Dave,
I'm not sure what the motivation for 4x4 or other interleaves is. Usually I would think of it as a means to reduce assembly cost. I suppose it would make for a little more air gap effect than full interleave if one wanted less sensitivity to DC. But it really looks more like a smoothing choke where the easy magn. pathes will saturate early leaving the gapped pathes at higher flux (or current) levels.
This could have the effect of flattening the effective perm. for the core since the M6 has low perm at low current. The shortcuts would increase effective perm. at the low currents. But the resulting "flattened" perm level would just be the low air gapped perm.. Not really all that desirable except in SE designs with DC.
Don
I'm not sure what the motivation for 4x4 or other interleaves is. Usually I would think of it as a means to reduce assembly cost. I suppose it would make for a little more air gap effect than full interleave if one wanted less sensitivity to DC. But it really looks more like a smoothing choke where the easy magn. pathes will saturate early leaving the gapped pathes at higher flux (or current) levels.
This could have the effect of flattening the effective perm. for the core since the M6 has low perm at low current. The shortcuts would increase effective perm. at the low currents. But the resulting "flattened" perm level would just be the low air gapped perm.. Not really all that desirable except in SE designs with DC.
Don
hey-Hey!!!,
Also remember the pish-poor stacking factor of the high perm material. 10% perhaps?
The AC flux equ'n does not have a material property or core geometry save for cross section in it( RDH3, p.216). Number of turns in the denominator, applied voltage in the numerator...volts per turn.
cheers,
Douglas
Also remember the pish-poor stacking factor of the high perm material. 10% perhaps?
The AC flux equ'n does not have a material property or core geometry save for cross section in it( RDH3, p.216). Number of turns in the denominator, applied voltage in the numerator...volts per turn.
cheers,
Douglas
The AC flux equation I think you are refering to is the voltage transfer equation. And, yes, the voltage determines a net flux, and net flux determines output voltage indeed.
The problem with permeablitity variation comes in when you consider magnetizing (or excitation) current. (ie, the parasitic winding inductance draws current from the source, and permeability determines how much current. Any source resistance present and the effective voltage applied to the primary drops along with the resulting secondary voltage) So one really has to take into account source resistance and permeability for real transformers.
Don
The problem with permeablitity variation comes in when you consider magnetizing (or excitation) current. (ie, the parasitic winding inductance draws current from the source, and permeability determines how much current. Any source resistance present and the effective voltage applied to the primary drops along with the resulting secondary voltage) So one really has to take into account source resistance and permeability for real transformers.
Don
hey-Hey!!!,
That is why I am going to go and test output TX's in the small signal domain. I will apply a source impedance, and drive a load while measuring distortion. Should be a small hump, and how high, and ending at what power should be some useful numbers. that is part of why I want to try three TX's with different cores, but the same coil.
What I'd like for my own stuff is 49% Ni with a gap to drop initial perm to just over good Si steel. This is for PP iron. reduction of remnance, and smoothing out the huge gain in realized perm as the core comes on the pipe( far different effect from a TX winder comming on the pipe or dooobie).
cheers,
Douglas
That is why I am going to go and test output TX's in the small signal domain. I will apply a source impedance, and drive a load while measuring distortion. Should be a small hump, and how high, and ending at what power should be some useful numbers. that is part of why I want to try three TX's with different cores, but the same coil.
What I'd like for my own stuff is 49% Ni with a gap to drop initial perm to just over good Si steel. This is for PP iron. reduction of remnance, and smoothing out the huge gain in realized perm as the core comes on the pipe( far different effect from a TX winder comming on the pipe or dooobie).
cheers,
Douglas
Even though I used DC magnetization curves, they can be fairly misleading if you are dealing with transformers with no or very little DC in the core, which I believe is the case with Peerless 20-20 Plus transformers. At midrange frequencies, Permalloy and M6 have lower permeability that remains fairly constant within the normal operating range and the astronomical figures quoted for Permalloy don't really apply. At 400Hz and 1000G, the permeability for 6 mil Permalloy is 30,000 and for 14 mil it is only about 21,000. M6 is 12 mil GOSS, and I don't have any data for it at midrange frequencies, but for 4 mil GOSS at 1000G, the permeability is about 6000 and it is about 4500 at 500G. The wonderful thing about Permalloy is the permeability remains almost perfectly constant from about 1G to almost 500G, and so if one had the money and was so inclined, he could theoretically build a (very large) virtually distortion-free output transformer.
John
John,
Very interesting data, thanks. Perhaps I have been too alarmist about the pinstriping if permalloy perm. is not so astro. high in practice.
Thinking about how to analyze the material combo leads me to think that the best approach is to simply visualize a ramping up current (rather than voltage) and from that one can determine (with datasets ) the permeablility of each material. Then one can compute effective core permeability from Mu times the area of each material. Most definitive of course would be some real measurement of excitation current versus AC voltage for an actual pinstriped xfmr. Hoping that Douglas will come up with some data here to put this to rest.
Some other interesting exotic approaches are to make the M6 (or whatever matl.) display higher initial Mu by means of a HF AC bias current. Much like tape recording technology.
There used to be a permalloy coated wire made for plated wire memory technology. If one made the permalloy (or other magnetic) coating rather thick and enameled it (insulating) it could be used to wind a magnetic wire core. HF or even DC current passed thru the plated wire could be used to eliminate hysteresis by generating an orthogonal field around the wire plating circumference. Magnetic domain flips from audio windings are avoided that way, only 90 degree domain rotations occur.
Ferrite pot cores with a central bolt hole can have a bias winding put thru the bolt hole to generate a bias field at right angle to the main winding field. Unfortunately, the ferrite materials are too limited in Bmax for audio use. Another scheme is to use two toroidal cores, side by side. The audio windings get wound around both cores and the bias windings get wound in figure 8 fashion (by two separate windings actually) so as to not couple with the audio.
A very extreme approach would be to heat the core material of the xfmr to near its Curie temp. which causes hysteresis to disappear via thermal agitation. But Bmax plummets too.
Maybe someone will just come out with a low cost microcrystalline material. Or a plated film with low hysteresis.
Let's see, if we could just drop an M6 core into a nuclear reactor for a while it would probably disrupt the crystaline structure to become amorphous. (Kenpeter: got a reactor at work, beside that X-ray machine?) Glowing xfmrs, very novel. I suppose leaving it under the X-ray machine for a few years might do the same without the residual glow side effects. Talk about expensive though. (Hmmm.., they must have a lead beam stop in the X-ray machine somewhere, just set some M-6 cores on it for a few years. )
Don
Very interesting data, thanks. Perhaps I have been too alarmist about the pinstriping if permalloy perm. is not so astro. high in practice.
Thinking about how to analyze the material combo leads me to think that the best approach is to simply visualize a ramping up current (rather than voltage) and from that one can determine (with datasets ) the permeablility of each material. Then one can compute effective core permeability from Mu times the area of each material. Most definitive of course would be some real measurement of excitation current versus AC voltage for an actual pinstriped xfmr. Hoping that Douglas will come up with some data here to put this to rest.
Some other interesting exotic approaches are to make the M6 (or whatever matl.) display higher initial Mu by means of a HF AC bias current. Much like tape recording technology.
There used to be a permalloy coated wire made for plated wire memory technology. If one made the permalloy (or other magnetic) coating rather thick and enameled it (insulating) it could be used to wind a magnetic wire core. HF or even DC current passed thru the plated wire could be used to eliminate hysteresis by generating an orthogonal field around the wire plating circumference. Magnetic domain flips from audio windings are avoided that way, only 90 degree domain rotations occur.
Ferrite pot cores with a central bolt hole can have a bias winding put thru the bolt hole to generate a bias field at right angle to the main winding field. Unfortunately, the ferrite materials are too limited in Bmax for audio use. Another scheme is to use two toroidal cores, side by side. The audio windings get wound around both cores and the bias windings get wound in figure 8 fashion (by two separate windings actually) so as to not couple with the audio.
A very extreme approach would be to heat the core material of the xfmr to near its Curie temp. which causes hysteresis to disappear via thermal agitation. But Bmax plummets too.
Maybe someone will just come out with a low cost microcrystalline material. Or a plated film with low hysteresis.
Let's see, if we could just drop an M6 core into a nuclear reactor for a while it would probably disrupt the crystaline structure to become amorphous. (Kenpeter: got a reactor at work, beside that X-ray machine?) Glowing xfmrs, very novel. I suppose leaving it under the X-ray machine for a few years might do the same without the residual glow side effects. Talk about expensive though. (Hmmm.., they must have a lead beam stop in the X-ray machine somewhere, just set some M-6 cores on it for a few years. )
Don
for all the doom and gloom regarding "pinstriping"...
I'm reminded of the many applications that have used this technique with great success over the years.
As just one example, consider that tens of thousands of the Heath W5M model amplifiers were built with the Peerless 16309 output transformer which used this core technology. Several other models produced by Peerless also used this technique--- amongst these are outputs that have been used by various amplifier manufacturers which have been prized by knowledgeable collectors and listeners for many years.
Consider also that Sowter of England currently offers several of their transformers with an admixture of nickel and M6. Anyone fimiliar with the history of research in the nickel alloys will surely appreciate the early pioneering role that the senior Mr. Sowter played in the analysis and documentation of these high-perm materials in audio applications.
I also believe that Electra-Print currently offers on some of their models a "nickel trim" to enhance performance.
Re: testing and exciting current. At the following url is a textual reprint of the Peerless Exciting Current Test which some may find useful;
http://www.audioasylum.com/forums/magnequest/messages/243.html
MSL
I'm reminded of the many applications that have used this technique with great success over the years.
As just one example, consider that tens of thousands of the Heath W5M model amplifiers were built with the Peerless 16309 output transformer which used this core technology. Several other models produced by Peerless also used this technique--- amongst these are outputs that have been used by various amplifier manufacturers which have been prized by knowledgeable collectors and listeners for many years.
Consider also that Sowter of England currently offers several of their transformers with an admixture of nickel and M6. Anyone fimiliar with the history of research in the nickel alloys will surely appreciate the early pioneering role that the senior Mr. Sowter played in the analysis and documentation of these high-perm materials in audio applications.
I also believe that Electra-Print currently offers on some of their models a "nickel trim" to enhance performance.
Re: testing and exciting current. At the following url is a textual reprint of the Peerless Exciting Current Test which some may find useful;
http://www.audioasylum.com/forums/magnequest/messages/243.html
MSL
Darn,
my copy of "Magnetic Circuits and Transformers by the staff of Massachusetts Institute of Technology, pages 161, 268, and 281. " is 800 miles away at the moment. Does it show results for a striped xfmr? And it would need to show results versus AC amplitude to settle the Mu question.
Don
my copy of "Magnetic Circuits and Transformers by the staff of Massachusetts Institute of Technology, pages 161, 268, and 281. " is 800 miles away at the moment. Does it show results for a striped xfmr? And it would need to show results versus AC amplitude to settle the Mu question.
Don
jlsem said:
Very good point about listening. If I recall--- the original poster asked about subjective qualities and impressions and was seeking comments from folks who have actually heard such critters as he was inquiring about.
And then it dawned on me---- that many of us have heard amps that use "pinstriped" cores---- which is the subject of another post that I made a short while ago.
MSL
Don,
I just took 4X4 as an example and the effect should be measurable with a 1X1 stacking or any other core structure with two distinctly different gap sizes Ultimately i prefer butt gaps (even for AC only situations) but will go to alternate stacking as a compromise if I need a boost in inductance. I'll also add I mostly deal with nickel
Just for kicks I took the same bobbin and measured with three different core configurations. The goal was to see linearity wrt applied flux and the measurements were made at 100hz with levels of 50mv, 250mv and 1V
With Butt gapped 80% nickel the inductance was essentially 50hy's at all three settings.
With 1X1 stacked M6 I measured 12, 15 and 20hy's
If I simply stacked the core with only the M-6 E's the measured inductance only dropped to 10, 13 and 17hy's.
dave
I just took 4X4 as an example and the effect should be measurable with a 1X1 stacking or any other core structure with two distinctly different gap sizes Ultimately i prefer butt gaps (even for AC only situations) but will go to alternate stacking as a compromise if I need a boost in inductance. I'll also add I mostly deal with nickel
Just for kicks I took the same bobbin and measured with three different core configurations. The goal was to see linearity wrt applied flux and the measurements were made at 100hz with levels of 50mv, 250mv and 1V
With Butt gapped 80% nickel the inductance was essentially 50hy's at all three settings.
With 1X1 stacked M6 I measured 12, 15 and 20hy's
If I simply stacked the core with only the M-6 E's the measured inductance only dropped to 10, 13 and 17hy's.
dave
smoking-amp said:
They do that to open up small air gaps in PP xfmrs to avoid DC saturation if the VTs in both phases aren't pulling the same Q-Point currents. I have Hammonds with a 6 X 6 stack, and a NOS Stancore with a 2 X 2 stack.
Since toroidals don't have any gaps, these are very sensitive to DC current imbalance. Since my PP designs allow for independent bias adjust, it's not a problem.
A couple of useful readings (maybe known)
http://www.sowter.co.uk/pdf/GAVS.pdf
http://cinemag.biz/application_notes/AN-104.pdf
Ciao
Gianluca
http://www.sowter.co.uk/pdf/GAVS.pdf
http://cinemag.biz/application_notes/AN-104.pdf
Ciao
Gianluca
Hi Miles,
I have not found alternate stacking in groups to be very effective at combating DC and it nets you some fairly peculiar behavior. This can easily be seen by looking at the Q of the inductance as you increase the flux. I have noticed that the first indicator of saturation is a marked drop in Q and after that drop is seen the inductance then starts to drop. If you measure an alternately stacked coil at increasing flux levels, you will get an initial high Q reading and then it will drop sharply at some low level. Then the inductance drops until at some point the Q increases again and the inductance levels out. Finally the Q will again drop sharply and the inductance will quickly follow.
This rollercoaster ride the Q takes can eaily be explained by a core structure with two distinct gaps and the first drop in Q is the smallest gap saturating and giving way to the remaining structure with a larger gap. The second drop in Q is that structure hitting saturation.
As a final note, Tape wound Toroids do technically have airgaps.
dave
I have not found alternate stacking in groups to be very effective at combating DC and it nets you some fairly peculiar behavior. This can easily be seen by looking at the Q of the inductance as you increase the flux. I have noticed that the first indicator of saturation is a marked drop in Q and after that drop is seen the inductance then starts to drop. If you measure an alternately stacked coil at increasing flux levels, you will get an initial high Q reading and then it will drop sharply at some low level. Then the inductance drops until at some point the Q increases again and the inductance levels out. Finally the Q will again drop sharply and the inductance will quickly follow.
This rollercoaster ride the Q takes can eaily be explained by a core structure with two distinct gaps and the first drop in Q is the smallest gap saturating and giving way to the remaining structure with a larger gap. The second drop in Q is that structure hitting saturation.
As a final note, Tape wound Toroids do technically have airgaps.
dave
After sleeping on the pinstriping permeability issue overnight, I now think I have been missing the real issue. ANY magnetizing current is out of phase or even harmonic ratio (hyst. dist.) with the signal, so higher permeability is ALWAYS better. High permeability should trump constant permeability, but having the permalloy perm. fade gracefully into the M6 peak perm. would be the best of both worlds. So some measured data would still be most desireable.
So my deep appologies to the pinstriping afficionados for being overly alarmist on the constant permeability issue.
Never being satisfied with the status quo, I now begin wondering if some type of SS servo winding could be used to null out magnetizing current for any core. Perhaps with the dual toroid scheme, one could arrange for the servo field to be non-interacting with the audio field.
Hmmm.....
Rough idea:
The servo Op. amp would monitor the tube current for variation from the tube voltage signal (ie, differencing of V and I signals) and a correction would be generated to put on the figure 8 winding of the dual toroids. (The audio windings being wound conventionally around both toroids) This would make for a non direct interaction between the audio and servo correction fields. Only interaction from permiability variation or magnetizing current would be possible. Whether this can successfully fix the magnetizing current I don't know.
Don
So my deep appologies to the pinstriping afficionados for being overly alarmist on the constant permeability issue.
Never being satisfied with the status quo, I now begin wondering if some type of SS servo winding could be used to null out magnetizing current for any core. Perhaps with the dual toroid scheme, one could arrange for the servo field to be non-interacting with the audio field.
Hmmm.....
Rough idea:
The servo Op. amp would monitor the tube current for variation from the tube voltage signal (ie, differencing of V and I signals) and a correction would be generated to put on the figure 8 winding of the dual toroids. (The audio windings being wound conventionally around both toroids) This would make for a non direct interaction between the audio and servo correction fields. Only interaction from permiability variation or magnetizing current would be possible. Whether this can successfully fix the magnetizing current I don't know.
Don
Here's some test data re: L vs unbal dc with several different stacking arrangements.
These tests were done by Freed Transformers--- the same guys who made professional test equipment including incremental inductance bridges.
The unit tested is a 10 watt output with a nominal 12,000 ohm CT primary.
Stacking L (zero unbal dc) L (4 madc unbal)
1X1 333 hy 99 hy
5X5 257 hy 104 hy
10x10 186 hy 97 hy
20X20 99 hy 74 hy
tests were done at 80vrms @ 60 cycles (approx 533 mW)
core material was M6 throughout.
MSL
These tests were done by Freed Transformers--- the same guys who made professional test equipment including incremental inductance bridges.
The unit tested is a 10 watt output with a nominal 12,000 ohm CT primary.
Stacking L (zero unbal dc) L (4 madc unbal)
1X1 333 hy 99 hy
5X5 257 hy 104 hy
10x10 186 hy 97 hy
20X20 99 hy 74 hy
tests were done at 80vrms @ 60 cycles (approx 533 mW)
core material was M6 throughout.
MSL
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.