Forums :: Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Jul Wed 08, 2020 2:08 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 31 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Tue 07, 2020 5:14 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
Getting ready to rebuild the auto-transformer from a Colonial set. Here is the simplified schematic:
Attachment:
colonial_autotrans.jpg
colonial_autotrans.jpg [ 85.58 KiB | Viewed 1002 times ]

The original transformer had a DC resistance of roughly 10Kohms. We only were able to measure 1/2 the winding, so this is approximate. The original used AWG41 wire and I'm not inclined to count the number of turns in what is left....:)

Testing my understanding of how this works:
I assume that the load on the transformer is dominated by the 400Kohm grid resistors for the 45s. One is driven directly, and the other through the (1:1) transformer, so I think the equivalent is 200K.
The plate resistance of a 24A is roughly 500Kohms. Ignoring the 20K resistor, the signal attenuation is roughly 200 / (500 + 200) = 2/7. Assuming an ideal transformer, the DCR of the winding does not com into play.

So, if this above is correct, the only question is the number of turns. The original was quite obviously designed for the maximum number of turns within the available space---and assumed a masochistic coil winder that enjoyed working with #41 wire. The construction is the classic flat winding with a sheet of very thin paper between each layer.
My method will be the so-called "scramble wind". Modern wire has no need for interlayer insulation, so I can go up at least one wire size and get roughly the same number of turns as the original.

The only question is what drives the number of turns??---I would guess that it is simply frequency response. If so, we can compromise and maybe use #39 wire--trading bass response against reliability-- and abuse of the coil-winding operator (me).

Any words of wisdom before I order ~ 3 miles of wire?

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Tue 07, 2020 5:29 pm 
Member
User avatar

Joined: Aug Tue 28, 2018 9:22 pm
Posts: 2128
Location: Sanford Fla 32771 (USA)
I would like to follow this "rebuild". Can you post a picture of the original autotransformer with some size reference ?

_________________
Paul of Florida ….I had my patience tested. I’m negative.
https://paulsironhorse.smugmug.com/


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Tue 07, 2020 5:48 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
pauls.ironhorse wrote:
I would like to follow this "rebuild". Can you post a picture of the original autotransformer with some size reference ?

Will do!!

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Tue 07, 2020 6:47 pm 
Member

Joined: Nov Tue 14, 2017 5:09 am
Posts: 2704
Location: Austin, Texas
The actual number of turns is not very critical as long as you get enough inductance. Looking at the specs of typical audio interstage transformers, you should try for something like 70 H.

I would wind it bifilar using two spools of wire. That way you are sure of getting the same number of turns on each side without having to do a careful count.

Jay


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Tue 07, 2020 7:32 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
I thought of the bifilar approach...I'm not clear what the tradeoff is with respect to distributed capacitance. There's no significant voltage gradient so no issues there.
The other issue is the spooler---I have enough trouble spooling just ONE spool of #40 with the wire breaking. But then my setup is a bit "shade tree".....

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Tue 07, 2020 7:48 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
Here's a partial reassembly showing the overall architecture. When fully assembled, one leg of the core is 1/2" X 9/16"
Attachment:
colonial_autoxfmr.jpg
colonial_autoxfmr.jpg [ 126.66 KiB | Viewed 967 times ]

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Wed 08, 2020 12:37 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 14289
Location: S. Dartmouth MA 02748-1225 USA
Magnet wire sold in larger "mill" spools will be loaded in a round flange spool of which the flange is considerably rounded. During the high-speed winding the reel is flat and the wire pays off of the side of the reel. The reel is supported on friction reducing bearing if it needs to turn...

GL

Chas

_________________
List' & I will Enchant Thine Ear


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Wed 08, 2020 3:33 am 
Member

Joined: Jan Thu 16, 2020 12:29 am
Posts: 574
pixellany wrote:
Getting ready to rebuild the auto-transformer from a Colonial set. Here is the simplified schematic:
Attachment:
colonial_autotrans.jpg

The original transformer had a DC resistance of roughly 10Kohms. We only were able to measure 1/2 the winding, so this is approximate.

Any words of wisdom before I order ~ 3 miles of wire?


Here are some words, possibly wisdom:

First way to estimate the total number of turns:

It is fantastic that half of was still intact, that is all you need (I hope its its still intact). You can measure the inductance of it easily (not with an inductance meter,won't normally read high enough or accurately enough, but some can) Do a resonance test on it with minimal loading, apply a signal generator to it with minimal loading, use two series 1 meg resistors in the earthy end and connect the scope across the lower resistor. Add a parallel capacitor across the coil that is large enough to swamp the self capacitance out by a factor of 20 at least, so use at least a 0.01uF capacitor. Find the resonant frequency and with the usual formula calculate the inductance.

(After that a resonance test without the parallel capacitance added; will then also allow you to calculate the approximate self winding capacitance of the winding half that is still working)

Since inductance is proportional to the square of the number of turns, you will then know the approximate inductance of the whole thing which will be 4 times what you have calculated from the above experiment.

Make a temporary former and wind some number of turns thin onto it, say 100 o 500 turns and document the turns. Measure the inductance on the inductance meter of your test coil (or with the resonance method) Again, keeping in mind that inductance is proportional to the square of the number of turns, calculate the total number of turns of the original winding.

Second method:

Of the winging half that is still working, look up the Ohms per foot of the wire. Make the best possible estimate of the average length of a turn, by looking at the winding cross section. (A turn will be shorter closer to the core and longer further away) you need what would be the rough average, so it would be a turn that bisects the centre of that cross section. Measure the length of the winding and divide into that the diameter of the wire (including enamel) or unwind one layer to find out.

The average turn length x turns per layer x n = total length. Total length you get from the Ohms per foot of the wire.

n is the number of layers you can now calculate from that data.

And therefore the total number of turns is turns per layer x number of layers.

Then with this data you can wind a test coil with some other number of turns and calculate the expected inductance of the winding half you are working with, and therefore the inductance of the whole thing which will be 4 times higher.That should roughly agree with the resonance test data.

Without the paper insulation there will be a little more space for slightly thicker wire, but only by a little:

For a fixed volume bobbin, the number of turns you can fit in there is inversely proportional to the square of the wire diameter (or radius). So if you double the diameter of the wire, you will only fit in 1/4 the number of turns on the same bobbin, using all the space.

In addition; the resistance of the winding has a much steeper relationship (one of the steeper ones in physics, like some of the laws of fluid mechanics). This is because, two things are at play. If the number of turns that you can fit in some fixed sized bobbin cross sectional area is inversely proportional to the square of radius of the wire, AND the resistance of the wire is pl/A (p is resistivity, l length , A cross sectional area of the wire, or pi.R^2) then the resistance of the winding you can make fit in that fixed cross sectional area is an inverse 1/R^2 x 1/R^2, or 1/R^4 relation.

Meaning that, if you fill the same bobbin up with a wire twice the diameter, you would have 1/4 the number of turns you had before and surprisingly 1/16 the resistance of the original. If you have 1/4 the number of turns, you will only have 1/16 the inductance you had before because inductance again is proportional to the square of the number of turns.

So what I'm getting at is the effect of changing the diameter of the wire, affecting both resistance and indutance is significant working in a fixed sized bobbin and has a 4th power relationship. Another example, Lets say the wire thickness is increased by a factor of only 1.2, and the bobbin is full, then you will have about 1/1.2^4 about 1/2 the resistance of before, and about 1/2 the inductance too.

So initially at least I would go to only one thicker sized wire for the rewind, even if it is jumble wound with no inter layer insulation.

Hugo.


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Wed 08, 2020 10:14 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
Thanks for all the inputs!!
Based on this and my own calculations, the plan is to go up one size to AWG40 and put on as many turns as possible.
I'm still considering the bifilar option, but a little voice is saying keep the design similar to the original.

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Wed 08, 2020 10:31 pm 
Member

Joined: Jan Thu 16, 2020 12:29 am
Posts: 574
pixellany wrote:
Thanks for all the inputs!!
Based on this and my own calculations, the plan is to go up one size to AWG40 and put on as many turns as possible.
I'm still considering the bifilar option, but a little voice is saying keep the design similar to the original.



I agree. Bifilar winding is helpful to keep the leakage inductances between the two primary halves low, say if it was two halves of a primary in a switch-mode psu and that prevents overshoots on switching and it makes the resistances of the windings exactly the same. However, in this application though, it is not an issue and if the resistance of one primary half (to think of it that way) is higher than the other, it will have little effect as the input resistance to the 45's is very high. The voltages presented to the two grids will be very close to equal & opposite.


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Wed 08, 2020 10:44 pm 
Moderator

Joined: Jan Thu 01, 1970 1:00 am
Posts: 30590
Location: Livermore, CA
That method of winding sounds good but doesn't work for audio interstage transformers. Both AES and PTOP tried it and high frequencies didn't come through. Too much capacity between windings along with high impedance.

_________________
Norm


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Thu 09, 2020 9:10 am 
Member

Joined: Jan Thu 16, 2020 12:29 am
Posts: 574
Norm Leal wrote:
That method of winding sounds good but doesn't work for audio interstage transformers. Both AES and PTOP tried it and high frequencies didn't come through. Too much capacity between windings along with high impedance.



Inter-stage audio transformers for vintage radios is an interesting topic.

I have researched this extensively and found that the equation that predicted the frequency response of a valve inter-stage transformer was:

1) Not published in any classical textbook and

2) not on the internet (as a result of 1)

So , using an equation set from Terman's Radio Engineers Handbook I was able to derive the equation.

If you are interested in this equation (and the implications of it), the factors which affect the response, have a look at page 15 of this article:

http://worldphaco.com/uploads/THE_GREBE_MU-1.pdf


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Thu 09, 2020 7:53 pm 
Member

Joined: Sep Thu 23, 2010 6:37 am
Posts: 12437
Location: Powell River BC Canada
What is the model of the set?
What is the core area?
What do(es) the surface(s) of the core laminations look like ?

_________________
de
VE7ASO VE7ZSO
Amateur Radio Literacy Club. May we help you read better.
Steve Dow
ve7aso@rac.ca


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Thu 09, 2020 8:48 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
radiotechnician wrote:
What is the model of the set?
What is the core area?
What do(es) the surface(s) of the core laminations look like ?

--Colonial 36
--1/2" X 9/16"
--smooth

-I'll be back to this after finishing some chores....

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Thu 09, 2020 9:15 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 14289
Location: S. Dartmouth MA 02748-1225 USA
P. P. Input transformer, do not use the primary. Or, rewire the circuit to use the entire transformer and nix the grid coupling condensers :roll: .

_________________
List' & I will Enchant Thine Ear


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Thu 09, 2020 10:24 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
Chas wrote:
P. P. Input transformer, do not use the primary. Or, rewire the circuit to use the entire transformer and nix the grid coupling condensers :roll: .

A valid point...the decision in this case was to preserve original appearance and performance. With regard to the latter, i have decided to duplicate the original winding configuration as closely as possible (while using #40 instead of #41 to minimize the chance of breaks)

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Thu 09, 2020 11:18 pm 
Member

Joined: Jan Thu 16, 2020 12:29 am
Posts: 574
It is a much better idea doing what they have done using a centre tapped choke (auto-transformer) as a phase inverter and driver device rather the a coupling transformer or another tube stage. The reason is there are no issues related to the leakage inductance between the primary and secondary (which the coupling transformer has) which have a significant effect on the audio frequency response. This is why, back in the 1920's many chokes were used as "output transformers" as well. Typically' such as in this amplifier I made with Samson transformers a tapped choke rather than a transformer is used as a push pull output device.

http://worldphaco.com/uploads/UX-171-A_ ... ifier..pdf

The advantage of the typical audio inter-stage transformer though, especially in class A or AB1 operation is that you have DC isolation, eliminating capacitors, and voltage magnification, typically 1:3 or 1:5 and you get that voltage gain. It relies on a few things though, especially that the tube being driven does not ideally go into g1 grid current or toward class B operation. The frequency response (as shown in the equation in the article I cited on the previous post) is dependent very much on the plate resistance of the tube driving the transformer primary and the self capacitance and inductance of the secondary, these are the dominant factors. For example if the plate resistance driving the primary is lowered, the frequency peaks up in the upper audio range. The voltage magnification is helpful in simple radios, running short on tube numbers, where every scrap of gain came at a premium.

On the other hand, audio coupling transformers for class AB2 or B operation are typically more like 1:1 or step-down types as some grid drive power is required in this mode. If you are already at 1:1 then an auto-transformer (tapped choke) is very attractive as a push pull driver device as it eliminates many of the frequency response issues a coupling transformer could have and gives a flatter audio response. It is also fine for driving tubes in class A push pull with minimal grid current too or class AB2 & B. You get a flat response in the audio, no voltage magnification (as the trade off) and no DC isolation, but the latter is solved with coupling caps and there my be plenty of gain available in the tube stages.

Class A push pull was once quite popular in the 1920's but the designers rapidly got on to the notion of the extra power output and efficiency that could be attained with class AB1 & AB2 & B, to the extent that tubes, with very low grid bias requirements were specifically designed for class B operation, for example the type 49, which runs in class B with zero grid bias obviating the requirement for a bias battery or bias supply. If you were tempted to make a push pull amp with these it is important that the driver transformer for them is a step-down type, which most class-B driver transformers were in that era.

One other thing I could have mentioned about the auto-transformer used here. Due to the very high inductance and the not insignificant self capacitance, which can be calculated for half the winding from the resonant frequency experiment on my previous post - whatever the self resonant frequency is you get for that working half, the self resonant frequency of the whole thing (when working) will be lower and close to 1/root2 or about 0.7071 of the value you get with just the self resonance of the working half. The reason is that for the chokes as a whole the capacitance has halved across the end terminals but the inductance increased by a factor of 4 as the turns doubled.

This was in fact answered some years ago with the exam question: If you have two near identical windings, very tightly coupled on the same core each of the same inductance L and the same tuning capacitance C across each, what will be the resonant frequency and the answer is 1/2.pi.root(LC) x 1/root(2) In other words the resonant frequency has dropped compared to there just being the one winding and one capacitor. It is actually much more complex if the coupling coefficient is not close to 1. The easiest way to visualise how this comes about is to link the two resonant circuits together, with just one wire (like the center tapped choke), so it does not create a new circuit and you have an inductor with 2 times as many turns & 4 times the inductance and half the tuning capacitance.


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Sun 12, 2020 2:56 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
I finally got this thing deciphered.....

Here is the "inner" half of the winding, after removing a large pile of fine wire and wax paper. It is electrically intact and measures 5Kohms as expected. Based on the physical measurements, it decodes to 14,000 turns of #41 with a "K" of 3.2.
Attachment:
inner.jpg
inner.jpg [ 91.42 KiB | Viewed 653 times ]

"K" is my version of a safety factor. They way I am doing it, it means that the winding height is 3.2 times the formula for "jumble-winding"**. This is consistent with the fact that it was wound with a layer of wax paper between every layer of wire.

So, the objective is to get a minimum of 28,000 turns for the whole coil, while fitting the available volume. The solution I'm going with does this with #40 "double-build"*** wire, with a K of 1.5. With a crude test with my winder, I can get K of around 1.2, so this give me some margin. The plan will be to wind to 1/2 of the available height, note the # of turns, put in the leadout for the center-tap, and then add the rest of the winding.

So, I'm off to order 1.54 miles of #40......



**The formula I am using for "jumble-wind": height = K * d^2 * N / W, where d is the wire diameter, N is the number of turns, W is the width of the winding, and K is my fudge factor. I have found the basic formula in several places.

***Modern magnet wire comes in a variety of insulation types. I use the wire that is rated at 155C and has a solderable polyurethane insulation. The "double-build" variety adds a nylon layer which improves the abrasion resistance without losing the solderability. Considering the "shade-tree" nature of my winding setup, this provides a margin of safety. The one compromise is that the supplier (Temco) offers only the even AWG sizes when you get about 30-something. Ideally, I would have chosen #39 double-build, but I can--with a bit of patience--wind #40 without breaking it......

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Sun 12, 2020 8:19 pm 
Member

Joined: Sep Thu 23, 2010 6:37 am
Posts: 12437
Location: Powell River BC Canada
Thank you for sharing the photo.
From the hookup diagram on the Riders page, have you determined
which end goes to which terminal, for both primary and secondary
with the core assembled exactly the same as the original mounting ?

Of great interest is what the laminations are made of, and how they
were treated.


Also of interest in that radio is whether the tapped choke circuit
permitted the 24-A power detector to drive the 45s, without a
need for a separate tube (i.e. , 27 )

If you get around to it, a photo of the speaker would be nice.

_________________
de
VE7ASO VE7ZSO
Amateur Radio Literacy Club. May we help you read better.
Steve Dow
ve7aso@rac.ca


Top
 Profile  
 
 Post subject: Re: Rebuilding an interstage AUTO-transformer
PostPosted: Apr Sun 12, 2020 8:40 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 28398
Location: Annapolis, MD
radiotechnician wrote:
Thank you for sharing the photo.
From the hookup diagram on the Riders page, have you determined
which end goes to which terminal, for both primary and secondary
with the core assembled exactly the same as the original mounting ?
.

If I'm reading it correctly, it's a 1:1 transformer---actually NOT an auto-transformer since the center-tap could be bought out a s 2 separate wires without affecting operation. I think it is exactly the same functionality as the common recommendation for conventional interstage transformers with open primaries--to wit: hook up the signal to one end of the secondary, using the secondary to generate the opposite phase.
Lamination material: How would I find out?
Speaker: I have only the coil

_________________
-Mark
"Even if you don't understand Ohm's Law, you are still required to obey it."


Top
 Profile  
 
Post New Topic Post Reply  [ 31 posts ]  Moderators: Marcc, Norm Leal Go to page 1, 2  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: Google [Bot], lorenz200w, Lou deGonzague and 15 guests



Search for:
Jump to:  


































-->


Privacy Policy :: Powered by phpBB