Lifting Ground Loops in a 6G15 Stand Alone Reverb

[tubeswell]

I posted this as a reply in one of the BYO cabinet threads, but since it is really changing the topic of that thread, I'm starting a new thread, so...

here is the guts of it:

1) Disconnect the Mains AC ground wire from its connection directly to the chassis.

2) hook the following parallel circuit between chassis ground and the power line ground wire.

a) A 10 ohm 1/2 Watt resistor
b) a 0.1 uf 400 volt orange drop cap
c) two rectifier diodes like 1N4007 that are wired back to back.

That's 4 items, all in parallel.

The 10 ohm resistor provides the partial ground lift which gets rid of the hum. The cap is the high frequency AC bypass which gets rid of other noise. The 2 diodes clamp the voltage difference between chassis and line ground to one diode drop (0.7 volts), for safety.

The only thing I wonder is will 1A diodes be sufficient? or should they be more robust? The scheme calls for a 3/4 amp mains fuse.

(Also I take it that if I build it with a 240-0-240 secondary going to a full-wave diode rectifier, and a 6V6 instead of a 6K6, the current draw is still going to be about the same? or should I use a 2A mains fuse?)

[R.G.]

1) Disconnect the Mains AC ground wire from its connection directly to the chassis.
...
2) hook the following parallel circuit between chassis ground and the power line ground wire...
The 10 ohm resistor provides the partial ground lift which gets rid of the hum. The cap is the high frequency AC bypass which gets rid of other noise. The 2 diodes clamp the voltage difference between chassis and line ground to one diode drop (0.7 volts), for safety.

The only thing I wonder is will 1A diodes be sufficient? or should they be more robust? The scheme calls for a 3/4 amp mains fuse.

(Also I take it that if I build it with a 240-0-240 secondary going to a full-wave diode rectifier, and a 6V6 instead of a 6K6, the current draw is still going to be about the same? or should I use a 2A mains fuse?)

If I were you,
(a) I would not do this, as it can be quite dangerous in some circumstances
but if I absolutely had to do it,
(b)I would make the diode's current rating be large enough to carry the AC power breaker that it will be plugged into.

In the case of a power fault inside the amp your ... life... depends on the diodes being able to carry enough current to let the full AC power line current go from chassis to the safety ground wire. If there is a fault and the current is high enough to open the solder joints on the diodes, then the chassis becomes fully hot to the AC line, and a ready electrocution hazard. It then depends on where in the AC power wiring the AC power line fuse is. If it's in the hot/line side, then the fuse MAY open in time to save you. If it's in the cold/neutral side, it's impossible for it to open on a line-to-chassis fault.

Although the idea of a "safe" ground lift, as described on Rod Elliot's ESP page is attractive, it carries the risk of electrocution if not implemented perfectly.

It would make far more sense to do the ground isolation on the front end of the amplifier, not on the safety ground of the power supply.

[tubeswell]

If I were you,
(a) I would not do this, as it can be quite dangerous in some circumstances

It would make far more sense to do the ground isolation on the front end of the amplifier, not on the safety ground of the power supply.

Thanks for your feedback R.G.. I take your point about safety very seriously. By way of clarification - I take it by your "far more sense" option, you mean to use a cable from the output socket of the reverb unit going to the guitar amp input that has the shield grounded at one end only?

While I am on this topic, I note that the Fender RI reverb unit has 2 x 6A4 diodes (CR5 and CR6) back-to-back in parallel with a 15R flameproof resistor (R23) on the 'ground-return' side of the HT secondary, yet one of the filter caps (C14) appear to connect straight to ground in front of these diodes. (- see attachment) What function does that serve?

[R.G.]

Thanks for your feedback R.G.. I take your point about safety very seriously. By way of clarification - I take it by your "far more sense" option, you mean to use a cable from the output socket of the reverb unit going to the guitar amp input that has the shield grounded at one end only?

Actually, what I had in mind was (a) using nothing at all unless hum is a really pervasive problem. My experience is that there is a lot to be done inside the amp itself before hunting for external sources that would require a ground lift; and (b) when needed, using an external isolator box to isolate the input from the incoming ground. This works very well, and can be done two ways, by a small transformer or a differential amplifier set to remove the common mode ground noise.

Frankly, there would be almost no issues with input hum on amplifiers if we used one 12AX7 as a differential amplifier at the input instead of half of one as a single ended amp. But there's no way to go back and do almost a century of amp electronics right now. Like it or not, the single ended coax-and-shield with a phone plug on it is how amps take their inputs.

While I am on this topic, I note that the Fender RI reverb unit has 2 x 6A4 diodes (CR5 and CR6) back-to-back in parallel with a 15R flameproof resistor (R23) on the 'ground-return' side of the HT secondary, yet one of the filter caps (C14) appear to connect straight to ground in front of these diodes. (- see attachment) What function does that serve?

Scary as it looks, it may be that Fender was trying to get a ground lift in the same way you proposed. I can't tell without scanning the rest of the schematic.

However - I believe that if that is actually true, then the amp in question will not pass UL, CE, CSA, and probably other electrical safety standards. There is a flat requirement - user accessible metal parts must pass a 25A ground current to the AC wall safety ground plug while only raising ?? X millivolts?? as a result of the current. It may be that it's a USA model. There is no federal law in the USA requiring safety certification. In the EU, Canada, Australia, etc, etc, it is a legal requirement.

By doing a ground lift, one is deliberately UNdoing about a century of learning about how to use AC power lines without dying. The ugly thing about this is that almost all of the time, it works just fine, and even gets past quite a number of fault conditions OK. So it is a great way to begin believing all is well and then forgetting about it.

Until lawyers for the estate of the guy you sold it to sends you a subpoena to appear in court ten years later. Or your late-teens-son or daughter finds the relic in the closet and begins playing.

[Enzo]

Tubes, you are misusing the term ground here.

First many Marshalls use that 10 ohm, two diodes and a cap thing, but not like you think.

Ground means many things. We often use the term when what we really mean is circuit common. And sometimes we use it interchangably with chassis. They are not all the same.

Look at the Fender drawing. The filter caps all go to the power supply common, which is also the circuit common here. We call it ground but it isn;t. They used the symbol wuth the triangle shape. Those diodes are not in the return to common. Look through the circuit for all the parts refernced to that triangle ground. The symbol that looks like a garden rake refers to chassis. The chassis should be grounded - earthed if you like - through the mains connection. Those diodes and resistor are between circuit common and the chassis, NOT between mains earth and chassis as you proposed earlier. They lift the CIRCUIT from the mains ground/chassis. Note also the plastic jacks do not connect common to chassis.

Look at a recent Marshall with the similar network. The circuit common is lefted from the chassis by the diodes, the mains ground connect directly to chassis as it should.

I make little boxes with small trasnsformers in them for ground loop busting. Break the ground path between the signal connections on the two pieces and the grond loop disappears.

[tubeswell]

Tubes, you are misusing the term ground here.

I make little boxes with small trasnsformers in them for ground loop busting. Break the ground path between the signal connections on the two pieces and the grond loop disappears.

Thanks for your inestimable assistance yet again Enzo. Would you be so kind as to let me know if you have a schematic of one of those little boxes circuits handy please? I would be interested in building one.

[tubeswell]

Further to this theme, my thoughts have run back to this over the last 2 months, and today's discussion with Sir Cuitous here:

http://music-electronics-forum.com/s...ad.php?t=10460

- has prompted me to revisit this, and wonder whether an adaptation of the Tore-T mod displayed on the Hoffman site would eliminate/reduce hum.

See here for Doug Hoffman's reference to the origi-mod on the revibe

http://www.el34world.com/Hoffman/ima...maticToreT.gif

I have prepared this (crudely plagiarized yet somewhat original - if you can call any old idea original) schematic kinda figuring how I would go about it.

Basically it is turning the CF stage into a VA stage and padding down the output of that stage before blending it back into the mixer. Will this work to reduce ground loop hum? (or did they only do that Tore-T mod to the revibe because they added the vibrato circuit to it?)? The voltages are guestimates based on hunches.

If anyone has some comment, critique or whatever remark, please shoot me down.

Cheers

[MWJB]

Why not use a XLR style microphone ground lift adaptor between the reverb tank & amp (essentially the same thing that Enzo mentions)? It will solve your ground loop problem and just just takes a few feet of cable, a couple of 1/4" & XLR plugs.

[Steve Conner]

That Tore-T mod won't have any effect on ground loops, since they're caused by external influences, not the amp's own circuit.

I've never found a guitar or hi-fi amp circuit where I couldn't get hum below the noise floor with careful layout and grounding. (And in the case of Champs, a huge choke in the power supply ) If you only remember three things:

Make sure that the ground path between input jack sleeve and AC cord green wire doesn't pass through other circuitry. (For instance connect the green wire straight to the sleeve lug of the input jack, this being in firm electrical contact with the chassis.)

Ditto for the ground path between the power transformer's HT winding center tap and the negative of the first filter cap. Keep this ground wire separate and don't ground other stuff through it.

Either use a humbucking pot on the heater winding, or center-tap it and elevate the center tap with some DC voltage.

If you do these things, any remaining hum is from your guitar or reverb tank

I also agree that an audio transformer is the best way of breaking ground loops. Just stick one in a box with a 9v battery and a JFET or op-amp buffer. (A guitar doesn't have enough juice to drive many audio transformers by itself.)

However, bear in mind that audio transformers themselves can pick up hum. In my recent Crown restoration, I decided to go with transformer balanced XLR inputs, and I got shielding cans for the transformers and spent a while finding the best place for them in the chassis.

Oh and R.G. nobody's stopping you from making an amp with a balanced front end Why not change to XLR guitar leads and 200 ohm pickups while you're at it?

[TD_Madden]

I fought this problem by plugging the unit into an Ebtech HumX......no more ground loop hum.

Not exactly cheap, but works really well.

[tubeswell]

Why not use a XLR style microphone ground lift adaptor between the reverb tank & amp (essentially the same thing that Enzo mentions)? It will solve your ground loop problem and just just takes a few feet of cable, a couple of 1/4" & XLR plugs.

Hi MWJB

Do you (and Enzo) mean a couple of boxes like this? Or is there a simple one-box solution? Is there any ideal spec on the tranny(s)? (BTW What to the two dots designate in the tranny schematic?) - (Actually I just found an article on passive DIs which explains that "the transformer is 1:1 ratio, and needs to be rated for 600 ohm operation (or higher)" http://sound.westhost.com/project35.htm - I guess 1:1 is what the dots mean?

[tubeswell]

Hi Steve

When you said "I also agree that an audio transformer is the best way of breaking ground loops. Just stick one in a box with a 9v battery and a JFET or op-amp buffer" did you mean something like this?

[MWJB]

I didn't use any boxes, just the in-line xlr ground lift adaptors & relevant cable connections, any degredation in tone was virtually unnoticable.

[tubeswell]

I didn't use any boxes, just the in-line xlr ground lift adaptors & relevant cable connections, any degredation in tone was virtually unnoticable.

Ohhhh I (think) I see... Is that like some pint-sized gizmo that is a male/female xlr two-ended extn plug type thingy that contains a switch, which you plug a phono-XLR cable in one end of, and a XLR-phono cable at the other end of (that goes between the reverb output and the amp input)? (I've never seen one of those - no wonder I'm confused).

[Enzo]

1:1 means 1 to 1. The dots on the transformer symbol indicate start end of winding. SO in the box on post #11 to top of both widings is the start, so the top end would be + on both. This is so you can keep track of phase. Is that what you refer to as the dots?

The whole point of busting ground loops is to open the common ground connections. You can break off the ground pin on the power plug and the one won't suffer, but the equipment is no longer grounded and thus unsafe.

It is usually the act of connecting two pieces of gear together with a signal cord that causes the loop and resultant hum. BY inserting a 1:1 transformer in a box in the middle of that cable (electrically) we can transfer any signal without a direct connection. SO we wire it up without a ground from side to side.