I've seen 3 different recommendations for routing filter grounds to reservoir

[nickb]

Balanced signals and transformers are devices for preventing hum from ground loops. In this situation you don't have any since the input is an ungrounded guitar. I don't see the upside, but I could be missing something (and not for the first time).

An audio transformer, unless it is a very expensive mu-metal shielded type, will probably introduce hum. Also they may not be well suited to high impedance loads due to poor low frequency response ( low magentizing inductance). A better approach would be a low noise active pickup.

[Mike Sulzer]

The input stage only give you about 6dB more power supply rejection that a conventional common cathode stage. Just doubling that stage's filter cap will give you about the same effect for no added complexity. Of greater concern to me is that the noise performance will be significantly negatively impacted due to complete lack of gain in the first stage.

I think the 100K in the plate of the 12AX7 contributes most of the noise, but not its full amount because of the ra of the tube.

[SoulFetish]

Balanced signals and transformers are devices for preventing hum from ground loops. In this situation you don't have any since the input is an ungrounded guitar. I don't see the upside, but I could be missing something (and not for the first time).

An audio transformer, unless it is a very expensive mu-metal shielded type, will probably introduce hum. Also they may not be well suited to high impedance loads due to poor low frequency response ( low magentizing inductance). A better approach would be a low noise active pickup.

There may be no benefit. it was more an exercise in exploring the use of a bipolar supply. Having said that, I wouldn’t presume that the only thing at the input of an amp would be an ungrounded guitar. There are plenty of opportunities for noise coupling and loops to develope using any number of devices between a guitar and amp input.
I started reading a bit of Op Amps for Everyone (5th), and some of the info on the noise performance of modern ICs operating rail to rail on ridiculously low voltages, and using things like pseudo random clocks is pretty AWESOME. But, I still have deep seated, ill founded prejudices against active pickups and hate the idea of using them still. It’s kind of like how I feel about open tunings - Its cheating.

[catalin gramada]

Having said that, I wouldn’t presume that the only thing at the input of an amp would be an ungrounded guitar. There are plenty of opportunities for noise coupling and loops to develope using any number of devices between a guitar and amp input.

Nothing should be ungrounded otherwise huge noise will occur. Is just a matter of choice WHERE you choose to route it. One ground one spot. A guitar pickup or any pickup which bear signal return by screening shield is not the best. Route the guitar signal return by dedicated wire directly at first stage local grounding point and guitar shields at the very first point of the amplifier case through the cable screen just using stereo cable and non isolated jack's
An ungrounded guitar is just a muted guitar ,just ground it by dedicated wire at first stage local ground point. Don.t run returns through the screens. If you have care to run returns in your amp through dedicated bus or whatever wires you have in a star and not through the chassis of the amp it is hardly to understand why to run input return through the shields and cable screen and not through dedicated wire directly at the local ground point...

[Malcolm Irving]

I think there might be an advantage in a balanced (three conductor) connection to a guitar.

An unbalanced screened cable eliminates magnetic interference by the signal conductor and screen being virtually in the ‘same place’ so that the loop area is negligible. Hopefully the electrical things at the guitar end (pickups etc.) don’t open up the loop area too much. Screened cable is used inside the guitar, so we mainly have the pickups to worry about, and we have the possibility of humbucking to help with that.

For interference from electric fields, the same principle applies, the signal conductor and screen are virtually in the same position, with regard to any electric field passing through. Any induced voltage will be the same for both conductors and will be cancelled at the first stage of the amplifier (which amplifies the difference between the two). However, at the guitar end we have the strings and other metal bits of the guitar connected to the screen side only, without any ‘balancing’ conductor connected to the signal side. Electric fields can then induce interference voltages on those metal bits, which are not balanced out.

EDIT: Am I saying that a guitar would buzz less if the strings etc. were not connected to the screen side ????

[Malcolm Irving]

... the signal conductor and screen are virtually in the same position, with regard to any electric field passing through. ....

I suppose it can be argued that an electric field cannot 'pass through' a perfectly conductive screen, so maybe this is where I am going wrong?

[catalin gramada]

EDIT: Am I saying that a guitar would buzz less if the strings etc. were not connected to the screen side ????

Yes like other metallic parts, also to the shields/screen side not to return path side, otherwise annoying pops each time touch/untouch the strings

[Mike Sulzer]

The advantage of running both pickup wires to the amp with guitar metal parts connected to the amp chassis separately through the cable shield is this: For the shielding of a low frequency ac signal to occur, current must flow in the cable shield. (The guitar metal is alternatively positively and negatively charged wrt the chassis ground.) This current causes a small voltage to develop in series with the shield, and if this is also the signal connection, this voltage appears in series with the amp input. However, this is a small effect compared to the magnetic hum sensed by the pickup, even with hum canceled as well as possible with a hum bucking arrangement. So I think it is not worth the trouble to go "two conductor + shield".

[Malcolm Irving]

Thanks Mike. That has helped to clarify it for me.

Some up-market guitar cables have a screen which is connected at the amplifier end only, and a twisted-pair for the signal. This is still unbalanced, as one of the pair (the signal ground) is connected to the cable screen at the amplifier end. This arrangement allows a 'shielding current' through the screen to the amplifier ground (and then to mains ground) without producing any voltage drop on the signal ground conductor. This seems to be a good compromise without needing to go for a fully balanced connection.

[catalin gramada]

You will still get some noise rejection even is not a balanced one. Just try it.

[Mike Sulzer]

Thanks Mike. That has helped to clarify it for me.

Some up-market guitar cables have a screen which is connected at the amplifier end only, and a twisted-pair for the signal. This is still unbalanced, as one of the pair (the signal ground) is connected to the cable screen at the amplifier end. This arrangement allows a 'shielding current' through the screen to the amplifier ground (and then to mains ground) without producing any voltage drop on the signal ground conductor. This seems to be a good compromise without needing to go for a fully balanced connection.

But it really does not do all that it should; the metal parts of the guitar that are connected to the low side of the signal still have to charge + and -, and the current to do this must still flow on the wire that carries low side of the signal. Small effect, but still there.

[Malcolm Irving]

But it really does not do all that it should; the metal parts of the guitar that are connected to the low side of the signal still have to charge + and -, and the current to do this must still flow on the wire that carries low side of the signal. Small effect, but still there.

Yes, I agree, the fancy cable helps the screening of the cable itself, but does nothing for the screening at the guitar.

[catalin gramada]

A screen is a screen and should never have care anything else but noisy EMI garbage induced currents.there are may other usefull options to improve noise rejection even in unbalanced device, used by professionals, like pseudo-balancing, ground compensation and so on ..supposed can use a balanced input



Twisted pairs didn.t help to reject common noise but offer some extra shielding performance