6G15 build ground loop hum question

[dai h.]

that was a suggestion for if you tried all the other stuff and it didn't work, but I suppose you could just try it anyway. The idea would be to lift the ground of the output of the reverb with a little bit of resistance. I have no idea of an exact value (I think it should be low) but if you look at circuits doing this such as the reiss. Fender reverb(albeit not signal on this one), DOD 270 A/B box, some other things i've read of, the values are something like 15, 51, 10 ohms--so around there I would guess. One idea is to use a pot(VR wired), start at zero (with the ground loop hum present), then raise the resistance until the hum goes away.

the other thing though, the idea in theory is to hook things up so the redundant ground connection doesn't matter(makes no difference)--just like the stereo equipment example doesn't seem to, even though you make a loop (i.e. make a redundant connection).

the other idea sometimes seen/suggested is to use a big fat wire to connect the two pieces of equipment and (I guess) force them to be at the same potential).

these aren't things I've worked out in practice (yet), but seem correct (or at least a sensible direction) from what I know of the theory (which is not that much)

[Enzo]

And how would that bring the ground potentials closer? And are you going to build a new one each time you connect two pieces of gear together that disagree on chassis potential?

[dai h.]

another easy to think of, should be common, non-esoteric example: a cheap mixer (or mixer section say on a "Portastudio") with a bunch of 1/4 inch IN/OUTs. You hook up an efx. device, efx. out to input. That's ONE(ground connection between the two electronic devices). Coming back to the mixer, (the efx. has a stereo out avail.) you hook up LEFT, and RIGHT OUTs. Total of three ground connections (TWO extra redundancies). Do you typically get objectionable hum and buzz when you do this? I don't think so. So the answer seems to be to create the same situation between the reverb unit and amp. Strictly speaking I think you are supposed to keep the loop area (the loop that will be created from the redundancy i.e. ground loop) small by keeping the wires together, but with the reverb signal return/shield(the cable to the amp IN) and safety ground (in the AC power cord) you wouldn't since putting the AC cords parallel with the signal wire should induce hum and buzz. Also, sometimes in discussions re: this problem you might see suggestions of connecting the AC cords from both devices to the same outlet, and the idea is related, that is the purpose is to keep the potential difference small(in consequential) by making sure the safety grounds are close.

[Enzo]

But he already has the two amps into the same outlet circuit.

A Portastudio has no ground connection, so it can;t make a ground loop. Adding several signal cables is not going to alter the ground potential differences between chassis.

When we have a ground loop, the loop path is common earth ground, up through earth lead on mains cord for item A, through the signal cable ground from item A to item B, and back down the earth lead on item B to common earth ground. You can break the loop by lifting the earth ground on one if the mains powered items, or you can break the ground path through the signal cable between them. The reason you have the loop in the first place is that the natural chassis voltage with respect to earth ground is not always zero. We wish it were, but that will never be always the case. That is because the mains cords do not provide a zero ohm conection to the common earth from each chassis. If one chassis has a tiny 1mv on its chassis and the other also has a tiny 1mv on its chassis, they both agree, and there is zero difference and no ground loop.

Now if one item has a chassis potential of 22mv and the other a potential of 12mv, then a 10mv difference exits between chassis. COnnect the two chassis together with a signal cable and 10mv of AC flows through that signal cable ground. With respect to the signal destination, the signal source has a ground bouncing up and down at 60Hz to the tune of 10mv. The signal coming out of the source device will have its signal sitting atop this 10mv of 60Hz. So the receiving device - the amplifier - will see a signal with 10mv of 60Hz added.

THAT is what we are fighting. BAttery or wall wart powered things don;t have an earth connection so they won;t participate. They may well have ground problems, but those are not ground loops and are not solved the same way.

[dai h.]

the point suggested earlier, to make the circuit ground to chassis connection at the signal input end is to try to ensure that that mV difference cannot occur so that the redudancy won't make a difference. If both devices have circuit ground to chassis connected the same way (suggested) then it looks more likely to work rather than if other higher current return portions of the circuit were connected and going through the chassis(more likely to create the problematic possible potential difference being referred to).

I think a major potential problem is that lots of amps have multiple circuit ground connections through the chassis(filament center tap, B+ center tap, etc. etc.). Making the redundant connection screws up the grounding scheme, so the idea is to have things so it doesn't matter(keep all the heavy currents off the chassis).

[tubeswell]

In all three amps I have a split ground where the pre-amp grounds (including pre-amp filter/decoupling cap ground) are all grounded via separate wires at one of the input sockets, and the output valves' (and in the case of the reverb unit - the pentode valve's) cathodes and grid bias resistors and filter cap grounds etc are grounded at one of the PT bolts.

If I adopted dai h.'s suggestion of star grounding, it might reduce the likelihood of loops.

So what if I had a star ground system with a 100R resistor between the 'star' connecting point and the chassis? Would that stop ground loop noise getting into the audio return path? (and would I have to do that in all the amps for it to work?)

[dai h.]

the easiest to try would be taking the higher current stuff off the PT bolt and instead of the chassis (which is being used now to make the connection to the other grounds) use a wire instead to connect to the other grounds. This has to be done in the correct way off the chassis but connected to the other grouped grounds (and for both reverb and amp).

the way it is now, I see the reverb out signal traveling from the reverb unit out to the amp in, then back through the shield (it's a shield and a return path) BUT since there is (I assume) a safety ground it also sees that path to return back to the reverb out, and since the PT bolt (on the chassis) is between the input jack and the safety ground the signal could be compromised there. Anyway the output signal sees the other path through the chassis, can also go through the PT bolt, then sees safety ground, then to the power plug, then power plug of reverb unit, then up through power cord, to inside the chassis (then if the grounds here are wired the same with the heavier current portions tied to the PT bolt or to the chassis), the same problem of picking up noise(developing an unwanted potential differnce) could happen yet again before traveling back toward the signal source.

so, by keeping the noisier parts of the ground line off the chassis so it doesn't see a common impedance with the parts of the ground line you don't want that return signal to share with, that should reduce the problem.

plugging both into the same AC outlet helps reduce the potential difference, makes the loop smaller. Imagine if they were plugged into two different outlets and the return had to make its way home via a really long physical path. More potential difference, more chance of a problem. There is a bit re: this subject in Dave Funk's book, and he has a suggestion to also power one of the devices from a courtesy AC outlet in the other--helps make the length a bit less.

[andrew_k]

There are a couple of threads about this same problem on the Hoffman forum, in both cases the final solution was to disconnect mains earth in the reverb unit
Here's Doug Hoffman's advice on the matter:

Don't use a 3 plug cord for reverb units.
When you plug your guitar cable from the reverb unit to the amp, the two chassis are connected togaeth and you have a proper ground.

If you plug a in 3 wire power cord and you plug a guitar cable from the reverb unit to theamp, you have two paths to ground and that is a ground loop.

That is assuming that the out jack on the reverb unit and the in jack on the amp are connected to the chassis.
Some people isolate the jacks from the chassis and that will not work.

I'm far too inexperienced to help, but I thought I'd pass on what I'd read elsewhere while researching this circuit

[dai h.]

it'll work to solve the audio problem because you break the loop but creates a safety problem. When the cable is unplugged from the amp, you have no safety ground. The way safety grounds are designed, it does not appear that it was meant to be something to be switched on and off at will. More an integral component whenever a device is connected to an AC source. It's not as if you'll instantly die as soon as you defeat the safety ground of course and lots of people get away with it without harm, but I think of it more like safety belts and lifejackets--that is you appreciate that they are there when something goes wrong--the safety ground is for those occasions.

[andrew_k]

Ah, so mains earth on a reverb unit is like a car airbag that breaks your nose when deployed.

[tubeswell]

the way it is now, I see the reverb out signal traveling from the reverb unit out to the amp in, then back through the shield (it's a shield and a return path) BUT since there is (I assume) a safety ground it also sees that path to return back to the reverb out, and since the PT bolt (on the chassis) is between the input jack and the safety ground the signal could be compromised there. Anyway the output signal sees the other path through the chassis, can also go through the PT bolt, then sees safety ground, then to the power plug, then power plug of reverb unit, then up through power cord, to inside the chassis (then if the grounds here are wired the same with the heavier current portions tied to the PT bolt or to the chassis), the same problem of picking up noise(developing an unwanted potential differnce) could happen yet again before traveling back toward the signal source..

Hi dai h.

I'm fairly confident there's no problem with internal grounding in any of the amps by themselves. The reverb unit is as quiet as a mouse with the 5G9 amp - no noise at all. And both the 5E3 and the 5G9 are quiet on their own account. FWIW the reverb tank shield connection is via the reverb return wire on the chassis only and that RCA socket gets its 'ground' potential from the chassis - not from any other wire inside it. The RCA send socket on the amp chassis is also grounded to the chassis but isolated at the pan (input) end.

But once again thanks for the thoughts. (Hope I didn't miss something). The courtesy mains plug is a possibility maybe.

[Enzo]

Hi Andrew, welcome to the place. When airbags break you r nose, then the airbags needs to be altered instead of removed. Or maybe wear a helmet.


You guys are missing the point. This is not ripple return hum. Ground loops are occurring between pieces of equipment here, NOT from something within the chassis. Dai, you are trying to invoke chassis curents betwen parts of the reverb circuit or something. TW can sit two amps side by side with NO reverb unit, and if he connects them together with a signal cord, a hum will result. SO that eliminates anything in the reverb chassis, since there is no chassis or reverb.

If we could get all chassis at the same potential WITH RESPECT TO THE EARTH, then you;d have quiet. Where does the chassis potential come from? I don;t know. I'd have to say some sort of coupling from the PT. Leakage capacitance to something? In the case of TW, the two amps have different amounts of it, but the reverb has the same as one of them.

[txstrat]

The reverb unit could be fitted with an output jack with a switch that lifts the ground connection. That way the 3 prong cord could do it's job, as long as no cable is plugged in.
With a cable plugged in, the reverb unit would get ground over the amps chassis ground.
Hope I'm not missing anything.

[dai h.]

Hi dai h.

I'm fairly confident there's no problem with internal grounding in any of the amps by themselves.

yes but connecting together changes things in a bad way (because another return path is created due to the presence of the safety ground and higher current returns are in the path since they are on the chassis between the input end and safety ground end--I'm assuming there is a safety ground) and that is what I'm suggesting to change. If one can't see the change in the grounding scheme from one amp by itself to two together then I don't think what I'm saying will be understood.

-the safety ground is not necessary for the electronics to function
a.primary function of the safety ground is for safety
b.secondarily, it can help reduce noise

-presence of the safety ground is creating an audio problem

-the audio problem is due to a difference in potential

-eliminate the difference, eliminate the problem

-hence the suggestion to re-sort the grounds so the only way the input jacks and safety grounds see ea. other is through reference portions of the grounding scheme where there are currents flowing but they are so tiny it (shouldn't matter) even if that redundant connection exists

You guys are missing the point. This is not ripple return hum. Ground loops are occurring between pieces of equipment here, NOT from something within the chassis. Dai, you are trying to invoke chassis curents betwen parts of the reverb circuit or something. TW can sit two amps side by side with NO reverb unit, and if he connects them together with a signal cord, a hum will result.

yes, same problem same suggested solution, that is to move the noisy parts of the ground out of the way so the redundant path by way of the chassis and safety ground can't (or finds it harder to--maybe there are other things such as capacitively coupled noise or something)

If we could get all chassis at the same potential WITH RESPECT TO THE EARTH, then you;d have quiet.

that's what I'm trying to suggest with the re-wiring. IOW try to ensure the potentials are the same by taking the higher current returns off the chassis

[tubeswell]

Okay thanks all you good people. I'm going to take a day (or two) to absorb this.

Cheers :-)