Hey all,

So, I have this project that involves running a fairly meaty amount of current through a very low Z coil (and array of coils sometimes).

meaty ~ 10-40 A

low Z ~ 10-40 mOhms at DC

L ~ 70uH

I was able to accomplish some great results at 60Hz with a variac and an old rackmount UPS power supply xFormer. However, I need to put higher freqs through it, 1 and 2 kHz would be nice. Sine wave.

I don't really care about THD or anything like that. And I'm definitely heating this coil up, which is actually the point.

Long post short, I was going to use an audio amp with some dummy resistors in series. But I'd like to make something more specific and basic. And I'd rather not destroy said audio amp if I can help it. maybe something like a tank circuit but for these lower f. I'm a little rusty though so I figured I'd ask.

I'm debating FETs vs. BJTs, maybe the equivalent of a common-collector arrangement....


Any thoughts?

 

70 uH at 1 kHz is 440 mOhms so you will be dominated by the reactance, although you might have more R at high frequencies from the skin effect - it depends on the construction of your coil.

I know of audio amplifiers designed to drive 2 Ohm loads so you're not that far off - only a factor of 4. An 800 Watt subwoofer amplifier (true 800 W rms) would deliver 20 amps into 2 ohms, it could drive a 440 mOhm load without extra resistance so long as you're careful you don't exceed the current that would flow into 2 Ohms i.e. the 20A.

You may be better off building/buying a transformer to match your 440 mOhm to a standard 8 ohm audio amplifier of a few hundred Watts, and even if you make your own driver consider a transformer output to get it up to a sensible impedance. You're in the territory where you can still use an iron cored transformer.

Personally I would wind a transformer with a centre tapped primary and a single secondary, with perhaps a 4:1 turns ratio - that would transform your impedance to around 8 Ohms. I would then drive it push pull with two bipolar transistors from a supply of a few tens of volts. At those frequencies you wouldn't need many turns. It depends which end of your range you want, 10 amps or 40, 500 Hz or 2 kHz. Decide what you really need and I could be more specific.

 

If you can use a single frequency, then you might try resonating the coil with a parallel capacitor. For example a 220μF motor run capacitor (non-polarized) would give a resonant frequency of 1.28kHz with you 70μH coil. With that the parallel resonant circulating current will be much higher than the input current, which an audio amplifier could likely supply. Or use that LC as part of an oscillator to automatically operate at the resonant frequency.

 

I like crutschow's suggestion of resonating out the L, and if you make a self-oscillating circuit that eliminates the need to tune your driver to match the resonance (useful if you want to switch in more or less C to alter the frequency).

I would suggest you use caution in your choice of capacitor though. You have several tens of amps passing though the capacitor that may have significant ESR and limited ability to dissipate heat due to its layered construction.

ESR information for motor run capacitors seems to be rare, but consider that the suggested 220uF capacitor would only pass 20 amps at 60 Hz 240V. It would probably cope with 40 amps (and 4x dissipation), but a 220 uF 110V motor run capacitor capacitor might not as it will have been designed with lower current in mind.

 

........................

I would suggest you use caution in your choice of capacitor though. You have several tens of amps passing though the capacitor that may have significant ESR and limited ability to dissipate heat due to its layered construction.

ESR information for motor run capacitors seems to be rare, but consider that the suggested 220uF capacitor would only pass 20 amps at 60 Hz 240V. It would probably cope with 40 amps (and 4x dissipation), but a 220 uF 110V motor run capacitor capacitor might not as it will have been designed with lower current in mind.

That's certainly a valid concern. I suggest monitoring the temperature of the capacitor by periodically touching it and if it gets too warm, shut off the circuit.

 

Great suggestions guys! I am partial to the ease of the tank circuit. I figure that the most important freq is 1 kHz with 2 kHz being a close second. However, I might need to do a 20-25 somewhere down the road....

So, maybe I can make a modular(ish) amplifier stage and then swap out or switch output components as needed.

I am leaning towards a pair of TO-3s I have around. 100V/30A

I also thought about running a motor backwards and generating current that way. But, there are some inherent caveats with that as well.

I'm gonna be playing around this evening. I'll let y'all know if I had any success. And if you don't see a post from me in the next day or so. I've probably 'crossed over'.

Regards

edit: p.s. I think 10A is a good starting point. I can get a lot done with that. Any higher is a bonus.

 

If you go to 20 kHz you'll bring the impedance up to where you can drive it directly without trouble.

The only motors you could reasonably get a kHz out of would be stepper motors, and you'd need a very large one to deliver the power you need. I can't see that being practical for even 10 amps, especially given you can only use half the windings (the other half being at 90° electrically).

There is an option that combines the transformer with the resonant capacitor. If you do end up with a transformer-output amplifier you can put an extra winding on the transformer, say with 10 times the turns of the output winding. You can then put a much smaller (but higher voltage) capacitor across that. You'd need 2.2 μF but with a rating of a few hundred volts. The transformer than converts the -72j Ω of that small capacitor to the -0.72j Ω you need to resonate out the +0.72j Ω of the coil at 2 kHz.

This is useful because a 400 V 2.2 μF will often be cheaper and easier to find than a 220 uF 40 V that can take the ac current - The suggested motor run capacitor would be an excellent choice but rather expensive. It's probably not worth doing though unless you're already winding a transformer as part of the amplifier.