Clarification on Pedal Grounding

[Big Monk]

It starts from the fact most pedals switch the power using the input socket.

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

I wonder if switching power on the footswitch would be better. You could easily source a 4PDT stompswitch and dedicate 1/2 to true bypass switching while the other half switches the LED circuit and battery circuit. Or you could simply do what some of the major purveyors of vintage style PNP fuzzes do and put a small switch near the DC jack that allows the user, particularly one with the pedal integrated into a pedalboard, to manual switch the battery connection off.

That way you could just use a mono jack for input and output.

As I am now seeing pretty clearly the way forward here, now I need to visualize and draw up how all the connections route in my enclosure.

[anotherjim]

Power switching with bypass causes big thump noises. There are plenty of small chip amplifiers with soft power-on muting and any muting circuit could be added to an existing design, but we don't want the short silence during changeover either.
Everybody thinks of that idea at some point! 

[EBK]

You could also choose to skip power switching entirely.  I haven't built a pedal with a battery option in many many years, so I build my pedals so they are always powered on when the power supply is present.  Much simpler. 

[Big Monk]

You could also choose to skip power switching entirely.  I haven't built a pedal with a battery option in many many years, so I build my pedals so they are always powered on when the power supply is present.  Much simpler.

Well, I am not a fan of batteries. Even for a PNP circuit, i'd prefer to just use a charge pump voltage inverter to get the required voltage for the PNP circuit as opposed to running a battery.

An easy option, of course, it to simply use a SPST switch to turn off power. I see this all the time on many of the boutique vintage fuzz clones. Doesn't have to be operated when the pedal is engaged and simpyl allows the user to remotely switch the battery off without unplugging anything from the pedal board.

[Big Monk]

Well it's settled: The Tayda "Copper Hammer", pre-drilled enclosures i'm using don't have enough space for a battery so I don't have to worry about power switching!

[Rob Strand]

You could also choose to skip power switching entirely.  I haven't built a pedal with a battery option in many many years, so I build my pedals so they are always powered on when the power supply is present.  Much simpler. 

I've notice pedals which don't have batteries still switch the power.    If you think about it, it's not very useful and seems more like a tradition if anything else.   If you have a string of pedals powered from an adaptor/PSU you aren't going to be playing with the input sockets at all.   Once you get rid of the input socket switching RG's star point on the PCB is very appealing.

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.

I had an Ibanez bass which switched the +rail with on of this *style* of jack, they were common back in the day.  They didn't last long.  I eventually gave up and put in a standard stereo Switchcraft socket (and switched the -rail).    I'm not sure about the style you posted or the complex isolated ones.  IIRC switch craft might have switched types.



Another way to do the switching is to use a transistor/MOSFET to switch  one of the supply rails *on the PCB*.   The switching is done with a low current control signal wired to standard stereo input socket.   That way there's no real power going through the input jack.   A few pedals with charge-pumps and DC/DC converters do this to stop the DC input current pulses going through the input jack and causing noise/whine.

[antonis]

Everybody thinks of that idea at some point! 

Some of us at more than one point..

[duck_arse]

you need to be careful using those ^ style jacks Rob shows ^ above, because the single throw type sometimes is N/C when you want it N/O contacts.

[anotherjim]

Some while ago I'm sure somebody here made a detection switch for the plastic Cliff/Re-an type sockets. The plastic covering the end over the plug tip was removed so a microswitch lever could be operated by the inserted plug. Clever, but I can't remember why it was necessary.

[Big Monk]

By ruling out the use of a battery, and the subsequent removal of power switching from consideration, i'll be able to follow R.G.'s steps, with a few minor exceptions, to the letter. I just need to revise my board design to accept, in a logical and elegant manner, the concept of the star ground being the ground "rail" of the circuit board.

[Rob Strand]

Some while ago I'm sure somebody here made a detection switch for the plastic Cliff/Re-an type sockets. The plastic covering the end over the plug tip was removed so a microswitch lever could be operated by the inserted plug. Clever, but I can't remember why it was necessary.

Cool idea nonetheless.

you need to be careful using those ^ style jacks Rob shows ^ above, because the single throw type sometimes is N/C when you want it N/O contacts.

I got a feeling i saw some switchcraft ones which went the other way.   I think Peavey had some similar looking to what I posted but they had open backs like the Boss connectors.  Those had switches.   I suspect the Peavey ones are a hell of a lot more reliable.   The prong things like my pic are hopeless.

By ruling out the use of a battery, and the subsequent removal of power switching from consideration, i'll be able to follow R.G.'s steps, with a few minor exceptions, to the letter. I just need to revise my board design to accept, in a logical and elegant manner, the concept of the star ground being the ground "rail" of the circuit board.

RG's scheme is what's used in larger scale equipment (equipment where doing silly things like switching power never comes into it).  It's works well.

[Big Monk]

So I've set up my PCB layout and done my enclosure planning based on the following:

1.) Input jack will be mono plastic Cliff style with tooth washer and washer with solder tab. Input sleeve will have a wire to the solder tab washer and wire to ground rail on board.

2.) Shielded wire will run, with shield grounded at input jack to the switch then another pieces of shielded wire from switch to board. Shield has a dedicated spot on board connected to ground rail.

3.) LED will connect in the normal fashion and a ground wire from the 3PDT will go to dedicated pad on ground rail. (This one I'm up in the air about, as LED transients seem to be an issue of concern and grounding at the board is something I'm not sure about).

4.) Non switching DC jack will run to dedicated +/- pads on board, with ground dead side running directly to the ground rail.

5.) Output will have same arrangement as input.

Sound alright? I believe this upholds what has been discussed here and besides some unknowns/misgivings surrounding my connection of the LED, I feel pretty good about it.

[anotherjim]

Reduce LED transients by using high-efficiency/high bright LED's so you can use a large series resistor.

[Big Monk]

I finalized the PCB last night. I laid out the ground rail to accept connections from all the off board mechanical parts and positioned the pads in the logical places where they will make connection. Controls pads are measured to be directly in line the pots, input and output connections are on the side and can accept the shield wire ground at the point of entry, there are ground pads for the input and output jack sleeves and also a pad for the switch ground. Power input has its own ground pads as well and I'll be taking the advice on high powered LEDs and large limiting resistors into account.

I'm using Neutrik unswitched mono jacks with ferrule and shoulder washer and have solder tab washers to make the Chassis connection at the input jack.

Thanks for all the help! I need to put eyes on the layout a few more times with an eye toward ensuring I didn't miss anything and then I think I'll order a limited run.

[Big Monk]

Like a fool I never attached a photo:

[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.

Circling back on this after setting up my "Prototyper" breadboard environment for proper grounding: So even though there is continuity between the signal grounds and the enclosure, ground current won't flow through the enclosure? My concern is I have continuity from the ground rail on my breadboard to the enclosure.

I'm likely missing something elementary here but even as an EE, i've never been afraid to look stupid if it will help people out or help me learn something!

[iainpunk]

Circling back on this after setting up my "Prototyper" breadboard environment for proper grounding: So even though there is continuity between the signal grounds and the enclosure, ground current won't flow through the enclosure? My concern is I have continuity from the ground rail on my breadboard to the enclosure.

I'm likely missing something elementary here but even as an EE, i've never been afraid to look stupid if it will help people out or help me learn something!

its generally a good idea to use a real soldered wire connection to the ground lugs, since you can't trust the jack's connection to the enclosure.
i personally use one grounded jack and the rest are isolated ones, to avoid ground loops.

cheers

[EBK]

Circling back on this after setting up my "Prototyper" breadboard environment for proper grounding: So even though there is continuity between the signal grounds and the enclosure, ground current won't flow through the enclosure? My concern is I have continuity from the ground rail on my breadboard to the enclosure.

I'm likely missing something elementary here but even as an EE, i've never been afraid to look stupid if it will help people out or help me learn something!

I could make this explanation more complicated and complete, but let's think of it this way:

Imagine you have some wires connecting two points in your circuit A,B and you make a single connection to your enclosure at point C in the middle.  We can model your enclosure as if it is just a long wire connected at that one point, like this:



The Xs indicate that there are no additional connections to the enclosure.  Thus, there are no current sources or current sinks anywhere along that enclosure "wire", except through point C.  Therefore, no current can flow to/from anywhere along the modeled enclosure.  You should have continuity between any point on the enclosure and point C because current could flow if you added a source, such as the test current provided by your multimeter probe.  Is that any clearer?

[Big Monk]

Circling back on this after setting up my "Prototyper" breadboard environment for proper grounding: So even though there is continuity between the signal grounds and the enclosure, ground current won't flow through the enclosure? My concern is I have continuity from the ground rail on my breadboard to the enclosure.

I'm likely missing something elementary here but even as an EE, i've never been afraid to look stupid if it will help people out or help me learn something!

I could make this explanation more complicated and complete, but let's think of it this way:

Imagine you have some wires connecting two points in your circuit A,B and you make a single connection to your enclosure at point C in the middle.  We can model your enclosure as if it is just a long wire connected at that one point, like this:



The Xs indicate that there are no additional connections to the enclosure.  Thus, there are no current sources or current sinks anywhere along that enclosure "wire", except through point C.  Therefore, no current can flow to/from anywhere along the modeled enclosure.  You should have continuity between any point on the enclosure and point C because current could flow if you added a source, such as the test current provided by your multimeter probe.  Is that any clearer?

It is!

So in short, there is no issue with the sleeve being connected to the enclosure AND the sleeve going to signal ground on the PCB/Breadboard/Etc. since the enclosure only connects to the sleeve of the input jack and has no other path by which to "receive" (for lack of a better descriptor) current, i.e. a single connection to the enclosure is shunting rather than conducting current.

I don't believe I have any ground related noise anyway, but was mostly just trying to square this idea in my head.

[EBK]

...i.e. a single connection to the enclosure is shunting rather than conducting current.

I prefer to think of it as the enclosure does not "participate" in the basic operations of the circuit.  Shielding is a "bonus" that the enclosure provides in this configuration, which I like to simply enjoy and take for granted.