Creating my own transformer

[Sadlercomfort]

I would like to create my own transformer for practical purposes. I would need a square shaped core that is laminated.

I can't find where I could buy one, that is large enough and just a square. I'd like to have it around 20 cm x 20 cm. Does anyone know where to buy one?

 

[KrisBlueNZ]

You will also need a bobbin.

What frequency will it be operated at?

If it's for mains frequency, I doubt you can buy cores and bobbins. I could be wrong. But it might be simplest to find a suitable looking transformer and unwind it.

 

[Sadlercomfort]

I'm not sure what frequency is best tbh, but I was hoping to use a low alternating triangular waveform. If that would even work.

What's a bobbin?

 

[KrisBlueNZ]

The bobbin is the plastic part that you wind the wire around. You can't really wind it directly on the core (unless you use a ring core).

What do you want to make with it? Some kind of high-voltage source? A high-voltage DC supply? If so, a ferrite core operating in the 30~200 kHz range is probably best.

 

[davenn]

start from the start and tell us all about your project
voltage, frequency and current capability of the source

what output voltage and current you need
what sort of load you want to drive

with that info help will be easier to give

Dave

 

[Colin Mitchell]

Making a transformer is much harder than you think.
Yours will simply buzz and annoy you.

 

[hevans1944]

20 cm x 20 cm x whatever thickness is a pretty big core. If its 20 cm x 20 cm x 20 cm it will weigh a lot. You should be able to purchase grain-oriented silicon steel laminations, stamped for either an "E" core or a "C" core, of that size. Just Google transformer cores. You need a bobbin to protect the enameled magnet wire from the sharp edges of the laminations. You will also need to insulate the laminations with shellac from each other as you assemble the core onto the bobbin. At the end of the assembly process the laminations must be a very tight fit in the bobbin. You will need to wind the primary and secondary winding(s) onto the bobbin before you insert the laminations. And you need to calculate how many turns of wire to use on the primary and secondary winding(s) and then make sure the completed bobbin will fit around the laminated core.

If you can accept that the tips of the triangle waves will be rounded off a bit, they should work just fine. They get rounded because of the finite high-frequency response of the transformer.

Do tell us why you need to wind your own transformer.

 

[Sadlercomfort]

I've studied transformers alot, so understand why the bobbin is used and the core is laminated to prevent losses. The purpose of this is to learn more about transformers practically. I have a square transformer core which is already laminated but with a slight dent on one edge . I also have transformer or magnet wire, so if that core is fine I just need to invest in a bobbin.

Let's say for example I experimented with a triangular or sinusoidal ac, of around 12 volts or 24v peak to peak. I'd like to see the effects of stepping up the voltage to say 240v. All the time studying the circuit methods of driving the transformer and smoothing the output efficiently.

Thanks for the reply guys, this has no real purpose. I just want to learn.

 

[hevans1944]

Can you upload a picture of the transformer core that you have?

I used to salvage transformers from old television sets and try to rewind them... but that was fifty-something years ago. The hardest part was separating the "E" and "I" laminations because they were shellacked together, each "E" lamination alternating with an "I" lamination. Then, once they were separated I had to clean the shellack off of each lamination, polish it up with steel wool or sandpaper, make sure each piece was flat, and then re-assemble them onto a new bobbin already wound with magnet wire. You can make your own bobbin by winding and gluing kraft paper around a mandrel and gluing rectangular plastic end pieces to it.

My hand-wound transformers never performed as well as manufactured transformers, but at my young age I couldn't afford a manufactured transformer (I was still messin' 'round with vacuum tubes back then). Then along came transistors and cheap silicon rectifiers, which changed the game completely. No need for a +300 V DC B-supply for the plates, or a -100 V DC C-supply for the grids. Radio Shack started selling filament transformers, 6.3 V AC @ 1 A (or thereabouts) at a price I could afford, so I built low-voltage DC power supplies and started messin' 'round with transistors.

I hardly ever looked back at vacuum tubes after that, except when I built my amateur radio Novice transmitter, basing the output "final" on an RCA 6146B, and when I assembled a Heathkit SB-301 receiver which used vacuum tubes. I built the power supply for the transmitter using salvaged TV transformers and 5R4-GY vacuum tube rectifiers. No re-winding required. It had a voltage-regulated output using a power pentode vacuum tube and, IIRC, 6SN7GT-B dual triodes for the error feedback amplifier. It was a thing of beauty with illuminated push-button ON, STANDBY, and OFF switches, a professional aluminum chassis, and rubber feet. It weighed about fifty pounds, while my transmitter weighed about five pounds or so.

Today I operate and maintain a linear particle accelerator that uses two high-power tetrode vacuum tubes in a push-pull power oscillator configuration. Just about everyone else who owns this particular model accelerator has long ago replaced the vacuum tube version with a solid state oscillator. But if it ain't broke, I don't believe I should try to "fix" it.

Good luck learning more about transformers! A dent on the laminations won't hurt anything, and you can straighten it out after you have disassembled them. If you are real lucky there will be bolt-holes through the laminations that you can use to tighten the lamination stack and mount the transformer.

BTW, learning is what we are all here for.

 

[Sadlercomfort]

Interesting you saying all that. Still can't quite believe the size of those vacuum tubes to the size of a small semiconductor today. Lecturers talk about them now and again, but it's not in our course guide anymore.

I have this transformer:
http://tinyurl.com/msh48wb

I separated the E and I a while ago, hence a small dent. The laminations are still pretty tight.

 

[davenn]

I have this transformer:
http://tinyurl.com/msh48wb

I separated the E and I a while ago, hence a small dent. The laminations are still pretty tight.

OK cool, that a good start
OK what do you know about turns ratios between the primary and secondary and how that affects the output voltage ?

Dave

 

[Sadlercomfort]

First of all I know that in an ideal transformer the input power and output power are approximately the same. Since P = VI, stepping up or down the voltage respectively decreases and increase the current. Which maintains the same power. Of course there are losses.

I know that if you have twice as many turns on the secondary or primary. You are respectively doubling the voltage or halving the voltage. 1:2 and 2:1. Or maybe 12:24 and 24:12.

I have also been taught this basic relationship, correct me if I'm wrong:

Np/Ns = Vp/Vs = Is/Ip


Now the problem is nothing is ever 'ideal' and there are losses practically. I know a little about saturation and eddy currents/hysteresis. But not much.

 

[davenn]

OK that's good

from that info you will be able to start to work out what input voltage you need for a given output voltage

Efficiency of the transformer is directly related to the core and the frequency
you will be able to experiment with that. with a steady input voltage say 12V AC sine
you can graph the output voltage generated verses frequency

I wont tell you what happens ... will be good for you to do the experiment and see what you get and post your results here

And if anyone gives the answers I will delete their posts BE WARNED

Now the problem is nothing is ever 'ideal' and there are losses practically. I know a little about saturation and eddy currents/hysteresis. But not much.

OK so there is a reason why we use laminated cores it has to do with eddy currents
what can you come up with as an answer ?

Dave

 

[Sadlercomfort]

Okay, lol. Well I think this might be linked to an electromagnet wave and it's counterparts. An magnetic field produced by the inductive means of an electromagnet or the transformer windings. And the opposing electric field produced around any nearby particles containing electric charge.

So when there's a change in the magnetic field, and electric field opposes the change and produces current flow.

I don't understand though why the laminations are places horizontally. Is this because the electric field produced is perpendicular to the magnetic field lines? If this was the case wouldn't the eddy currents naturally flow vertically, aligned with the electric field? So in theory vertical laminations are useless because it would assist the flow of eddy currents? I don't know lol.

I read somewhere that an higher frequency causes more eddy current losses, why I don't know. I'm sure the flow of eddy currents produce heat losses too.

=)

 

[davenn]

a good try

the magnetic field generated by current flowing in the Primary winding ( and to a lesser extent in the Secondary)
generate eddy currents in the iron core. The purpose of breaking the core up into insulated laminations is to
minimise the amplitude of those eddy currents.

There are 2 main losses due to the core
Hysteresis losses and eddy current losses --- eddy current losses result in heating of the core
you can google those 2 for more info

here's a good pdf file on the subject of increased freq causing higher eddy losses in a core
http://www.ti.com/lit/ml/slup197/slup197.pdf

cheers
Dave

 

[Sadlercomfort]

Wow, it is a good pdf file. What a strange phenomenon, I've only read the first few pages. I will read time rest tomorrow.

Thanks Dave, this deserves printing for a hard copy =)

It's strange to think that the current only flows at the surface of the conductor at high frequency. I'll read the rest and ask more questions if that's okay.

 

[davenn]

It's strange to think that the current only flows at the surface of the conductor at high frequency.

This is called Skin Effect. The higher the frequency the more it has to be considered
When I am working on my microwave radio projects from 1GHz to 24 GHz it becomes strongly relevant
and silver plating of conductors becomes the norm to produce the lowest resistance in a thin skin layer

Dave

 

[hevans1944]

This is called Skin Effect. The higher the frequency the more it has to be considered ...

Skin effect never goes away, even at low frequencies such as 50/60 Hz power line frequencies. Utility high voltage overhead transmission lines typically use a steel core (for strength) under aluminum cladding. I am sure they would use silver or even copper cladding if that were economically possible, but aluminum is dirt cheap compared to alternatives.

The skin depth at power line frequencies is deep enough to penetrate the thick aluminum cladding but not deep enough to significantly penetrate the underlying steel. Litz wire (Google that) is used in some transformers to reduce skin effect, but Litz wire becomes ineffective around 100 kHz or so. At radio frequencies the use of hollow copper tubes, often plated with silver, is common for air-core inductors. The hollow tubes also allow (de-ionized, non-conductive) water to be circulated for cooling at higher power levels. Even with silver plating, hollow copper tubes can get very hot with hundreds or thousands of amperes of current flowing on their surface, silver-plated or not.

 

[Sadlercomfort]

Hi Guys,

I've read more or that pdf document. Altough it's nice to understand more about eddy currents, I have more questions lol.

So at higher frequencies the eddy current flows at the outer edge of the conductor. And different frequencies produce different penetration depths.

I know that they're using a conductor as an example, but when the eddy currents are produced in the transformer core. What is actually happening in the laminations?

Is the current flowing in the directions relevant to Flemmings left hand rule? Also, how does the thin laminations reduce eddy currents, is it because you've effectively limited the penetration depth which prevents eddy currents from flowing.


Looking at these eddy current losses, has gotten me confused about the main principles of the transformer:
http://tinyurl.com/osrxg8j

There is an induced voltage through the core, which is produced by a 'changing magnetic field'. So the current and eddy currents are different right? The current is induced and the eddy currents which oppose it in the centre are just a result of higher and higher frequencies right?

Haven't even touched hysteresis yet. But I'm getting there =)

 

[Colin Mitchell]

Each of the laminations acts as a SHORTED TURN. This is something that no-one has ever explained before.
Even though a shorted turn is considered to be a turn that has both ends connected together, each of the laminations is exactly like a shorted turn and to make this clearer we place a solid core in the centre of the winding.
Now the current produced around the edge of the core will be very high and the transformer will no work. All the primary current will be absorbed by the core and it will get hot.
But if we cut the core into layers, the length of the circuit (short-circuit) becomes longer and the flux that passes through the centre is reduced so that the result is cold laminations.
It is only the flux that passes through the centre of the lamination that produces the current in the lamination.
We are referring to cutting though all the laminations and seeing the thin cross-section of each lamination.
As the frequency increases, the number of times the short-circuit current flows, will increase and the losses will increase.
This current is called EDDY CURRENT and is the secret to picking up aluminium cans out of your trash, along with the metal cans. The eddy current in the aluminium makes the aluminium stick to an electromagnet.