UT 3670 Autotransformer

[Deang]

I called Mark Cooper to place an order, and in passing I asked, "You know Mark, I'm curious about something -- if you were going to build one these things for you, how would you build it?"

He began with a story about PWK and his Dad. I like those kinds of stories.

 

It seems that in the early days, PWK would visit them folks over in Texas, and he had some very strong opinions as to what he wanted to see in the chokes and autotransformers he was getting ready to buy. Mark told me that PK and his Dad would talk for hours, late into the night sometimes.

 

PWK wanted the DCR for everything as low as he could get it. Now this of course relates to money, because you can only get those numbers as low as you can afford to pay for them. I got the impression that PWK ended up with what he needed, not necessarily what he really wanted. Mark intimated that PWK really wanted the DCR numbers on both the 2.54mH choke and the T2A autotransformer to be lower. When I asked how much lower, Mark said he would have to ask his Dad. As you will see shortly, this may be a moot point.

 

The second thing PK wanted was for everything to be dipped in wax. Now, we all know about this one already, since it's called out on many of the schematics, and of course all of the original parts are packed with it. Mark told me that they quit using wax after Klipsch stopped doing business with them, and that customers normally prefer varnish because It makes for a better looking end product. Varnish is the sealant of choice, which protects the windings and prevents any exposed steel from rusting. It's an industrial varnish, and after the units are dipped, they are heated to 200 degrees. Varnish is used on everything from some of the larger transformers to all of the small audio parts wound for Bob, Al, and myself. When Al originally contacted them about winding the 3619, varnish was, and still is, the industry standard -- filling that tank with wax wasn't exactly at the top of their to-do list.

 

"So let me tell you about that tank."

 

"Okay."

 

"When we first built it, we poured water into it, and then heated it up to test for leaks, and it checked out okay. That's when we learned that wax is thinner than water."

 

"I don't understand."

 

"When we filled the tank with wax and brought up the heat, we started losing wax through some of the seams."

 

"So wax gets into places where water can't (I'm a quick learner), so, you gonna fill that tank up for me?

"I suppose we could do that."

 

"Anything else we should do?"

 

"Bigger wire. 1" frame instead of the 3/4" frame. That'll bring the DCR down to where, well ..."

 

Now everyone can argue about whether it was worth the trouble or not. Dipping in varnish mostly just seals the outer windings, where the idea is to prevent moisture from getting in. Wax on the other hand, does the same, plus removes any air gaps between the windings (units are submerged until there are no more air bubbles) and provides damping which prevents the windings from moving and/or "singing".

 

Mark Cooper sent me what I'm pretty sure he believes is what what PK initially wanted to see in the T2A, but couldn't really afford. This new unit of course incorporates the use of three input taps and the necessary output taps to allow for 18dB of attenuation in 1dB increments.

[Deang]

Pics.

[cradeldorf]

Nice!!! where are they heading? I tore apart an Altec N1600 crossover network once, it was potted in beeswax. They did things right back then.

Edited May 17, 2014 by cradeldorf

[Arash]

nice. I don't know if beeswax is any better than shellac.

I have an old compressor container. I fill it with shellac and put OPT or inductors or whatever I wind into it and put everything under 60PSI pressure. shellac will fill every molecule. just for 1 minute and then I put them in a chemistry laboratory center-fuse with moderate speed rotation. all extra shellac will be wiped out then put it in a warm place to dry.

[toolz]

Dean, Thanks for the information. Very informative.

[awsjr]

hi Dean... whats the cost of a pair of these ? ... thanks

[Deang]

Testing first.

 

I'm concerned that the reduction in resistance might require the use of different tap settings to achieve the same result. If so, it's not a big deal, but I need to know so I can account for it.

 

It's also large enough that it will have an impact on most layouts. Between there not being enough of a measurable difference between the other units, and its size, I may end up being the only person using it.

 

Though I'm a strong supporter of measurements, I've always believed that most simply don't have the resources and equipment to measure and test everything that the ear/brain mechanism can detect.

 

@Arash, shellac, like varnish, does not have a viscosity that makes it thinner than water. If your windings are loose enough, your method might accomplish the desired results. However, I've seen your work, and even though your air cores end up being scramble-wound, they appear to be nice and tight. I don't think shellac will be as effective at removing air gaps as paraffin.

 

@Cradeldorf, welcome to the forum. The unit on the left is basically what all of us builders are using right now. It is a better built unit than the original units used by Klipsch. It is bifilar wound, has additional input taps and output taps, and this single unit can be used to rebuild any Klipsch network, including any DIY effort utilizing any combination of horns and drivers - it's extremely versatile. The new unit is an attempt to improve upon it, so the first stop as for "where it's heading" will be Bob Crites' bench, and then possibly John Warren's if he's interested.

 

There has always been some debate as to whether the new units provide any real audible benefit to the ones we are replacing. We've collectively decided "probably not", yet we continue to replace them. I've been content to reuse all of the original units that are part of any PCB mounted network, even telling people there is no reason to pay for the replacement of perfectly good parts (I've felt the same way about coils). Any build from scratch gets a new pair of units by default.

All of the small differences add up. I don't think we should chuck things off to the wayside just because they don't make a meter jump.

[Deang]

Unrelated but related.

Think about this in the context of a coil inside a loudspeaker cabinet being exposed to insane levels of mechanical and acoustical vibration.

One might think not such a big deal with networks that sit out in the open under the top section, so I invite anyone to crank their system up and then stick their head back there.

I just received a return on a Universal I built. The midrange was no longer working. Once I verified that the capacitors were okay, I figured out that it was the autotransformer. So, I replaced it. When I tested the network, the midrange was still dead. I was staring at the board like an idiot, and started checking solder connections. While wiggling on a joint, the entire grounding bar started to move. What I discovered was that every screw had vibrated loose, and since polarity is reversed for the midrange -- there was no contact for the wire supporting the connection to the autotransformer. This blew me away, because that grounding set up is awesome, and I really tighten those screws down -- yet there was enough vibration to shake every screw loose. I'm grateful that I only did about a dozen this way, I went back to the old school method because I liked the idea of being able to crush/deform the buss wire under the square plate that's attached to the screw. Live and learn.

Edited May 17, 2014 by DeanG

[WMcD]

Not that I really know anything.

But it strikes me that transformers in power supplies and output transformers in tube amps might well get to a high enough temperature to melt wax just because of the heat that they generate internally. Also, in old type gear, the gear gets hot from vacuum tubes.

Therefore, varnish would be used as a potting compound for transformers since it has a higher melting point. Also, the wires have varnish insulation to begin with. And I think the soft iron sheet which make up the core, like the I-E pieces, are insulated from each other using varnish.

OTOH, the autotransformers in Klipsch speakers are not handling too much power and are not mounted on metal chassis of vacuum tube gear.

But why would wax potting have an advantage over varnish in the electrical performance?

WMcD

[Deang]

Hi Gil,

I only know what Mark Cooper shared with me: Paraffin is much thinner than varnish (which mostly just seals the outer windings), and displaces any air between the windings. This does a better job at protecting the windings, and increases damping - which reduces potential for resonating. Any advantage in electrical performance is probably limited to that.

 

Of everyone here, I was kind of hoping that you would be the one that might be able to shed some additional light on this. PK had a choice, and he was apparently pretty adamant about the wax thing.

I thought the video was pretty informative.

[Arash]

really good job. can you provide any details about turn ratio? what EI lamination size did you use? I always use Hi-B EI for autotransformers I think they are better with lower hysteresis and lower foucault current due to less thickness stacks

[Deang]

I'm not a transformer guy Arash, in fact, the more I read about these things, the more I realize how little I understand them and some of the problems associated with them. At any rate, I can probably get that information, but to be honest, I'm not really all that thrilled with the idea of sharing too much technical data related to this particular unit.

[Schu]

in...

Edited May 18, 2014 by Schu

[Marvel]

nice. I don't know if beeswax is any better than shellac.

I have an old compressor container. I fill it with shellac and put OPT or inductors or whatever I wind into it and put everything under 60PSI pressure. shellac will fill every molecule. just for 1 minute and then I put them in a chemistry laboratory center-fuse with moderate speed rotation. all extra shellac will be wiped out then put it in a warm place to dry.

Arash, I think a vacuum container would work better, and cause all the air bubbles to escape.

Bruce

[djk]

Most motor repair shops can do vaccume varnish, or using a pressure cooker at home is the next best.

[WMcD]

DeanG,

Thanks for the complement but I don't see anything to make me think that wax would improve any electrical performance.

There is a Dope From Hope where Paul is anticipating concerns about his shift to air gap inductors (and probably air-gap autotransformers). It seems that critics favored his earlier use of inductors without iron, thinking that saturation of an iron core would certainly cause distortion. Of course air wound inductors don't have a core to reach magnetic saturation.

It is my understanding that there is a tradeoff in using an air core versus iron or ferrite core. Without the iron core, you need more windings and that increased d.c. resistance. But if you use a core, eventually it will saturate. But in a good design, there is plenty of overhead before that happens. An air-gap iron core reduces the chance of saturation but might be reducing the overall effect of the core.

I did read something long ago. We know that voice coils have their own inductance of about 1 mH in woofers. The winding is sitting in a sort of iron core formed by the field magnet and pole pieces, which form a magnetic circuit.

Apparently, the strength of the magnetic field is regulated, or it maxes out, because of the pole pieces are saturated by design. This saturation causes a limit to the magnetic field in the gap where the voice coil sits. So the magnet is designed to max out the pole pieces by some margin. If the magnet loses strength, the field is still the same as long as the field pieces are saturated.

It seems to me that the pole piece and magnet is forming an air gap inductor, all on their own.

The other interesting thing, if I understand correctly, is the inductance of the voice coil is lower than otherwise expected otherwise. This is because the field caused by current through the voice coil can not cause a variation in the overall magnetic circuit. This is because the circuit is already saturated.

Just looking around the web, I see descriptions that old time transformers were potted in pitch / tar. It might be a melting point issue. I'd think that pitch could not really get into the winding to damp out movement and maybe varnish can't either. Therefore wax would do better for the mechanical issues.

WMcD

[DrWho]

Winding capacitance could be another reason - although it was probably more about cost.

[Deang]

Good stuff, I enjoyed that, though I can't say with all honesty that I understood all of it, and how it relates to our little wax issue.

I've spent the last week, on and off, searching through Google Books, reading old technical manuals.

 

Where heat was an issue, the wax would be mixed with special adhesives, or the coil would be vacuum impregnated with the wax, and then dipped in adhesive. The greatest concern was that of the coil "singing along" with the circuit, which required a great deal of rigidity and damping. So the stability of the mechanical structure (frame, laminations and winding) was paramount, and the wax was used to hold things in place. The electrical specifications and the behavior of the coil in the circuit was also dependent on the mechanical integrity, and again, wax was a factor here as well, because once it solidified, it provided the much needed damping. This applied equally to the air cores. I have dozens of Klipsch air cores in my work room, and they are all the same -- the wax feels like concrete, and the windings are immovable.

 

Some email with Bob today, which looked like this:

 

Dean > Bob: "None of the steel laminates or iron cores on the market are dipped in anything, and based on all of my reading, that's not a good thing. It may not be as big a deal for the big three, but it is for any crossover that's mounted inside a cabinet. Now, I'm making an assumption here, because according to Mark, it was the Klipschorn that was under discussion between PK and his Dad when they were spec'ing out the T2A, and PK called out wax, which was used in everything that came later. The laminations are subject to, and affected by vibration. They are also affected by temperature. The laminations and the wire have different expansion properties, and they tend to rub against each other a little. If things aren't solid, they move, and when they vibrate, they "sing" or resonate -- this was the biggest known problem with them in all of the old books I was reading (google books, old tech manuals)."

 

Bob > Dean: "I haven’t thought much about loose windings in inductors and transformers in a long time. I did have some of that training in school back in the 60s, but have lost most of those thoughts somewhere along the line. We used to have to think about that a lot in AM radio transmitters. If you have ever been around one of those, all of the parts in the transmitter output section sort of sing along with the audio that is being transmitted. All the coils had to be really tight, or they would rub the insulation off of the windings from rubbing together. I don’t think they were secured with wax though. It was hotter than blue blazes inside those transmitters, and I think wax would have just melted out."

 

Then of course we have also seen the potting of entire crossovers. I recently received an email with a question about a network that had been completely potted in epoxy. Is this taking things too far -- I don't know, some old crossovers were buried in tar.

[Arash]

I think that "ringing" or "resonance" means a piezoelectric effect. I have precise equipment at hand that can measure piezoelectric vibrations. inductors without wax or shellac are buzzing devices. coils move in microscopic dimension but they literally "move". movement in an inductor means the inductance factor varies. but this change is very small. and it's effect on frequency is even smaller. I never could trace it on a scope or something.

one thing is firm, magnetic coils move! even though they are dipped into shellac, wax, glue, concrete or whatever and then berried 1 km under surface of planet! molecules move.

Edited May 21, 2014 by Arash

[DrWho]

I could see EV potting stuff so it didn't break when a speaker was dropped / thrown around by touring guys.