Power transformer noise in tube circuits

[beatnik]

Is it safe to assume that the noise radiated from tube equipment power transformers is mainly from the heater winding since it normally has much higher current ratings compared to the HT ?

For tubes with series / parallel heater option like the 12A_7 series does it make sense to use the 12.6V option whenever possible since it has half the current requirement ?

I guess the power transformer should radiate less noise with a lower current and also be of smaller size but I am wondering if the 6.3V option has an advantage in terms of less noise coupling into audio because of the lower voltage ?

Probably the noise creeping in the HT from the heater winding should be taken in consideration as well, is that in some way related to the voltage of the heater winding ?

 

[SteveF]

The magnetic field strength is directly proportional to current so from that standpoint you would think less noise. However, the 12.6 volt winding would have twice as many turns as the 6.3 volt winding, so the resulting field strength (radiated noise) would be the same. Voltage itself never comes into play. At least that's how it looks to me.
I can see where using the lower current option (12.6V) would be useful outside the transformer. I've seem many circuit boards burned up by manufacturers trying to run 10 amps of heater current through those little traces and cheap connectors.

 

[CJ]

There are also electrical fields to worry about. The higher the voltage, the more problems from the e field. That is why they put a copper belt around most power transformers.

I have an amp on the bench that was humming. The heater wires were not twisted but laying flat on the chassis at opposite ends of the tube sockets. I rewired the heaters with elevated twisted pairs and the hum stayed the same.

The power transformer is like a Marshall. Lay down type with no end bell on the inside of the chassis. There is no copper belt around the lams and coil. No real way to add one so I am stuck with the noise unless I change out the XFMR and I do not want to do that as it is a rare vintage amp, a Premier Twin 12 made by Multivox.

Transformer shielding in low level audio transformers sometimes consists of nested cans, alternating between copper and mumetal. The copper shields the electrical fields but not the mag fields, the mumetal shields the mag fields with it's hi nickel content but does not shield the electrical fields. So the combination of copper and mumetal does a good job on both fields. Sometimes these shields will consist of 3 cans of each.

 

[abbey road d enfer]

The magnetic field strength is directly proportional to current

That is true, however...
Stray field from a power xfmr is almost constant whatever the actual power it delivers, as long as its limits are not exceeded.
It's the core magnetizing flux that counts.
The only way to decrease the magnetizing flux is to use lower nominal flux. Xfmr winders tend to optimize their use of iron by running at a flux of 1.5-1.9 Tesla. The higher the flux, the higher the nominal power.
When I had the luxury of having custom-built xfmrs, I specifically asked the winders to use a core a couple of notches above what they would have naturally chosen, which reduced significantly flux (down to about 0.8 to 1.0T), which resulted in stray magnetic fields about one order of magnitude smaller.
Of course it has a cost and takes more space, but it's worth doing. Of course, easier for low-power devices (FX, comp, EQ...) than for power amps. A few boutique amplifier manufacturers have gone this route, ad produce amps that weigh like a boat anchor.
When the magnetic material approaches saturation, the magnetic permittivity decreases significantly (look at the slope of the magnetic cycle that becomes flatter and flatter). It means it loses its capacity at concentrating flux lines, so stray field increases , and particularly on the peaks of he waveform, so not only it increases, but it also distorts, which makes it more audible.

 

[abbey road d enfer]

Transformer shielding in low level audio transformers sometimes consists of nested cans, alternating between copper and mumetal. The copper shields the electrical fields but not the mag fields,

That is 100% correct, however...

the mumetal shields the mag fields with it's hi nickel content but does not shield the electrical fields

The NiFe alloy actually also perfectly shields electrostatic field. It does not take much conductivity to provide electrostatic shielding; See how some cables are shielded with conductive plastic or guitar cavities with carbon-loaded paint.
What the copper does, it creates a shorting ring around the xfmr, and since copper is a much better conductor than NiFe, it's more efficient at that.
Shorting ring decreases stray flux by creating a current that opposes to the excitation flux (Lenz's law). This results in heat dissipation, but is often negligible.

. So the combination of copper and mumetal does a good job on both fields. Sometimes these shields will consist of 3 cans of each.

Indeed. As I mentioned in another thread, sandwiching 2 magnetic shields of thickness x is better than only one of thickness 2x.