Making a transformer 5

[BobLWeiss]

I need to make a air-core type transformer that can produce 6 volts DC (rectified) and at least 600 ma's. What I am doing is winding the secondary coil with 27 awg magnet wire with a diameter of 1" about 150 turns. The primary will be same wire but 100 turns and about 1 1/4" diameter. The secondary coil will need to be able to spin freely in the middle of the primary for my project. I plan on sending 12 vdc at 200Khz through the primary coil to exite it. Am I on the right track or would I need more/less windings on either coil? What about heat buildup? I am not that experienced with transformers to figure it out and was hoping someone here had some knowledge? Thanks!

 

[ScottyUK]

Hi Bob,

When you are doing you tests for secondary voltage, are you loading the secondary winding? If you are testing with an air core, you will have poor mutual coupling of the coils, giving the low voltage, and the voltage will drop very quickly as you load it because the leakage impedance is very high giving poor regulation. Improving the mutual coupling and reducing the leakage are the two key elements for success in this task.

If you have any high-school physics books to hand, these often have a reasonably easy to follow study of magnetically coupled air-cored coils.

Good luck!


----------------------------------

If we learn from our mistakes,
I'm getting a great education!

 

[Warpspeed]

Excellent information Scotty.

It is very important to keep the air-gap as small as possible, better still run no air-gap. Ferrite is extremely hard and smooth and should run o/k as a dry bearing.

Scotty is also correct in pointing out that the drive waveform must be symmetrical. The best way to do this is to drive a flip flop that divides the frequency by two. That way, the positive and negative half cycles will always be exactly the same time period, even if the frequency changes. Some of the switchmode power supply control chips have an oscillator, a flip flop, and an output stage that might have enough grunt without using anything else.

It may be advisable to connect a large non polarised capacitor in series with the primary. That will prevent the driver stage from blowing up if the oscillator stops.

To get suitable bobbins to wind your primary and secondary, three segment plastic bobbins should available to suit the particular pot core. Just cut out the centre section leaving two narrow bobbin halves that will each occupy around one third of the internal space. It is about the only way you can do it to get mechanical clearance between the rotating and non rotating parts.

I still believe a pair of slip rings would be easier, smaller, cheaper, more efficient, and have less rotating friction.

 

[BobLWeiss]

Would the output performance be any better if I was to just put 16v AC through the primary instead of pulsing DC? What kind of power output could I expect using that? I know its hard to measure but a range would be good enough for me at this point. Would increasing the current on the primary side also help?

Thanks!

 

[Warpspeed]

If you use a ferrite pot core, it must run at relatively high frequency to keep the flux density well below the saturation knee. As Scotty says, 300mT is a good safe maximum design figure to aim for.

Very high frequency sinewaves are much more inefficient to produce than squarewaves, so why would you want to do that ???

 

[ijl]

If you still prefer an air-core transformer, you might try the following:

Instead of a “plain” transformer, use the primary coil together with a capacitor over it as a parallel resonant circuit. The pulsating DC cannot be connected directly over the primary. Instead, add a tap (wire out) at about 1/10 of the total number of turns, and connect your pulsed source between this tap and the lower end of the coil. No "freewheeling" diode needed! A first guess for the number of turns in the secondary might be about half of the number of turns in the primary. Adjust the frequency of your source so that the circuit is in resonance.

The idea behind this suggestion is the following: The coupling between the primary and the secondary is weak in a coreless transformer. This means that you need a large primary current in order to get the desired secondary voltage (and current). The current in a good resonance circuit builds up to a value that is much larger than the current it takes from the source.

I made an experimental transformer of this type. I used the components I found hanging around, without any optimisation. The main parameters are the following:

Primary winding has about 230 turns (in four layers) of 0.6mm enamelled wire on a core with a diameter of 53mm. The length of the winding is about 50mm. The inductance of the primary winding is about 2.6 mH. The tap for the pulsed source is at about 30 turns from the lower end. The secondary has 100 turns (in one layer) of 0.25 mm wire on a core with a diameter of about 45 mm. The capacitor over the primary winding is 47 nF, so that the resonance frequency is about 15 kHz.

I connected a 20 ohm resistor directly over the secondary, without any rectification. With a DC source voltage of 9V, the AC peak to peak voltage over the resistor was about 15V, so that the transmitted power is about 1.4W. It was difficult to estimate the efficiency, but it was rather low, perhaps only about 30%. It should be possible to increase the efficiency by a careful design and/or experimentation, but you have to experiment a lot. In any case, this resonant circuit-transformer worked much better than a plain transformer made of the same coils.

Finally, to your question of the use of 60Hz (or 50Hz) AC. I do not think that it is a good idea. Because the inductance of an air-core transformer is small, the idle current will be large at this low frequency.

 

[ijl]

BobLWeiss
One more thing. I checked your link (the "example") first now. The transformer is defined as "20:25" in the link. My understanding is that this means that the primary has 20 turns, and the secondary 25 turns. This is the information you are after, isn't it !

The thickness of the wire is not very important. Something around 0.2...0.5 mm should be fine, I guess.

 

[BobLWeiss]

Yes I saw that and wasn't sure if he meant 20 turns or a ratio or 20 to 25. Like 100 turns to 125 turns. Also he lists how much current he is able to get from it and thats not enough for what I need. I need something like 600 ma on the secondary. I saw someone else make one and they did it a little different. They made the primary and secondary both the same size (2" diameter) and put the secondary on top of the primary seperated by a piece of plastic that was spinning. The distance between the 2 coils was about 1/8". Do you think that way is better or is it better to put the secondary in the middle of the primary coil?

Thanks!

 

[BobLWeiss]

I built a circuit to power the primary and it is driven by a 200Khz oscillator and then fed into a pair of transistors (PNP/NPN) and then fed into a 8A MOSFET. One end of the coil is wired directly to a 12vdc source and the other end is hooked to the MOSFET. The drain of the MOSFET is hooked to ground. The problem I am having is that the MOSFET is getting real hot right away. Should I use a resistor to limit the current, or properly heatsink the MOSFET as this is normal? The circuit that I am building is:

Circuit[/LINK]

He does not use a resistor and mentions that nothing runs hot...so I am not sure what I am doing wrong. I am not a beginner to electronics and I know the circuit is wired correcly.

 

[ijl]

1. One coil inside the other or on top of the other?
Well, I have no direct experience here. My _guess_ is that putting a well prepared flat coil on top of another flat coil might be better, because the gap can be smaller.

2. Running hot?
Heat sink yes, resistor no. Have you a load (a resistor, about 10 ohms, 10W) in the secondary circuit? You should. Is the FET switching properly on and off? (just guessing). This might boil down to the often mentioned weak coupling between the primary and secondary.

 

[BobLWeiss]

1) The coil is on top of the other seperated by a 1/8" pc. of plastic. They are not necessarily flat but more like a thick rope shaped in a circle.

2) I didn't have a resistor on the secondary and I didn't think that putting a load on it would do anything towards the MOSFET getting hot but now I can see why.

 

[GOTWW]

While we are talking high freq's and narrow gaps, why not just use a air dielectric capacitve feed through. Much simpler in principle.

 

[ijl]

A capacitor, or actually two, why not. But a high frequency and a narrow gap would still be needed. And a high voltage, because the capacitance would be small, of the order of 100 pF, depending on the physical size and construction. A capacitance of 100pF and a frequency of 1 MHz, for example equals to 1.6 kOhm. 600 mA at 6V are needed. Thus, a step-up transformer to around 600V and a step-down transformer back might do the trick.

 

[GOTWW]

Headlines: electrocution by toy clock!

 

[ijl]

Yes, that is "why not".

 

[ScottyUK]

Totally different idea: reduce the load power consumption. 3.6W is a fair bit with modern low power IC's and high efficiency LEDs. What is the load comprised of?



----------------------------------

If we learn from our mistakes,
I'm getting a great education!

 

[BobLWeiss]

I am using 2 pics (low power versions), 1 Real time chip, 43 High Brightness LEDs, 3 mosfets and 1 IR module. I figured it to be around 600 ma consumption but thats on paper. The LED's are only on (at not all at one time) for 1/2500 RPM. (estimate). Some are on (about 5) are all time.

 

[GOTWW]

If you have a 10MHz crystal source driving a 5W narrowband amplifier, into ijl's 100pF feed-through (rotory) cap, and on the rotor a series 2.533uH inductor for series "warpspeed" resonance (The L-C disappears!!). Then into a small bridge and cap for your source. Plus no need for rubber gloves (ie. 600V) when trouble shooting, the rotor does not even need to be turning to probe other problems!