6G15 build ground loop hum question

[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).

bit of a correction: according to Randall Aiken in ye olde thread, the filament center tap (if there is one) has no heavy currents (the other ends which go through the tube heaters do) but should be connected to the "noisy end" because the heater winding (in the power transformer) picks up noise capacitively and by induction from the other windings in the PT

[dai h.]

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

Cheers :-)

-with the two-point grounding the chassis is being used as a part of the grounding scheme (this is a point that needs to be comprehended and realized--so the higher current return portions of the ground line can be removed out of the way of the redundant path due to the presence of the safety ground)

-the circuit should be set up in a way (basically) where the chassis is not being used for returns but it is still connected at one point from the circuit ground (from the "quiet end" i.e. where the very least amount of currents are flowing) to the chassis (and the chassis connects to safety ground since the safety ground should be on the chassis)

-do not just use the input jack. Not because it won't work, but because input jacks can come loose. You DO NOT want the circuit ground to become disconnected from the chassis (and thus the safety ground). Make sure the circuit ground to chassis connection is secure. From what I've seen in production amps, I'd say at least use a Nylock nut(makes it really hard to come loose). I think you should be able to have the input jack direct to chassis, then another redundant connection very close by (side or bottom) without the redundant connection (i.e. since the jack to chassis and circuit ground to chassis make two physical connections to chassis) creating a problem but if you want to be really stringent with having one-point only, then you could always use an insulated jack.

-a different result (possible problems) that can come about from the introduction of the safety ground to amps designed at a time where there wasn't one probably needs to be considered when cloning (appropriate adjustments made)

-another thought is to bring the safety ground all the way towards the input(long extension, to its own point on the chassis--the idea here is still the same, to "unconfuse" the redundant return path). I don't know if it is an acceptable practice though so if it's not then it should not be done.

[Enzo]

Or simply do what we have always done, break the ground through the signal cord with a small transformer.

If your auto seatbelt was uncomfortable to wear, would your advice be to not wear it? Or might it be better to change the way the seatbelt fits?

We NEED the safety features.

All this stuff about heater currents and things are a distraction. You can turn the power switch off on one of the amps and the hum would remain.

[tubeswell]

-with the two-point grounding the chassis is being used as a part of the grounding scheme (this is a point that needs to be comprehended and realized--so the higher current return portions of the ground line can be removed out of the way of the redundant path due to the presence of the safety ground)

I understand the 1st part of the statement (up to the double hypen), but I don't quite get what you are saying about removing high current return portions 'out of the way of the redundant path due to the presence of the safety ground'

-the circuit should be set up in a way (basically) where the chassis is not being used for returns but it is still connected at one point from the circuit ground (from the "quiet end" i.e. where the very least amount of currents are flowing) to the chassis (and the chassis connects to safety ground since the safety ground should be on the chassis)

Gotcha. The only thing I was wondering was - aren't there more currents flowing from the output stage returns (including plate and screen node filter cap grounds) than there are in the preamp stage and input returns, therefore should this not have its own ground point away from the 'lower current' return paths?

-do not just use the input jack. Not because it won't work, but because input jacks can come loose. You DO NOT want the circuit ground to become disconnected from the chassis (and thus the safety ground). Make sure the circuit ground to chassis connection is secure. From what I've seen in production amps, I'd say at least use a Nylock nut(makes it really hard to come loose). I think you should be able to have the input jack direct to chassis, then another redundant connection very close by (side or bottom) without the redundant connection (i.e. since the jack to chassis and circuit ground to chassis make two physical connections to chassis) creating a problem but if you want to be really stringent with having one-point only, then you could always use an insulated jack.

Gotcha

-a different result (possible problems) that can come about from the introduction of the safety ground to amps designed at a time where there wasn't one probably needs to be considered when cloning (appropriate adjustments made)

Gotcha

-another thought is to bring the safety ground all the way towards the input(long extension, to its own point on the chassis--the idea here is still the same, to "unconfuse" the redundant return path). I don't know if it is an acceptable practice though so if it's not then it should not be done.

Take the safety ground off the PT bolt and run a separate wire through the amp and bolt it to somewhere near the point where the signal ground is connected?

[tubeswell]

All this stuff about heater currents and things are a distraction.

Yu're darn tootin'

(Nevertheless fun to chew the fat over)

[dai h.]

I understand the 1st part of the statement (up to the double hypen), but I don't quite get what you are saying about removing high current return portions 'out of the way of the redundant path due to the presence of the safety ground'

okay, I've tried to draw out what I'm thinking so hopefully it makes more sense:

http://www.flickr.com/photos/7878340@N05/3253570670/

http://www.flickr.com/photos/7878340@N05/3252782853/

Gotcha. The only thing I was wondering was - aren't there more currents flowing from the output stage returns (including plate and screen node filter cap grounds) than there are in the preamp stage and input returns, therefore should this not have its own ground point away from the 'lower current' return paths?

yes but the point though is to not use the chassis to make the connection

Take the safety ground off the PT bolt and run a separate wire through the amp and bolt it to somewhere near the point where the signal ground is connected?

I wouldn't do this unless it's confirmed to be a safe practice first. But that might be easier if it is okay.

[dai h.]

Or simply do what we have always done, break the ground through the signal cord with a small transformer.

If your auto seatbelt was uncomfortable to wear, would your advice be to not wear it? Or might it be better to change the way the seatbelt fits?

We NEED the safety features.

All this stuff about heater currents and things are a distraction. You can turn the power switch off on one of the amps and the hum would remain.

1. yes transformers can break ground loops but if you can work things out without one, why not try?

2.I don't think I've suggested defeating the safety feature. More trying to satisfy both.

lastly since this isn't something I've actually worked out in practice, thar' ain't no guarontee hoss!

however, according to steve c. and R.G.'s posts plus in addition the bit in Dave Funk's book, it does sound like it can, so from there the attempt to find and illustrate the logic and see where (in tubeswell's reverb to amp) the practice is conflicting with the theory (and hopefully) get them to be happy together audio-wise while at the same time not compromising safety.

[tubeswell]

Thanks for those sketches dai - Now I see what you mean. (I really appreciate the effort you have gone to in order to produce the artwork)

[Enzo]

Dai, I may be mistaken, but I think you are trying to solve internal grounding hum issues when what TW has is a real ground loop. People around here are referring to ground problems within their amps as ground loops when they are not loops. If an amp sits there making hum by itself it may have grounding issues inside, but that isn;t a loop. A ground loop ocurs between two pieces of gear. SOlving problems like having tube heaters current running through the chassis along with a signal is not the same issue as the overall chassis potential compared to the earth.

In your second drawing you shor return to earth from one and return to the second chassis . That is not what you want to equalize. There may be some potential between the earth and the chassis at return 2. That means the chassis compared to earth is Xmv hot. Making the difference along return 1 the same means there is that same Xmv flowing between the units. Ideally you need to measure a "return 3" that I am adding. That is between earth and the chassis on the left. When return 3 equals return 2, then there will no longer be any difference along return 1. It is any difference between ends of return 1 that cause ground loops in the first place.

Remember when he had two amps side by side, simply connecting the input jack grounds together made a ground loop hum. A test here is illustrative: He plugs the amp and reverb into the same outlets, and when he connects them with a signal cord, they hum. This is his classic ground loop. Chancces are he can then turn the power switch off on the reverb and it will still hum. With the power switch off, all those internal currents are gone.

[tubeswell]

Chances are he can then turn the power switch off on the reverb and it will still hum. With the power switch off, all those internal currents are gone.

Yes that is indeed what happens. The 5E3 amp still hums even if the reverb unit is switched off (but still plugged in to the guitar amp). The hum only goes away if I switch the 5E3 amp off (or unplug the signal lead from the reverb unit to the amp).

But if the hum is caused by voltage differential between the two chassis that (in turn) causes current to flow between one chassis and the other, then could I not stop that current flowing between the two chassis with resistance in the path of the signal cable shield? (maybe with a pot wired as mixer/divider? - with the case of the buffer box grounded via its own mains ground wire to the same point as the guitar amp and the reverb unit were grounded) Can AC be impeded (or balanced?) like that?

[Enzo]

Look at your guitar cord, and consider the tiny currents the pickup sends down it. How much do you think a 47 ohm or 470 ohm or 4700 ohm resistor in series with it would affect the amount of signal the amp gets? The 68k series resistor at the amp input doesn;t knock it down much. In short, it will have no effect.

By still having the ground loop hum even with the reverb turned off, we can see that the internal grounding arrangements are not the deal. It is more about minute leakage in the chassis wiring and induced voltage from mains wire placements.

Remember, the ground loop is a relationship between the two units, ther is nothing wrong with either one of them. They disagree on chassis potential, you have three pieces of gear (in the discussion), and two of them have the same chassis potential and one does not. How do we determine which one is "wrong?" What if we got five more amps and three more reverbs and discovered that the odd unit out now actualy agreed with more of them than the other two?

There is an experiment for you: your two amps disagree and if conected together they hum. You can turn the one off and it still hums, what happens if you turn the other one off instead? I bet it still hums out the one left on. Obviously that won;t work on the reverb-only since it has no speaker. But on the two amps?

As long as we have a difference in potential, it doesn't really matter which thing is which, it is the difference that matters.

The hum only goes away if I switch the 5E3 amp off (or unplug the signal lead from the reverb unit to the amp).

Or pulled the mains plug for the reverb out of the wall socket or lifted its ground. Remember there has to be a loop for a ground loop. The loop is up one earth lead, across the signal cable ground, and back down the other earth lead. Break any of those legs and the loop disappears.

Get a wireless guitar setup, and plug the belt pack into the reverb out and the receiver into the amp input. No more loop.

[Sir Cuitous]

Or try something like this:

http://pro-audio.musiciansfriend.com...or-?sku=180014

It's probably just a buffered circuit with an isolation transformer.

[Enzo]

I doubt it is buffered, it is passive. It is just a couple transformers in a box. and it is exactly what we would use to solve this problem when it arises in real life. We discussed transfomers earlier in this matter. I am hoping the light bulb of OH I get it comes on so we can all understand WHY the problem happens.

This gadget - I myself build something similar - breaks ground loops and stops ground loop hum, but it won;t cure hum from other sources such as heaters.

[txstrat]

If you lift all the circuit grounds on the 6g15 off the chassis except on the output jack. Then fit the output jack with a switch that lifts ground as soon as a guitar cord is plugged in, wouldn't the ground from the 5e3 be the ground for the 6g15 then? And therefore stop the loop?

[dai h.]

Remember when he had two amps side by side, simply connecting the input jack grounds together made a ground loop hum. A test here is illustrative: He plugs the amp and reverb into the same outlets, and when he connects them with a signal cord, they hum. This is his classic ground loop. Chancces are he can then turn the power switch off on the reverb and it will still hum. With the power switch off, all those internal currents are gone.

mmm... I don't think so. The internal currents are still in the amp since the chassis is being used for parts of the higher current returns. That's where(why) I'm thinking a difference is created and why I'm saying try the re-wire so the chassis is basically just a shield with only the tiniest of currents flowing (at least DC wise--maybe there is some other stuff from capacitive coupling or induction possibly). Maybe I'm not explaining well enough but I'm not really sure if my logic is being understood.