Barkhausen noise in audio transformers

[georgewalker]

So I'm trying to wrap my head around core losses in HF power transformers for work, and ran into "domain wall pinning."  I had no idea what that meant, and ended up finding out a little bit about Barkhausen noise, which (if I understand this correctly) to be the noise generated by magnetic materials being magnetized in discrete bits (domains) rather than in a continuous fashion.  It sounds like the kind of noise that would give audio people fright since it is conceptually like magnetic quantization noise.  (People pontificate about "transistor grain," but I've never heard anyone talk about "input transformer grain," which sounds more physically plausible to me at this moment -but please feel free to correct me.) 

My extremely slippery grasp on this makes me suspect that Barkhausen noise would be minimized by using ultra-high permeability cores like mu-metal, since this keeps B fields very low up to the point of saturation.  But I could very well be misunderstanding things.

Anyone know if this is ever relevant for mic-and-line level audio?  Has anyone here heard Barkhausen noise?  Is this the kind of thing people like Deane Jensen and Mark Fouxman measure(d) while evaluating their transformer designs?


Best regards,
George

 

[abbey road d enfer]

My extremely slippery grasp on this makes me suspect that Barkhausen noise would be minimized by using ultra-high permeability cores like mu-metal, since this keeps B fields very low up to the point of saturation.  But I could very well be misunderstanding things.

Barkhausen noise could very well be less in high-perm material, but not because the induction B is low; the operating induction in high-perm materials is not vastly different than the induction in more common magnetic materials (M6 silicon steel). The smaller size and more regular orientation of the magnetic domains is the reason for different (lower?) noise.

Anyone know if this is ever relevant for mic-and-line level audio?

It is. That's part of the quality assessment of cores used in audio.

Has anyone here heard Barkhausen noise?

I would say everybody has heard barkhausen noise, but it is generally hidden by the signal; it is similar to noise modulation so it is not identified as something peculiar. It takes a specific experience to put Barkhausen noise in evidence. It's relatively easy to do, involves a coil wrapped around a magnetic core. the coil is connected to an audio amplifier and the magnetic structure is perturbed by approaching a magnet. A typical noise can be heard when it happens but it is similar to white noise so it may come unnoticed. I have not personally heard B. noise, I'm only referring to what I heard in class.

Is this the kind of thing people like Deane Jensen and Mark Fouxman measure(d) while evaluating their transformer designs?

I can't speak for them, but transformer designers who use their ears as much as their lab equipment do listen to their transformers and B. noise is a component of their sound signature.

 

[georgewalker]

Doe!  Jfgi: http://www.groupdiy.com/index.php?topic=25941.0

 

[Jonte Knif]

My experience is that Lundahl amorphous core (which have very small magnetic domains) transformers sound more detailed compared to their silicon iron core models, I don't use them for the fun of spending more money. I've done a blind test in which one of my customers clearly succeeded, but others had a bit difficulties. It depends on how you listen to. If you get into very detail and ambience oriented mode, it seems to be easy to spot the amorphous models, but in this test no advise of any kind was given. Well, in any case with AM cores I am able to manufacture tube line amps with no observable detail loss, whereas silicon iron models are pretty easily spotted, often for effects that are wanted.