Transformers from Heaven! Or Brazil.

[CJ]

Ahh yes. That's the bunny. I have to build one of these guys. Hopefully I can get some good iron out of this batch and send you a few. It looks like this vacuum design has hints of things to come in the way of transistor design and iron. Same core, balanced circuit, etc. I bet the lead in the pic that goes to both coils is that centertap. Look at the cathode circuit on the first stage. Looks just like the AM 16. Same with the feedback. Where do those X znd Y arrows go to?



Looking forward to that input!

I have tried the thinner method also, and it just makes things worse. Melts all the paper, but not the sticky icky.

Tonite. Can't wait to un wrap those bad boys!

Yeah, you guys butter me up in hopes of me winding vintage ebay iron for pennies on the dollar, I know what your up to folks! :green:

 

[rafafredd]

Where do those X znd Y arrows go to?

This is just to test plate balance between triode sections, I believe...

 

[CJ]

OK, it goes to the meter. I see.

Raf, to you have an amp there that you could look at? It might help if you could match up the color codes on the wire with the terminal numbers.
Not abosolutely necessary, but the more information the better.
Thanks!
cj

 

[rafafredd]

Sorry CJ... When I took those cards home, the output transformers were all out, and the amps were working with a strange mod, transformerless outputs.

 

[CJ]

Thats right. I remember the pics.

No biggy.

I am making slow steady proress. These outputs really got cooked by those BH7's. The secondary winding keeps breaking every 30 or 40 turns. The cooked enamel, interlayer rice paper, and impreg compound mixed together to form a real nasty goop, but it's coming apart. The finish to coil A is that red/yel splice that goes to both coils, so this is so far very mych like the AM 16 strutcure.

0.0030 - 0.0035 wire, about 242 turns per layer, I am on my 6 th layer, giving the eyes a break. Back and forth machine wound, not bi-filar, yet. Maybe the pri's, which I can't wait to get to, because it will be larger, non cooked wire that won't break. Have to heat coil up after each layer is un-wound.
But life could be worse. I could be living in Florida. :green:

 

[CJ]

OK, I cracked this nut on the first try. Coil A at least.

This was by far, the gooooiest take apart I have ever done.
Any baseball players out there? You know that stuff you use on the bat to keep a grip? I think it's called pine tar. George Brett got busted for too much of a while back. Remember? Good one, Billy!
Well, I think Langevin's x-former designer was George Brett.
Man what a mess. Anyway, I did a hand count of each turn.

Instead of cool color graphics, it's late, so a description until I get something in a pdf or jpg.

Twin coils on a C-Core.

Coil struture: 1/2 Pri - Sec - Sec- 1/2 Pri

Coil A

- Inner pri. - 2857 turns - 12 layers of 0.0025 ( whatever gauge that is back then)

- Inner Secondary (remember, this is an output, so the pri has all the turns to match up with the plate and the sec is the line winding )

734 turns of 0.0050 wire 5 layers

- Outer Secondary 396 turns 3 layers.

- Outer pri. - 2857 turns 12 layers of 0.0025 wire

- Outer Primary Finishes from both coils are brought together and used as the center tap.

The secondary windings are split between coils. In other words, The low Z tap for a sec is on Coil A, the hi z tap for that same sec is wound on Coil B. And there are two sec windings available, so that's four sec windings total. Splices connect the sec sections.

There is no reverse winding or bi-filar windings on this output transformer.
Just a lot of pine tree sap.

So, 5714 total pri turns, 1130 total sec turns, = 5.05/1 ratio.

Just call it a five to one output transformer.

Sec has 5 Henries with everybody lined up, so figure the pri has 5 time 25 equals 125 Henries.

Coil B just un raveled and it is the same as coil A.

 

[rafafredd]

Great CJ!

You are the biggest friend of DIY!

A drawing would be cool if it´s not asking too much.

Thanks for this, man!

 

[CJ]

Yeah, gonna get all your starts, finishes, splices and color codes dialed in.

Gotta get a winding machine so I can try to send some of these back!

Neat thing is, they use the exact same core as the AM-16, so you could use the bobbins and core and the can, and wind an AM 16 output.

Can't wait for the input!

 

[CJ]

Here are a few pics of the output dissection that I took last Saturday.

Lead breakout:

Note that the green lead goes to both coils:

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/langevin_raf_20.jpg

You have to cut some leads to split up the coils for reversing.

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/langevin_raf_22.jpg

Margins are the distance from the winding to the edge of the insulation. They are usually listed on a typical transformer winding blueprint to tell the winder how far in to start the winding. Most of the margins I have seen have been symetrical. They are a good thing to keep track of as far as measuring, if you want to submit plans to a winder.

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/langevin_raf_24.jpg

Winding voodoo? Maybe. Talk to John Hall. Or whoever did this work for Langevin.
The funny part about this layer is that it looked exactly the same on Coil B as it did on Coil A. We will never know why this was done. The winder who did this is long gone, no doubt.

The beezwax, or whatever the potting compound used was, did not penetrate all the way to the center of the winding, as can be seen by the interlayer insulation. Some winders do not like varnish or other compounds to be used with the interlayer insulation, as it can change the dialetric properties of the insulation.

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/langevin_raf_30.jpg

OK, doing a winding chart if anyone interested.

On to the input!

 

[CJ]

OK, here goes the input to the Langevin Tube Pre.
I was sent a TF-432B from our buddy in Brazil. It was the hardest dissection I have ever done, but not from a complictaed structure point of view.

It has a very nice 1 khz sq wave pattern:

One of the secondaries was open, but I did manage to get a voltage ratio of about 1:6, that is both primaries in series, and reading the voltage on the lone secondary.

Pri inductance was about ten henries per primary, which is about 40 henries with them in series. Pretty healthy. That means secondary henries for one winding will be 1:6 . 1:36 imp ratio which is then 36 times 40 which is 1440 henries per secondary. That's pretty healthy also. Actually it's downright crazy considering the core it was wound on.

Here it is, the Blob #2.
That's a copper shield around the coil/core structure. Underneath the copper is another mu can, so pretty good shielding on this guy.

Another pic of the cans:

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_7.jpg

Opened the can up:

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_8.jpg

This gorilla snot was a pain in the backside:

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_10.jpg

Thats the core exposed with some orange mylar tap around it. Lead plate...

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_11.jpg

 

[CJ]

Nice solid core ground ....

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_14.jpg

It even had a strip of copper surrounding the lam to solder to. Only time I have seen this extra care taken. Lam type 24-25 EE. It is in the magmet catalog. Most likely 80% Ni. 0.014 mil thick. 26 ea. A very popular lamination used by Sowter, Jensen, API, DuKane, you name it. Nice small magnetic path for low core loss and low capacitance due to keeping the turn length to a minimum.

Heat was applied to get the lams out. Ready to dissect the coil. Pri leads are on top, sec's below...

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_15.jpg

Leads mapped out again...

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_16.jpg

Removed the outer coil wrapper. It was stamped with a number. Not much of a winding length, eh? Couldn't be many turns on a winding that is less than 3/8 inch tall right?

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_17.jpg

A secondary winding here. Pretty thin stuff. #48 wire folks. Most winders give you a hard time about winding with #44. This is #48!!! 0.001 inch dia. or less!!! A real pain in the ass!!! And I don't mind saying so!!! And it's impregnated with a high temp wax that is the stickiest junk you will ever encounter!!! This means that the miniscule break strength of the #48 wire will prevent unwinding of the coil unless the wax is heated up every two seconds. Which is not normally a problem. But, there are 36 layers of #48 on the outside of the coil, and 36 more layers next to the core!!! And 172 turns per layer!!! Thats 12,384 turns of wire that is thinner than you hair and cemented together with a potting compound from Hell!!!!!

But hey, this is the Lab, bi-atch!

The primaries are wound together with a much easier to work with bi filar winding of #44, or 0.0015 inch dia. wire. :twisted: I thought I would catch some slack when I got to the primaries, but no! Actually, after dealing with 6000 turns of #48 breaking every two seconds, #44 was like finding an oasis in the Sahara Desert. And there were only 512 turns on ten layers, so it was a piece of cake. But then there was another secondary inside, so.... Here is a shot of the bi filar primaries. (bi filar just means two wires wound at the same time, side by side).

http://vacuumbrain.com/The_Lab/TA/Langevin/Raf/Input_Transformer/langevin_input_20.jpg

Here is a shot of the inner secondary. A pencil and exacto blade for referance:

OK, get to the bottom line you say:

Single coil on 24-25 EE 80% Ni lams-0.014-26 ea.
Structure: Sec - Bi Filar Pri's - Sec
Pri. - 51 turns of 0.0015 wire - 10 layerrs for 510 turns each pri.
Sec. - 172 turns of 0.001 wire per layer, 36 layers each sec. for 6192 turns each sec

Pri starts - Red and Black - Terminals 3 and 4 respectively
Pri Finishes - Orange and Brown - Terminals 1 and 2 respectively

Secondary Starts - Yellow and Grn/Wh -
Secondary Finishes - Blue and Green

Inner wound secondary, next to the core, is Yel - Blu
Outer wound secondary is Gr/Wh - Green

Note:
With this winding structure, the DCR of the primaries are 3.92 K inner, and 5.66 K outer, for a delta of about 1.74 K ohms. That's a pretty big DCR imbalance, but it does not really matter, as the secondarys hit the grids of a tube, which have a cvery high impedance.

So there ya go.
Any questions? Call 911 cus I'm out.

cj

 

[rafafredd]

WOW! :shock: This is really impressive stuff. I´m speechless...

 

[12afael]

HEY WHERE YOU GET THAT GIGANT PENCIL AND EXACTO BLADE!!!!!

liliput work...
Great Work CJ

Rafa do you think is posible to find good lamination on Brazil? some good for this?

 

[rafafredd]

Yes, there is very good C-core M4 material for the output transformer in Brazil. Better than what was originally used by Langevin, that is probably M6...

But nothing for that input one. Who would wind one of those inputs anyway... :?: :sad:

But I´m going for a run of 10 outputs with the M4 C cores... Guilherme will wind it for me.

 

[CJ]

The output of the tube and solid state Langevin pres both use an MH-40 (mag met) core, which is 4 mil.

 

[Snatchman]

Dang... I had thought about "dabbing" into some DIY projects until I saw this thread! :shock: ..I'll leave this stuff to you all geniuses.. I'm off to build me a house. Call me when you make something, I'll buy it from you all!.. :grin:

 

[Bluzzi]

CJ -

I hear Raphael would like the transformers he lent you back when you get a chance...!

(sorry couldn't resist).

Jim

 

[rafafredd]

:?: :roll:

 

[Bluzzi]

Raphael - That was supposed to be a joke!

Jim

 

[rafafredd]

Sorry, I just don´t get it... :? Sometimes I´m bad about jokes , specially in a forum, and in a foreign language :wink: But I´m trying. I hope I didn´t sound too unpolite in my last post.

Now, CJ, send me those fucking vintage units back now!