Rewinding Bush DAC90 Audio Output Transformer

[David G4EBT]

I have absolutely no expertise in this area, but out of curiosity, I'm wondering why single insulated wire wouldn't be up to the task in this application, especially given the list of 'typical applications'? As a lay person, it seems to me that the wire needs to be capable of carrying the expected current, that the insulation will withstand the likely operating temperature, and that the breakdown voltage of the insulation is high enough to withstand the potential difference between turns, and layers of turns.

Where windings are machine wound and laid neatly side by side, the potential difference between turns - even on a mains transformer primary - will be a minute fraction of a volt. When we wind them ourselves, however neatly we try to lay many hundreds of turns of fine wire, they won't be side-by-side and some will come into contact with turns maybe two or three layers lower down, so the P.D. between turns may be greater, but even so, maybe no more than a few volts at worst. Surely that is only a fraction of the likely rated breakdown voltage of the insulaion?

Under normal operating conditions, though the transformer may get a little warm, it will surely be far far below the 180C (356F) maximum rated temperature. 40C (104F) is almost too hot to touch, and for an OPT to be getting anywhere near that, something would be very wrong.

What's more, I dare say that the polyester insulation of modern wire is superior to the original (wound maybe 70 years ago).

There may well be reasons why single coated wire isn't ideal in this application, but in my ignorance that eludes me so some explanation of why that might be so would be helpful. On the few transformers I've re-wound or wound, or field coils, I've only ever used solderable enamelled wire from Wires and Co (retail arm of Scientific Wire Co), which I assume is single coated. I didn't know there was such a thing as double coated, but I'm old enough to remember double silk and double cotton!

Chris has got 14.3 KMs (8.8 miles!) of wire to go at. A heck of a lot of primaries.

Good luck with the transformer 'MOT' Chris - 'moment of truth'!

[ChristianFletcher]

Thanks David

Well I will look for the grade 2 wire in the future now I have been made aware of it. I guess its part of the learning curve etc. I will post some pictures later today and a link to some video footage for anyone that is interrested.

Thanks Everyone for the help and advice.

Chris

[kalee20]

Originally Posted by David G4EBT

Where windings are machine wound and laid neatly side by side, the potential difference between turns - even on a mains transformer primary - will be a minute fraction of a volt.

Depends on the transformer. Perhaps 6 turns per volt, so 166mV between adjacent turns.

However, think of a turn at the end of a layer. As it is wound, the turns will traverse across the bobbin width, and then back again. So by the time wire is laid on to of our subject turn, it will have gained two widths-worth of volts, which is considerably more.

Bear in mind that older transformers will have a thin layer of paper between layers, which will fix the interlayer voltage situation, but Chris isn't doing this.

Originally Posted by David G4EBT

What's more, I dare say that the polyester insulation of modern wire is superior to the original (wound maybe 70 years ago).

It is. Which is why the interlayer insulation can be dispensed with.

The spec of the wire used is fine, although Grade 2 enamel is my personal preference. Grade 1 enamel can allow a wire size larger to be used, with the same overall diameter, which is important if the ultimate lowest resistance losses are required. FYI, my tables show 0.1mm wire has maximum OD = 0.129mm with Grade 2 and 0.121mm with Grade 1. The extra enamel means that it is slightly more forgiving to any scratches arising in the winding operation; slightly better resistance to cut-through damage, and of course slightly higher insulation breakdown voltage.

Don't forget that although most of the time the operating voltage is low, loudspeakers do occasionally get disconnected and people wind up the volume wondering why they can't hear anything, and the transformer can then develop well over 1kV!

[ChristianFletcher]

Thanks Kalee

Very helpful and useful information and comments. If I was doing this again I would certainly use the grade 2 wire and I will make a point of looking for this in the future. I did put some insulation in between layers but it was only every 1000 turns,. I only did this as a practice rather than for the insulation. It would have been better to have used the slightly thicker grade 2 wire and left out that insulation.

I guess my next question will then be once you have wound the transformer what testing do you do. I was thinking of insulation resistance at 500v but dont want to punch through the insulation. The way I am looking at this its just an experiment so I am in for the journey but hopefully the desination will ok.

https://youtu.be/QJM00hMW5jQ

Anyway here is my primary wiring first trial link above to YT

Chris

As the output valve is essentially constant current (being a pentode) if (a very big if, modern wire is very good) it does go short all that will happen is reduced (or no) volume. Give it a go...

[ChristianFletcher]

Thanks Merlin

Regards Chris

[joebog1]

Beautiful job for a first wind !!! I might say its very similar to how I wind a transformer such as yours. EXCEPT My Glow winding machine ( made sometime in the 30,s) runs the bobbin in the other direction. I feed the wire from behind the chuck/bobbin. With a little care you will wind exact side by side windings. AS far as insulation goes I think its mostly been said already. Modern insulation is generally 500 volt rated.

Once again, excellent job.

Joe

[ChristianFletcher]

Thanks Joe

I think I need to find a way to slow down the speed of my winding machine. It is converted from a old Dewalt battery drill. At the moment its set on the high speed and takes of a bit too quickly. The drill does have a gear box and its set at fast unfortinately the way i am clamping the motor makes the lower speed impossible to select. I think I may try to home brew my own pulse width modulated controller to try and give a slow jog function at startup etc.

Chris

[deliverance]

Excellent work Chris.

[David G4EBT]

PWM low voltage speed controllers that operate from 6 - 28V are widely available on e-Bay Chris and cheap too. I used one for my homebrew PCB saw.The pic shows the one that I used, mounted in a small ABS project box. It works very well and will slow to a snail's pace if needs be. )My own coil winder uses a sewing machine motor and foot controller).

Here are a couple of links - the controller in the second listing is reversible:

https://www.ebay.co.uk/itm/Regulatin...18a14b6c23441b

https://www.ebay.co.uk/itm/6-24V-DC-....c100005.m1851

Hope that might help.

Great video by the way - we'll done on winding the primary!

[kalee20]

Originally Posted by David G4EBT

Great video by the way - well done on winding the primary!

Second that! Holding the wire with soft cloth to prevent sweat and skin oils getting on is a good precaution - you might as well minimise the chance of it slowly corroding through over the next 60 years!

[TrevorG3VLF]

For speed control I have used a sewing machine foot pedal plugged into my lamp limiter and feeding a simple old battery charger which then feeds a DC motor. If I have a lot of turns which will take a long time, I add a Variac so it can run gently for hours.
I have plans of putting the PSU and counters in a battery charger case.

[joebog1]

My old winding machine is also powered by a 1/4 horse squirrel cage motor. It has a clutch VERY similar to an industrial Singer sewing machine. More pressure on the pedal, more speed to the chuck. I have a De Walt drill I would wager is same/similar to yours. I am even more impressed with your work as the De Walt drill is not exactly "fine tuned", more like cobbled together. When I hand wind I use cheap cotton gloves available from hardware stores for a few bucks, they are washable and they do keep the wire clean.

"Most" commercial winding machines have a "grain breaker". That is a series of pulleys that breaks the grain in the wire so that it forgets its spent years rolled up. For very fine wire like the guage you used on the primary, its probably not necessary, but will definitely help with the secondary.

The suggestion from David is a very good one. You can obviously use your hands, so to rig up a pot (entiometer) to a foot controller will be easy for you. Have a look on Ebay for winding machines. Most Chinese ones offered today are made using stepper motors and some software connected to a tiny computer. This is just to add to your education of winding machines.

Joe

[trobbins]

Chris, are you going to modify the original circuitry to alleviate the chance of damage to the output transformer going forward ?

[ChristianFletcher]

Many thanks for all the excellent advice and positive comments.

I am using a foot pedal and I used the original speed controller built into the drill. Thanks for the links to the other DC speed controller it probably worth me buying one to see if it works any better than the original one within the drill.

It can do a slow start it’s just very hit and miss, like the speed controller lacks resolution but it could also be the start up inertia of the armature. I was think it may be better to roll my own controller that has a wide low frequency pulse width modulation to try and jog the motor at low speed then switch over to conventional pulse with modulation at higher RPM. It’s a shame I cannot fit a encoder at the back of the motor and run it closed loop which is how the higher quality motor controller work in servo applications. Or I could just swap it out for a stepper motor which would likely be easier for me.

This coil winder is one of those ongoing tinkering project and sometimes good enough is good enough. I have to admit I have really enjoyed winding these transformers and I am thinking of importing one of the Chinese winding machines that can put the windings exactly side by side. I’m not planning on going into transformer production but I like tools and toys and think I would like to do a review. I have had my eye out on old winders on eBay but they come up infrequently and cost more than a fully automatic Chinese one for around £380.

Regarding modifications to the original circuit. It’s not my project the transformer is for my friend Graham AKA the radiocruncher. But Graham will be replacing all the grid coupling capacitors and the Zobel capacitor. I have a link below for anyone that wants to follow the story of this radio and transformer

https://youtu.be/xCCpUokGHJA

I must say I’m not total happy with the winding ratio of this transformer it doesn’t appear to match my math but I’m sure it will still work as valves aren’t fussy. I will post another comment below with regards this as I don’t have the math to hand.

Thanks best regards Chris

[ChristianFletcher]

Just reviewing my work and I am concerned in the diffference between the calculated winding ratio and the actual ratio I measured when I unwound the transfromer. I built a jig to unwind the transformer so I am quite confident that I counted the turns correctly.

I calculated a turns ratio of 31:1

I measured the turns ratio as 42:1

Can anyone help explain the difference between measured and expected values. I have attatched a sheet showing my math. Did the manufacture jus install whatever they had at hand etc.

Thanks Regards Chris

[Leon Crampin]

The Bush service data for the DAC90A gives the output transformer ratio as 40:1.

Leon.

[ChristianFletcher]

Thanks Leon that’s useful to know that my sample transformer as removed from a DAC 90 is in the ball park for how it’s been wound. I am trying to understand why it differs from my calculated value. Did the person winding the transformer have a tolerance would they have all been wound the same etc.?

Chris

[ronbryan]

My Mullard Data Book 1965-66 lists the Ra of the UL41 at several different HT voltages. For 200 Volts Va, it specifies Ra as 4,300 ohms, which moves the turns ratio more in the direction of 38:1.

Ron

[ChristianFletcher]

Thanks Ron

Thats funny I was speaking to a really nice guy this week at the vintage valve museum. He also said his research showed it was closer to 4300. But 3000 is a widely published figure

Running Rons numbers I get 37:1


errrrrm?

Thanks
Regard Chris