Help needed with small transformer design

[nctnico]

When using stranded wire: make sure to put the strands together first and then wind all of them together. If you wind multiple windings which are intended to be connected in parallel one by one you may forget a turn and effectively have a shorted turn.

[akis]

I read now that when you only have one or two turns you may/must use copper foil and not copper wire because the foil occupies the whole window whereas it is very hard to wind one or two turns  spread out over the window area. As I do not have copper foil at home, I will experiment with aluminium foil to see if it makes any difference.

You definitely want to get Kapton tape or Polymide tape from other manufacturers. Electrician's tape is bad news for transformers in general, and for you, its thickness will result in loss of winding coupling.

For the secondary, you should be able to get a wire gauge that will allow just one layer of winding which will help with the coupling. Foil, or multiple finer gauge wires in parallel for the primary will help. With skin effect, about 28 AWG is the largest you want to use at 200KHz. Above that the center of the copper is doing nothing, and the thickness of the 1mm wire is detracting from good coupling. Aluminium foil may work, but you may be better off trying to use about 6 strands of the wire you are currently using for the secondary. The wires need to be flat in a single layer, and you really have to be precise to minimize losses at 200KHz.

I think the primary is the most challenging, so here is what I would do. Do a test to set a performance benchline. Take your current secondary wire and wind a perfect single layer - how even many turns that is.

One layer of Kapton tape. If you do not have Kapton tape, use a layer of sticky tape just for the test. Ban the electrician's tape.

Another full layer of the secondary wire with the same number of turns as the first layer. You now have a 1:1 transformer. Check the regulation. that is probably about the best regulation you will do with that core.  If it is good, take the last layer off, and try and get your 4 ohm primary to regulate as well.

It could be that a large part of the regulation problems are losses in the ferrite at 200KHz, and if so, then you need to search for better ferrite or ferrite geometry.

Richard.

I had heard that electrician's tape cannot withstand the potentially high temperatures in transformers but had never thought about its thickness being a problem.

My ferrite core is EPCOS N87/N97 sold specifically for these frequencies. Specs show 100mT for 100KHz. Supposing we use 50mT for 200KHz, Vin=2.828 V rms, F=200KHz, A=1.6cm2 -- this calculates as just 0.44 turns for the primary. So it seems that by using 10, 12 and 2.5 turns for the primary I have been using too many turns.

For the secondary I use 0.5mm and for the primary 1mm wire thickness.

Taking into serious consideration your advice and the calculations, I will aim for just 1 primary turn. The window size is 14mm.

The plan is to wind two 0.5mm wires for the secondary in parallel to get 14 turns in total to fit precisely inside the bobbin. Then I will add the special tape. Then I will use multiple of the 0.5mm in parallel to effect just one turn.

Will have to do this tomorrow morning - there must be something about winding transformers, I hate doing it, they are so fiddly.

[akis]

When using stranded wire: make sure to put the strands together first and then wind all of them together. If you wind multiple windings which are intended to be connected in parallel one by one you may forget a turn and effectively have a shorted turn.

I really hate winding transformers...

[akis]

You want only 1.225W resitive load at 200kHz you should only need a small core maybe an efd15 or EFD20 these are pretty commonly available Newark and mouser both stock EFD cores. For an efd20 core I get 6 turns primary you could fill the bobbin with paralleling windings and you should be using around 28- 30 AWG. Much smaller then 30AWG and it snaps to easy at least for me even with 30AWG you have to be careful.

If you go through the link or links there are equations for area product which will give you an idea of the core size required the ones you have seem large for such a tiny power output. I would suspect that it would require a smaller core then even a efd15 but you have to consider available winding area and how much of a pain in the ass it is to wind and insulate tiny transformers. Can you say carpol tunnel syndrome. Going to small would be too difficult to wind but EFD15 or 20 are easy to work with and relatively cheap.

If you would have caught me a couple months ago I could have tested out an EFD15 core at your 200kHz and load. I was fully set up when I was working on an AD9834 DDS signal generator. I was considering a broadband transformer for matching and taking advantage of the differential output but decided instead just to use an OPAMP. Unfortunately at the moment my bench is fully cluttered with my own projects.

What is this transformer for?

Using aluminum foil isn’t likely a good idea “shorted windings” insulating with tape without tearing up the AL would also be a problem I would think.

I think using a bigger size core would result in better regulation? I have already ordered and expecting to arrive tomorrow an RM14 which is a size up from my RM12. Hoping to get better regulation this way.

And I tried the aluminium foil and yes you are right it is too fiddly. I will use multiple strands of 0.5 mm wire.
The transformer is to step up a 4V peak into a 50V peak. Cannot be done electronically (easily) since the power supply is +/-6V (batteries). I had thought the transformer would be a 5 minute job, but it seems it is a small science in itself.

[akis]

I made lots of experiments and built at least 10 transformers using varieties of windings for the primary and the secondary. Even with using the large RM14 core, it seemed there was no way I could get better than 18% regulation. I never had problems with transformer getting hot, it was always a problem with regulation being too high.

Today I took a different approach. I suspected that with a turns ratio of 1:13 it is impossible to get better regulation. So I changed the design and the requirements and now the Vp is higher at 9V pk (was 4 V pk) and to achieve 50-55V pk I only need a turns ratio of 1:6 rather than 1:13. I quickly wound a transformer like that, and tested it, and voila! regulation solved, I did not even detect the curve changing on the scope when off and on load. For primary I used just one turn, and for secondary there are 6 turns. The secondary is wound in first, covering all the available area, then there is PTFE tape (plumber's tape), then just one turn on the primary.

Tomorrow I will try using the smaller core, the RM12, because I only need about 1.8W power.

I have read a ton about transformers and I think I am getting a good idea by now, but I still have some missing points. I will make a new post for this, as it is a more general question.

[T4P]

When using stranded wire: make sure to put the strands together first and then wind all of them together. If you wind multiple windings which are intended to be connected in parallel one by one you may forget a turn and effectively have a shorted turn.

I really hate winding transformers...

I think you would hate dewinding a toroid with >3000 turns ... i better get back onto dewinding a shot toroid i have (I realized why it's shot ... the primary leads disconnected from the wire)

[SeanB]

Try rewinding motors........

[T4P]

Try rewinding motors........

I did ... my hands were tired as ****

[szhighstar]

Hi friend, you do not hate winding transformer, I am sorry I kike, it is very interested if you deep study, I reply your question.
(a) ferrite cores , becasue power is less than 2.4W and 200K frequency , so you only need RM4 if you like RM configuration,EF13 or T14 are also good.

(b) number of turns on primary (and therefore secondary) , according to formula: V=L*di/dt, V is minimum DC voltage, I is Irms, t is time of flow, you have frequency and duty cycle, you can know t value.

(c) wire gauges (only rough ideas using Ohm's law)  for power transformer, current density is 3-5mm/A, you know Irms and current density, you can know wire diameter,which type will you select depend on your design standard,EN61558 or EN60065 or EN60950 or other.

(d) size of core to be able to transfer the energy required  , please check (a)

(e) primary or secondary wound first (innermost)  , not special requirement, however you should be consider low leak inductance and capacitance

(f) direction of windings,  all winding should be keep same direction normally

Addition, will transformer applied either flyback and forward, if it is flyback transformer, it must open air gap.
If you have other question, I am glad to help you.

[szhighstar]

sorry, I just printed wrong in last post, right word is I like