Use ground wire on input, output or both?

[moogatroid2000]

My understanding is the input has a ground wire and the output is grounded through the enclosure, if the output had a ground wire then the pedal would have a ground loop.
Is this correct?
Where are the gotchas?

[markr04]

While some people have used the enclosure as a ground, I've never heard of  problems the other way. My grounds all come together on the circuit... ins, outs, battery, you name it. 'Course I paint the inside of the enclosure too, and the paint I use is non-conductive - so there's no ground to be found there even if I wanted to use the chassis as a ground.

[phillip]

I always ground both the jacks to the circuit board for reliability purposes, especially on my commercial pedals.  Hasn't caused any problems as far as I can hear, even in higher gain distortions.

Phillip

[Paul Marossy]

I'll second what phillip said.  

[Dave Z]

But what about when you have a dual-voltage opamp, with a virtual ground....not common I know in stompers but I'm doing a headphone amp, has a virtual grnd at 4.5V

[Paul Marossy]

But what about when you have a dual-voltage opamp, with a virtual ground....not common I know in stompers but I'm doing a headphone amp, has a virtual grnd at 4.5V

I would isolated jacks in that case.

[Mike Burgundy]

or good quality caps that are large enough to let everything you need through.
Virtual ground at 4.5V is very common in stompboxes: it's done with practically all opamp ones.

[Paul Marossy]

Hmm... I thought Dave Z was saying that he had a bipolar power supply.  :? In the case of a bipolar power supply, wouldn't you want to isolate signal ground from the power supply ground or does it matter?

For opamp based 9V single power supply circuits, that voltage divider trick to get 4.5V (virtual ground) is pretty standard.

[Dave Z]

Well, in a bipolar design, your ground is at 4.5v, and the opamp VCC-is at -9V. For instance, pin 4 on a dual opamp like the jrc4558 is at -9V.

In most stompers, the pin4 is at ground, and pin 8 is at 9V....I'm thinking power supply ground should be separate as Paul mentioned, and separate signal ground.

Which brings up the question: for RF rejection do you want to ground the enclosure to power supply grnd?

[niftydog]

Best practice is to wire all grounds to a single point.

But what about when you have a dual-voltage opamp, with a virtual ground

I think you're getting a bit mixed up. A bi-polar op amp circuit has a (real) ground AND it may also have a virtual ground. A virtual ground tied to ground will result in ZERO signal! Virtual grounds have nothing to do with grounding. They are just a concept used in circuit analysis.

Virtual ground at 4.5V is very common in stompboxes

A virtual ground is no longer a virtual ground if it's at 4.5V. I know what you mean, but technically it's not a virtual ground (because it's not at ground potential), it is just a bias point for the op amp.

In the case of a bipolar power supply, wouldn't you want to isolate signal ground from the power supply ground or does it matter?

No, you need that ground. it's the reference for the op amp!

your ground is at 4.5v

Huh!? Wait wait wait wait wait... (maybe my understanding of virtual ground is unique???)

Virtual ground is a term used in bi-polar op amp circuit analysis. It refers to the fact that the non-inverting terminal is tied to ground, and that the inverting terminal always attempts to maintain itself at the same potential as the non-inverting terminal. Hence, the op amp is trying to keep both inputs at ground (zero) potential. The signal has no DC offset and passes through ground.

The concept of virual ground allows us to make approximations and simplified calculations around the op amp.

The common circuit for stompboxes is a biased, single supply op amp. There are NO virtual grounds in a single supply op amp circuit. The non-inverting terminal is biased at half supply, but there is still a ground and it's still at zero potential. The signal is now DC offset above ground and the signal does NOT pass through ground.

Provided you have DC blocking caps in the appropriate places you should be able to ground the enclosure regardless of the op amps bias point.

Ground is ground is ground.

Virtual ground is NOT ground.

Power supply ground should be the reference for signal ground. (in most cases, this means they are simply connected together.)

[Paul Marossy]

Thanks niftydog. I still have a few things to learn about opamps.  :oops:
Actually, I think it would be better if I had said that the 4.5V supply fools the opamp into thinking it has a bipolar power supply, correct?

[niftydog]

that's basically it.

Like I said to Mike - "I know what you mean". As in; I can see how the term "virtual ground" could be useful to describe the 4.5V bias point.

But, the term already has a meaning, and it's quite different!

I use the terms; bias point, DC offset, reference point etc etc.

In fact, you could use the bias point as the circuits ground (and call the real ground -4.5V).

BUT!

The problem with this comes when you attempt to interface your stompbox with another stompbox or an amplifier. Your stompboxes "ground" is not at the other equipments ground.... and when you ground your "ground" you're stuffed!   :?

[brett]

Hi.  Yep, that *Vb* on the schematics is a big hint about what the 4.5V is.

I always ground both the jacks to the circuit board

Maybe I'm fussy, but I use multi-strand speaker wire (light guage) to ground both jacks direct to the *power supply*, which in my pedals is a 2.1mm socket.  I don't get earth loops and I don't want them.    For 2c worth of wire (47c per metre :wink: ), why do it any other way?
cheers