Clarification on Pedal Grounding

[Big Monk]

My late night researches have been focused on grounding for the past few nights. I have been looking for ways that provide a more idealized (from a technical standpoint) grounding scheme than what I am used to (Star Ground at input jack). What I came across was a post from R.G. that has intrigue (and confused) me:

So how it should be done is:

1. The input jack sleeve should be connected to the enclosure right where it punches through the enclosure. This can be done by just the screws holding it into the box, but as noted paint, and eventually just dirt and time will make this not so reliable. This ought to be done with a star/toothed lockwasher under the fixing nut to bite into the enclosure metal for long term reliability. This makes the input ground a solid RF/shield ground to bleed incoming RF from the wires into the surrounding enclosure and keep it out of the circuits. This can also be done with a small ceramic cap of maybe 0.001 from the input sleeve right to the chassis if for some other reason you don't want to - or can't! - DC-ground the input sleeve right there. The cap leads have to be as short as you can possibly make them to be effective.

In the ideal case, this connection to the chassis at the input jack should be the **only** conductive connection to the chassis.

2. The incoming DC adapter jack and battery leads, etc. grounded sides should not connect to the chassis where they come in or elsewhere. They should run directly to a point on the circuit board where traces split out to power the circuit.

3. The input ground should connect to the circuit board ground right at the input circuitry.

4. The output jack should be isolated from the chassis, and a signal ground wire run from its sleeve back to the ground common point on the PCB. This way, only output signal currents run through that ground wire. Hooking its ground to the input jack ground terminal forces both input signal currents and output signal currents to flow in the same wire resistance. This makes an explicit shunt feedback loop around the whole pedal. As long as gain and impedance are low enough, it doesn't matter, but when you get enough, it oscillates from the feedback.

5. The input jack should not be wired for the input-power switching trick by using a stereo jack and running power negative to the ring terminal. This forces 100% of the power supply current for the whole pedal down the input signal ground wire, and the currents used by the circuit are fed back into the input. When current, gain, and/or impedance get high enough, this can make it oscillate. The output jack is a better place for this, as the currents are higher and it's not automatically a feedback issue. But it does risk putting signal out  on the output line for the next pedal. The best way to do jack-based power switching if you want to do it is to use a bipolar transistor to switch power and a resistor from its base to the stereo jack ring terminal. This can only contribute a DC offset, and a very small one, instead of a signal.

6. LED ground returns should go either to the place power is brought onto the board from the power jack or back to the power jack. This keeps the large-ish current changes off the signal reference ground wires.

Most of this makes perfect sense to me from an implementation standpoint but my brain is stuck on what I am used to and what i am used to is tying all grounds to the input jack sleeve.

In the scheme shown above in R.G.'s quote, where are the rest of the grounds connected to provide independence from the enclosure ground provided by the input jack sleeve?

[anotherjim]

Sometimes, I use mixed jacks. A plastic body pcb mount that also plays a role securing the board, and a metal one to ground the case. Only suits small projects as the board is L shape to clear the metal jack and the plastic one should be mounted at a slight tilt on the board so it remains level due to the enclosure sides being angled.

[Big Monk]

I guess my main concern would be, in the absence of the Input jack functioning as the global "star" ground point for the circuit, where do the other circuit grounds make their connection?

[GibsonGM]

Monk, open the .pdf here where it says "Click Here" to download.  This shows the standard ways of wiring up stompboxes...all work fine, most of us use the 3pdt option but there are exceptions.    Each shows how the grounds work!  They all DO connect together...


http://www.tonepad.com/getFileInfo.asp?id=76

[Big Monk]

I'm looking more for clarification on R.G.'s quote above and how it is applied practically.

I'm comfortable with traditional offboard wiring. I guess I am just wondering how R.G's grounding methods above are incorporated with normal offboard wiring.

[EBK]

I do something in my builds that is similar (I believe) to what R.G. proposed, but I use the output jack sleeve terminal instead of the input jack sleeve terminal as my enclosure shield ground connection.  Essentially, treat your enclosure as a shield, not a ground current conductor. Connect only one spot on the enclosure to ground.  For me, this means I use insulating shoulder washers on my input jack so the sleeve terminal doesn't contact the enclosure.  The output jack is mounted as usual, and it's sleeve provides the shield ground connection.  It doesn't much matter (usually) where you decide to connect your circuit grounds together, but I try to make a star ground connected as close as possible to the DC jack (my input jack, output jack, stomp switch LED, and circuit board each get their own ground wire; this does mean I'm not keeping my input ground "with" the input signal technically).
I will never tell you that my method is "necessary", "proper", or "better" though.  There are plenty of noises from plenty of sources in guitar signal chains.  You'll often hardly notice any difference between sloppy builds that work and super neat by-the-book wiring.  It is more a pride in craftsmanship issue, I believe.

[Big Monk]

I do something in my builds that is similar (I believe) to what R.G. proposed, but I use the output jack sleeve terminal instead of the input jack sleeve terminal as my enclosure shield ground connection.  Essentially, treat your enclosure as a shield, not a ground current conductor. Connect only one spot on the enclosure to ground.  For me, this means I use insulating shoulder washers on my input jack so the sleeve terminal doesn't contact the enclosure.  The output jack is mounted as usual, and it's sleeve provides the shield ground connection.  It doesn't much matter (usually) where you decide to connect your circuit grounds together, but I try to make a star ground connected as close as possible to the DC jack (my input jack, output jack, stomp switch LED, and circuit board each get their own ground wire; this does mean I'm not keeping my input ground "with" the input signal technically).
I will never tell you that my method is "necessary", "proper", or "better" though.  There are plenty of noises from plenty of sources in guitar signal chains.  You'll often hardly notice any difference between sloppy builds that work and super neat by-the-book wiring.  It is more a pride in craftsmanship issue, I believe.

Maybe this is my hangup: If we use either one of the jacks as the shield ground connections, through it's sleeve connection, how do we then run a ground from it's sleeve terminal and not have issues?

As an FYI, i'm a pretty sharp guy, it's just that i am having a mental block here.

[EBK]

Maybe this is my hangup: If we use either one of the jacks as the shield ground connections, through it's sleeve connection, how do we then run a ground from it's sleeve terminal and not have issues?

You don't have "issues" because you only have one path for the ground current for that jack, i.e. the wire from the sleeve terminal to your circuit board ground connection (or wherever you choose to run that to).  The ground current won't flow over the enclosure itself because there isn't another electrical connection to the enclosure it can flow to/from.  If you ground both jacks and don't insulate one from the enclosure, then you would have two electrical connections to the enclosure, and current can flow between them.

As an FYI, i'm a pretty sharp guy...

I hope you didn't read my comment as implying otherwise.

[Big Monk]

Ok. I plan on using Cliff style jacks but with a raised tooth washer and also a 3/8" washer with a solder tab on the input jack.

I hope you didn't read my comment as implying otherwise.

Absolutely not. I was just trying to highlight the fact that this is me genuinely trying to sort this out on my own first rather than just pinging people to do it for me. You've been extremely helpful and I appreciate it.

[Rob Strand]

The wiring is about problem sources and counter-measures.

The problems are:
- External RF sources
- Ground loops on the signal wires  (from amps, pedals, or other external devices).
- Ground loops relating to power (possibly only at high frequencies)
- Ripple currents in the grounds due to unregulated power supplies putting charging filter caps on the effects PCB.
- Currents pulses on the grounds originating from devices on the effects boards (like charge pumps, DC converters, even LFO).
   Remember these currents will appear between the effects board and the DC power.
- RF originating from the pedals digital circuit getting out of the box. 

The worst wiring will work some of the time and the best wiring will work most of the time.  Most good wiring will be close to the best except for a few specific problem scenarios.

The normal way of using the input socket as the star point and grounding the output socket is one good way and is probably the most economical way of getting good performance    RG's is another way and EBK's is also a good way.   All have small cracks in the armor.   Peter Cornish's method is probably the most evolved in the sense of trying to remove the problems.  However a non-isolated input socket might be better.  The output socket not grounded with the mounting bolt might not be as good for RF.   

Peter Cornish's methods is:
- Isolated input socket.
- Input switches input power.
- Input socket wired to chassis with short wire from the input socket ground to grounding lug on the enclosure.
- The input and output cables are shielded inside the enclosure.
   [clarification: the cable shields must connect at both ends.]
- The output socket is isolated.
- Output socket wired to chassis via a capacitor

The reason the input socket is used as the star ground in the normal case and in Peter Cornish's case is because the input socket is used to switch the battery and/or DC power in let.   The requirement for input socket to switch the both battery power and/or DC adaptor power is a constraint.   RG's solution for example doesn't lend itself to input socket switching as undermines some of the good intentions of having the star point on the PCB.

Some builders might fob-off some of the problems or say "I don't do that" but the truth is when a pedal will goes out into the world people will used them without considering any constraints.


EDIT:

Here's some picks of the Cornish method,
https://aionfx.com/news/tracing-journal-cornish-ss-3-soft-sustain/

[Big Monk]

The wiring is a about problem sources and counter-measures.

The problems are:
- External RF sources
- Ground loops on the signal wires  (from amps, pedals, or other external devices).
- Ground loops relating to power (possibly only at high frequencies)
- Ripple currents in the grounds due to unregulated power supplies putting charging filter caps on the effects PCB.
- Currents pulses on the grounds originating from devices on the effects boards (like charge pumps, DC converters, even LFO).
   Remember these currents will appear between the effects board and the DC power.
- RF originating from the pedals digital circuit getting out of the box. 

The worst wiring will work some of the time and the best wiring will work most of the time.  Most good wiring will be close to the best except for a few specific problem scenarios.

The normal way of using the input socket as the star point and grounding the output socket is one good way and is probably the most economical way of getting good performance    RG's is another way and EBK's is also a good way.   All have small cracks in the armor.   Peter Cornish's method is probably the most evolved in the sense of trying to remove the problems.  However a non-isolated input socket might be better.  The output socket not grounded with the mounting bolt might not be as good for RF.   

Peter Cornish's methods is:
- Isolated input socket.
- Input switches input power.
- Input socket wired to chassis with short wire from the input socket ground to grounding lug on the enclosure.
- The input and output cables are shielded inside the enclosure.
- The output socket is isolated.
- Output socket wired to chassis via a capacitor

The reason the input socket is used as the star ground in the normal case and in Peter Cornish's case is because the input socket is used to switch the battery and/or DC power in let.   The requirement for input socket to switch the both battery power and/or DC adaptor power is a constraint.   RG's solution for example doesn't lend itself to input socket switching as undermines some of the good intentions of having the star point on the PCB.

Some builders might fob-off some of the problems or say "I don't do that" but the truth is when a pedal will goes out into the world people will used them without considering any constraints.


EDIT:

Here's some picks of the Cornish method,
https://aionfx.com/news/tracing-journal-cornish-ss-3-soft-sustain/

Thank for this. I will read through it more thoroughly and internalize it but one thing that jumps out at me is the use of the PCB itself as the star ground point. I hadn't really thought of this.

So essentially I'd pick a point for Chassis ground, likely the input jack, and then take all other grounds to the PCB? Obviously in a well thought out manner and not willy nilly.

[Rob Strand]

Thank for this. I will read through it more thoroughly and internalize it but one thing that jumps out at me is the use of the PCB itself as the star ground point. I hadn't really thought of this.

There's motivations behind the normal method and RG's method.  The biggest change in RG's method is the star ground on the PCB.

So essentially I'd pick a point for Chassis ground, likely the input jack, and then take all other grounds to the PCB? Obviously in a well thought out manner and not willy nilly.

Yes that's pretty much it.   Boss pedals have this method nailed.    IIRC, some of the Boss schematics explicitly show the wiring - but it's not as clear as having a pic or actual unit showing where all the wires go.

[Big Monk]

Thank you to everyone who replied. This was one of those things where the pieces were floating around ready to click into place with some constructive advice.

Much appreciated.

[Big Monk]

Thank for this. I will read through it more thoroughly and internalize it but one thing that jumps out at me is the use of the PCB itself as the star ground point. I hadn't really thought of this.

There's motivations behind the normal method and RG's method.  The biggest change in RG's method is the star ground on the PCB.

So essentially I'd pick a point for Chassis ground, likely the input jack, and then take all other grounds to the PCB? Obviously in a well thought out manner and not willy nilly.

Yes that's pretty much it.   Boss pedals have this method nailed.    IIRC, some of the Boss schematics explicitly show the wiring - but it's not as clear as having a pic or actual unit showing where all the wires go.

Just to put the fine point on this: let's say I choose the input jack to make my Chassis ground connection. I would bring a wire from the sleeve to Chassis in some way (many ways to skin this cat) but also bring in input ground (signal ground) from the sleeve to the star point?

[Rob Strand]

Just to put the fine point on this: let's say I choose the input jack to make my Chassis ground connection. I would bring a wire from the sleeve to Chassis in some way (many ways to skin this cat) but also bring in input ground (signal ground) from the sleeve to the star point?

In the simplest case you use the input socket sleeve terminal as the star-point.

A second configuration would be a to use a large ring terminal as the star-point, mounted on the socket itself.     In this case there's no harm wiring the sleeve terminal to the start-point.   The advantage is all the connections are soldered and not passing through the chassis.

A third configuration would be to use a separate grounding lug for the star-point.  It *must* be located as close as possible to the input socket.  In this case the reasoning follows the previous case.

To wire or not is (or should be) a minor point as you want the connections to the chassis to be good (by using star washers).    Also when you consider the output socket, soldering wiring the sleeve of the input socket is a little hypocritical because the output socket often uses the case connection to get the ground.   You probably don't want power running through the ground connection though, so that's a good motivation to solder the input socket sleeve to the star point. You can run a wire wire from the output socket sleeve back to the star point but then the output socket has two ground point which are not close to each other.    To avoid the output socket dilemma you would use an isolated output socket, then you go down the rabbit hole of needing to use shielded cable and have cap to ground like the Cornish method.   

You can debate all the fine points for hours but in practice the difference between the Boss method and the Cornish method is pretty small.     I'm with EBK in that  I'm not enforcing what I think on others.   For a small box like a pedal the differences are small.

I remembered this thread which went through all the same issues,

https://www.diystompboxes.com/smfforum/index.php?topic=123930.0

[Big Monk]

Maybe this is my hangup: If we use either one of the jacks as the shield ground connections, through it's sleeve connection, how do we then run a ground from it's sleeve terminal and not have issues?

You don't have "issues" because you only have one path for the ground current for that jack, i.e. the wire from the sleeve terminal to your circuit board ground connection (or wherever you choose to run that to).  The ground current won't flow over the enclosure itself because there isn't another electrical connection to the enclosure it can flow to/from.  If you ground both jacks and don't insulate one from the enclosure, then you would have two electrical connections to the enclosure, and current can flow between them.

So, if I'm processing this correctly, I can make a single connection from the input jack sleeve to the enclosure (in any suitable way) and also connect the sleeve to the star ground. In essence, one provides my shield ground for the whole circuit while the other provides the signal ground for the input jack.

[Fancy Lime]

Is there a reason to use the input as the star point rather than the output? I would not expect that to make a huge difference but if it does, wouldn't the output be closer to earth (assuming the effects chain ultimately goes to an amplifier with ground=earth)? The guitar end (input) is just floating with no real ground, no? Just curious why most people seem to star the input.

Andy

[Rob Strand]

Is there a reason to use the input as the star point rather than the output? I would not expect that to make a huge difference but if it does, wouldn't the output be closer to earth (assuming the effects chain ultimately goes to an amplifier with ground=earth)? The guitar end (input) is just floating with no real ground, no? Just curious why most people seem to star the input.

It starts from the fact most pedals switch the power using the input socket.   There's a slight advantage going with the input because the input circuits are more sensitive to junk and if you have a gain pedal it will boost that junk.

In many cases you could try hard as you like to create a bad set-up and it still works.  Then you get some crazy cases like the guy with the 1stop power supply issue a couple of weeks back.   He was getting buzz in *bypass* mode with a true bypass setup.  Worked fine of battery but acted weird with the PSU.   He had filters on the PSU rails  and a star ground at input.   A tough one and it was not resolved.

EDIT: FWIW, running the power through input socket isn't actually a good thing but it's what pedals do!

[anotherjim]

You can get jack sockets with isolated switching if you want to keep a power connection away from signal. Up to DPDT. Commonly used in rear panel patch connections on fancier guitar amps. They are nearly always PCB mounting. I'm not certain if the pin size/spacing is right for matrix board. I've got an open frame one that was used to switch resistor networks to run headphones off main speaker amplifiers. It looks quite a contraption.
https://www.tubeampdoctor.com/en/marshall-1/4/6.3mm-jack-pcb-with-dpdt-switch
Not really suitable for panel/case mounting though.

[EBK]

So, if I'm processing this correctly, I can make a single connection from the input jack sleeve to the enclosure (in any suitable way) and also connect the sleeve to the star ground. In essence, one provides my shield ground for the whole circuit while the other provides the signal ground for the input jack.

Yes.