Tube filament wiring question

[R.G.]

Wouldn't a steel (Magnetic) tube be needed to be really effective?

Chuck's right, copper or other high conductivity stuff will work pretty well. It's great for capacitive pickup, and pretty good for magnetic.
There are two mechanisms of shielding, magnetic shunting and conduction/eddy currents. Well, three really, but let's not go into the fields equations. Magnetic materials like iron, nickel, and cobalt work by being several thousand times as easy for an M-field to "conduct" through than free space. Any M-field preferentially jumps into ferromagnetic metals, but some still gets through. It's' like shunting a 10K resistor with a 10 ohm resistor; most of the current gets shunted around the 10K, but some still goes through it. Conductors work by the M-field causing circulating currents in the conductor that make opposing M-fields that try to force the M-fields to stay out. The better the conductor, the better it forces the M-field to stay out.
Iron/steel/etc. are still conductors, but not great ones. IIRC steel is about 6 times more resistive than copper, but that's better than nothing.
There is a fields-theory thing that goes on too. The fields formed on a conductive enclosure force the magnetic and electrical fields to stay in the conducive enclosure walls, not on the inside of the enclosure.This is more effective with higher frequencies, but it happens. A conductive tube isn't a complete enclosure, but it approximates one away from the ends.
For critical uses, like microphone transformer shielding, the makers use alternating steel and copper layers, one inside the other, several layers depending on how much they want to spend on shields.

[Tom Phillips]

...Regarding the 100R CT resistors, they’ve always been soldered to the pilot light. What do you think about using the filaments from the first power tube and soldering them to pin 8 of that tube?

Thanks.

David,
I have experimented with the location of the 100 Ohm resistor virtual center tap for the heater supply. I found no detectable hum reduction performance between different chassis ground locations. Therefore, I think that you should feel free to put the resistors at the most convenient location in your build.
To take it a step further, I find that a humdinger pot, when adjusted correctly, will allow an improvement over the use of the fixed resistors.
Cheers,
Tom

[R.G.]

+1, Tom! No difference, basically undetectable even if you wire it up as wrong as you can.

I declared myself a snow day after the local temps dropped under 0C for an hour or two - we're on the same latitude as Morocco here - and I started writing up some stuff about wiring for low hum in guitar amps. One of the things that fell out of this is that you can just out-and-out calculate hum injected by ground wiring by Ohm's Law.

EDIT: THE STUFF IN THE QUOTE IS WRONG. MY HEAD WAS TOO FAR INTO OTHER WIRING ISSUES.
I really need to have my full ration of coffee before squinting circuits.
Mismatch in the 100-100 string does give you hum, but it's not from current offsets. Rather, it's from the unbalanced voltages around the fake CT junction. Think of Fender's use of a fake CT pot as the divider.

A 100 ohm CT maker string conducts 3.15/100 = 31.5ma through the resistor. Each resistor conducts equal-and-opposite currents to the CT, so the net current from the CT is zero. If the resistors are 5% tolerance, and mismatched the maximum amount, then the combined current mis-match is 10%, or 3.15ma that has to go somewhere from the CT.

#24AWG hookup wire is 26.2mOhm per foot, so a foot of hookup from the CT to signal ground is 0.00315*0.0262 = 0.00008253V. That is, 82.53 microvolts. So even if you do the worst case of grounding the first input grid to one end of the wire and the cathode of the stage to the other end of the wire, you would get 5mV of hum signal at the first plate. That's the worst-worst case for this. More reasonable cases of less mismatch and not-insane wiring would cut that probably by a factor of ten, maybe to 500uV of induced hum at the first plate.

Even better, if the 100-0-100 CT has a single wire going only to the negative of the first filter cap, then the induced hum goes to zero because the current in the CT wire goes to zero - it's only connected to circuit ground on one end, so the heaters themselves wobble up and down by 82.5uV. This is invisible inside the 6.3V of self-wobble on tube filaments.

[Helmholtz]

A 100 ohm CT maker string conducts 3.15/100 = 31.5ma through the resistor. Each resistor conducts equal-and-opposite currents to the CT, so the net current from the CT is zero. If the resistors are 5% tolerance, and mismatched the maximum amount, then the combined current mis-match is 10%, or 3.15ma that has to go somewhere from the CT.

I don't understand what you mean.
The two 100R resistors are wired in series across a 6.3V source.
So their currents must be equal (there's only one common current), no matter what resistance tolerance.
Furthermore the 6.3V winding is floating, so there's no current from the virtual CT to ground (except for a tiny µA capacitive current due to inter-winding capacitance within the PT).

[Tom Phillips]

Following is a method one can use to hear first hand the hum introduced by the AC heater supply.
You can disconnect the heater supply! If you do so after the amp is up and running, you will instantly hear the amount of hum contributed by the heater wire routing and leakage internal to the tubes.
The effect is instantaneous when the heater supply is disconnected and the amp continues to operate for a few seconds because of the time lag for the actual heaters to cool below effective operating temperature. The method I use, with the amp line power OFF, is to disconnect one side of the 6.3 heater winding at the first junction near the power transformer. Then reconnect it with a clip lead that can be easily disconnected while the amp is running with the speaker connected.

This type of experimentation should only be done by individuals who know how to work safely on live high voltage equipment. For extra safety and to prevent problems such as the live end of a clip lead or wire inadvertently contacting something it shouldn't touch, two clip leads can be used to bring the connect/disconnect point outside the chassis. All contact points should be stabilized and protected to prevent shorting or inadvertent disconnection mishaps.

[R.G.]

I don't understand what you mean.
The two 100R resistors are wired in series across a 6.3V source.
So their currents must be equal (there's only one common current), no matter what resistance tolerance.
Furthermore the 6.3V winding is floating, so there's no current from the virtual CT to ground (except for a tiny µA capacitive current due to inter-winding capacitance within the PT).

You're right. I edited my post. It's not the current mismatch that causes hum at all. It's the offset of the heater voltage from balanced around the CT. Thanks for catching that.

[Helmholtz]

It's not the current mismatch that causes hum at all. It's the offset of the heater voltage from balanced around the CT.

Interestingly with heater balance pots (humdingers) I found that lowest heater buzz is not always achieved with perfect heater voltage balance wrt ground.
I assume that heater-cathode leakage can vary from end-to-end of the filament.

But main point is that as the heater circuit doesn't cause/inject a ground current, it doesn't matter where the CT is grounded.

[Tom Phillips]

Interestingly with heater balance pots (humdingers) I found that lowest heater buzz is not always achieved with perfect heater voltage balance wrt ground...

YES! That was a learning experience for me too. I have had some amps where the humdinger was set very far to one side to achieve the lowest hum. The best performance setting is also affected by the exact tubes installed. Therefore, I re-adjust the humdinger whenever any tubes are changed. Perhaps, a re-adjustment would be useful as tubes age. Just my guess so maybe someone can chime in with their experience.

[Helmholtz]

A humdinger is always superior to a heater winding CT (real or artificial).

[Tom Phillips]

For those who would like to read about the physics related to heater induced hum there is interesting information beginning on page 1196 of the Radiotron Designer's Handbook 4th Edition. A free download of the book is available at http://www.tubebooks.org/books/rdh4.pdf .
Due to the rather large index section at the beginning of the book, page 1196 begins on page 1236 of the downloaded PDF.

[R.G.]

My best understanding of the humdinger lowest hum not being exactly in the center is that (1) mechanical center may not be electrical center; these are not precision parts mechanically and (2) the humdinger can trickle in hum in opposing phases, depending on which direction it's offset from center. That means that it cancel some hums from other sources, given that the other hum sources are not shifted far from the AC mains.

All this has been on my mind, as I kind of fell into explaining why star grounding is good, and creeping featurism has expanded that into a general discussion of wiring practice, shielding, hum reduction, and chassis layout, with diagrams. I'm up to about 40 pages now, and thinking about putting in a couple of pages of "just do this" for people who can't take the time to read.

[Davebassman]

Next step. Main board mounted, and started wiring tube sockets. Any advice will be welcome.