SY said:
It sounds like you really gave it an effort to duplicate the effect so I commend you. Actually the A/B listening test would probably be more effective with normal music. If the effect is worthwhile it should show up there. There's all kinds of differences the ear can hear that instruments don't really know how to measure (yet). And if part of the effect that creates the SE signature is caused by small signal hysteresis, Menno suggested, that "should" show up in listening tests.
I'm not sure what to A/B. Two different transformers? Too many variables. I won't object if Menno sends me some 200W toroids to try, but I don't think I can rack up any audible differences to the lack of "stiction." Extra windings with ultrasonic bias? Again, that introduces some unwanted variables, like feeding back ultrasonics into the circuitry.
Personally, I don't like the way SE amps typically sound (I find well-design push-pull to be more neutral), so it's probably wasted on my philistine ears anyway.
Personally, I don't like the way SE amps typically sound (I find well-design push-pull to be more neutral), so it's probably wasted on my philistine ears anyway.
Too many variables. The HF signal shooting back into the circuitry and radiating around can have an effect.
If I want my amp to sound like an SE, I can stick a resistor in series with the output to lower the damping and change the frequency response near the crossover points and unbalance the phase splitter to get me a nice fat'n'forward second harmonic signature.
If I want my amp to sound like an SE, I can stick a resistor in series with the output to lower the damping and change the frequency response near the crossover points and unbalance the phase splitter to get me a nice fat'n'forward second harmonic signature.
SY said:
Hey Stuart, that's a bit of an oversimplification. But I am with you, a good PP amp has nothing to be ashamed off and the best amps I've heard have NOT been single-ended. It also strikes me that tube amps with lots of even order distortion sounds slow and boring. I want decent dynamic shadings and real life excitement.
Joe R.
SY said:
Whew! Been a busy few days....
Wehrmacht? I think I would have remembered it if that name had been pointed out to me. Maybe you can refresh my memory. But knowing your kin and I also belong to a certain minority that was treated in like manner, wouldn't some find that name offensive? The Wehrmacht - and I am a History addict - disbanded in 1945 and was set up by Hitler circa 1935 - lasted ten years, included the SS etc and became the largest standing military force of the times peaking at over 16 million. Why would anyone think that to be a desirable name?
BTW, Eric, "exeric" have exchanged emails with me and discussed the lecture Menno gave that I videoed (I have a vague memory that you were sitting on my left or just behind my left shoulder?). I have made it into a 305MB avi file and it zipped into 233MB. If I could find somewhere to "park" it and post a link here? Guido is also seen at the end wrapping things up.
I think the video could be of interest and also basis of discussion. Also see Menno and Guido in person. BTW, did you not raise a question when Menno invited audience?
Cheers.
Joe R.
SY said:
Assuming a pure sine wave here.
exeric said:
SY said:
With a sine wave, the crossing point will occur at the same point in every cycle (correlated), which should make the measurement the easiest and most sensitive. With music, the zero crossing will be uncorrelated with any given frequency component. To the extent correlation occurs, it will be only with pure frequencies, with few other tones or overtones present - not so common in music.
Changing to opinion: If the crossing is uncorrelated, I would think that it would be harder to perceive.
Sheldon
Sheldon, yes, it was a pure (>0.002%THD) sine wave. I was also looking for the predicted loss of LF extension, which I did not see.
Joe, Menno's lecture at ETF was on... ummm... something slightly more esoteric. "Wehrmacht" was actually my nickname- it was a WWII-era setup using a single driver, set up in the basement at Biezenmortel. I remember going in there repeatedly with John Broskie for relief from all those godawful horns. It may not have been "real," but it was incredibly easy to listen to.
Joe, Menno's lecture at ETF was on... ummm... something slightly more esoteric. "Wehrmacht" was actually my nickname- it was a WWII-era setup using a single driver, set up in the basement at Biezenmortel. I remember going in there repeatedly with John Broskie for relief from all those godawful horns. It may not have been "real," but it was incredibly easy to listen to.
This is another interesting transformer related thread here on DIYA. Is it me or is there a recent increase in transformer threads here? Anyway, it got me interested and I found the term "Barkhausen effect" which I think describes what you are talking about: http://en.wikipedia.org/wiki/Barkhausen_effect (This effect seems to be how non-destructive testing of suspension bridges, etc are performed)
And this quote from the Bel Canto SET 40 description:
And here I was thinking that using a PS toroid as an OPT in a SE amp could be interesting ala http://www.diyaudio.com/forums/showthread.php?s=&threadid=129030
It would be interesting to hear from some of the more openly communicative magnetics experts here?
And this quote from the Bel Canto SET 40 description:
And here I was thinking that using a PS toroid as an OPT in a SE amp could be interesting ala http://www.diyaudio.com/forums/showthread.php?s=&threadid=129030
It would be interesting to hear from some of the more openly communicative magnetics experts here?
Ummm, I don't think I can agree with most of that Bel Canto quote. Even for a SE DC biased xfmr, flux domains have to flip in order for the flux to change. The DC bias acts just like the internal material coercive forces to pin domains down in one direction. Lowering the permeability as a result. No energy is supplied by the DC field, it would violate energy conservation since no energy is put into the DC field except resistive losses in the winding.
Certainly putting a DC bias on a core will quiet down the Barkhausen noise, but its most likely in direct proportion to how much the permeability has been reduced as well. So the noise reduction is just a good measure of how useless the core has become. High quality materials like permalloy have low internal coercive forces so the domains can move easily. Putting a DC field into the core is equivalent to reducing the material quality.
I have nothing but disgust for SE xfmrs, they are the worst of every engineering quality factor I can think of. The only thing "useful" they do is cause odd harmonic core distortion to appear as even harmonic distortion due to the DC field causing operation (and incremental permeability) to be lop-sided with respect to the signal.
Don (just makes me feel sooOOO good to pi----ssss on SE xfmrs.!)
Certainly putting a DC bias on a core will quiet down the Barkhausen noise, but its most likely in direct proportion to how much the permeability has been reduced as well. So the noise reduction is just a good measure of how useless the core has become. High quality materials like permalloy have low internal coercive forces so the domains can move easily. Putting a DC field into the core is equivalent to reducing the material quality.
I have nothing but disgust for SE xfmrs, they are the worst of every engineering quality factor I can think of. The only thing "useful" they do is cause odd harmonic core distortion to appear as even harmonic distortion due to the DC field causing operation (and incremental permeability) to be lop-sided with respect to the signal.
Don (just makes me feel sooOOO good to pi----ssss on SE xfmrs.!)
Don, why dontcha tell us how you really feel?
The quote is marketing...
But I have a question. I am seeing the large-signal hysteresis and
the small signal activation flux as perhaps two different things.
Are they in fact controlled by the same variables?
Re: magnetic zero crossings... I think hysteresis occurs at the
ends of the B-H curve as typically displayed. Which is where the
direction of the voltage change reverses i.e. increasing to decreasing,
not the polarity of the voltage. So PP and SE would appear to
only differ in this respect by the inductance of the winding.
Don't line toroids use permalloy cores? If there is a serious drawback
to using these cores, it should show up as high "activation energy"
or whatever you call it.
Just a few random thoughts.
Michael
PS maybe the gap changes the hysteresis or activation energy?
The quote is marketing...
But I have a question. I am seeing the large-signal hysteresis and
the small signal activation flux as perhaps two different things.
Are they in fact controlled by the same variables?
Re: magnetic zero crossings... I think hysteresis occurs at the
ends of the B-H curve as typically displayed. Which is where the
direction of the voltage change reverses i.e. increasing to decreasing,
not the polarity of the voltage. So PP and SE would appear to
only differ in this respect by the inductance of the winding.
Don't line toroids use permalloy cores? If there is a serious drawback
to using these cores, it should show up as high "activation energy"
or whatever you call it.
Just a few random thoughts.
Michael
PS maybe the gap changes the hysteresis or activation energy?
Well, 1st, I'm no expert on magnetics. But I do have a bunch of very good reference materials available to check when needed.
"But I have a question. I am seeing the large-signal hysteresis and
the small signal activation flux as perhaps two different things.
Are they in fact controlled by the same variables?"
My guess is that these are different effects but with some relationship still. The Barkhausen noise is a factor due to finite size magnetic domains jumping suddenly. This causes tiny sudden glitches in flux variation, which induces tiny voltage glitches. I would think that material makeup or processing could alter this. Like microcrystalline or amorphous materials, presumeably would have smaller domains.
The hysteresis is due to energy barriers preventing the magnetic domains from flipping. Usually this is attributed to internal magnetic fields in the material, crystal anisotrophy (ie, preferred local magnetic axis's), crystal defects, impurities....
High permeability materials, like permalloy, generally have low energy barriers giving low hysteresis losses (for comparable % of saturation flux variation) and high permeability. Permalloy (or Supermalloy or other super Mu materials) should give excellent results for line toroids. And would be great for output xfmrs too if they weren't so expensive.
I'm not sure about permalloy etc for Barkhausen noise. Some of the factors like low defects, large crystals, high temp. annealing, which improve the Mu (and lowers hysteresis) could cause very large domains. The hysteresis curve for the material may be some indication. Some materials or anneals are known for producing a square hysteresis curve. I think this might be an indication of the material behaving with large domains. But I'm out of my knowledge base here. Usually for linear applications the non-square hyst. materials and anneals are preferred.
Using a larger core with fewer turns would seem useful for minimizing Barkhausen noise, since each flux jump will induce a smaller voltage. An expensive solution though. Also, a larger core will allow operation to a lower max. flux level. The smallest domains generally are supposed to flip first I think, with bigger-noisier- ones flipping at higher flux (higher energy barrier to overcome). There is a very low flux region where the domains do not have to even flip to change core flux. They just tilt some angle less than 90 degrees. This is sorta like elastic operation, no jumps occur.
Its pretty simple to measure Barkhausen noise. Just connect a high turns winding up to a sensitive amplifier input and bring a permanent magnet up slowly to the core. Makes a noise/hiss like sound as the magnet approaches or recedes.
"stop sitting on the fence and tell us where you stand on this!"
Are you referring to the HF AC bias scheme? I think one needs to determine whether there is a problem to be solved first. As Sy mentioned. If you can't hear it, it ain't broken. Maybe for some types of xfmrs there could be effects. Simple enough to test with the magnet scheme. I would think that smaller xfmrs (like line xfmrs) would be most susceptable. But they also use the best materials.
Don
"But I have a question. I am seeing the large-signal hysteresis and
the small signal activation flux as perhaps two different things.
Are they in fact controlled by the same variables?"
My guess is that these are different effects but with some relationship still. The Barkhausen noise is a factor due to finite size magnetic domains jumping suddenly. This causes tiny sudden glitches in flux variation, which induces tiny voltage glitches. I would think that material makeup or processing could alter this. Like microcrystalline or amorphous materials, presumeably would have smaller domains.
The hysteresis is due to energy barriers preventing the magnetic domains from flipping. Usually this is attributed to internal magnetic fields in the material, crystal anisotrophy (ie, preferred local magnetic axis's), crystal defects, impurities....
High permeability materials, like permalloy, generally have low energy barriers giving low hysteresis losses (for comparable % of saturation flux variation) and high permeability. Permalloy (or Supermalloy or other super Mu materials) should give excellent results for line toroids. And would be great for output xfmrs too if they weren't so expensive.
I'm not sure about permalloy etc for Barkhausen noise. Some of the factors like low defects, large crystals, high temp. annealing, which improve the Mu (and lowers hysteresis) could cause very large domains. The hysteresis curve for the material may be some indication. Some materials or anneals are known for producing a square hysteresis curve. I think this might be an indication of the material behaving with large domains. But I'm out of my knowledge base here. Usually for linear applications the non-square hyst. materials and anneals are preferred.
Using a larger core with fewer turns would seem useful for minimizing Barkhausen noise, since each flux jump will induce a smaller voltage. An expensive solution though. Also, a larger core will allow operation to a lower max. flux level. The smallest domains generally are supposed to flip first I think, with bigger-noisier- ones flipping at higher flux (higher energy barrier to overcome). There is a very low flux region where the domains do not have to even flip to change core flux. They just tilt some angle less than 90 degrees. This is sorta like elastic operation, no jumps occur.
Its pretty simple to measure Barkhausen noise. Just connect a high turns winding up to a sensitive amplifier input and bring a permanent magnet up slowly to the core. Makes a noise/hiss like sound as the magnet approaches or recedes.
"stop sitting on the fence and tell us where you stand on this!"
Are you referring to the HF AC bias scheme? I think one needs to determine whether there is a problem to be solved first. As Sy mentioned. If you can't hear it, it ain't broken. Maybe for some types of xfmrs there could be effects. Simple enough to test with the magnet scheme. I would think that smaller xfmrs (like line xfmrs) would be most susceptable. But they also use the best materials.
Don
Goldman's book doesn't seem to have anything useful on Barkhausen noise. The old Bozorth book has quite a bit. It seems that the biggest glitches are caused by avalanches of domain flips. Usually around the very steep part of the hysteresis curve. Square curves being the noisiest.
Quiet materials have their domains oriented at right angles to the operating field direction by annealing. They just do the elastic twist thing then. But this causes the material to have a lower Mu.
The steep, square shaped hyst. materials have their domains oriented along the operating field direction, and so do domain flips. Very square hyst. materials can produce total material flips in one avalanche, these are the materials used in old computer cores.
In general, the magnetization details are quite complex, and vary between different materials and processing. Beside transverse annealing, one can also apply a DC field at right anges to the operating field direction to quiet things down. That also lowers the Mu like the transverse annealing does. Ironically, the parallel DC field in SE transofrmers would make the B. noise worse, not better. The reason for the illusion of being better is just due to the effective Mu being dropped by the DC, so not much flux change is occuring any more. Truly a disfunctional design. (now imagine the sound of the last remaining SE xfmr banging down into the dumpster )
Don
Quiet materials have their domains oriented at right angles to the operating field direction by annealing. They just do the elastic twist thing then. But this causes the material to have a lower Mu.
The steep, square shaped hyst. materials have their domains oriented along the operating field direction, and so do domain flips. Very square hyst. materials can produce total material flips in one avalanche, these are the materials used in old computer cores.
In general, the magnetization details are quite complex, and vary between different materials and processing. Beside transverse annealing, one can also apply a DC field at right anges to the operating field direction to quiet things down. That also lowers the Mu like the transverse annealing does. Ironically, the parallel DC field in SE transofrmers would make the B. noise worse, not better. The reason for the illusion of being better is just due to the effective Mu being dropped by the DC, so not much flux change is occuring any more. Truly a disfunctional design. (now imagine the sound of the last remaining SE xfmr banging down into the dumpster
)Don
Hey Don,
I was joking when I said "stop sitting on the fence and tell us where you stand on this!" but thanks for coming back anyway. As you say if the problem doesn't exist (from Sy's testing results) then no point. But it would be great if Joe Rasmussen put up a LEM for testing or pitched in here.
Here's the Menno paper: http://mennovanderveen.nl/eng/index.html
I was joking when I said "stop sitting on the fence and tell us where you stand on this!" but thanks for coming back anyway. As you say if the problem doesn't exist (from Sy's testing results) then no point. But it would be great if Joe Rasmussen put up a LEM for testing or pitched in here.
Here's the Menno paper: http://mennovanderveen.nl/eng/index.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.