables also.

 

[chas.wahl]

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

View attachment 1076081

If by "what not to do" a split power|preamp ground is meant, I don't know how that fits in the context of this discussion. Sorry if that appears to be dense on my part.

Part of the problem with the way most schematics are done is that they just include a lot of ground symbols without making plain any intention about how/where they're terminated. And it seems to be an issue that a) doesn't seem to matter in many cases and b) makes a huge difference in many others.

 

[sds1]

If by "what not to do" a split power|preamp ground is meant, I don't know how that fits in the context of this discussion. Sorry if that appears to be dense on my part.

The discussions you cited in starting this thread deal with avoidance of using the chassis as a return path.

The layout diagram he cited uses the chassis as a return path.

 

[gabasa]

Here's the opposite; an example of how to avoid the issue above. If the filter cap grounds are reconnected, and the power amp ground lug is separated from the chassis, then the new current path is shown below. No current can possibly flow through the input jacks and chassis back to the PT because no path exists in that direction after having lifted the ground bolt off the chassis.

This is relevant to the question in the original post, "why not wire this (or these) termination directly to the ground reference where the mains ground is fastened to the chassis?" In the first layout I posted above, all the return current from both preamp tubes goes back to the PT via the input jacks and chassis. In this second layout, return current cannot flow through the input jacks and chassis.

Once there's pulsating DC flowing through the chassis, it can induce noise into the circuit, as well as to other guitar amps in proximity.

 

[King Fan]

To quote the immortal Bluto Blutarsky, “Food fight!!!”

Seems like we're now arguing mainly over terminology and the misunderstandings that arise from it. When folks seem to confuse “power amp ground” with household power ground (“safety earth, please”) we're in trouble…

Rather than go further down that ravine, lemme climb back up on the mesa and survey the countryside. Most of us can’t speak with the insight of Merlin or Aiken. They don’t actually disagree that much; to the extent they do, there are likely informative reasons, like the upstream devices in Merlin's thinking. BTW, neither endorses sharing circuit ground with household ground / safety earth.

For a refreshing and logical third view, read R.G. Keen; he’s good on theory and practical about practice. Sorry. Link tool seems to be MIA. Google “geofex grounding” or copy this URL…

http://www.geofex.com/article_folders/stargnd/stargnd.htm

Now let's take off the space helmets and breathe the good Earth air. We're mostly building simple amps on time-tested layouts. The theoretic risk of *noisy* ground loops exists, but on a standard 50s-60s circuit, it's rare. Merlin is the first to admit we'll often get away with the theory violations Fender, Marshall, and others did.

We're hearing here from super smart TDPRI builders who don’t follow the pure theory and have nice quiet amps. I’ll just guesstimate that of experienced TDPRI builders, 80% don’t isolate input jacks, 70% don’t group their grounds by node, 60% don’t isolate output jacks, 30% or more run their bus on the pots — and well over half use a Rob-style split bus. A significant number still use a brass plate and say it works fine.

Is any of that best practice? Well, Rob tells us some smart builders have found a split bus quieter *in certain amps.* Other than that, those are all Merlin/Aiken no-nos. I’m not saying best practice isn’t a good thing. I’m saying these tiny arguments about exactly how near the input jack we 'must' ground are theoretic angels dancing on the heads of invisible pins.

Me, I try to go best practice when I can. *Most of the steps are dead easy.* I've seen elegant best-practice builds by our earth-wise friend @2L man . @Lowerleftcoast often helps me on ground order. But the one audible change I’ve made? It's another one that (again, guessing) 75% of builds don’t follow, Merlin’s HT CT to first filter negative. If you’re designing amps, or building big complex high gain amps, novel layouts, etc. “might as well have the best.” But in the meantime, put down the mashed potatoes…

 

[Scottcurry13]

Good post @King Fan. I would add that with anything we discuss, circuits, mods, layouts, components, etc., sometimes it’s best to try them out for ourselves and evaluate the results with our own ears. I have not isolated jacks, but have the washers to try it if ever needed. The last two builds I’ve done have not had hum issues so I didn’t have to go that route. But I do enjoy these discussions because I learn from them.

 

[gabasa]

“Food fight!!!”

If I came across that way, I apologize to the OP.

Merlin's chapter on grounding states that, "Many old guitar amplifiers use more-or-less random grounding to the chassis ... The chassis should be considered nothing more than a metal box which protects the user from the circuit, shields the circuit from interference, and provides a solid support for the construction. Allowing it to become part of the audio circuit by treating it like ‘one big fat wire’ is asking for trouble." He also states that, "The bus bar should be connected to the chassis at one point only, at the input end of the amplifier."

Regarding the, "Ground-to-Chassis Connection," Merlin says, "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."

So basically, connecting the PT center tap to the chassis is a no-no. Even the speaker jack must be isolated. Understanding how to implement this all (as in the last layout drawing I posted) was probably the biggest light-bulb moment I ever had in my amp building journey.

Aiken generally agrees with all this:
"If you instead ground the buss at the input jack, which is usually best for EMI/RFI you don't need to use an isolated jack."
"Do not connect both the input jack and the first power supply ground to the chassis (or any other point, for that matter) or you will get low-level ground loop hum."
"Under some circumstances you might be able to get away with using the chassis as the buss ground (instead of an isolated, heavy-gauge buss wire), but this almost always leads to low-level ground loop hum problems, and should be avoided."

I was previously just trying to provide visual layout examples to show what happens to current flow when you follow Merlin's, Aiken's, or 2L man's suggestions, and what can happen when you don't. Here's a drawing from his grounding chapter. The only connection to earth ground is at the input jack. The reservoir cap end of the bus does not touch earth ground.

I hope that this clarifies where I was coming from in my previous posts.

 

[gabasa]

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."

Just for clarification, he (Aiken) says the opposite. If you ground at the power supply, you need to isolate your input jacks and use the capacitor. If you ground at the input jacks, you don't need to isolate your jacks, nor do you need the capacitor.

 

[Scottcurry13]

If I came across that way, I apologize to the OP. Perhaps I inadvertently threw a banana across the lunchroom.

Merlin's chapter on grounding states that, "Many old guitar amplifiers use more-or-less random grounding to the chassis ... The chassis should be considered nothing more than a metal box which protects the user from the circuit, shields the circuit from interference, and provides a solid support for the construction. Allowing it to become part of the audio circuit by treating it like ‘one big fat wire’ is asking for trouble." He also states that, "The bus bar should be connected to the chassis at one point only, at the input end of the amplifier."

Regarding the, "Ground-to-Chassis Connection," Merlin says, "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."

So basically, connecting the PT center tap to the chassis is a no-no. Even the speaker jack must be isolated. Understanding how to implement this all (as in the last layout drawing I posted) was probably the biggest light-bulb moment I ever had in my amp building journey.

Aiken generally agrees with all this:
"If you instead ground the buss at the input jack, which is usually best for EMI/RFI you don't need to use an isolated jack."
"Do not connect both the input jack and the first power supply ground to the chassis (or any other point, for that matter) or you will get low-level ground loop hum."
"Under some circumstances you might be able to get away with using the chassis as the buss ground (instead of an isolated, heavy-gauge buss wire), but this almost always leads to low-level ground loop hum problems, and should be avoided."

I was previously just trying to provide visual layout examples to show what happens to current flow when you follow Merlin's, Aiken's, or 2L man's suggestions, and what can happen when you don't. Here's a drawing from his grounding chapter. The only connection to earth ground is at the input jack. The reservoir cap end of the bus does not touch earth ground.

I hope that this clarifies where I was coming from in my previous posts.

View attachment 1076273

This is what is confusing to me. The first filter cap has a lot of voltage in it. So its potential to ground is high. Having that current go back to the input jack seems counterintuitive to me. The input and first gain stage is the most sensitive to noise. It’s seems we would want to get any potential ground current as far away from the input as possible.

 

[sds1]

This is what is confusing to me. The first filter cap has a lot of voltage in it. So its potential to ground is high. Having that current go back to the input jack seems counterintuitive to me. The input and first gain stage is the most sensitive to noise. It’s seems we would want to get any potential ground current as far away from the input as possible.

If you do it right the current can only flow on the bussss bar, and the chassis connection at the input jack is merely a voltage reference.

 

[Badside]

This is what is confusing to me. The first filter cap has a lot of voltage in it. So its potential to ground is high. Having that current go back to the input jack seems counterintuitive to me. The input and first gain stage is the most sensitive to noise. It’s seems we would want to get any potential ground current as far away from the input as possible.

I got confused by that too at first, until I realized the chassis connection when done only at the input jack is merely a 0v reference. No current goes through there.

Gotta make the distinction between ground and earth.

In this approach, you control where high current returns happen and therefore can keep it away from the first stages. It just returns straight to the PT center tap without a trip through the chassis.

 

[sds1]

I got confused by that too at first, until I realized the chassis connection when done only at the input jack is merely a 0v reference. No current goes through there.

Gotta make the distinction between ground and earth.

In this approach, you control where high current returns happen and therefore can keep it away from the first stages. It just returns straight to the PT center tap without a trip through the chassis.

If it helps folks...

Purely in terms of not using the chassis as a current return, it is irrelevant where the single connection to chassis is physically located. Could be at the input jack, or at the opposite side of the chassis, the amp's current flow will not be affected by this decision either way.

The preference to locate the single chassis reference at the input jack is actually part of a different discussion regarding interference with other devices nearby sharing the same earth ground.

Two different design objectives combined into a single recommendation.

 

[Scottcurry13]

Thanks @Badside and @sds1. I keep thinking of it as flow because I’m a retired chemical engineer when it actually isn’t like that. Got to study up on that more. Now if this was plumbing……

 

[2L man]

Amplifier secondary 0V "ground" does not have to be hard wired to Chassis / Safety Earth. Electrically it function just fine.

Mains Safety Earth must always be hard wired to chassis using own bolt/nut mounted crimped lug because thru the SE "the wall" fuse protect in case amp own fuse fails to protect. This wall fuse protects the wires inside the building not causing fire.

A circuit called "ground loop hum eliminator" can be used to tie secondary 0V to SE softly! This make the use of two amp setup hum less when other amp has this circuit.

It consists four parallel components: Two diodes which are connect opposite each others and which current rating is at least three times the HV fuse rating so HV fuse burn thru these diodes if HV leak to Chassis. A 10 ohm 3W resistor and it is which does the softer connection to SE. And 100nF X-class capacitor which "shorts" high frequency noise.

 

[King Fan]

Amplifier secondary 0V "ground" does not have to be hard wired to Chassis / Safety Earth. Electrically it function just fine.

I always respect your insights, @2L man . Of course you're right. Merlin says the same thing, with a practical coda:

“Although the terms ‘earth’ and ‘ground’ are often used interchangeably, the audio circuit ground does not necessarily have to be connected to planet Earth. The entire amplifier circuit could be built ‘floating’ inside the metal chassis, with no connection to the chassis at all. However, in reality the circuit will be connected to chassis at some point since this ensures the amplifier’s working voltages are properly defined with respect to zero volts, and that the chassis acts as a shield against electric fields.”

 

[King Fan]

Meanwhile, back at the ranch, er, um, thread title. @chas.wahl , do you feel you got your question answered?

 

[Peegoo]

“Although the terms ‘earth’ and ‘ground’ are often used interchangeably, the audio circuit ground does not necessarily have to be connected to planet Earth.

Yes...there are two different symbols for 'ground'.

 

[Lowerleftcoast]

"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.

This is what is confusing to me. The first filter cap has a lot of voltage in it. So its potential to ground is high. Having that current go back to the input jack seems counterintuitive to me. The input and first gain stage is the most sensitive to noise. It’s seems we would want to get any potential ground current as far away from the input as possible.

We assume everyone is on the same page but we are on different levels of understanding what is happening in our little amplifier circuits.

There are three sources we are discussing here. The pickups in the guitar make a small AC voltage (signal voltage), the AC from the wall, and the DC which is rectified AC. (There are also the heater, bias, and speaker voltages but for the most part they are not coming into this discussion.)

When we are discussing *ground bus*, *using the chassis for ground return*, *one reference to ground at the input*, two point ground scheme*, *power amp ground and preamp ground*, *grounding to back of pots*, *brass plate*, etc, we are discussing various paths for the (B+) DC current to travel.

DC in our amplifiers is rectified AC. The DC source and return (+ and - on a battery) are within the Power Transformer/Rectifier/Reservoir Capacitor. (It takes all three to make B+.) This DC source and return (+ and -) is not returning current to the Mains AC ground. The return current is returning to the DC source.

Because the DC is rectified from AC, there is AC content (Ripple current) in the DC. The Reservoir and Filter Capacitors are there to reduce the AC content. The DC should have the least AC content near the first gain stage (input) of the amplifier because it has been through all of the filter stages (Nodes). This AC *content* does not flow to the Mains AC ground.

Pure DC will not induce hum in the *Signal AC* from the pickups. There must be AC or DC with AC content to cause hum and it must *ride* with the signal voltage to be amplified by the gain stages. When a ground scheme does not follow the technically correct ideal, it can be perfectly quiet if the Signal voltage is not contaminated with *other* unwanted signals.

 

[chas.wahl]

This is what is confusing to me. The first filter cap has a lot of voltage in it. So its potential to ground is high. Having that current go back to the input jack seems counterintuitive to me. The input and first gain stage is the most sensitive to noise. It’s seems we would want to get any potential ground current as far away from the input as possible.

To me, the most instructive graphic aid for this is the one I included in post #16, Merlin Blencowe's figure 15.6 from his Grounding article, showing the current flow in B+ and ground bus, with magnitude decreasing as the circuit progresses to the preamp (or the reverse, if imagined from an electron's point of view):

He doesn't include the input(s) per se; they're just part of the "load" on the right, but my presumption is that by the time one gets there in the circuit, current is basically spent -- as I think I said, the only thing left is the AC signal from the instrument at the input. All the rest has gone back to the PT-rectifier.

Meanwhile, back at the ranch, er, um, thread title. @chas.wahl , do you feel you got your question answered?

Let's put it this way; pretty much so. I understand that it's a Bad Idea to connect the circuit to or terminate it at the mains ground point. What confuses me still are a) the notion that the circuit could just be "built ‘floating’ inside the metal chassis, with no connection to the chassis at all"; b) the complication that in the case of "other devices" like stompboxes or effects that have grounds to earth, it makes practical sense to connect the bus to the input to minimize ground loops, and c) my inference that doing so somehow causes the ground reference to be set to the ground potential of earth -- presumably because of its indirect connection to the mains ground through the chassis metal. Ignoring, for a moment, the principle that the mains ground to chassis should be dedicated to that purpose alone, for reasons of safety, it still seems to me that electrically, the most efficient way to ensure that would be to connect the bus to the mains ground point -- though @Peegoo's admonition to avoid long wire runs might rather simply put the kibosh on that notion.

I don't have a problem with the orthodoxy, I just am looking for a way to understand why it is that way. And as you (@King Fan) have pointed out, there's plenty of heterodox practice anyway. Sometimes it works fine, sometimes it doesn't, apparently.

 

[sds1]

He doesn't include the input(s) per se; they're just part of the "load" on the right, but my presumption is that by the time one gets there in the circuit, current is basically spent -- as I think I said, the only thing left is the AC signal from the instrument at the input. All the rest has gone back to the PT-rectifier.

The inputs are not part of the DC load, which is what that graphic depicts. Generally speaking the load furthest from the power supply is the V1A triode, and the current returns at V1A cathode. That's as close to the input of the amplifier as it gets.

No DC current will flow anywhere near the input jacks unless your grounding scheme allows/requires it to.