Question about toroidal power transformers

[Pedro Vecino]

I have made three prototypes of different amps. All are completely silent without the guitar connected even at maximum volume. None has added shielding in the lower part. But one of them (the third, with 6V6) is very prone to pick up noise. The typical produced by single pickups very close to the amp. All use different toroidal power transformers.

Could it come from an inappropriate transformer orientation? Mounting it with inverted top and bottom can affect it? Or could it just come from horizontal turn? Can these possibilities exist? My experience with toroidal transformers is minimal.

Thanks.

[R.G.]

It's unlikely that it comes from mounting with top/bottom reversed or from horizontal rotation. Toroids work by making the magnetic field rotate back and forth around the circle of the toroid. Leakage is minimized by having the iron be nearly full circles, with air gaps only at the boundaries of the (typically) wound strip of iron making up the core. This means that the air gap is very widely distributed, not at abrupt boundaries like the ends of the E's in EI transformers, so leakage is generally less and much less orientation sensitive.

I can imagine a poorly built toroid with primary or secondary or both improperly clustered on one side of the circle as opposed to distributed all around the circle that would have worse leakage than a correctly built one, but that would be very unusual and probably hand-made, not machine wound.

There's a good way to test. The amps ought to work, if not be ideal, if you swap the transformers around. Or you could do the old test for leakage we used to do in power supply design - hook a circular loop of wire to a 'scope probe and use it as a mag field sniffer.

[Helmholtz]

Could it come from an inappropriate transformer orientation? Mounting it with inverted top and bottom can affect it? Or could it just come from horizontal turn?

Because of the toroidal symmetry of the PT, I don't expect any improvement by rotation or flpping it over.

Typically toroidals have a very low leakage field. But leakage will increase if the core is operated close to its magnetization limit, caused by poor transformer design.
Another possibility could be a DC component on the secondary as would result e.g. from half-wave rectification.

[R.G.]

Transformer action does not transform DC, as the voltage on the secondary is proportional to the rate of change of the magnetic field, not the actual value of the magnetic field. So it is not possible to saturate a transformer by a DC component in the secondary.

You can only saturate a transformer by putting too large an AC voltage into the primary, or providing a DC offset in the primary.

Up until recently, I would have said (and did) that it is impossible to saturate a transformer from the secondary by any kind of rectification.

I have recanted that based on the following argument: it is possible in rare circumstances to get core walking on the primary by half wave rectification by the fact that half wave rectification causes resistive losses in the primary in only one of the half cycles. This makes the primary AC voltage smaller on the half-cycle where the rectifier conducts, causing a DC offset on the primary.

It's difficult to make this happen. It takes high currents through a high resistance primary, so it's most likely to occur in very small transformers. I've never seen it happen, and I've been working with transformers of all kinds, including with half-wave rectification and small transformers, since the 1970s. But I will admit a possibility.

I think it's unlikely to result in this case. It's also not a reason to assume that all half wave rectifiers saturate the core. I see on the web the admonition that half wave rectification saturates cores much more frequently than I see half wave rectifiers. We should follow the advice that doctors give one another: when you hear hoofbeats, think horses, not zebras.

[Helmholtz]

Well, of course I never said or meant that transformers can transform DC. I didn't even mention saturation either.

But using half-wave rectification on the secondary results in a net DC secondary current which produces a DC pre-magnetization of the core, resulting in reduced effective core µ and primary L. This causes the primary magnetization current to increase. In sum these effects are likely to increase the stray field of the transformer - especially with "economical" PTs that already operate close to the saturation "edge" (as can be seen by the waveshape of the no-load mains current).

[pdf64]

Maybe the OT in that amp has a particularly bad radiated field, ie it's nothing to do with the PT?

[R.G.]

Well, of course I never said or meant that transformers can transform DC. I didn't even mention saturation either.

But using half-wave rectification on the secondary results in a net DC secondary current which produces a DC pre-magnetization of the core, resulting in reduced effective core µ and primary L. This causes the primary magnetization current to increase. In sum these effects are likely to increase the stray field of the transformer - especially with "economical" PTs that already operate close to the saturation "edge" (as can be seen by the waveshape of (no-load) the mains current).

OK. I understand what you meant.

However, no, that's not what happens. Secondary currents do not affect core magnetization that way. Here's why:

The core is a large inductor, and is pumped up to a certain alternating magnetization field intensity by the primary voltage. The core magnetization causes a counter EMF in the primary windings that limits primary current to just balance keeping the M-field going at the net resulting voltage. The secondary sucks energy out of the core's changing field, instantaneously removing some of the back EMF and that lets more current flow into the primary. The energy goes into the M-field, instantly replacing what is going out on the secondary. In effect, the T-model of an ideal transformer is correct: the primary inductance is a side effect of energy flowing from primary to secondary. As such, the instantaneous current flow in the secondary does not cause any change of substance to the core's field. At all.

The only exception to this is the idea that if secondary current flows only on one of the half-cycles, it causes the effective primary voltage to be lower on that half-cycle due to primary wire resistance. That does constitute an offset of the primary voltage, and that can push the core - a little - to one side. So the secondary may cause it, but it's not from DC in the secondary affecting the M-field, pre-magnetization (?) of the core, or a change in L or u. It's not in the core at all. It's an offset on the primary voltage. I'd have to do some more thinking about whether leakage inductance can get into that act or not.

[Helmholtz]

Not sure, if I understand correctly (language barrier).

But I tend to disagree with your conclusions.

The core "sees" the difference between primary and secondary fluxes. If there is a net DC component in the secondary current, this will cause a resulting DC flux in the core. Any DC flux premagnetizes the core and reduces AC-µ and primary L.

The ability of the core to confine the magnetic field inside depends on its effective AC-µ.

So far my explanations are purely speculative, as we don't know if the OP actually uses half-wave rectification.

[Pedro Vecino]

Many thanks for your explanations and advices. My idea is that the symmetry in these transformers was complete, so there are few or no possibilities, but before disassembling or testing others it´s better to ask.
The problematic amp (3) is the one with the least complexity: a Champ with an added preamplifier step (Princeton Reverb style). It has a rectifier bridge and more intense filtering. The form of construction is practically the same in all three.
In this amp I did not want to select an oversized OT, but a standard one (Hammond with 8K).
I only find two differences with the others: this has the PT primary and secondary wires entering the chassis through the same hole and I see in one of the pics some proximity in the signal that goes to master volume and the input of the transformer wires. The latter, easily correctable.

When I have a little time I will continue with the tests. Thanks again.

[catalin gramada]

Didn.t read the thread. What I suggest to try is just unscrewed from chassis and raise it little bit up (put a book under it and chasiss). See if humming gone or it was consistent reduced.
Cheap toroids have lot of leakage (didn.t experienced hq.ones) and easy coupling into OT even spaced at opposite extremes of long marshall type chassis. (You can get noise in speaker even without the amp powered on )Electromagnetic shielding with some Fe Si band is effective,an found is a compulsory feature for toroids, as well can try to raise the transformer over chassis I suggested before.

[nevetslab]

If I understand this, the noise pickup (hum field) only happens when the guitar (single coil pickups) is connected. And, I'd also presume in close proximity to the chassis. While Toroids are by far the lowest stray field xfmrs, they do still radiate, and, in some cases, you'll find M6 grain oriented steel shields wrapped around the outside of the core (taped on). Looks like steel chassis construction. There can be some minor variation in the radiated stray field as you move around the circumference of the toroid. Your lead dress from the xfmr thru the grommets to the insides of the chassis will, of course, limit just how much you can rotate it, in trying that.

For M6 Grain Oriented Steel, you'd have to find a vendor who can cut you a few strips....width same as the overall height, length would be greater than the circumference. Typically you'll find tape wrapped over the edges, then taped tightly around the outside of the xfmr.

Anyway, just a thought......I've had to use the M6 shields around 800VA Toroids when I had signal processing electronics in close proximity to them, and usually had to rotate them somewhat for minimal field pickup.

[R.G.]

Not sure, if I understand correctly (language barrier).

But I tend to disagree with your conclusions.

I long ago gave up on thinking I knew everything, so it's possible that I am wrong about this. We should probably discuss this in a different thread.

I do agree that it's probably not the OP's problem with hum.

[nickb]

Eyeballing it, the 6V6 version PT has the biggest cross-sectional area. The PT likely all have similar working flux densities. Therefore the 6V6 one has the highest total. It follows that it has the highest leakage flux.

The big question is, how close do you have to be for this is be a problem?

[Helmholtz]

While Toroids are by far the lowest stray field xfmrs, they do still radiate, and, in some cases, you'll find M6 grain oriented steel shields wrapped around the outside of the core (taped on).

Yes, that or mu-metal foil wrapped around the circumference or even a steel cage for the PT would reduce the stray field.

[J M Fahey]

To add still another seasoning to the sauce, we may also have an improperly designed or cheaply made PT.

Lower quality steel and/or less than optimum amount of turns on primary can largely increase core saturation with the secondary effect of much increased radiation, and "ugly waveform" , I mean a very distorted *current* waveform (although mains voltage will look fine) which in due turn will induce an ugly waveform magnetic field.

Full of harmonics which will be easily picked up by a single coil pickup,not as deep smooth 50/60Hz hum but as its buzzy cousin.