Winding my own output transformers

[jonasnoble]

Hi everyone! i've done my first transformer using the Delatsch app.
i've tried to make a 10kct:600 transformer, but i don't really know how to interpret the center tap. i have to split the 10k into two 5k or as a two 10k windings?
thank you!
i've bought everything from italtras, they're great, they respond fast to emails and very fast shipping here in Italy. i strongly suggest them

I'm sure somebody here knows more than me, but I'm pretty sure it's 5000/5000. If it was me, I would wind the first half primary, then the secondary, then the second half of the primary.

 

[xazrules]

thank you jonas! i've done like you described, and the freq is super flat, tomorrow i try to record some graph with roomeqwizard.
this was my very first trafo and is easy with the delatsch app, but for sure things can improve a lot, especially not over enginering so cheaper costs for production. i calculated how much it costs me a 10kct:600 5W at 50hz with EI66 lams, 66grams of 0.16mm copper 3 euros, M6 grain oriented EI66 lams for a weight of 0.7 kilograms 7 euros.
i've remained with a lot of free space, i could used a bigger wire or do more turns, let's see how it works as a tube output

 

[jonasnoble]

thank you jonas! i've done like you described, and the freq is super flat, tomorrow i try to record some graph with roomeqwizard.
this was my very first trafo and is easy with the delatsch app, but for sure things can improve a lot, especially not over enginering so cheaper costs for production. i calculated how much it costs me a 10kct:600 5W at 50hz with EI66 lams, 66grams of 0.16mm copper 3 euros, M6 grain oriented EI66 lams for a weight of 0.7 kilograms 7 euros.
i've remained with a lot of free space, i could used a bigger wire or do more turns, let's see how it works as a tube output

That's great! I'm right behind you, as I've been doing the prep work but haven't wound one yet. It sounds like you have been successful on your first attempt, so I may reach out if I have any questions. First, I need to buy some bobbins, but I haven't decided the size yet.

Have you put square waves into it yet? I am curious about the best way to test transformers to predict audio quality.

 

[abbey road d enfer]

I'm sure somebody here knows more than me, but I'm pretty sure it's 5000/5000. If it was me, I would wind the first half primary, then the secondary, then the second half of the primary.

It may be counterintuitive, but when you CT a 10k winding, it consists now in two 2.5k windings.
That's because the inductance varies with the square of the turn number.
Inductance Lnom=Al.N²
So if the calculation gives Nturns for X Henries, when you split the winding in two halves, each section now measures 1/4 of Lnom.

 

[xazrules]

I call it a success only if I see no smoke once hooked to plates..
I think the proper way to test them is to have an oscilloscope and a function generator that can output tons of voltage. I have none, I’m going to square wave it once in the circuit to see no ringing, but in this case even the termination is very important.
There’s a cool video on YouTube of a tech that test a trafo, search for audio transformer distortion or something similar.
The size of the core dictates the wattage you can squeeze out of it and the size of the wire you can use, the bigger the window height and width the easier it is to keep it inside the bobbin, but it is more expensive because of the weight of the material you need. Next time I try with an EI60 and EI48

 

[xazrules]

It may be counterintuitive, but when you CT a 10k winding, it consists now in two 2.5k windings.
That's because the inductance varies with the square of the turn number.
Inductance Lnom=Al.N²
So if the calculation gives Nturns for X Henries, when you split teh winding in two halves, each section now measures 1/4 of Lnom.

That’s a great lesson abbey, so if I need to turn 1270 turns for a 10kct, i do two 1270 windings or I split 1270 in half? Then interleave them with the secondary

 

[jonasnoble]

It may be counterintuitive, but when you CT a 10k winding, it consists now in two 2.5k windings.
That's because the inductance varies with the square of the turn number.
Inductance Lnom=Al.N²
So if the calculation gives Nturns for X Henries, when you split teh winding in two halves, each section now measures 1/4 of Lnom.

Thanks for clarifying, I never thought about it like that. I was talking in terms of splitting the turns, but spoke in terms of impedance, which was wrong. I love this group.

 

[Disco Volante]

i've remained with a lot of free space, i could used a bigger wire or do more turns, let's see how it works as a tube output

I've read you're supposed to fill the "winding window", to utilize the core most efficiently and get less stray field?
Still honing my transformer skills myself...
Congrats on your first transformer! yeah!
Cheers, V!

 

[abbey road d enfer]

That’s a great lesson abbey, so if I need to turn 1270 turns for a 10kct, i do two 1270 windings or I split 1270 in half? Then interleave them with the secondary

As the name "center tap" implies, you split in two 635 turns windings.

 

[abbey road d enfer]

I've read you're supposed to fill the "winding window", to utilize the core most efficiently and get less stray field?

Filling the winding area is mainly about minimizing DC resistance. Stray field reduction is obtained by sandwiching.
For a xfmr with equal turn windings (for example for a 1:1, 1:2, 1:3 or 1:4) bifilar, trifilar or quadfilar winding is recommended.

 

[xazrules]

Filling the winding area is mainly about minimizing DC resistance. Stray field reduction is obtained by sandwiching.
For a xfmr with equal turn windings (for example for a 1:1, 1:2, 1:3 or 1:4) bifilar, trifilar or quadfilar winding is recommended.

Exactly like that jonas, for minimizing the dc resistance you use bigger wire but you need a bigger window, so it is a compromise everywhere, you have to take into considiration the skills too. Maybe with time we can use smaller cores and end up with more efficient transformer in perfomance and cost production.

 

[CJ]

anything but copper in the window = leakage
so they try and cram as much copper in there as possible.

 

[CJ]

Window Utilization Factor Ku

Is influenced by

1. Wire Insulation
2. Wire lay (fil factor)
3. Bobbin area
4. Insulation required

A high voltage xfmr will waste more window due to the extra insulation needed.

A toroid will waste window area due to the clearance needed for the shuttle.

Energy stored in the leakage flux can be detrimental because when it is released it can cause ringing.

You can minimize leakage inductance by

1 Minimizing turns
2 Reduce build of coil
3 Increase winding width
4 Minimize insulation between windings
5 Use bifilar windings

Unfortunately corrective measures taken to minimize leakage and capacitance work in opposite directions.

So it is all about juggling tradeoffs.