Three wire twist for filaments?

[Chuck H]

I just saw a schematic in another thread that caught my attention. The filament leads were indicated as twisted and there was a third lead twisted with them. This third lead was simply grounded at one end. What does this arrangement do? My best guess is that the grounded lead helps to induct differentials to ground and further reduce hum over the two lead twist. Does this seem like a good idea? Would it work?

[soundguruman]

yes the third wire is creating a capacitor to ground from each of the heater leads.
two wires side by side create a capacitor.
Twisting the wires makes that capacitor more efficient. Twisting the AC out of phase heater wires is cancelling radiated noise...
twisting the third wire creates evenly distributed capacitance, over the entire length of the wire...
It's making a shield. (UK guys: "it's making a screen.")

Bloody clever.

[Chuck H]

I see that, I really don't see how that small capacitance would be useful on a heater circuit. The impedance is low and the current is high. What difference would a few tens of pf's make to 60Hz circuit like that?

I was actually considering inductance as the purpose. My thought being that the extra lead might, via inductance, help to neutralize differentials that aren't canceled by the usual twisted pair.

For that matter, what if you elevated the winding CT, as you might for an elevated DC reference, and take the third lead to the CT rather than ground? That may even work better.?. Like NFB for differentials on either filament lead. Similar to how shunt filters to CT work on output transformers.

[Mick Bailey]

I had an old 60s Japanese amp for repair that had a similar idea. A normal twisted pair for the heaters and a flat-section copper wire coarsely spiralled over the pair and grounded at one end. My view was that it was used to couple EM to ground - similar to running heater wiring close to chassis corners.

[Chuck H]

I suppose that may be the case. The schematic showed a short section of the three leads with an oval around them and below the oval it said "twisted". The grounded lead only extended a little further with no termination while the other two went on to the filaments. So the actual nature of the grounded lead was unclear. I got the most obvious impression that it was three wire leads twisted together, but that may not be the case.

[Merlinb]

I suppose that may be the case. The schematic showed a short section of the three leads with an oval around them and below the oval it said "twisted". The grounded lead only extended a little further with no termination while the other two went on to the filaments. So the actual nature of the grounded lead was unclear. I got the most obvious impression that it was three wire leads twisted together, but that may not be the case.

A complete shield would protect the audio circuit from the electric field of the heater wires. But of course, a third wire is a pretty pathetic excuse for a shield, so you don't get much of this advantage.

Increasing the capacitance between each heater wire and earth will, however, improve the common-mode rejection of the heater:
High frequency hash on the mains supply couples through the transformer via stray capacitance and will appear (to some extent) as common-mode hash on the heater. Capacitance to earth will reduce this. Whether this has a significant effect on the noise in the audio circuit, I don't know.

[Chuck H]

I've always associated heater circuit noise with hum. If there is HF hash there I don't see why that couldn't be amplified just as easily. But considering the impedance of cathode bypass caps at HF it doesn't seem like it would be an issue. I dunno.?. The circuit is probably as Mick describes.

[Mick Bailey]

Would have though just a cap to ground would be as effective. Seems a lot more work to twist an additional wire in or wrap over the heater pair. Maybe it does a little of both - a bit of EMR coupling, a bit of capacitance to ground.

Any hash would need to be removed before it reached the 1st preamp tube, so is this the reason it only goes partway, I wonder?

[Jazz P Bass]

Isn't this the same thing as a shielded cable?

[Merlinb]

Isn't this the same thing as a shielded cable?

It's a question of surface coverage. If the grounded wire is a wide flat copper strip, which covers a decent amount of the surface of the heater wires, or if it is a thin wire that is densely wrapped around the heater wires to get decent coverage, then yes, it provides shielding.

[dai h.]

seems like twisting (and relative orientation to other loops) would take care of hum, and the added wire (which would ordinarily be copper) would help to some degree with buzz (the harmonics of 50/60Hz). From physical schemes I've seen one in a Roccaforte (and some old tube gear--the type of which I forget--I saw described over on ax84 forum some time ago) seemed like it would be effective. Essentially, you make holes for the heater wires in the chassis (or subchassis) and run them with the chassis(or subchassis) metal between them, and their connection to the tube socket is through holes nearby (and the HTR wires exposed to the other tube pin wires are kept very short up to the pin connection). (On another forum) some guys had an issue with long-ish runs on the outside of the chassis being a possible safety hazard (high current winding, seems unfused often). I suppose that is a possible concern.

[pdf64]

The Sound City amps I've seen had their heater wiring runs outside the chassis, just popping through by the tube sockets.
It makes me cringe because the chassis holes used weren't grommited!
To do this with a reasonable degree of safety robustness, the heaters must have a very solid (eg winding center tap) ground reference, otherwise a tube short could blow referencing resistors and pull the heaters up to VB+.
Sound City amps have that and the arrangement was used for heads (wiring behind a panel) and combos (wiring accessible).
Pete

[Chuck H]

I have a design with the first preamp tube located very close to the input jack. All other tubes are toward the rear of the chassis as usual. To get the filament leads to the first tube they exit the chassis at the second preamp tube and re enter right at the first. It was that or run them right under the preamp board. My holes aren't grommeted either but they are large enough that the leads aren't even touching the edges (which are beveled round to break sharp corners). The consideration of a short causing HV on those leads didn't even occur to me at the time. I'm thinking that I'll take my prototype and see how it goes moving those leads inside the chassis under the preamp board. Perhaps through a copper tube or something.

[soundguruman]

I have a design with the first preamp tube located very close to the input jack. All other tubes are toward the rear of the chassis as usual. To get the filament leads to the first tube they exit the chassis at the second preamp tube and re enter right at the first. It was that or run them right under the preamp board. My holes aren't grommeted either but they are large enough that the leads aren't even touching the edges (which are beveled round to break sharp corners). The consideration of a short causing HV on those leads didn't even occur to me at the time. I'm thinking that I'll take my prototype and see how it goes moving those leads inside the chassis under the preamp board. Perhaps through a copper tube or something.

Cool strategy. Yeah, should use bushings or grommets to protect wires. Even on conduit fittings, the ends have wide rounded edges to protect wire.
Flared ends would be good on tubing. Tubing has occurred to me also, but DC is saving me some effort instead.
I don't know if tubing would really get all the noise, but it would probably reduce it quite a ways.

I mean, some of this noise is radiated far and wide, some of it is right at the tube itself, and some of this winds up at the control grid.
It's hard to avoid all of it, even by shielding it; as far as the actual tube the heater is powering --- is also amplifying it's own heater, to a very small degree.

It's the high gain of the amp that creates the necessity.

A tiny heater noise at the input = a big heater noise at the output.

[Chuck H]

There's another trick too. I've seen it on some Mesa products (your favorites ) as well as others. You tack a few inches of unterminated lead to one of the filament leads in the preamp then move it around and trim it until you find a location that phase cancels the most hum/noise. Then tack the lead in place right there. I thought that was pretty clever when I saw it. I wonder if Randall has a patent