Getting to see what a new OT will or wont do

[pjd3]

Hello.

This is just a special interest post I guess, if nothing else it will be interesting to see (hear) what happens.

I have a Musical Power Supplies output transformer OT22PP coming to replace the one in my 6V6 single channel plexi. It is described as a "over-sized deluxe reverb output transformer) - 6.6K primary impedance, 30VA.

When I gutted my old Bogner Alchemist to build a Sluckey 6V6 20 watt plexi, the only component I kept and used from the original Alchemist was the output transformer. It is one with barely no information to be found, name - Chuang Meei.

Of course, since this output transformer was used for a 6L6GC pair, it was designed with a 4K primary in mind so, I connect my 16 ohm V30 to the 8 ohm tap so that the 6V6's see about 8K. It has worked fine, and I've received alot of compliments on my tone. But, when I recently sparked up one side of my current build stereo guitar amp, same circuit as my 6V6 plexi, what I heard was alot more of what I "think" I want to hear - richer sounding mids, and a wider set of high frequencies that to me amounted to a less harsh and piercing tone and a more rounded and robust tone. Even my wife noticed a difference without my even asking. "Easier on the ears" she said. While I realize there are a number of reasons for this tone difference than simply an output transformer, I did get very curious what would happen if I replaced the "Chuang Meei" with the Musical Power supply "appropriate" OT that was part of the "better" sounding 6V6 plexi in my current project. My current project used MKT1813 caps, nearly all CF resistors and I did as an experiment use those MKT1813 caps for the treble and bright cap, which could very well explain the large difference in the presentation of high frequencies. But, I figured 63 bucks plus shipping wasn't a very expensive experiment even if it didn't improve anything. I can also move the "Chuang Meei" to my design board for a KT77 project I'm working up..

So, in a few weeks I'll be back with the report. OT's really are one component that seem to have big camps of opposing opinions on just how much difference an OT can, or would make in certain circumstances. Now, I get to find out a little of that myself and pass on whatever my findings happen to be. I really am looking forward to doing this.

I'll be back!
Best,
Phil Donovan (pjd3)

[R.G.]

Good plan.

There was a lot - a LOT - of interest in OT performance in the Tube Golden Age of hifi. It produced the usual stuff about "deeper sound stage", "presence", "clarity between the notes", etc., as well as some detailed science/technical looks into OT performance. The subjectivist camp did what they always did and still do. The objectivist/scientific camp dug into what and why with some very interesting conclusions.
1. Transformers have their own frequency limitations. Well, duuh, R.G. But it's calculable, based on measuring primary inductances, leakage inductances, differential and common mode capacitance. One counter to leakage inductances is interleaving. Interleaving reduces leakage by the square of the number of interleaves.
2. Transformers have distortion, mostly third order. This is an effect of the nonlinearity of the B-H magnetization curve. The closer you drive the iron to bending over into soft saturation, the more it limits peaks and the more distortion it causes to the volt/current/magnetic zoots that the secondary gets. This is countered by very, very low peak flux density in the iron, done by more turns and hence bigger primary inductance than needed for frequency response, and by introducing air gaps with a variety of stacking patterns. Any air gap also makes you meed more primary turns and this also increases leakage inductances... YIKES!!! This is primarily a bass effect, because the impedance of the primary inductance increases linearly with frequency, so, for instance, to get the same magnetizing distortion at 200Hz as at 100Hz, you have to drive the primary twice as far. It's mostly a bass effect.
3. Son of Transformers gots distortions: leakage inductances from half-primary to half-primary in push-pull circuits cause crossover distortion independent of and adding to AB biasing distortion. To get lowest distortion from this effect, you have to not only interleave primary and secondary, you need to interleave or multifilar wind the two half-primaries with each other.

So yeah - two OTs not of the same manufacture are likely to sound different.

My foray into designing guitar OTs wound up convincing me that probably no modern guitar OTs are going to be wound in enough sections and inter leaves to sidestep these issues. There's too much money to be saved by using less complicated windings.

[pjd3]

Hi R.G.

Thanks again very much for your typical comprehensive and historical coverage on OT's.

The aspects that you describe that can be implimented in the design of an OT for the specific things that its attempting to eleviate or gain certainly tell me that there are a good handful of things that are going to shape the response and percieved quality of an output transformer (as opposed to those that site a few elements of an OT that justify a "little difference" between most of them - I suppose at least for the improvements that are left out to save money for guitar level amps.

I while back when I took on my first tube amp build someone made apparent to me that, "What you are really hearing from an amp is the power supply, and that is just being maniputated by a signal that is but a remenant of the original guitar signal". That did make me wonder just how many things of a power supply, and in what manner the sound would be changed should those elements of the power supply be changed. I almost wish I had proof that the "Chuang Meei" was the cheapest piece of shit that ever floated over the ocean, so that I could with deeper judgement conclude just how much "better" or "not better" the Musical Power Supply OT22PP is, since I do know that many have in fact sang the praises if Matts work, materials, and performance of his transformers. By the sound of it he does use good materials and processes and shows good consistency. That was enough for me to trust grabbing one to replace the one thats their now.

So, in most peoples eyes (ears) I'm a fairly picky (sometimes aggrivatingly) about the smallest things in tone, and I will certainly be taking that to a good level when assessing the differences between these OT"s. I hope it does strike be as better because, well, better is better, and 63 bucks plus the light to moderate effort it will take to install it isn't quite enough to force a confermation bias onto me. You can bet I'll be watching out for that monster!

Thanks again RG, I'll be back sometime with the report!
Best,
Phil

[R.G.]

Yes, there are a bunch of overlapping and sometimes conflicting things going on in an OT. Juggling them to get a nice, smooth, wide bandwidth response is tricky. I have read in the hifi and guitar literature from the 1960s that using a hifi OT in a guitar amp was a noticeable step up in sound. Of course today we don't have the option to do that very often, other than by cannibalizing.

I think the idea that what you're hearing is the power supply is a little off. The measure of "goodness" in a power supply is how close it comes to making pure DC with all other artifacts suppressed. There's probably a case to be made that you hear the power supply's defects, though.

[...] I almost wish I had proof that the "Chuang Meei" was the cheapest piece of shit that ever floated over the ocean, so that I could with deeper judgement conclude just how much "better" or "not better" the Musical Power Supply OT22PP is, since I do know that many have in fact sang the praises if Matts work, materials, and performance of his transformers. By the sound of it he does use good materials and processes and shows good consistency. That was enough for me to trust grabbing one to replace the one thats their now.
There are some measurements that can give you clues from the outside. If you have the means to measure inductances, measuring the leakages from half-primary to half-primary, full primary, and primary to secondary can tell you a lot. Just showing that these are much bigger in one versus the other tells you it is less well interleaved, and therefore less attention and time was paid to winding it up.

If I had an Audio Precision rig to play with, I'd look at the distortion analysis across the bandwidth - but then, I don't have access to toys like that; Yes, I have test equipment envy.

I will enjoy reading your findings.

[pjd3]

Thank you R.G.

What I do have in my work lab (Zoll Medical, Defibrillators, Chelmsford, Ma.) is a rather expensive LCR bridge. I love that thing, measure inductance, capacitance/dissipation/Q, inductance, and impedance. I always bring caps in to look at their accuracy, consistency and dissipation, just to see how those stats compare, and how they differ from cap type to cap type. I don't know if that's enough to help assess all the specifications that you mentioned (could certainly measure inductance) OT performance but, it is here for the taking (measuring). I could take the time out to learn how to say, assess leakage inductance as that seems to be an element that is brought up quite frequently, here in work as well, too. I bet one of the engineers would be able to describe a method. Some of them are older and were very involved with transformers, and have even designed them for a living.

OK, Thanks again, RG, good things for me to think about and work towards better understanding their impact on these darn guitar amps!

Best,

Phil - I'll be back with my report, and I will bring them into work for some measurements on the LCR bridge.

[R.G.]

Wow! Access to the company's lab!

Yes - grizzled engineers that remember transformers will be a huge help. They were to me 50 years ago when I started trying to figure out what the devil transformers were doing.

Here's how to measure leakage inductance. Any winding when hooked up to an inductance measuring device will show just the inductance of that winding. Works the same on a transformer; any winding with all of the others open-circuited will show the inductance of just that winding. If you have for instance a three winding transformer, you can drive one winding from the inductance measuring rig and short either one of the other two windings. The magnetic fields of the core force all windings on it that share the core's field to have the same voltage per turn. Shorting a winding forces all the other turns, including on the other two not-shorted windings to have the same, essentially zero volts per turn. So if you drive winding 1 from the meter and short winding 2, the inductance you measure will be the leakage inductance between the driven and shorted winding. Winding 3, being open, has the same 0V per turn, but can't affect the measured leakage.

So you can measure three leakages: W1 to W2, W1 to W3, and W2 to W3. Drive W1, short W2, write that measurement down. Now open W2, short W3 and you get the leakage from W1 to W3. Now open W1, drive W2 and short W3. That's the leakage from W2 to W3.

What happens if you drive W1, short W2, then drive W2 and short W1? That's right, the leakages measured will in general not be the same. That's because leakages, like wire resistance and secondary loads are transformed by the square of the voltage ratio (which is the turns ratio). If W1 has 2X the turns of W2, then the leakage seen from driving W1 and shorting W2 will be four times the leakage measured by driving W2 and shorting W1.

On a typical OT, it can get complicated. You need to know the leakage from the whole primary to each tap of the secondary. You need to know the leakage from half-primary to the other half-primary. It's interesting but not critical to know secondary tap to secondary tap.

The best sound will usually be considered to be the lowest of all leakages. Exceptions might be people who have come to like the funny-ness of higher leakages, or the response dips and peaks that can intrude when you have higher leakage causing phase shift in the feedback for amps that use feedback.

With a primo LCR measurement device, you can measure the self capacitance of windings, and the cross-capacitances from winding to winding. These are generally small enough that they were neglected, even in the tube OT Golden Age, but they did get some notice when people tried to solve equations perfectly. For guitar, the response of pickups falling off to zip at a few kHz means that it's less important. Maybe for your favorite distortion, like say a Univox SuperFuzz or a Fox Tone Machine you would hear some differences, but guitar speakers typically don't extend up over a few kHz either. >>SPEAKERS<< are a huge effect on tone that sometimes people forget. Do your comparisons on the same speaker cabs.

[pjd3]

Hello,

Well the Musical Power Supplies OT22PP "oversized deluxe reverb" output transformer arrived yesterday, so I took it into work today. There wasn't much time, lots to do, but, I managed to slip over to the LCR bridge to just see what the inductance was for the primary's, Blue/Brown, Blue/Red, and Brown Red. The measurments also included "resistance @ 1MHz I believe it was, and one other measurement who's units I cant recall right now.

That was just more of a curiosity initiation but will be finding time to carry out the procedure that you left last post, jot them down, and when I replace the Chaung Meei, bring that in to do the same. Of course, they are of different power ratings and primary impedances but, I don't care, I still want to see what differences there are, and it there are any results that could tell me something about the status of their leakage inductance..

I'll at least get to the new OT this week.

Thank for those procedures! Im looking forward to measurements and what they can tell me.

Best,
Phil Donovan

[pjd3]

OK, little update here.

While I'm still working to understand the why's and how's this inductance leakage test works, I tried to follow through with what I "think" the instructions mean. I could be wrong but, here is some measurements of my new MPS OT22PP 30VA output transformer to replace the unknown Chinese I used for the 20W plexi I build a couple years ago and gig with.

Straight primary inductance (no shorting of other windings) (LCR bridge settings @ 1Khz/1.00V)

Brown/Blue - 5.6H
Brown/Red - 2.69H
Blue/Red - 3.00H


And here is applying the LCR bridge to Brown/Blue primary while shorting out each individual secondary tap with respect to secondary black common: (LCR bridge settings @ 1Khz/1.00V)

Green/Blk - (4 ohm tap) - 13.418mH
Yellow/Blk - (8 ohm tap) - 11.12mH
Org/Blk - (16 ohm tap) - 8.64mH

I also did a few more - the tap-to-tap shorting, but, I'll leave it here if this is enough.

So, I gather that when I remove the "Chuang Meei" and replace it with the new MPS OT22PP, I'll measure the same format on the Chuang Meei. Now, while I don't expect similar values since the Change Meei was designed for a 6L6GC pair @ 40-50 watts, perhaps I will be able to, with a little math look for a ratio comparison of "Inductance to Leakage inductance" as the various taps are shorted? I'll get to that when the time comes.

Thanks RG for getting me going on this, its pretty interesting already, and I look forward one day soon to see if any "quality" assessment can be made in comparing results, and if there might be any correlation of that to tone. It's a good insight into transformer performance and some good education for me.

Best,
Phil Donovan (pjd3)

[pjd3]

Hello!

Question. In the above test outline for inductance leakage, I believe it describes half-coil to half-coil leakage inductance for the primary. I'm a little baffled by how that is measured. In a typical OT primary, there are the outer wires, Brown, and Blue, and then there is the center-tap, the Red wire. If you short say, the Blue-to-Red wire, I assume you must apply the LCR device to the Brown, but then would the return of the LCR just go to the shorted ct and blue wire?

I ask because since it is only 3 wires, it presents a different relationship than measuring from Primary to any of the secondaries. In the latter case you have a separate return for both the primary LCR device, and a separate return for the taps and the common.

If you could clarify if I'm looking at the half-to-half primary leakage measurement correctly, that would be great. I'd just bring the transformer back in here and do that measurement as well.

Thank you for hanging in with me here!
Best,
Phil Donovan (pjd3)

[nuke]

Looking forward to more data. I have only a digital handheld DE-5000 LCR meter. It's pretty good for a number of tasks and a great value for the money, but doesn't do as much as a lab-grade, old-school LCR bridge does.

On the hi-fi history, output transformers were indeed the hot topic. McIntosh with their bifilar winding. Then the ever famous David Hafler Dynaco/Dynakit tube amplifiers, which were an exercise in well-deisgned, simple amplifier stages and very well made output transformers and were really great examples. The famous Stereo-70 and the Mk-III monoblocks are quite decent hi-fi amps even today.

Look forward to more readings.

I dug up RG's post on geofex about reverse-engineering an output transformers.

http://www.geofex.com/Article_Folders/x ... former.htm

Some good food for thought about taking measurements of existing transformers without specific lab tools.

[pjd3]

Not totally finished but little update,

New Musical Power Supply "over-sized 30VA Deluxe Reverb" OT is in.

It sounds good. I'm not sure if I'd call it "better" but it is different, in a good way I guess. The EQ actually seems a little different, hard to explain. But one thing it seems is that the amp was a little louder with the mismatched Chuang Meei in there, not that I need more volume but, its somewhat noticeable.

I'll be back with the "leakage inductance comparison". I did follow RG's outline best I could and will post an A/B comparison of leakage inductance values. I just need to bring the Chuang Meei in to measure its values on the LCR bridge. I asked one of the longtime EE's here about the method of measuring leakage inductance on a transformer. He stated the same method and interestingly, he described the action of shorting the secondary as akin to shorting the primary, as they somewhat become part of the same circuit, reflectively speaking, and that whatever inductance now measured would have to be that which is kind of lost. Something like that. But, it made a little more sense to me about how that measuring technique is actually working.

Thanks for hanging on this,
Best,
Phil D

[R.G.]

I asked one of the longtime EE's here about the method of measuring leakage inductance on a transformer. He stated the same method and interestingly, he described the action of shorting the secondary as akin to shorting the primary, as they somewhat become part of the same circuit, reflectively speaking, and that whatever inductance now measured would have to be that which is kind of lost. Something like that. But, it made a little more sense to me about how that measuring technique is actually working.

Hmmm. Smart guy. I never heard it described that way, but it's a completely valid way to describe it.
Shorting a secondary makes the voltage across every turn in the secondary be nominally zero. The magnetic field forces the voltage on every turn to be the same, neglecting only the amount of magnetic field that doesn't couple every turn.
So the primary is effectively "shorted" excepting only the non-coupled parts of the magnetic field through its own wire turns. All that remains is the magnetic field "lost" from coupling to the core.
I'd buy your EE a beer if I could.

[martin manning]

When I first ran across that virtual short explanation of leakage inductance measurement it cleared things up for me. The series resistance across the windings isn't really zero, so the short is not perfect, and there is some capacitance in there too. I assume the LCR meter can sort out the resistance and reactance, but I believe the reactance would be made up of the combined leakage inductance and capacitances.

[pjd3]

Thanks for coming by guys,

I believe that LCR meter does also measure Z and series resistance, or some related resistance. I'd bet it's the most important one to know in general.

In terms of volume differences, I hadn't mentally yet taken into account that there is a different NFB scheme now. Before this OT change I had implemented an adjustable NFB with 15K being about the smallest NFB resistor and I could get it up to around 100k or so. Now, it's a three-way switch to the 8 ohm tap - 22K, 15K and no feedback. Where this amp is a pedal platform of sorts, I kept the NFB generally on the lower side (high side, actually), plus a "No NFB" selection. That actually gets the amp in the ballpark for "fake early Beatles Vox type response". I just finished a 30 concert Beatles band tribute series so that was handy for early Beatles stuff. But surely the varying NFB levels have a very notable shift in volumes, and with the previous rendition of the amps NFB, I had no real way on the fly of knowing just how much NFB I was using - did it by ear.

I'll get the Chuang Meei in there next week for some leakage inductance measurements to see how it measures up. Now that you mentioned Z, which would have included resistance and capacitance, I wish I had measured those on the new OT but, I will measure that on the Chuang Meei just to see how it weighs in.

Thanks again guys,
Best,
Phil D.

[R.G.]

Yeah, the short-the-secondary process does leave in resistance and capacitance.
The ideal-transformer model lets you refine this pretty easily if you want more accuracy. You can measure the DC resistance of the primary and secondary windings, and reflect the secondary resistance into the primary, and compute the resistive part pretty accurately.

The capacitive part needs a little bit of math, as there are three capacitances to be taken into account; the primary self capacitance, the secondary self-capacitance, and the primary-to-secondary bridge capacitance. You can come close to measuring the bridge capacitance by shorting both primary and secondary and measuring LCR from primary to secondary. The result is the bridge capacitance in series with the primary leakage and the reflected secondary leakage. Self capacitance is probably most easily measured with either and LCR bridge or just a sweep generator and looking for the resonance.Induc

In fact, maybe the best way to really measure things instead of estimating them is to put a pure-resistive load on the secondary and drive the primary from a known-impedance source. You sweep the source frequency and look for the rolloffs. Some backwards math tells you the results.

In practice, even in the Golden Age, they ignored the capacitances in most circumstances, since the leakage inductance was the dominating factor in high end response. Not so for ultrasonic and RF work, but audio is pretty low frequency.

All of this is pretty much linear components modelling. In fact, the primary inductance is not constant, even within a single cycle of signal. Both the inductance and a core-loss loading resistance equivalent vary with the instantaneous volt-time value of the core material as it transits its B-H curve. Incremental inductance drops as the core approaches saturation, and core loss goes up. You might think that this would cause distortion. Well, yes, it does. There is a general third order distortion caused by M-field and saturation effects in the core. This mostly expresses itself as a nonlinear resistance in parallel with the primary inductance, where the resistance drops as you push it to higher volt-second levels. This is mostly a low-frequency effect, as at high frequences, you don't have as many seconds in the volt-second integral, even if you drive it to higher voltages. This is the basis of a transformer having a much higher power ability as frequency goes up.