isolate the ground?

[joesatch]

What happens if i isolate the ground from the chassis on the PT secondaries and forward? The only ground touching the chassis will be the green AC ground wire. The rest isolated from the chassis on separate ground planes.  The rectifier ground, the filter cap grounds, the preamp grounds , none of these grounds touching the chassis.  Okay? i ask this because everytime i build an amp i have to ground lift the signal to eliminate hum.

[pdf64]

Yes, see section 15.10 http://www.valvewizard.co.uk/Grounding.pdf

[joesatch]

Can i run a ground scheme like this? will this be an issue?  My goal is to isolate the ground of the amp from the chassis. The only ground touching the chassis will be the green wire from 3 prong AC plug. The rectifier will be grounded to this isolated ground buss wire as well. I know the valve wizard shows a "Hum-loop block network" but i have trouble understanding the diagram.

[PRR]

Try it. It worked for Western Electric and other sophisticated workers. Especially in totally-balanced systems. For guitar, you probabaly DO want chassis to common at one place. If you truly isolate and have buzz that varies with direction and proximity, jumper-clip chassis to common at the input jack (usually) and see what changes.

[joesatch]

what is this line? a solid jumper?

[PRR]

Oddly, yes.

He wants two BIG diodes series, and both ways. That connection on a 4-diode bridge rectifier is equivalent, and easier.

[astronomicum]

Without one connection to ground, would there not be a safety issue if voltage appears on the 0V side of the supply, and consequently, to the guitar strings? I would think at least the input jacks should be grounded.

[mresistor]

there is a ground on the bottom of the bridge .

[astronomicum]

How does that ground get the earth if not through the amp?

[joesatch]

the rectifier in the diagram is not the rectifier for the amp HT. That's another rectifier. Where would i ground the rectifier for the HT? to the ground buss or chassis?

[joesatch]

Without one connection to ground, would there not be a safety issue if voltage appears on the 0V side of the supply, and consequently, to the guitar strings? I would think at least the input jacks should be grounded.

input jack would be grounded to the ground buss but isolated from the chassis.  I have come to realize almost every amp that grounds the circuit to the chassis and connects to outboard gear is going to suffer ground loop noise. I want to eliminate this by lifting the ground from the chassis

[astronomicum]

How does that ground get to earth if not through the amp?

Sorry. I was asking about the ground on the bridge of the guitar. I should have quoted to be clear.

The bridge would ground to the 0V rail of the amp but would not reference earth. If the 0V rail does not reference earth, it can have a potential other than earth, and the strings will follow. I am calling it 0V because it is no longer referenced to earth (ground). What would happen if one on the main power supply caps or diodes shorted? Would the 0V rail still be 0V referenced to earth? And if I am touching the strings and I touch something referenced to earth? If this is not the case, what am I missing?

[tdvt]

And if I am touching the strings and I touch something referenced to earth? If this is not the case, what am I missing?

Interesting question. I would think that for practical purposes, the upstream-side a 10R resistor is essentially at ground. Merlin mentions a "compromise" to mitigate hum, not necessarily eliminate it.

The text that accompanies that hum-blocker circuit from the Valve Wizard:

"Ground Lift"
"To combat the problem of creating a ground loop via mains earth when
connecting two devices together, some appliances will offer a ground-lift switch.
This disconnects the audio ground from the chassis, breaking the loop. However, the
connection between chassis and mains earth is NOT broken. The chassis must
remains earthed at all times!

Completely breaking the connection between audio ground and mains earth still has
some safety implications, so a compromise is to use a hum-loop block network.
instead. This consists of a 10Ω resistor in parallel with a 100nF to 470nF capacitor.
 
When a ground loop is created it will now have a 10Ω resistance in series with it, which should reduce power-supply ground-loop currents to negligible levels. The capacitor ensures audio ground still appears to be connected directly to chassis as far as high frequencies are concerned,
so we still have good shielding against radio interference.

The resistor should be a power device, say 5W, so that it can withstand fault
currents. A pair of high-current diodes should also be connected in anti-parallel to
bypass more serious fault currents, thereby ensuring any fuses will reliably blow.
 
A>6 amp bridge rectifier package is quite convenient for this, as shown in fig. 15.16
(it does not need to be a high-voltage rectifier). A silicon diode will not turn on until
the voltage across it reaches about 0.6V, so the ground loop still appears to be
blocked as far as small hum voltages are concerned. In low-level preamps, mixers,
and other noise-sensitive equipment, a switch may also be included as shown in the
figure, but this is not necessary in a power amp."


addendum:
I have implemented the "hum-blocker" circuit in a few builds with very good results. ALL of the amp grounds are isolated (from the chassis) with just one "earth" connection to chassis, near the incoming AC safety ground.

[PRR]

...What would happen if one on the main power supply caps or diodes shorted?...

That current would not flow back to dirt, but to the power supply.

In hi-gain audio, you really want your circuit to be near the average electric potential of your body, and of the room, for least induced hum. With unbalanced systems and hum-adverse users, it is almost essential. Before 3-pin outlets, we would flip the 2-pin plug for least hum, hoping one side was cleaner than the other.

[joesatch]

additional question.  the 6L6's cathode to ground and the bias ground. Both these connected to the same isolated ground buss?

[pdf64]

Yes.
Pretend the chassis is made of wood.
That forces you to - shock horror - use wire to provide a path for 0V common.

[pdf64]

... What would happen if one on the main power supply caps or diodes shorted? Would the 0V rail still be 0V referenced to earth? And if I am touching the strings and I touch something referenced to earth? If this is not the case, what am I missing?

The worst case is that there’s about 1.2V DC difference between the amp’s common and its chassis and mains earth.
That shouldn’t last for long as a fuse would blow.
1.2V isn’t an electrical hazard, eg we don’t feel a tingle from 1.5V batteries.

[astronomicum]

... What would happen if one on the main power supply caps or diodes shorted? Would the 0V rail still be 0V referenced to earth? And if I am touching the strings and I touch something referenced to earth? If this is not the case, what am I missing?

The worst case is that there’s about 1.2V DC difference between the amp’s common and its chassis and mains earth.
That shouldn’t last for long as a fuse would blow.
1.2V isn’t an electrical hazard, eg we don’t feel a tingle from 1.5V batteries.

Thanks. Just to be clear, this is worst case using the Hum Loop Blocking Network?

[dwinstonwood]

I have converted some of the amps I've built to this grounding scheme.
Power cord green wire bolted to chassis; Heater CT connected to a cathode; and one single ground point at the input jack for everything else.
So, a total of only two connections to the chassis.
It's proven to be a very quiet method for me.

[JPK]

I have converted some of the amps I've built to this grounding scheme.
Power cord green wire bolted to chassis; Heater CT connected to a cathode; and one single ground point at the input jack for everything else.
So, a total of only two connections to the chassis.
It's proven to be a very quiet method for me.

Are your pot's isolated from the chassis?

[dwinstonwood]

Are your pot's isolated from the chassis?

JPK, no, I just attach them right onto the chassis. I've never bothered to isolate them.

[joesatch]

the pot case touches the chassis but the pot lug that normally gets soldered to the back of the pot instead goes to the isolated ground

[HotBluePlates]

Without one connection to ground, would there not be a safety issue if voltage appears on the 0V side of the supply, and consequently, to the guitar strings? I would think at least the input jacks should be grounded.

How does that ground get the earth if not through the amp?

IMO, the real reason to ground the strings is not for safety.  It's for nixing hum/buzz by applying the buzz your body picks up out of the air & adding that to the "0v signal reference" and thereby canceling that same buzz elsewhere in the circuit.  Which is why your guitar might buzz until you touch the strings (and that string-ground).

One could build a guitar amp in a plastic box, and never have "ground" anywhere.  No 3-prong cord either, because there's no conductor (chassis) for positive voltage to touch that the user can also touch.


Such a "plastic-box amp" would still need a "reference/return circuit" to work.  We might call that Reference "0v" and measure other voltages in regards to its voltage (whether truly 0v compared to dirt, or not).  This would be no different than a plastic-bodied flashlight.

[Williamblake]

I am very happy with shield and ground separated on the guitar. Neil does it.

[pdf64]

I am very happy with shield and ground separated on the guitar. Neil does it.

How does that work, eg stereo jack socket?

[Williamblake]

Yes stereo socket. Shield for guitar bridge etc and twisted pair for signal and ground. Guitar cable shield goes to chassis and my amps and reverb units are grounded at he input. I don`t play venues with lot`s of lights and have humbuckers, although i do split them very often. Just mentioning this because somebody suggested a symmetric input. Don't know where my wiring sits on the scale of diminishing returns but i am happy with it.