What exactly makes a transformer work?

[Red Squirrel]

I know the basics, the first coil produces electromagnetism and the second coil basically picks it back up, the same way electricity is generated when a magnet passes by a coil.

However, there is most likely more to it than that, as every transformer I've seen is made pretty much the exact same way. The E metal plates interlined together. If you just take a metal bar and wrap wires around it and try to make your own, it wont work. it seems the design that is used is very specific to why they work. Just curious what exactly is going on inside a working transformer.

 

[MongGrel]

...

 

[stormkroe]

Squirrel, you absolutely can make your own transformer using the method you just mentioned, you just need 2 or more insulated conductors. The ferrous core is there to increase inductance, it's made if many laminated plates to reduce eddy currents.
For instance, take an iron bar, wrap 100 turns of insulated #24 around one end, and 1000 turns of insulated #24 around the other. Congratulations, you've just made a (very inefficient) 1:10 transformer.

 

[Red Squirrel]

Squirrel, you absolutely can make your own transformer using the method you just mentioned, you just need 2 or more insulated conductors. The ferrous core is there to increase inductance, it's made if many laminated plates to reduce eddy currents.
For instance, take an iron bar, wrap 100 turns of insulated #24 around one end, and 1000 turns of insulated #24 around the other. Congratulations, you've just made a (very inefficient) 1:10 transformer.

Hmm maybe I just don't have enough turns on the primary then (only like 10) since I tried and it did not work. Though, does it have to loop back around or can it just be a bar? Totally random experiment with zero purpose while I was bored:



Or could it just be that it's so inefficient my meter can't get a reading? Was using a ~30v AC adapter on the short coil.

Though I guess the other coil should be beside and not on top.

So basically the more metal that is around the coil increases efficiency as there's more mass to "store" the magnetism?

 

[videogames101]

Hmm maybe I just don't have enough turns on the primary then (only like 10) since I tried and it did not work. Though, does it have to loop back around or can it just be a bar? Totally random experiment with zero purpose while I was bored:



Or could it just be that it's so inefficient my meter can't get a reading? Was using a ~30v AC adapter on the short coil.

Though I guess the other coil should be beside and not on top.

So basically the more metal that is around the coil increases efficiency as there's more mass to "store" the magnetism?

More metal? As in the core, or the wire itself? The core being metal has to do with permeability of the medium in which your generating a magnetic field. Iron has a high permeability and is cheap, so we use it to generate a stronger field and therefore a higher efficiency transformer. The wire just depends on the number of loops you make, and to a smaller degree the quality of the winding. Mass has little to do with it.

http://en.wikipedia.org/wiki/Permeability_(electromagnetism)

 

[Red Squirrel]

Meant the core. In my case it's just a flimsy little PCI slot spacer so maybe why it does not work, not enough metal to induce a decent field. If I put DC or AC current on the bigger coil it will act as an electromagnet though. A weak one mind you.

 

[Mark R]

A transformer works by the process of mutual induction.

A time-varying current in a coil of wire will produce a time-varying magnetic field. That time-varying magnetic field will induce a voltage in a neighbouring coil of wire.

At the same time, the magnetic field will induce a voltage in the 1st coil of wire - this voltage will serve to oppose the supply voltage, and limit the current in the 1st (primary) coil.

If a current is drawn from the 2ndry coil, then it will produce a magnetic field which opposes the magnetic field generated by the 1ry coil. As a result, the voltage reflected into the primary is reduced, and the 1ry current can increase (as the supply voltage is less counteracted).

Quite how well this works depends on the inductance of the coils, and the degree of linkage (mutual inductance). This depends quite strongly on the strength of the magnetic field (or more precisely, the flux).

Magnetic flux, like electric current, can be thought of as flowing in a circuit (although there are no specific start/end pos/neg points, just a continuous loop). The same "magnetizing force" will produce a higher "magnetic flux" in a circuit with a higher magnetic permeability.

Iron has a very high magnetic permeability, so produces a high magnetic flux from a modest magnetomotive force. Air has a very low magnetic permeability, so any air in a "magnetic circuit" will greatly reduce the magnetic flux for the equivalent MMF (amps * turns). This property can be used to modify the properties of a transformer - for example, an arc welder may construct the iron core out of a "C" and an "I", the two parts are then mounted on an adjuster screw, so that they can be pulled apart or pushed together, allowing a small air-gap between them. The introduction of a tiny air-gap allows the overall magnetic permability of the core to be adjusted, changing the properties of the transformer (adding an air gap will result in have a current limiting effect).

If you try to make a transformer from just a rod, then it doesn't work very well, as the magnetic circuit has to complete the loop in air. The low permeability of air results in a small magnetic flux (relative to input current). Without a strong magnetic flux, you don't get good transfer of energy to the 2ndry coil when used as a "transformer".

A large air gap "transformer" can be useful, and is more usually called a flyback inductor. This used to be used for the generation of HV in CRT displays. However, it is widely used in cheap SMPS and as a HV pulse generator for vehicle spark ignition. In this setup, a DC current is applied to a coil on a low-permeability core. The current is allowed to ramp up (the inductance will slow the current ramp). As the current ramps, energy is stored in the core. The DC input is then shut off. The energy stored in the core's magnetic flux then has to go somewhere, and a voltage will be generated that will try to force the current through the coil in the same direction. This is "back EMF". This back EMF will be generated in ALL coils on the same core, and the pulse voltage is proportional to the number of turns. So, in a car ignition system, the primary side might have 100 turns, and the 2ndry 10,000 turns. When the ignition driver turns off, it might reflect 250 V back into the primary side, but because of the turns ration, the 2ndry will have 25 kV reflected into it.

The opposite turns ratio might be used for a phone charger - in order to assist with step down. By having electronics monitor the voltage reflected back into the primary, it is possible to determine the voltage on the 2ndry side (and therefore regulate the output voltage with feedback to the SMPS controller) without needing to electrically connect the input and output (for safety reasons).

 

[William Gaatjes]

Nice explanation Mark R. Linear technology has a lot of those opto less smps ic's.

If memory serves me right, i used this IC in a design for 48Vinput to 3x 12V/100mA out.

http://www.linear.com/product/LT3512

 

[Mark R]

The E shape transformer is widely used because it is easy to build.

You simply take a bobbin, put it on a motor, and wind the coil of wire onto the bobbin. The you install the bobbin onto an E, then clamp the other E on to complete the transformer.

Transformer bobbins being made

The disadvantage with this design is that there is a tiny air gap between the two Es. This results in reduced transformer performance, especially for small transformers. For big transformers (like power company transformers) the air gap is very small compared to the size of the transformer, so it less relevant.

There is an alternative design of transformer, called a toroidal transformer. In this case, the core is a donut of iron. There is no air gap, it's just a continuous piece of iron. The windings then loop over the donut. This type of transformer has much higher performance (especially when small) because of the lack of air-gap. The disadvantage is the difficulty winding, because on each turn, you need to loop your spool of wire through the hole in the donut.

Once you get above a certain size (about 10000 W), it becomes impractical to make toroidal transformers, plus the benefits become much lower.

Note that in the toroidal transformer case I lied about the consturction slightly. It is a single piece of iron, but it's actually an iron ribbon which is rolled up in a spiral to form a donut shape.

This is because transformer cores are best made from thin strips of iron, not large chunks. The reason is because of eddy currents. If you have a big slab of iron, then because the iron is conductive, the iron itself can act like a shorted turn on the transformer. This shorted turn will absorb energy and lose it as heat. If instead of a single slab, you have insulated slices glued together, then you massively increase the resistance of this "parasitic" coil turn, so you don't get as much loss in the core.

Eddy current

 

[Red Squirrel]

Interesting info, so idealy the iron acts as a conductor for the magnetic flux, so the better your conductor the more efficient, which is very similar to how an electrical conductor works as well. With a tiny piece of metal the conductor is not really good and most of the magnetic flux is happening in the air.

 

[mindless1]

Meant the core. In my case it's just a flimsy little PCI slot spacer so maybe why it does not work, not enough metal to induce a decent field. If I put DC or AC current on the bigger coil it will act as an electromagnet though. A weak one mind you.

What were the other properties of the 30VAC power supply? With only 10 windings if it were not current limited at 30V it would probably melt the wire. If it is current limited then the voltage may be dropping very low or the PSU shuts down, or blows a fuse or other component.

 

[Piroko]

Hmm maybe I just don't have enough turns on the primary then (only like 10) since I tried and it did not work. Though, does it have to loop back around or can it just be a bar? Totally random experiment with zero purpose while I was bored:



Or could it just be that it's so inefficient my meter can't get a reading? Was using a ~30v AC adapter on the short coil.

Though I guess the other coil should be beside and not on top.

So basically the more metal that is around the coil increases efficiency as there's more mass to "store" the magnetism?

Yeah, you probably had problems registering a current with your Voltmeter (your typical Voltmeters need at least a small current for themselves to measure).

You need to keep in mind with your little experiment that:
U*I(primary) = U*I(secondary) (minus transformation losses)
...should result in a high voltage and a very small current with the number of primary and secondary loops you've chosen. Add that most power supplies can not sustain their rated output when you (almost) short them.

 

[Red Squirrel]

Oh that does make sense, could very well be what is happening in my case I'm just completely killing the source power supply when I put it in a dead short like that. The big coil is about 1 ohm even that is practically a dead short depending on how much voltage I give it. (ex: I would not plug that into the mains, or it will go poof) When I put it on the other side the voltage of the small coil is probably simply too small to detect with the multimeter.

Of course in a real transformer I'd be using much thinner magnet wire too, the coils being tighter (less insulation thickness between wires) probably adds to the efficiency as well.

I may experiment further, it seems I'm on the right track. Would be a fun project to build a super high voltage transformer from scratch to build a tesla coil or something. Sure you can buy one, but more fun to build yourself. :biggrin:

 

[Mark R]

The problem you've got is that current isn't being adequately limited in your coil.

It is the magnetic field in the core being reflected back into the primary that limits the current. If your magnetic flux is weak (because you have an air gap in the core) then very little gets reflected back and you get no significant current limit effect - the coil acts like a short circuit. (but see below)

You could make it work with a ton of turns, but a proper core would make a big difference.

Technically, any inductance has a current limiting effect when used with AC. However, there are limits. Inductance actually limits the rate of change of current (it makes rate of change of current proportional to voltage). In an AC circuit, when voltage is constantly changing and reversing, then this effect is to limit the current. Note that the "impedance" of an inductor with AC current is proportional to both inductance and frequency.

If you use a higher frequency, you can use fewer turns and a smaller core - which is how a SMPS is able to push 1000W through a matchbox size transformer, whereas a 60 Hz 1000W transformer is as big as a house brick and weights about 15 pounds.

If you use a ridiculously high frequency (100s of kHz, or MHz) then you can use a low-permeability air-core - and this is how a Tesla coil works. You needs such a low inductance, that an air core will work. (There is also the issue that you can't use iron or ferrite at these frequencies, because they magnetize too slowly).

I would abandon trying to make this work with a piece of iron, and find a scrap transformer; rip the windings off and then have a go winding your own on top of that.

 

[Thebobo]

Nice explanation Mark R. Linear technology has a lot of those opto less smps ic's.

If memory serves me right, i used this IC in a design for 48Vinput to 3x 12V/100mA out.

http://www.linear.com/product/LT3512

I need that with a 13 v input for my audio player in my car. It plugs into my stock CD deck input I hacked (tapped and put a swipe switch after the CD preamp and main audio amp.) Anyway stock 12 v to 5v .5 amps? volts to much noise coming from car.

 

[Red Squirrel]

On the subject of power supplies what would be a fun project is to convert a regular PC power supply into a -48v input one. Been reading up a lot on SMPS's and it seems like it would be doable without too many changes. Only thing is I'm not good enough at electronics to trust anything I make/modify for production use though. Even the transformer that's just something I'm doing for fun of it, not something I would actually use in production. I will probably retry it with a big bolt instead of that flimsy metal piece, and make way more turns.

If ever I do make a tesla coil I'd probably make something that runs off a low DC voltage just to be different. Pulse it at high frequency through a super step up transformer and go from there. I guess it would be closer to a taser at that point.

 

[edro]

Were you using AC to power the homemade transformer?
Where are you getting your 30v from?
You said you had a ~30v AC Adpater. "AC Adapters" output DC.
Transformers don't work with DC; they require the constant switching of AC current.

 

[mindless1]

^ True, AC-AC adapters aren't very common but I have seen a few around that voltage range, usually used on audio systems. It could also be the float voltage of a modular 24VAC nominal doorbell PSU or an industrial PSU for some random thing. Otherwise 30VDC aren't all that common either except for computer printers.

 

[William Gaatjes]

I need that with a 13 v input for my audio player in my car. It plugs into my stock CD deck input I hacked (tapped and put a swipe switch after the CD preamp and main audio amp.) Anyway stock 12 v to 5v .5 amps? volts to much noise coming from car.

Do you need electrical isolation between the 12V and the +5V ?
If not, this ic can be used and it can even be done on prototyping board under the condition that you create actual traces by soldering the wires on the islands and laying out traces. If you just wire stuff up, it will become noisy and you might hear it in your audio.
The NCP3063 from onsemi can be build up on a prototype pcb. Rock stable.
http://www.onsemi.com/PowerSolutions/product.do?id=NCP3063

The LT1375 from Linear Technology needs a bit more components but has a higher efficiency.
http://www.linear.com/product/LT1375

Since the 12V from an automobile is not rock stable and occasionally jumps a bit higher, the NCP3063 works up to 40V and the LT1375 works up to 30V.
Enough headroom to not get damaged by voltage spikes. Although a small input LC filter together with a transient voltage suppressor of around 24V on the input would not hurt.

If you really want it silent, you have to use an output LC filter as well to reduce the switching noise from the smps. something with a cutoff frequency around 1000Hz will work since that will attenuate enough (the smps switches at 150kHz and 500kHz respectively).


If you do need electrical isolation...
I do not know if the LT3512 will output 12V with a 12V input. I will see what kind of transformer you need.

 

[William Gaatjes]

For the LT3512 :
The flexible (air gap) transformer 749196241 or 749196141 from Wurth Electric will function just fine. It has 6 universal windings. You need to connect two windings together for the 12V input primary circuit. The dot end from one winding to the non dot end of the other winding. And the other 4 windings you can connect in parallel for the 5V output secondary circuit. Watch the dot on the windings for the secondary circuit. The dots must be tight together for phase reasons on the secondary circuit. After that, it is following the easy calculations in the datasheet of the LT3512.

http://katalog.we-online.de/en/pbs/WE-FLEX?sid=a20ca47332#vs_ct:1

http://katalog.we-online.de/pbs/datasheet/749196241.pdf

 

[Red Squirrel]

Were you using AC to power the homemade transformer?
Where are you getting your 30v from?
You said you had a ~30v AC Adpater. "AC Adapters" output DC.
Transformers don't work with DC; they require the constant switching of AC current.

And "AC" adapter outputs AC, a "DC" Adapter outputs DC. This is just a straight up AC adapter, no rectification. The 30v was AC. It came up to a bit higher when I was messing around with a capacity and rectified it. It's a "24v" adapter but without a load it seems to be closer to 30. It came with a humidifier that hooks to a furnace, but I just tapped into the "hum" terminal of the furnace when I installed it so I use this adapter for a low voltage source of AC for random messing around.

As a side note, a 63v capacitor does not like 170vdc, it makes funny noises. :biggrin: I'm sure it would have popped if I left it on more than a split second.