2002-10-21
Almost a Tesla Coil
Over the weekend I had a chance to catch up with my electronics stuff which I haven't touched for ages, work has been nuts. I was sitting there building an inverter to drive a fluro tube, when I got interested in the properties of the primary and secondary windings, they were misbehaving, pulling the oscillator to one frequency, then another. I guess it was their self-resonance frequencies they were settling in, so curious as I was I wanted to measure them.
I tried a few ways, external drive coils with the CRO across the the coil, driven directly measuring the absorbed current with a in-line resistor, and probing with a dipper. It turned out that the primary's was about 1.6 Mhz and the secondary was about 50 kHz. The results of all methods agreeing with each other to within a percent or so.
All this fooling with inductors got me curious, so I put aside the project for a moment and started testing every inductor I could find. After finishing the junkbox full and promising to make a jig or tool to do this useful array of tests in the future really easy, I started winding my own with the left over wire from the inverter transformer. I think I learnt more about parasitics in inductors that night than ever before.
I've ended up with a simple measuring system, the square wave generator on my bench. Just injecting it into the inductor directly gives you almost all the information about an inductor you'll ever need, its resistance, its capacitance, its core losses. Try it sometime you'll never just grab a molded choke and whack it in the circuit without thought again.
So anyway, I decided to wind a long solenoid on a cardboard tube and measure its properties. A few turns at one end allowed tight coupling for injecting RF and seeing what happened. Here is the test set-up, my half-watt RF amp boosting the output of my signal generator:
The solenoid has a self-resonance of around 730 kHz with both ends of open. On grounding the bottom end, this falls to about 715 kHz kHz. Those numbers surprised me a bit, although the wire is normal multi-strand hook-up wire of very poor quality, $DEITY knows what its dielectric properties are. Its self-resonance harmonics are not 'perfect harmonics'. This is normal for a coil but somewhat surprising at first. My amp doesn't cover its fundamental, so I fed it at its higher overtones. Grounded is on the left, open on the right.
Note that there is no connection to the frequency counter. The huge field around this coil fed at resonance is penetrating the shielding of the meter and inducing RF currents directly in its input circuits. The frequency was checked by direct connection to the counter and by reading the roughly correct generator scale.
Considering the input power is only about 500 mW the field is very strong. Strong enough to light a neon bulb held near the coil. Tuning is touchy, my body capacitance pulls the coil around as I approach so I had to tune the generator while I was near the coil. Strange effects could be witnessed, like going too close detunes the coil so much the bulb goes out and multiple anti-nodes in the field along the axis of the coil. The photo doesn't really do it justice, the flash has washed it out, the neon was glowing very brightly, like a mini road flare. You can also see the PSU's drop-out light is on, the RF is upsetting its ripple detection circuit:
It is possible to strike a tiny RF arc across the output terminals, perhaps 1mm long at best. I would loved to have had a fluro tube handy (considering I was building an inverter for one to begin with!) but all I had was that lousy neon that has seen better days in a ion counter.
With more RF power I'd imagine I would have had a small Tesla coil. It was clearly a very high impedance at the top when grounded at the bottom. My poor little half-watt amp just isn't up to job though. It got stinking hot running into this load (I was abusing it too, running it from 24 volts to get more power). I would like to build a dedicated driver amp or maybe a power oscillator that uses the secondary as its frequency determining part, that would solve the detuning as I load up the coil approaching it. I've always wanted to build a Tesla coil, and a solid state one seems achievable now. It shouldn't be too different from making a lowfer power amp, or class-D transmitter, hell it is almost identical to the original inverter, complete with its parasitic oscillation problems.
1 comment.