Eliminating Ground Loops

[AndrewCE]

My current build is an acitve mixer with 3 inputs and one output. My problem is ground loops.

My pedalboard and Mesa are connected to the same outlet. My Pignose is not. So when I run my guitar into pedalboard, into mixer(in progress), into Pignose, i get a ground loop. When using the Mesa instead, I get no ground loop. Interestingly enough, I do not have the ground loop problem going straight from pedalboard to Pignose, indicating that my mixer is causing the ground loop to happen(mixer is currently on a 9V battery, basically isolated).

How do i fix my mixer design to not give me ground loop problems? Should I implement a ground lift switch, and if so, how to I go about doing that, especially with no transformer in my mixer(i know that's how DI boxes do it)?

[Paul Marossy]

I wouldn't mess with the mixer. I would just run your offending piece(s) of equipment thru a DI box and into the mixer. That should isolate your grounds.

[petemoore]

When using the Mesa instead, I get no ground loop.
  This being connected to ground and the PB also ground, is the only invitation to ground loop I see from your description, other than daisy chained PS [?].
  The guitar, PB and Pignose share the cable ground only, the PB being the only point where the ground is referenced [as in third prong].
  Perhaps I'm misreading, but I don't see the 'invitation' to a ground loop problem with Guit>PB>Pignose, everything is floating except the PB. could it be something else?
  What is the power supply to the pedalboard, judgeing how 'good' it is could be by comparing it with batteries.
 

[R.G.]

How do i fix my mixer design to not give me ground loop problems?

If that's really the problem, it requires knowing the internals of the mixer very well. The design details necessary to un-hum things like mixers are not things you'll find on schematics.

Should I implement a ground lift switch,

No. Ground lift switches constitute one of the two steps necessary to be electrocuted with that kind of gear. The last step happens unpredictably. Is your music worth dying for? 

[AndrewCE]

Paul, well i dont wanna have to buy 2 more DI boxes. Not to mention that that would get me nowhere educationally as far as learning how to avoid the problem.

Pete, ok so maybe it's not actually a ground loop, idk, i'll run some more "tests"...

ok, update: There is significant noise going into both the pignose and the mesa. So for all i know that could just be crappy opamps or design or something. (what else would it be) Also, I disconnected the pedalboard and the noise is still there, with both amps. I'm fairly sure it's not a ground loop now (although i would still like to learn how to correct such a problem, if anybody still feels like answering). The problem is the noise, but that's not what worries me; what worries me is this teakettle sound i'm hearing from my Mesa when my guitars are passing through the mixer. It just keeps rising in pitch and sometimes it does this flangy wah type noise (oscillation, NOT an effect on my input signal) that sounds like the mixer is saying "bwoooooop" or "bwoooowwwoooop" kinda like the Jakovasaurs from Southpark. Maybe this is from capacitors??? idk

R.G, my design is basically inputs into buffers, then mixed with resistors, and buffered at the output. (all buffers are opamps) It's still on breadboard, so all ground connections are straight to the bus strip.
And btw, i'm assuming you are RG Keen of geofex, in which case i'd like to say thank you very much for your "The Technology Of" articles, they've been very helpful and educational. You should consider doing some more.

[petemoore]

(what else would it be)
every time I asked that [thousands of times BTW] I got answers that fall into two categories.
  Something, zalways something...
  And exactly 'this', after finding the problem, it's so much easier to reliably identify.
  The range/number of possible answers before discovering the answer or answers is staggeringly varied/high.
  I would take some voltage readings, verify every ground marking on the schematic matches your board and offboard circuit, then audio probe it [actually I use the thumbuzz technique instead, for 9v circuits at least].
 

[R.G.]

I'm fairly sure it's not a ground loop now

Good call. It's probably not a ground loop given your last description.

(although i would still like to learn how to correct such a problem, if anybody still feels like answering).

Ground loops are easy to correct in the conceptual sense. You just use low-impedance differential signals on shielded, twisted pair wiring between equipment and star-ground the equipment. The interconnecting cord shields are connected only at one end.

Of course, it's impossible to actually do that with the guitar equipment we have. Guitars and amps as we have them are fundamentally NOT differential signals. They're single ended signals that carry return current on the shields so you simply can't open one end of the shield. Worse, guitars as we know them are inherently high impedance signal sources, so you can't get rid of hum and noise that way. It is possible to modify all of your guitars and amps to be low impedance with differential interconnections, but then every piece of equipment is incompatible with everything else.

So the only practical course is to do the best you can. That means all ground connections clean and tight, all AC powered stuff plugged into the same outlet (or outlet strip) and as few and short guitar cords as you can possibly have. Sound bad? It is.

The problem is the noise, but that's not what worries me; what worries me is this teakettle sound i'm hearing from my Mesa when my guitars are passing through the mixer. It just keeps rising in pitch and sometimes it does this flangy wah type noise (oscillation, NOT an effect on my input signal) that sounds like the mixer is saying "bwoooooop" or "bwoooowwwoooop" kinda like the Jakovasaurs from Southpark. Maybe this is from capacitors??? idk

No, it's not from capacitors. If the amps are OK playing one guitar into just one of them at a time, my best guess is that your mixer is oscillating. If you post a schematic, we could look for design issues that would cause that; but grounding and power distribution issues are specific to the physical setup. Speaking of which:

R.G, my design is basically inputs into buffers, then mixed with resistors, and buffered at the output. (all buffers are opamps) It's still on breadboard, so all ground connections are straight to the bus strip.

It is possible that you're having breadboard problems. Breadboards are really handy devices for testing ideas quickly. They're also heartbreakingly prone to intermittent problems when the plug in connections are not tight, or when dust gets into the works. And they wear out. After a number of insertions, the contacts are sprung open and don't hold as well any more. Could be that.

Could also be power supply problems, or chip decoupling problems. The physical layout and connections matter for that.

And btw, i'm assuming you are RG Keen of geofex, in which case i'd like to say thank you very much for your "The Technology Of" articles, they've been very helpful and educational. You should consider doing some more.

Yeah, that's me. I do consider doing more every now and then, but I have a fairly hectic schedule now that I'm working for Visual Sound.

I also have to consider that with the bloom of thousands of new effects gurus selling hand-made pedals based at least partially on information from Geofex, what I put up there has to be things I really can give away freely without conflict of interest with my employment.  I have always taken conflict of interest seriously. That was not much of an issue when I was working for IBM, but it is now. I may be doing more of those, but it will be slower.

[Paul Marossy]

I have always taken conflict of interest seriously.

Yeah, when you ignore those, you can get yourself into some real messes.

[AndrewCE]

update:
with guitars straight into the mixer, the pignose gives no noise. but as soon as i put one guitar through a pedalboard, there is loud buzz. Even after disconnecting the pignose power supply.

i repeated the test with the mesa, and the 2 guitars give me an increased hiss, which could just be attributed to opamps. with one guitar through the pedalboard the result is exactly the same; increased hiss.

so really my only problem is running pedalboard-mixer-pignose (even with pignose on batteries). what could that be? 

r.g., yeah thats why the articles are so good; they take on "the classics"; not modern, secret-of-the-trade technologies. you could do something like the mxr phase 90, a classic compressor, small stone/clone, etc. idk, the point is if you do more articles, it should be about something classic that does not have secretive schematics and that everyone is familiar with. no conflict of interest.

as far as ground loops WITHIN the mixer, or any pedal for that matter, is it better to run all the offboard grounds to a star ground ring, then connect the ring to the board's ground in one place, or to run all offboard grounds to the board's ground and use the bus strip as the "star"?

also concerning general pedal builds(but doesnt have anything to do w/ ground loops), when i bias the front end of an opamp by tying it to Vr(4.5v) via a resistor, what is the best value for the resistor? I understand that too big a resistor will allow the input to stray too far away from the DC reference voltage, but too small a resistor will destroy the AC signal. Is there a theory as to what is the best middle ground?

[.Mike]

r.g., yeah thats why the articles are so good; they take on "the classics"; not modern, secret-of-the-trade technologies. you could do something like the mxr phase 90, a classic compressor, small stone/clone, etc. idk, the point is if you do more articles, it should be about something classic that does not have secretive schematics and that everyone is familiar with. no conflict of interest.

Off topic:

For the classic compressor, RG posted a "Technology of the Orange Squeezer" here back in 2007 that I don't think has made it to his site yet. I learned a lot from it when I built mine (thanks RG!). Here it is:

http://www.diystompboxes.com/smfforum/index.php?topic=54308.msg414842#msg414842



Mike

[AndrewCE]

Off topic:

For the classic compressor, RG posted a "Technology of the Orange Squeezer" here back in 2007 that I don't think has made it to his site yet. I learned a lot from it when I built mine (thanks RG!). Here it is:

http://www.diystompboxes.com/smfforum/index.php?topic=54308.msg414842#msg414842



Mike

thanks, that looks cool. im not personally interested in compressors, but the technology is interesting to study.

[AndrewCE]

here's the schematic as of now, breadboarded.

i don't have the cap values zeroed in on yet; i'll do that when i'm sure i have the right resistors.

questions:
For the inverting amplifier stage, i have the noninverting input going to Vr. I get no sound, or weak sound, when connecting it to ground. Why is this?

For the inverting opamp buffer, should i put a resistor b/t the noninverting and Vr(wikipedia says this eliminates input bias current errors. is this worth it?)

And if i do put a resistor Rz there, does that increase that stage's input impedance to [1M+Rz]?

Also, am i missing anything critical that would maybe reduce noise or make for more reliable operation?

[AndrewCE]

Ok, my mixer is still doing the teakettle/spaceship/jakovasaur noises(with either amp) indicating that it is oscillating. I have posted a schematic. R.G., you said you or someone else may be able to see a problem why it would be doing this?

The noises happen with multiple guitars. With a single guitar, the disconnected jack hums(i'll worry about that later).

[R.G.]

Ok, my mixer is still doing the teakettle/spaceship/jakovasaur noises(with either amp) indicating that it is oscillating. I have posted a schematic. R.G., you said you or someone else may be able to see a problem why it would be doing this?

The noises happen with multiple guitars. With a single guitar, the disconnected jack hums(i'll worry about that later).

Well to start with, "ground" only appears once in your schematic. Is that the only place that's grounded?

[Paul Marossy]

Wouldn't there be floating grounds causing some of the problems? Also, I see a lack of volume control on the three inputs for signal balancing.

If this were for use with three different guitars simultaneously connected, I also see potential problems with them not being switched on/off. The noise picked up from the guitars not being played can get into "the system", unless their volume controls are turned down. If you're playing thru high gain distortion pedals, those unused guitars could definitely start feeding back if the volume controls are left full on, since they appear to always be connected to the circuit.

[petemoore]

  Multiple cables and guitars sharing a common connection looks antenna like.

[AndrewCE]

Well to start with, "ground" only appears once in your schematic. Is that the only place that's grounded?

all the input jack ground lugs are connected to ground and so are the "V-" pins of the TL082's. Is there any other place that should be connected to ground? you say that like it would normally cause a problem.

[AndrewCE]

Wouldn't there be floating grounds causing some of the problems? Also, I see a lack of volume control on the three inputs for signal balancing.

If this were for use with three different guitars simultaneously connected, I also see potential problems with them not being switched on/off. The noise picked up from the guitars not being played can get into "the system", unless their volume controls are turned down. If you're playing thru high gain distortion pedals, those unused guitars could definitely start feeding back if the volume controls are left full on, since they appear to always be connected to the circuit.

You mean floating grounds by way of the fact that the mixer and pignose are on 9V batteries? seems like that would help to eliminate ground loops, why would it make problems. (pedals do a fine job with it) Also it should be noted that ultimately i plan to put a DC jack on this, to run it from my power supply.

And yes, I do plan to put volume controls on each channel, i'm just not there yet. (btw, i'm planning on a variable-gain-pot, or a post-buffer volume instead of a pre-buffer volume, as the latter would decrease input impedance) And possibly a switch, too. 

[Paul Marossy]

Wouldn't there be floating grounds causing some of the problems? Also, I see a lack of volume control on the three inputs for signal balancing.

If this were for use with three different guitars simultaneously connected, I also see potential problems with them not being switched on/off. The noise picked up from the guitars not being played can get into "the system", unless their volume controls are turned down. If you're playing thru high gain distortion pedals, those unused guitars could definitely start feeding back if the volume controls are left full on, since they appear to always be connected to the circuit.

You mean floating grounds by way of the fact that the mixer and pignose are on 9V batteries? seems like that would help to eliminate ground loops, why would it make problems. (pedals do a fine job with it) Also it should be noted that ultimately i plan to put a DC jack on this, to run it from my power supply.

And yes, I do plan to put volume controls on each channel, i'm just not there yet. (btw, i'm planning on a variable-gain-pot, or a post-buffer volume instead of a pre-buffer volume, as the latter would decrease input impedance) And possibly a switch, too. 

Well, looking at your schematic, it's leaving out details like how things are grounded. From your schematic, it looks like the inputs all have floating grounds. If everything is referenced to Vr only, then your artifical ground (Vr) is not the same as your signal ground. Do you see what I'm saying?

[AndrewCE]

Well, looking at your schematic, it's leaving out details like how things are grounded. From your schematic, it looks like the inputs all have floating grounds. If everything is referenced to Vr only, then your artifical ground (Vr) is not the same as your signal ground. Do you see what I'm saying?

yes, i see what you're saying. a couple weeks ago i tried to build a headphone amp and someone on this same site save me advice to connect all of the jack ground lugs to "actual" ground(what you see as "ground" on the schematic) and to use Vr only to bias transistors and opamps. I wasn't sure if he was right or not; the circuit worked both ways. Are you saying that it would be better to connect the jack lugs to Vr instead? What difference would that even make?