Ground loops in the context of stompbox construction - please explain...

[luap77]

Hi all,

I would really appreciate an explanation about what potential ground loops exist in stompboxes & how to wire/construct so as to avoid them. I'm confused as I see all sorts of things like:

* Separate earths for input vs output jack vs a common earth for both
* 1/4" Jacks isolated from diecast enclosure with plastic bushes or via use of plastic jacks
* Power DC jack isolated from enclosure
* Pots wired to individual ground pads on board vs a single ground pad and piggybacked tinned copper wire soldered along the backs of the pot
* etc etc

I'm sure that some others would appreciate some clarification on this too.

Thank you!

Paul E

[brett]

Hi
If you search the forum, or go to geofex.com, you'll find good resources.
The DC jack is usually isolated from the enclosure because the convention is for a +ve barrel, and so the box would short to ground if not isolated.

For a quick guide to earthing:
ground currents are very small in low power devices (like stompboxes) ( and unlike power amps)
boxes are usually highly conductive aluminium (not steel, like some amp chassis)
These two factors mean that, WITH LUCK, you can earth as many times in a box as you want, and get virtually no hum.  Assuming that the tracks on PCBs or vero are thick, pots can usually have separate earths without any problems.

However, for some circuits with AC power or oscillators (e.g. 555) or high gain (or all three), problems can arise, in which case a single earth / star-ground will help fix it.

Other people are more expert, and hopefully will add to these rules-of-thumb.

cheers

[GibsonGM]

In my understanding, having multiple grounding points on a PCB or enclosure allows the possibility of a potential differential to exist between them.  The length of conductor between the ground points represents a resistance, and therefore a different current level flowing at each ground point. Star grounding eliminates that.  Certain situations are more sensitive, as Brett was saying....oscillator ckts and the like are sensitive to differences in ground potential.   

A lot of the talk on this subject can seem confusing...a search of the forum will bring up lots more info!  In my experience, settling on a single ground trace on a PCB, where the V- connects (for negative ground ckts), and bringing all my grounds back to this has been reliable...I've done dist. ckts, a phaser, an envelop follower this way with no troubles.  Hope this helps!

[tcobretti]

Ok, here's my question for the experts.  I've worked with dc thus far, and have come up a on a ridiculous conundrum.  In an ac circuit, where do you connect the ground?  I'm experimenting with an AD633, and using a wall wart to power it.  I can't seem to figure it out or find the answer in the archives. 

Thanks in advance.

[R.G.]

Ground in an AC circuit is just like ground in a DC circuit - it's the place that you designate as being "zero volts".

Ground for AC power line circuits has a special meaning, because in most cases it is really, no fooling, the potential of the dirt and rocks under your feet - the ground. Or, as the Brits say, the earth.

What gets confusing is in non-powerline AC circuits. Powering a circuit from a wall wart may mean that neither side of the wall wart output is grounded. Both float, with no reference to any ground point. In AC signal circuits, the connection to ground is explicit. Your guitar, for instance, puts out a grounded AC signal.

So there is a lot of room for interpretation. Can you state your question more clearly. Is it AC power out of the wall wart you're asking about, or the signals in the circuit? Or both?

[luap77]

Thanks to all who have taken the time to answer. I'm still a bit uncertain about things after reading numerous old posts via the search function. A lot of conflicting viewpoints...

Has anybody written a definitive article on grounding in stompboxes? R.G., I noticed in one of you previous posts you mentioned that you might write such an article and post it on GEO site. I for one would be very thankful.

Paul E


P.S. As far as I can gather, the "consensus" is that:

- Currents in stompboxes are small enough to accomodate ground loops without much noise. i.e grounding rules not as strict as tube amps.
- Enclosure should be grounded
- DC jack does not need to be isolated unless it is shorting
- The pots should be wired to ground individually rather than piggybacked
- The input jack should be grounded through the enclosure only
- The output jack should be grounded to the PCB ground trace

Is this fair comment?

[tcobretti]

Specifically in this circuit:



Where do I run ground from the x/y inputs and pins 5,6,and 8 from the chip to?

Thanks again.

[markm]

Not sure how much I beleive in the star grounding deal.
I mean, wouldn't all the leads to the star grounding point then need to be equal in length or one would be at risk of having one grounding point different? 
No malice in this question, it just does not make very much sense to me as I have yet to have a ground-loup problem with any of my builds.
Luck I guess huh? 

[Meanderthal]

 In the context of dc stompboxes, star grounding is killing a fly with a sledgehammer. That's something more appropriate for point-to-point audiophile tube amp circuitry. If ground loops worry you, tree grounding is plenty good enough for what we do 'round here..

[RedHouse]

In the context of dc stompboxes, star grounding is killing a fly with a sledgehammer. That's something more appropriate for point-to-point audiophile tube amp circuitry. If ground loops worry you, tree grounding is plenty good enough for what we do 'round here..

ABSOLUTELY!

Though it's still a good idea to solder a single wire from the input jack ground lug, to the output jack ground lug and to the PCB ground.
(don't rely on mechanical connection of jacks, PCB and enclosure for grounds)

[R.G.]

I'm still a bit uncertain about things after reading numerous old posts via the search function. A lot of conflicting viewpoints...

That's because there are some slippery concepts involved, and because the concepts require discarding some of the simplifying assumptions we've all been taught to make in learning electronics. This makes for a case of a thousand blind men and a big, strange looking elephant.

Has anybody written a definitive article on grounding in stompboxes? R.G., I noticed in one of you previous posts you mentioned that you might write such an article and post it on GEO site. I for one would be very thankful.

It is in the queue. I apologize for the depth of the queue. There's just not enough time.

P.S. As far as I can gather, the "consensus" is that:
- Currents in stompboxes are small enough to accomodate ground loops without much noise. i.e grounding rules not as strict as tube amps.

Thae's trrrue. Stompboxes are more forgiving, with the exceptions I mention later.

- Enclosure should be grounded

This is to prevent RF and hum pickup. It's shielding, not grounding as the rest of this thread discusses it. That's the third way we used grounds.

- DC jack does not need to be isolated unless it is shorting

This depends. Most of the time you get away with it. But in positive ground pedals fed by normal-style plugs, it needs isolated for DC reasons. I guess that's the "unless it's shorting".

- The pots should be wired to ground individually rather than piggybacked

True.

- The input jack should be grounded through the enclosure only

Maybe, maybe not. It depends on how the enclosure gets grounded. Bearing in mind that in stompboxes you can get away with grounding murder a lot of times, the best setup is to let the input jack be what grounds the enclosure, not the enclosure what grounds the input jack.

- The output jack should be grounded to the PCB ground trace

Ideally, one and only of the input or output jack should be insolated from the enclosure and grounded with a wire back to the circuit board.

One ugliness we get a lot is the stereo jack power turn-on trick. We get away with it a lot, but most people put it on the input jack. That means that 100% of the sewer-ground current is traveling down the input ground reference. It's much better to do this on the output jack if you have an option, but most people are too ingrained in thinking it's only the input jack to change this now. A much better approach would be to put a PNP transistor on the effect board which switches the DC power to the effect, and has its base pulled to ground by a resistor to this switching line. The current which runs through the input ground is then much smaller and almost constant.

But mostly we get away with it the wrong way round.

Where do I run ground from the x/y inputs and pins 5,6,and 8 from the chip to?

Ah, that's better.

In this circuit, the power supply provides a "middle ground" which is both a reference and a sewer ground. The grounds from the decoupling caps on 5 and 8 are sewer grounds, and should go to the power supply ground on independent traces. Pin 6 is a signal input, and you want to hold it to a clean 0V, so that's a clean reference ground.

Grounds for the X and Y inputs depend. Strictly speaking neither of the X or Y inputs should be grounded. There should be a resistor of equal value from X+ and X- to reference ground (it's an input) and also from Y+ and Y- to reference ground. For highest accuracy, the + and - should be fed a signal balanced around reference ground. In many cases you can get away with making the + or the - be tied through its resistor or held directly to ground, depending on how much accuracy you need. For quick and dirty work, ground one of each of the X and Y. For accurate work, feed them a signal balanced around reference ground.

Not sure how much I beleive in the star grounding deal.

That's the funny thing about Mother Nature. She doesn't care in the slightest whether you believe in Her or not. 

I mean, wouldn't all the leads to the star grounding point then need to be equal in length or one would be at risk of having one grounding point different?

That's not the point. Well, it is in some high speed pulse and logic circuits. But star grounding works because only the currents for one localized bit of the circuit travel through the ground resistances. That way they can't cross-pollute.

Ground is many things - a clean, clear reference and a sewer for used-up power current among them. Where you get into trouble is using the sewer ground as a clean reference.

In the context of dc stompboxes, star grounding is killing a fly with a sledgehammer. That's something more appropriate for point-to-point audiophile tube amp circuitry. If ground loops worry you, tree grounding is plenty good enough for what we do 'round here..

That's generally true - stompboxes are so small and have such small currents that (a) the area of a ground loop is too small to pick up much radiation and (b) the currents are so small compared to the conducivity of the grounds that you have little trouble.

It is still possible to get into trouble with grounds in a stompbox. Ticking LFOs and LED-pop are two really common ways. You can also get oscillations from improper grounding in high gain, high impedance distortion boxes, especially with FET input devices.

But the small sizes let you get away with ground-murder.

[luap77]

Thanks to all who posted.

Thank you R.G. for taking the time to explain this in detail! Greatly appreciated.

Paul E

[Paul Marossy]

IMO, star grounding is a must for tube amps. It can make a HUGE difference.

[tcobretti]

Thanks for the help RG. 

[seanm]

One ugliness we get a lot is the stereo jack power turn-on trick. We get away with it a lot, but most people put it on the input jack. That means that 100% of the sewer-ground current is traveling down the input ground reference. It's much better to do this on the output jack if you have an option, but most people are too ingrained in thinking it's only the input jack to change this now. A much better approach would be to put a PNP transistor on the effect board which switches the DC power to the effect, and has its base pulled to ground by a resistor to this switching line. The current which runs through the input ground is then much smaller and almost constant.

But mostly we get away with it the wrong way round.

Huh, I never thought of that! I always use the input jack for the "turn on" for no good reason other than everybody else does and it seems to work. The only time I have ever seen the stereo jack on the output is when there are multiple input jacks. I may try switching this around in my next box.