Output Transformer Saturation Surpise

[nickb]

It makes as much sense as the $1500 used EC tremolux. One of those popped up at my local GC last week and it was gone in 2 days.

Yup. It's completely crazy. By an amazing coincidence I have a '64 Tremolux here that I just finished doing up after it's original owner passed away (huge collector) and it sat in an attic for a decade or so. Sounds nice with single coils, not so much with humbuckers. I think it's the closed back cab.. but now I'm wandering off topic.

[printer2]

Not a lot to add other than some manufacturer of guitar amp uses their 15W P-P transformer for their 5W SE amp but I can not remember which it is. Being English seems to ring half a bell.

[J M Fahey]

Maybe, but in that case I bet the transformer is gapped to allow dual use.

"Most Ungapped Transformer on Earth" title goes to Toroidal transformers, where the core is a ribbon of silicon steel , no gaps of course, except that ot starts atb one end and stops at the other, but after *many* turns whoch make that irrelevant.

And toroids are famous for being WAY TOO pesky about ANY DC .

Some ill advised Tube amp makers have made Toroidal OTs (what wereb they thinking?) and then go crazy trying to balance current, to thye extreme of addingb tight servocontrolled bias circuits, but even worse: even *Mains* toroidals , which by definition receive "pure AC"v suffer from the slight DC component which appears randomly because of waveform assymetry.

There is a whole section at DIYb Audio dedicated at correcting bthat (I think Bryston Audio started doing so) and nthat caught like wildfire.

Of course, they use nonsensical complex expensive "solutions" to what "should" not be a problem to begin with.



If you must buy 10000uF worth of caps just to avoid your Toroidal transformer buzzing and chugging, I suggest you look around for the nearest junkbin and throw it there for good.

There´s a few (*expensive*) DC blockers for Toroidals,

DC Blocker Trap Filter – Assembled in Case


or you can build your own:



so back to the original topic: IF ungapped *mains* transformers are problematic, think about a regular *Push Pull* amp where tubes may be (and usually are) unbalanced, then think about Single ended ones.

In all cases a gap is the practical solution.

[Gregg]

And toroids are famous for being WAY TOO pesky about ANY DC .

My experience with toroidal PTs and OTs doesn't confirm that at least not in guitar amps.
For the sake of experiment I've run toroidal OTs with as much as 10mA of imbalance at idle (for example one tube at 35mA the other at 45mA) without any catastrophic consequences as usually assumed. That particular OT also had different DCR primary halves.

[nickb]

My experience with toroidal PTs and OTs doesn't confirm that at least not in guitar amps.
For the sake of experiment I've run toroidal OTs with as much as 10mA of imbalance at idle (for example one tube at 35mA the other at 45mA) without any catastrophic consequences as usually assumed. That particular OT also had different DCR primary halves.

Not sure that I have this right but I think there is a world of difference operating a transformer from a high impedance source i.e. output tubes vs a low impedance i.e. power line. In the first case the average current is limited and therefore so is the flux. In the second the average current rise is limited only by the total circuit resistance, so the core can saturate. The toroid with no gap will have a high mu and therefore very little current is required to saturate it.

[Gregg]

I made many measurements back then with sine wave and spectrum snapshots and I was able to really saturate that OT (before reaching max power) only with 30Hz signal when it started buzzing. However it was a 4k, 40W, 60Hz LF cutoff OT per specs so I did that only for the sake of experiment. By the way I still have the amp so I can always make some more measurements.

[Mike Sulzer]

The toroid with no gap will have a high mu and therefore very little current is required to saturate it.

But it is the same current at dc or 20KHz. Of course, for a given primary voltage, the magnetizing current required to support transformer operation increases with decreasing frequency. If the transformer has no dc, then this magnetizing current is all that the core sees. The signal currents in the primary and out the secondary cancel as far as the core is concerned. If there is dc, then the flux from this current must be supported as well. Do you agree that in an SE class A amp, this dc current is equal to the magnetizing current at the lowest intended frequency of operation, using the usual definition where the impedance of the inductance at the lowest frequency is equal to the the V/I ratio of the primary signal? Or do I have that wrong? It cannot be off by too much.

[nickb]

But it is the same current at dc or 20KHz. Of course, for a given primary voltage, the magnetizing current required to support transformer operation increases with decreasing frequency. If the transformer has no dc, then this magnetizing current is all that the core sees. The signal currents in the primary and out the secondary cancel as far as the core is concerned. If there is dc, then the flux from this current must be supported as well. Do you agree that in an SE class A amp, this dc current is equal to the magnetizing current at the lowest intended frequency of operation, using the usual definition where the impedance of the inductance at the lowest frequency is equal to the the V/I ratio of the primary signal? Or do I have that wrong? It cannot be off by too much.

Complete agreement, except the discussion had moved on to consider why toroids can't tolerate much DC and why it is not a problem output transformers and why it is for power one.

[SoulFetish]

All the gap does is lower the effective permeability of the core, and so the voltage at the lowest frequency can be larger. (But the inductance is also lower and so the magnetizing current rises, and so the actual useful lowest frequency rises somewhat.)

Maybe the solution that I have suggested is wrong, but it is certainly wrong to say that a gap is essential. Any core can support some level of dc current without saturating. That is, the B field produced by a dc current must be less than the B field that causes saturation, and for the transformer to work, it must be less by enough to allow the sum of the dc and ac components to stay below that limit.

So if you guys want to show my solution is wrong, you have a lot more physics to do.

I can suggest a hint as to the misunderstanding here. It is true that the magnetizing current would have to rise to infinity to support a "zero frequency" transformer. But that is not what is happening here. The dc current is fixed to a finite value by the biasing in the circuit. It does not go to infinity as would be required for some kind of "runaway" saturation.

I think Mike is correct about this. Well, he's definitely correct that the purpose of air gap is to lower the relative permeability of the core. Using an air gap'd laminated core is by far the most common technique. But it isn't the only way to manage the DC current which must exist in the primary (and by nature, canno be transferred to the load). Another way to manage this is to add an extra primary winding, with only the DC component, wound to effectively cancel it out.
I imagine the reason you don't really see this is that they would rather not give up the valuable space in the winding area. Also, maybe it's too fussy a process to mirror the quiescent current of the tube for it to really be worth it.

[Mike Sulzer]

Complete agreement, except the discussion had moved on to consider why toroids can't tolerate much DC and why it is not a problem output transformers and why it is for power one.

Moved it has, but it is really the same thing. A toroidal power transformer, like any other power transformer, meets its specs to support the power line voltage at the power line frequency. Otherwise no one would buy it. You can figure out what the magnetizing current is, and that peak current is the dc current that could be supported with no ac current, and that determines any combination (sum) as well. The permeability of the core enters into the multi-variable optimization process that the engineer performed to design the product, but it does not tell you much by itself.