Idle question: Square or Sine wave for transformer?

[apelly]

I was just reading another thread, and I was wondering. Are transformers more efficient driven with a square wave or a sine wave?

You want maximum flux as soon as possible, changing as fast as possible. The core size and composition effects how fast and how much the flux can change.

It seems like you should jam all your electrons it at once on a sharp edge.

Am I thinking (and talking) out of my arse?

[Dave]

Well, a square wave is really just a bunch of sines on top of each other.
Here's the thing, a transformer is going to be most efficient at one particular frequency. If you only feed it a single tone at that frequency, it is surely going to more efficient than putting in higher harmonics, which aren't going to be transferred to the other side as efficiently.

Bear in mind that this is just the fruit of my thought process, putting 1 and 1 together, I haven't actually researched this.

[Vgkid]

Wouldn't putting a square wave into a transformer give you a lc time constant pulse as the output. I have never actually messed with this.

[rx8pilot]

I would think the transformer would turn it into a more of a sine wave and the 'destroyed' harmonics would be turned into heat.

Never tried it.

[apelly]

Well, a square wave is really just a bunch of sines on top of each other.

That is true.

Here's the thing, a transformer is going to be most efficient at one particular frequency.

Also true. And determined by the core.

If you only feed it a single tone at that frequency, it is surely going to more efficient than putting in higher harmonics, which aren't going to be transferred to the other side as efficiently.

Not sure. The tone doesn't transfer. The field does.

The core doesn't saturate immediately. More energy should cause faster saturation, no? Sure, we have no influence over how quickly it collapses. Or does the load effect this?

Anyway, there's surely no benefit in providing power as the field collapses? That's half of a cycle wasted.

[apelly]

I would think the transformer would turn it into a more of a sine wave and the 'destroyed' harmonics would be turned into heat.

Never tried it.

I agree the output is going to look sinusoidal.

Now I'm thinking that the efficiency is coming from the magnetic field resisting current flow. Therefore it makes no sense to to drive it faster than the core magnetises.

I think I agree with you.

[VK5RC]

I tend to think of a transformer as two close coupled inductors,  the higher frequency harmonics of the square wave just won't 'get in'  ie too high an impedance presented by transformer.

[apelly]

I tend to think of a transformer as two close coupled inductors

That's how I was thinking too.

the higher frequency harmonics of the square wave just won't 'get in'  ie too high an impedance presented by transformer.

Only after the magnetic field is established. Prior to that it's just a piece of wire.

Of course, as I said earlier, I'm only guessing. I'll google it later I suppose. I could just do some experiments, but unfortunately all my kit is unavailable temporarily.

[T3sl4co1l]

Losses generally increase with frequency, so yes, sine is more efficient.

The difference is small for most designs.

Switching converters don't have much choice, of course!

Tim

[Seekonk]

I have a 60z transformer that is operating at300Hz square waves, it was expedient to do so.   The core clearly heats up and the windings stay cool. Transformer was designed for a MSW converter.

[Stray Electron]

Well, a square wave is really just a bunch of sines on top of each other.

   Almost correct. A square wave is actually a sine wave of the same frequency but with all of the ODD-harmonics of that frequency added in. The magnitude of each odd harmonic decreases as the frequency rises.  Adding the odd-harmonics "files In" the corners of the sine wave to make it a square wave.

   Long story short, a 60 Hz square wave will contain a 60 Hz sine wave, a 180 Hz sine wave, a 300Hz sine wave, etc, etc etc.  Transformers tend to work only a specific frequency so the they won't react well the 180Hz and the other high frequencies.  The harmonics will be mostly filtered out and generate heating in the transformer. The transformer output will be 60 Hz with few if any of the harmonics so it will be a nearly pure sine wave.

   Note: transformers can be built to have wide frequency range, for example for audio use, but power transformers are usually designed for a specific frequency and will sharply attenuate other frequencies.

[bobcat]

Actually, the secondary will give spikes generated during the rise/fall of the square wave of current through the primary.
See Jeri Ellsworth's video demonstrating this:
https://youtu.be/1tAyvafIv04

 

[fubar.gr]

Another way to look at it, is that passing a square pulse through a traffo is essentially similar to passing DC. The core will saturate and the primary winding will become a short circuit.

But if the pulse duration is short enough (frequency is high enough), so that the core has no time to saturate, the traffo will work ok, as long as the core has no significant iron losses at that frequency.

This is essentially how SMPS transformers and signal transformers (eg. Ethernet) work.

[Dave]

Well, a square wave is really just a bunch of sines on top of each other.

   Almost correct. A square wave is actually a sine wave of the same frequency but with all of the ODD-harmonics of that frequency added in. The magnitude of each odd harmonic decreases as the frequency rises.  Adding the odd-harmonics "files In" the corners of the sine wave to make it a square wave.

   Long story short, a 60 Hz square wave will contain a 60 Hz sine wave, a 180 Hz sine wave, a 300Hz sine wave, etc, etc etc.  Transformers tend to work only a specific frequency so the they won't react well the 180Hz and the other high frequencies.  The harmonics will be mostly filtered out and generate heating in the transformer. The transformer output will be 60 Hz with few if any of the harmonics so it will be a nearly pure sine wave.

   Note: transformers can be built to have wide frequency range, for example for audio use, but power transformers are usually designed for a specific frequency and will sharply attenuate other frequencies.

Soooo... which part of your explanation invalidates my statement?