DIY Transformer, number of turns of the primary winding

[Phil Smith]

Hey guys!

I got a question. The one of my project is a DIY isolation transformer. I managed to find a pretty beefy toroidal core (211 mm external diameter, 51 mm height with crossectional area 24.5 cm2). I've been started to search some information related to this topic and found this video:

-

They got quite in-depth explanation how to calculate and wind a toroidal xformer. Also they suggest to check primary winding with a light bulb connected in series, like so -

http://wordpress.videorockola.com/wp-content/uploads/serie.jpg

So requiring number of turns for my primary is:

N = (42 / A) * Vrms = (42 / 24.5cm2) * 230V = 394.29, say 395 turns.

I've winded 395 and checked with the bulb - it turned on! quite dim though.. So I've winded all wire I had - 418 turns total - and it still turns on.. There is standard formula to calculate necessary number of turns -

N = Vrms / (4.44 * f * A * B),

where B is maximum flux density, say B = 1.5 Tesla.

N = 230V / (4.44 * 50Hz * 24.5e-4m2 * 1.5T) = 281.91, say 285 turns -

even less, than I've got before!

I suspected  short somewhere in the winding , but resistance of the winding is around 0.678R, that sounds ok for 93 meters (305 feet) of 13AWG wire (According to this table - http://www.interfacebus.com/awg-ohms-per-1000-feet.jpg - 13awg wire has resistance of 0.00203 Ohms/foot at 20degC; 0.00203 x 305 = 0.620 Ohm - close enough)

I've performed some AC current measurements with the 60W bulb circuit in series -
418 turns           -   139.00 mA AVG;
410 turns           -   143.78 mA AVG
399 turns         -   151.74 mA AVG
000 turns           -   251.75 mA AVG.
Just for kicks, I've measured a real working xformer with this circuit - and sure enough, the bulb didn't light up and average current consumption was just 6.53 mA.

A question is what is the problem? Perhaps not even wire placement due manual winding of 13AWG wire? Or maybe the core was a piece of crap?
Very appreciate any help!

Cheers!

[Belrmar]

I would suggest to check the inductance , maybe is too high and it could be blocking too much current when adding the reactance of the inductor to the wire resistance

[Phil Smith]

I've measured inductance of the winding (with my chipo LCR T4) - It's 11 H. Resistance of the wire (measured with Agilent U1272A DMM) is 0.678 Ohms.

[Belrmar]

Well assuming a  50hz intput and that inductance, your input inpedance is about 3.5 K , maybe thats the reason your output is so low, assuming that your core does not saturate.


(Edited bc i didnt think about core saturation)

[Phil Smith]

Hmm, can it be due uneven wire placement?

Here's a picture of the xformer - the ruler and DMM are for reference:
https://app.box.com/s/53w8wlmzcpjs3kq8k5thbx1ussfuoio4

[Belrmar]

Well im not an expert on trnasormers, not by far. But i would assume that you are using too many turns for the magnetic permeability of your core, creating a really high inductance and as a result z, you have too much impedance on the primary coil.

[Phil Smith]

Sounds promising Thanks! I'll unwind 30-40 turns and check if something will changed for a better result. If not, will continue unwinding.
Also I found a nice article - http://www.sklaic.info/forum/index.php?topic=175.0
If unwinding wont help - I probably should give it a try and test the core

[Belrmar]

Will you keep your output winding constant? i would suggest to try to keep the ratio more or less equal so testing is more precise

[Phil Smith]

Right now I have not any secondary winding. Finding the core permeability and requiring number of turns of the primary winding is paramount. So I've started slowly unwind the wire When Primary side will be ready, I'll calculate number of turns of the secondary winding

[Phil Smith]

I've been unwiding my xformer for a while and taking some measurements along the way. Here's a little report:

Permeability of the core, mu_c = (L * l_c) / (A_c * N^2)

where:
L - inductance of the winding (in my case it's measured with LCR T4 meter);
l_c - mean length of the core, l_c = (outer_dia + inner_dia)/2 * Pi = (211 + 115)/2 * 3.14 = 511.82 mm = 0.512 m;
A_c - area of the core, A_c = 48 x 51 = 2448 mm2 = 24.48 cm2 = 0.002448 m2;
N - number of turns.


N = 418 turns;
Bulb circuit current, I_b = 139.00 mA;
Inductance of the coil, L = 11 H;
Permeability of the core, mu_c = (11 * 0.512) / (0.002448 * 418^2) = 0.0132.

N = 410 turns;
I_b = 143.78 mA;
// I didn't measure inductance here.

N = 399 turns;
I_b = 151.74 mA;
// I didn't measure inductance here.

N = 300 turns;
I_b = 195.80 mA;
L = 6.57 H;
mu_c = (6.57 * 0.512) / (0.002448 * 300^2) = 0.0153.

N = 280 turns;
I_b = 197.81 mA;
L = 5.88 H;
mu_c = (5.88 * 0.512) / (0.002448 * 280^2) = 0.0157.

N = 250 turns;
I_b = 209.10 mA;
L = 4.70 H;
mu_c = (4.70 * 0.512) / (0.002448 * 250^2) = 0.0157.

N = 200 turns;
I_b = 219.80 mA;
L = 2.88 H;
mu_c = (2.88 * 0.512) / (0.002448 * 200^2) = 0.0151.

N = 150 turns;
I_b = 228.5 mA;
L = 1.30 H;
mu_c = (1.30 * 0.512) / (0.002448 * 150^2) = 0.0121.

N = 100 turns;
I_b = 237.15 mA;
L = 0.39 H;
mu_c = (0.39 * 0.512) / (0.002448 * 100^2) = 0.0082.

and the last measurement:
N = 50 turns;
I_b = 244.60 mA;
L = 0.054 H;
mu_c = (0.054 * 0.512) / (0.002448 * 50^2) = 0.0045.

[Belrmar]

The inductance look more or less linear , what suggets a good winding procedure... What kind of bulb are you using? maybe it has too much resistance? I gotta say that im getting quite confused . check the votlage across the light bulb and across the transformer, that should tell us what is the problem

[Phil Smith]

The inductance look more or less linear , what suggets a good winding procedure... What kind of bulb are you using? maybe it has too much resistance? I gotta say that im getting quite confused . check the votlage across the light bulb and across the transformer, that should tell us what is the problem

Alright, I've checked the voltages with 50 turns primary.
Supply voltage - 220.1 VAC;
Voltage across the primary winding - 11.47 VAC;
Voltage across the bulb - 208.4 VAC

I use a simple 60W incandescent bulb

[Zero999]

Just one question: why?

Transformers aren't expensive.

How are you planning to insulate the secondary from the primary? Do you have a hi-pot tester to ensure this is safe?

[Belrmar]

Ok , it looks liek our light bulb is becoming the main problem right now , i would suggest yo go back to 150-200 turns and putting a 1ohm shunt resistor and a 250v ac rated at about 2-5 amps(safety first). measure the voltage across your shunt and you will have an aproximation of the VA that your transformer will have (i think so , dont trust me too much)

[Belrmar]

Just one question: why?

Transformers aren't expensive.

We do these things not because they are easy, but because they are hard"-John F. Kennedy

and not i am not from the US, but i like that way of thinking

[Phil Smith]

Just one question: why?

Transformers aren't expensive.

How are you planning to insulate the secondary from the primary? Do you have a hi-pot tester to ensure this is safe?

I just got this core for 10-15 bucks back in the day, so I decided to give it a try. Anyway, I've learnt about transformers for these several days more, than for a whole year xD
Guys from that video in my first post uses layer of prespan paper + masking tape + copper shield cover in prespan paper + masking tape again to isolate primary from secondary
Sure, an isolation test is crucial, so I'm looking forward to find this hi-pot tester.

Just one question: why?

Transformers aren't expensive.

We do these things not because they are easy, but because they are hard"-John F. Kennedy

and not i am not from the US, but i like that way of thinking

Hah, indeed 

Ok , it looks liek our light bulb is becoming the main problem right now , i would suggest yo go back to 150-200 turns and putting a 1ohm shunt resistor and a 250v ac rated at about 2-5 amps(safety first). measure the voltage across your shunt and you will have an aproximation of the VA that your transformer will have (i think so , dont trust me too much)

Alrighty, I'll try to find some power resistors for this job

[Belrmar]

Alrighty, I'll try to find some power resistors for this job

a 1w resistor should be enough for the inductance you gav eme earlier on with 150 turns
 

edited bc i am a mess with quotes

[oldway]

I do not know what you do with this lamp in series, it's completely wrong. 

This series lamp can only be used to check if there is no short circuit between turns, nothing else.

Normally, a variac is used to test the transformers, not a lamp in series.

By choosing an induction of 1.2T or 12.000 Gauss, with 230Vac , 50Hz and a core area of 24.5 cm², you need 374 turns.
NB: I use a 0.95 factor to reduce core area because there is some isolating material.

You test the primary winding with a lamp in series ... If the lamp does not light up fully, it's good.

Now, power on the primary winding without the lamp but with a fuse and measure the primary current. This is the magnetization current.
The smaller the magnetization current, the better the transformer.

Then you wind up the secondaries.
To calculate the number of turns of the secondary, you have at primary 374 turns for 230V so 374/230 turns per volt or 1.63 turns / volt

To know the number of turns of the secondary, multiply the tension of secondary by 1.63

Again a quick test with the lamp in series to check that there is no short circuit between turns.

If the lamp does not light up or weakly, it is OK.

You can then power on the primary without the lamp but with a fuse.

You can then check the secondary voltages.

All that easy.....   

[Phil Smith]

I do not know what you do with this lamp in series, it's completely wrong. 

This series lamp can only be used to check if there is no short circuit between turns, nothing else.

Normally, a variac is used to test the transformers, not a lamp in series.

By choosing an induction of 1.2T or 12.000 Gauss, with 230Vac , 50Hz and a core area of 24.5 cm², you need 374 turns.
NB: I use a 0.95 factor to reduce core area because there is some isolating material.

You test the primary winding with a lamp in series ... If the lamp does not light up fully, it's good.

Now, power on the primary winding without the lamp but with a fuse and measure the primary current. This is the magnetization current.
The smaller the magnetization current, the better the transformer.

Then you wind up the secondaries.
To calculate the number of turns of the secondary, you have at primary 374 turns for 230V so 374/230 turns per volt or 1.63 turns / volt

To know the number of turns of the secondary, multiply the tension of secondary by 1.63

Again a quick test with the lamp in series to check that there is no short circuit between turns.

If the lamp does not light up or weakly, it is OK.

You can then power on the primary without the lamp but with a fuse.

You can then check the secondary voltages.

All that easy.....   

That's interesting!

This trick with a lamp I saw in this video:


They said the bulb should not light up, if it does, it indicates either a short or lack of turns - and that's what brought me off course..
Also, in some article I read this bulb current should be less than 50mA. Could you tell me please, what philosophy stands behind this procedure?

I'll rewind the primary more meticulously and isolate every layers of it (there will be three - 160 + 160 + 54 turns ) with masking tape.

Anyway, thank you, oldway! I know it's gonna be easy) Sort of

[oldway]

Anyone can post on the internet, no need to show his diploma, nor to prove his skills, experience and knowledge ..... So, we must be careful because there is plenty of stupidity on the Internet ....

Fortunately, there are real experts on this forum who can correct mistakes, which makes EEVblog forum one of the best forums in electricity and electronics. 

The bulb lights up or not .... First, did you ask yourself what bulb ? 25W, 40W, 60W, 75W or 100W? ....
If not specified, it smell bullshit... 

The magnetization current depends on:
- power of the transformer
- quality of the magnetic core (losses by eddy currents and hysteresis, gap, leakage flux, ....)
- induction level.
NB: the magnetization current is highly distorded, use only a true rms meter to measure it.

What is the normal magnetization current?
It depends on many factors, it can not be determined with a solution as simple as putting a bulb in series.

It is necessary to have a variac which makes it possible to apply a progressive voltage on the primary of the transformer, monitoring the current,  and even to boost it up to 265V to check if there is no saturation of the core.

When saturation is reached, the magnetization current no longer increases linearly, it increases rapidly for a small increase in voltage.

As for the level of induction, it must be chosen according to the magnetic material used ....

I was manufacturing cheap transformers with low carbon steel (not silicon steel), so I used 10,000 Gauss (1T) .... for high quality magnetic steel (cold rolled high silicon steel) ) we can go up to 16,000 Gauss (1.6T), but be careful, because the saturation in a toroidal transformer is very sharp.

The toroidal transformer has not only advantages, it also has drawbacks, including high inrush current and low short circuit impedance.

[oldway]

There are many other important concerns that should be considered, but it is impossible to develop them on a forum.

For example, as Hero999 recalled, the high pot safety test.

Also the quality of the insulation used and their resistance to humidity and high temperature.
Cartonboard and masking tape are not at all adequate .... it's low-level brewing.

It is also necessary to impregnate the transformer with special varnishes ..... Ideally, it would be necessary to use the vacuum + pressure process to impregnate a transformer correctly, otherwise the varnish or the resin does not penetrate well the windings and the insulating materials.

There are many books on project and manufacturing of transformers, if you want to deepen the subject, I recommend you to consult them.

Manufacturers often use a powerfull transformer design software as RALE software to optimize their transformers.

http://www.rale.ch/

[Phil Smith]

Anyone can post on the internet, no need to show his diploma, nor to prove his skills, experience and knowledge ..... So, we must be careful because there is plenty of stupidity on the Internet ....

Fortunately, there are real experts on this forum who can correct mistakes, which makes EEVblog forum one of the best forums in electricity and electronics. 

Hah, yeah, EEVblog forum is the place to be

Anyway, thank you oldway!
Could you recommend some books on this topic [project and manufacturing of transformers] please?
Yep, I've got a variac and will test transformer as you described.
Right now I've started to wind the primary. It's gonna take a while ))  Now 100 turns out of 374 are done.

[oldway]

Some books about transformers.....

Also in french, MACHINES ELECTRIQUES  by M. KOSTENKO et L. PIOTROVSKI, éditions MIR, tome II

[Phil Smith]

Thank you, quite interesting books!
PS. Winded 200 turns so far, keep on turning
PS. PS. Happy New year to everyone!!!! Cheers!

[Phil Smith]

Hey guys!
I've winded up 375 turns and performed some suggested measurements.
First off, the bulb circuit test (see Fig.01) - the bulb current consumption is around 250mA, with transformer it's 158 mA. Basically, result is pretty much the same as it was before rewinding.
Second, I've tester the primary with a variac, monitoring voltage and current (see Fig.02). Results see Fig.03.
As far as I can see from the chart, magnetization current slope drastically changes its angle after 160VAC, indicating the core saturation.


Inductance of the winding is 9.25 H (measured with LCR-T4 tester).
So permeability of the core mu_c is:
mu_c = (L * l_c) / (A_c * N^2) = (9.25H * 0.512m) / (0.00245m2 * 375^2) = 0.0137.
where l_c is mean length of the core.


So what is my next step?

Regards, Phil.

PS. The transformer designed power is 2200 VA.