Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

DIY transformer/flyback, measuring turns ratio.

Status
Not open for further replies.

Blueteeth

Well-Known Member
Hey,

(please note, this isn't about high voltage, thats why its not in the high voltage section).

After putting it off for months, I have just wound my first transformer as a practice run for future projects. Some are high voltage, some are low voltage and there are different topologies I plan to use.

This first one however, is a simple flyback transformer (or coupled inductor) wound on a small E-core ala 'flybacks for dummies':
Flyback Converters for Dummies

Since I didn't know the properties of the ferrite, I used an inductance meter I built a while ago to measure the inductance of 20 turns. Then, since I knew I would need an air gap anyway, I slowly filed down one of the centre legs of the E-core's to end up with a very small (maybe 0.1-0.2mm) gap. With 20 turns, surprisingly I got a nice 101.2uH!!

So, since I wanted a 1:12 step up transformer, thats 20 * 12 = 240 turns on the secondary. As I couldn't be bothered to count the actual turns as I did them, I wound one neat layer, counted the turns (roughly 40-50) then added 6 more layers. Now, the reason for the extra layer (50*6= 300, not 240) was because I would rather over-wind and have to 'unwind' it afterwards, than cut the wire, measure, then have to splice more wire and add more.

Now, I have read up on inductance and transformers, but I guess i'm missing something. When I measure the primary now, after I have wound the secondary over it, I get 308uH, instead of 101uH. Also, I measured the secondary inductance as 15.6mH.

Inductance ratio = 15600 / 308 = 50.65. Theoretically, the turns ratio is the square root of this, which is 7.12. I know I wound exactly 20 turns on the primary...so that would mean that, the secondary has around 20 * 7.12 = 142.4. On 7 full layers, that would be 142/7 = 20 turns each layer - I wound waay more than this. (I purposefully added far too many turns).

So, obviously either my inductance meter isn't up to the job (quite possible, although it seems to measure 'known' inductors very accurately). Or, my testing methods are flawed (very probable).

Could someone explain two things:

1. why the primary inductance increased when I added the second winding? Is it simply that the primaries magnetic field is better coupled to the core? since it is now surrounded by a coil.

2. Why are my results for measuring turns ratio (indirectly, by measuring inductance ratio) so off? as my measured results are pretty much half of what was expected. I mean, given my DIY inductance meter, and perhaps my poor experimental practice, I expected some error, but not this much...

I am sticking with 20 turns, but planning to gap the core futher to get 47uH (ish) and perhaps even lower if the saturation of the primary is under 2.5A. This is really 'trial and error', but its better than just going with pure theory :)

Also, if someone can provide a different method of measuring the ratio, preferably without dumping a 12V square-wave across the primary at a few kHz...just incase the ratio really is 1:14+, I don't trust my winding insulation layer to cope with the massive HV spikes it'll produce.


Thankyou, apologies if this really is basic stuff, I coverd it in university a few years ago, but that was all basic theory.
 
How does your inductance meter work? Does it apply a sine wave to the inductance to be measured, or does it use a square wave or pulse waveform?

If the applied waveform is not a sine (at a relatively low frequency, such as 1 kHz), the self resonance of the secondary is probably causing your problem.

Try applying a small (.1 volt), 1 kHz sine wave to the primary and then measure the secondary voltage.
 
Last edited:
TheElectrician,

Thanks for the response!

My inductance meter is a clone of this one: **broken link removed** Usually, I try and modify the circuits I clone for my own purposes, but since I wasn't sure exactly how this worked, I used the exact values as shown. It has been used on many capacitors (pretty damn accurate) and various 'known' inductors, as well as offline flybacks salvaged from old power supplies.

As a test, I found the datasheets for 4 offline flybacks, which give the primary inductance (+/-10%), and measured the transformers myself. It seemed to be well within 5% of what the datasheets provided.

As for the 'applying a small sine wave', good idea! I don't own a function generator, so I was trying to avoid a frequency sweep to get the 'true' reactance of both coils...but, that isn't as important to me as the turns ratio....which your suggested method should do rather well! So simple I didn't think of it....

I built a fairly accurate 1kHz sinewave test generator for audio circuits a while back (low THD), very low power, but with a series resistor, I'm guessing I can limit the current (10mA?), then use my scope to see the primary and secondary waveforms.

The only thing I can think of is, either the core is somehow non-linear (doubtful) or there is a short in the secondary somewhere (unlikely, but possible). As another half-arsed test, I measured the resistance of the secondary. Then, as I knew the gauge of the wire, I worked out the total length of wire used. (using ohms/metre for that gauge) Measuring the turn length at maximum (around the fully wound transformer), and minimum (around the primary) I got the average turn length. So, from that I got roughly 135 turns. - I realise that there are many errors in this result, but an error of 100% seems unlikely, as it sohuld be around 260 turns :)

I'll set my scope up now and give your test a try. Thanks! And apologies if I'm over complicating this and missing the obvious..its all experience.
 
Ok, that didn't work lol I forgot that because the primary is only 20 turns, its resistance is tiny. I can't measure it accurately, but we're talking 0.1 ohm or there-abouts. So of course, in order to get a half decent voltage across the primary, it'll need that test signal to provide an amp or more. Unless I increase the frequency to something more reasonable for a 300uH inductor.

I may just rewind the secondary in a dirty way. Start off with 20 turns, for a ratio of 1:1 just to see if its playing ball, or if I have some dodgy cores. The idea of doing 300 odd turns, AND insulating every other layer doesn't appeal to me though.

I still don't know why the primary inductance increased so much after I put on the secondary? by a factor of 3!
 
Would you please connect your scope to the primary winding at the same time that your inductance meter is connected, and see what frequency of signal the inductance meter is applying to the primary. I suspect that your problem is that the frequency is so high that the self capacitance of the secondary is causing a bad reading for the inductance.
 
Last edited:
I shall try that now. Thankyou! I think it works by measuring the frequency of an LC circuit, with the L, or the C, being the unknown. It picks up the resonant peak, so you could well be correct in your assuption that the secondary is interfering with the reading......this meter was deisgned for inductors..probably not coupled inductors.

Another option, whilst I have my scope out, is to use my setup for measuring saturation current...looking at the slope of the current flowing through the inductor, seeing whereit saturates and, by the slope, its inductance. If it really is 100uH, then that would make the ratio (15600/100)^0.5 = 12.5. MUCH closer to what I expected. In fact, you're idea seems to be the most likely, and I never would have thought of it myself (I assumed an open secondary winding would have no effect), so thankyou!
 
Just scoped the primary with the meter connected. It's a perfect sine wave which varies in amplitude (expected as the meter sweeps frequency to find the peak). And the reading remains a stable '308uH' on the display.

As for the 'saturation current' test. I compared the scope waveform with that of a 100uH 3.0A inductor. They are the same! (with some oscillation on the transformers waveform....leakage inductance). To make sure, I also tested a 47uH and 220uH inductor, the 47uH's slope being steeper, and the 220uH one shallower. Looking at the peak current, I got 1.1A for 10us, at 12V. L = (V/I)t = 109uH :)

Assuming the meter was correct when measuring the secondary (probably not, but it'll be close), 15.6mH, that gives the ratio 1:11.96. lol, as I was after 1:12, that really isn't that bad at all....since I didn't even count the secondary turns. Which I can now work out to be roughly 20 * 11.96 = 239. On 7 layers thats 34/35 per layer.

Note: this test was performed on the primary of a 1:12. I heard some really annoying whining from the transformer which I'm hoping is just the core rather than arcing on the secondary. As it was an identical setup to the 'flybacks for dummies' webpage, the flyback from the primary was clamps at 12v + vdrop = 12.7-12.9V. Am I correct in assuming this would clamp the secondary to 12 times this? at 155v ish? I can't see that arcing, even without insulation.

So yes, sorry for the long thread, but I thought I would post my progress in case it can be of use to someone.

Also, 'TheElectrician'. You were spot on in your assuption :D Seems the meter's method for measuring inductance doesn't cut it for coupled inductors/transformers with certain inductances...but the creator never advertised it as such, so its my own fault for assuming.

A very happy Blueteeth.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top