Ever heard of this...

Sure. Lots of core materials have lower permeability at low flux levels--it's the magnetic version of traps in semiconductors.

Cheers

Phil Hobbs

So there's no cure? I was pondering current drive, but it doesn't sound like that would help? ...Jim Thompson

Different core material, or an out-of-band bias signal, as in tape heads.

Cheers

Phil Hobbs

So, I read Jim's questions and think "ooh, I know the answer!", then Phil answers it. Twice. In a row. Hmph.

(Yes, this IS a contest).

If you're clever you can probably tailor the bias signal as needed, to reduce power drain and maintain sufficient overhead.

Don't feel bad--I'm stuck in bed getting over pneumonia, and am sufficientl y recovered to be excruciatingly bored, so I'm checking SED much more often than someone with other options.

If somebody asks something about systems with multiple inputs and outputs, I promise I'll be quiet as a mouse. ;)

Cheers

Phil Hobbs

Current drive sounds good as you can RSense and filter to get a control voltage representing the average flux in the core, and use that to control your out-of-band bias oscillator just summed into the signal, to keep the flux above some minimum constant

Real or approximate zero-flux mode would certainly help.

Knowing little about magnetics, could you turn on some DC bias at low signal levels?

George H.

DC could introduce distortion, maybe more easily than HF AC and will need some tweaking to get the best results It *was* used in many cheap portable tape recorders.

Maybe add another secondary winding, and use it for negative feedback in the circuit that drives the primary, a bit like old valve audio power amplifiers did.

You would want the feedback secondary and the real secondary to have similar coupling to the primary.

Why would this only affect low frequencies?

John

At higher frequencies less inductance is needed and the impedance of the transformer will still be sufficient despite the reduced permeability. Not so at low frequencies. It's the same with RF transformers such as baluns. There it is more extreme in that at really high frequencies the core wouldn't even have to be there.

Yes. The client found this...

...Jim Thompson

Interesting. (I've never used an audio transformer in real life--just pulse and RF.)

Cheers

Phil Hobbs

Nice oscillator circuits!

if you put a low frequency sine wave through the transformer, does it appear distorted, resembling crossover distortion?

or is it just that the client is making a measurement error?

the xformer attenuates low frequencies for all the usual reasons, but if you drive it harder, it distorts, and the clients measurements incorrectly indicate that the amplitude of the fundamental increased?

Mark