grounding: theoretical question

[barret77]

Hello all

I have a theoretical question, it does not regard any specific build...

I think there's some issues when designing "creative" enclosures regarding to grounding; I read a lot of problems about using wood and other material enclosures - noise, oscillation, hum... and people always refer that the enclosure is not providing proper grounding...

question 1) does the metal enclosure works like a magnetic shield ? Is its magnetic properties of the metal wrapping the whole circuit that count when they make a difference on the quality of the effect?

question 2) if not, if it is only its grounding properties, then could I have a "external virtual ground"? I mean, instead of star grounding a circuit to a point in a metal chassis, star ground a circuit to a jacket and plug this jacket to a 10 lb. iron block on the floor? (Living on an apartment, I can't bury a rod to make some real grounding rather the old-apartment-lousy-outlet-grounding)


thanks for any light in this question...

[R.G.]

question 1) does the metal enclosure works like a magnetic shield ? Is its magnetic properties of the metal wrapping the whole circuit that count when they make a difference on the quality of the effect?

Yes, and no. A pure copper box would provide very good shielding because it is highly conductive, and because one of the consequences of radiated electromagnetic waves (that's what carries the hum and radio interference) is that they are forced to the outside of a conductive surface. A hollow copper box has (theoretically) no external fields inside it because it forms a conductive sheet that "shorts out" the external fields trying to get in. This is sometimes referred to as a "Faraday cage" or "Faraday shield" after Michael Faraday, the first person to deduce the law that includes this effect.

The relative goodness of a Faraday cage is measured by how conductive the external walls are. Copper is second only to silver as a highly conductive metal; aluminum is third as I remember. Gold is close in there. Steel is more resistive than any of these, so it makes for a worse Faraday shield, and you have to use greater thicknesses for equal shielding. But that's OK because it's MUCH cheaper.

And steel is magnetic. "Magnetic" in this context means that it has a high permeability, which is to magnetic fields as conductivity or low resistance is to electrical fields. Steel is on the order of 1000 to 10K or more times as "conductive" of magnetic fields as free space vacuum or air is. For magnetic fields only, steel does a similar thing to what copper does for electrical fields, "shorting out" any magnetic fields that try to go through it.

So to a first order, all you need is the conductive metal. Having it be magnetic like steel is a benefit in some situations, but is not the primary "goodness" of a conductive box. Best would be a soft iron box that's been covered with a couple of mm of pure silver. But that may be too expensive for garage effects.... 8-)

question 2) if not, if it is only its grounding properties, then could I have a "external virtual ground"? I mean, instead of star grounding a circuit to a point in a metal chassis, star ground a circuit to a jacket and plug this jacket to a 10 lb. iron block on the floor? (Living on an apartment, I can't bury a rod to make some real grounding rather the old-apartment-lousy-outlet-grounding)

You are running head first into the most confusing issue around grounding.

The real thing that's slipping around in the shadows here is the question - "what's a ground?"

In the early days, grounding (or the British "earthing") referred to literally connecting the circuit to the planet electrically. This is not as easy as it sounds, because some earth materials like many stones are not conductive, so you need to connect your ground conductor to wet dirt, and that is why home electrical ground rods are driven 8' to 10' into the dirt. The earth itself serves as a convenient source of zero-point-zero volts and is as close to an infinite electron sink as we have. As a race, we cannot (yet) change the total charge on the planet.

But we are not too often interested in whether our guitar effect is a few volts different from the dirt under us. We are HIGHLY interested in whether a transistor in the front end of the circuit is a few volts different from one in the back end. So electronicists decided that for many things, "ground" means only a source of zero volts. And we can simply decide to call any point zero volts or "ground" because we could, in theory tie a big copper wire to that point and attach it to a grounding rod in the dirt if we wanted to. So "ground" means  "a reference voltage point".

The shielding of an enclosure does not depend on it being grounded to the earth in most cases. But since all of our common connectors connect electrically to the enclosure, we would like for the enclosure to be at zero volts, so we make a positive effort to tie our zero voltage reference point for signals to the enclosure. That "grounds" the box.  And "grouding" then becomes a problem, because the temptation is huge to use the box as not only a shield, but also as a current carrying conductor to connect internal circuit points that are supposed to be at 0V.

This is a BAD idea, and it is bad because the enclosure metal, even if aluminum or copper, have some resistance; and when current flows through them, a voltage drop results. So the two places in the circuit that are both supposed to be at the same zero voltage are now at slightly different voltages, and the different voltages depend on how much current is flowing through the enclosure metal.

Now you can have the case that an output-ish part of the circuit that uses big currents is returned electrically to the power supply source of current through the enclosure metal that's supposed to be a perfect conductor at 0V but isn't. And that "ground" potential is shared by a sensitive input circuit that amplifies the slight voltage drops across the enclosure by a million or more (almost any opamp will do this). That constitutes feedback, and Mother Nature being the ...er... lady... that she is, it's usually positive feedback and can make your circuit oscillate.

Not only that, but the way a Faraday shield works is that the impinging electrical fields have their energy eaten up in the shield by causing eddy currents as they try to cross the shield. Those eddy currents cause slight voltage drops and those are in turn amplified by those neato gain-of-a-zillion fuzz circuits, so we can then hear Radio Nepal on our Fuzz Blaster.

Same for hum. Magnetic fields cutting across electrical conductors like the box can cause electrical voltage differences. Bango, you get 60 cycle hum from the enclosure that's supposed to be protecting you.

Correct grounding for an effect is like this:
(1) the circuit itself is self-consistently grounded so that it works properly with no enclosure at all in a quiet environment. You do not depend on the box to carry "ground" from place to place for you.
(2) the enclosure is then connected to the circuit by one and only one conductor. That way the enclosure simply cannot be carrying current from one place in the circuit to another, and cannot cause feedback, radio pickup or hum.
(3) For really critical stuff like tube amps, the circuit itself should be internally star grounded or carefully buss grounded.
(4) whether the enclosure is actually attached to dirt ground or not is purely incidental if you have 1-3 correct. ... in most cases...

An enclosure needs to be a conductive metal. Steel is OK for this, and is used because it's cheap. The magnetic effects of steel are a benefit, but are not the only thing, because all aluminum boxes work as well. Connection to dirt is not mandatory.

[H.Manback]

Great stuff, and I think grounding is one of those 'mystery' subjects in electronics, where you can unfortunately get hit in the head very hard.

This stuff is really nice for a sticky btw.

[barret77]

WOW
thanks R.G., I'm really thankful you spent the time writing all that. Really, really, really nice. I hope other forum users will benefit from this colaboration as well.

[NaBo]

Great post, as usual, R.G.  

I think this came up before, but that was before I really knew enough to understand...  But about that "one and only one" connection point to the enclosure...

I remember reading how the stereo-input-jack-completing-grounds trick should ideally be employed using the OUTPUT jack... something about any noise resulting not then going through the circuit and being amplified a kajillion times...

Should I surmise then that the one connection should ideally be made through the output jack?  Or is completely isolating BOTH jacks and hard-wiring the circuit ground to the enclosure worth the trouble?

[d95err]

Great post, as usual, R.G.  
I remember reading how the stereo-input-jack-completing-grounds trick should ideally be employed using the OUTPUT jack... something about any noise resulting not then going through the circuit and being amplified a kajillion times...

I think the "use the output jack" thing is mainly about how to use the jacks as an on-off switch for the effect. I.e. power is off until you plug something in.

As far as grounding is concerned, you can solder a wire between the ground connectors on the in and out jacks, in order to ensure that the enclosure is not required for grounding.

(EDIT: Note that this is not from personal experience, this is just what I've read from other sources)

[NaBo]

err, yes, i'm aware of that much  :wink: ... I was just referring to how somebody (Mark Hammer I believe?) discussed how ideally the power-switching stereo jack should be the OUTPUT jack, as opposed to the INPUT jack, but because of practical reasons (unplug guitar instead of amp) and convention, it's always been the input.  The danger of this is that any hum picked up as a result of the solder joints or mechanical connections or interference or ANYTHING at this point will then be amplified by the circuit and show up a thousand times louder at the output.

My main query was about where the ideal SINGLE point to ground to the enclosure would be for the best Faraday shield with no ground loops... ie, whether it's necessary to use plastic washers or plastic jacks for isolated mounting of BOTH jacks and then run a wire to the enclosure, or if its acceptable to use the output jack as the connection point

[R.G.]

I remember reading how the stereo-input-jack-completing-grounds trick should ideally be employed using the OUTPUT jack... something about any noise resulting not then going through the circuit and being amplified a kajillion times...

Should I surmise then that the one connection should ideally be made through the output jack? Or is completely isolating BOTH jacks and hard-wiring the circuit ground to the enclosure worth the trouble?
*************
I think the "use the output jack" thing is mainly about how to use the jacks as an on-off switch for the effect. I.e. power is off until you plug something in.

As far as grounding is concerned, you can solder a wire between the ground connectors on the in and out jacks, in order to ensure that the enclosure is not required for grounding.

Grounding is one of those elephants us blind men have to feel of a lot to get a good picture.

Grounding gets mixed up with powering a lot because at least one of the power supply lines is almost always grounded, so we call that power supply line "ground" when it's really a power supply cable, not a true source of zero volts.

Power supply cables **do** show the "resistive voltage drop amplified a kazillion times" effect if you wire it up wrong.

So: d95err is correct, the battery-to-jack thing is for turning off power when the cable is pulled so that (a) you can be cheap and not buy or install a switch and (b) it seems clever. And things do all work if there is a wire between the sleeve lugs of the input and output jacks and then one wire to the circuit ground. In fact, this is the better, preferred way to wire these things up. You don't have to rely on the contact that the jack makes with the enclosure to "ground" the input or output jack. Some tube amps have severe problems when their jacks get a little loose and the jack grounding which relied on the jack-to-panel connection gets bad.  

Remember my  item 1 in correct grounding?

Correct grounding for an effect is like this:
(1) the circuit itself is self-consistently grounded so that it works properly with no enclosure at all in a quiet environment. You do not depend on the box to carry "ground" from place to place for you.

The wire from circuit to input and to output does this. Preferably two wires, each from the circuit board to one jack, but you can get away with one wire to a jack and then another jack-to-jack wire.

But back to power and ground. Mother Nature is a stickler for details, but sometimes you can get away with minor sins because the punishment is not too severe - like snitching cookies from the kitchen jar. If you "ground" the enclosure by connecting the input jack to it and the switch your power by connecting the battery minus to the ring contact of the input jack, where does the current go?

It runs down the battery minus lead, into the ring lug, onto the sleeve of the mono plug, into the sleeve of the jack from the sleeve of the plug, then down your *signal ground wire* to the circuit board.  Over at the circuit board, the input circuit sees the difference between what it gets from that ground wire (no parentheses - it really considers that wire to be zero volts) and the signal wire from the jack as signal and amplifies it.

If the circuit has a low gain like most phasers, flangers, tremolos, etc, then the circuit's used-up waste electrons that go down that wire back to the battery like a drain pipe cause whatever voltage they cause, but it's amplified by only the low gain, and that may be much smaller than the signal. So while it's still there, Mother Nature only taps your wrist with a microvolt signal, which you can ignore.

If the signal gain is a zillion, then Mother Nature insists that the zillion be applied to the microvolts too, and you have a problem.

If you wired the battery switch to the output jack, the same microvolt signal still runs on the "ground" wire, but now it's guaranteed to not be amplified by this pedal at least, so it probably won't interefere.

This is why I really don't like the simple form of jack ground switching. Consider how it works if you have wires from both input and output jack sleeves to the circuit, no power coming down them. Then you  wire the battery to the circuit board, both power and ground, and put a PNP transistor in the + battery lead. The base of the PNP is connected through a 10K resistor to a wire to the input or output jack ring lug.

Now the mono plug shorts the ring lug to signal ground. This pulls a current of battery voltage/10K through the PNP base and that current does indeed travel on the signal wire.

But it's both smaller than the raw power supply current by the gain of the transistor and *constant*. There's no signal component at all in the I*R voltage down the ground wire, so there's no AC signal to be amplified, and Mother Nature smiles approvingly on you even if you are amplifying with a Mond-O-MegaBlaster circuit. The waste electrons do not travel on wires that you are using for signal reference voltages.

And that is the entire trick to proper grounding - visualizing **where do the electrons travel?**