Quote:
Originally Posted by Sean Williams
... it seems to me to be a simple class c oscillator, using the secondary coil as a loosely coupled antenna.
|
Hi Sean, welcome to the club!
You're absolutely right. It's a classic class C oscillator.
The main difficulties arise with the grid bias and feedback circuits (and coupling generally) both of which are experimentally determined. And also, not getting flashovers, but as you have power RF design experience, this will be trivial for you.
I don't think that's the main reason why MOTs were popular. It was rather that they were easily available from junked ovens. They are designed, obviously, with a magnetron as a load and they get very hot very quickly in Tesla coil operation. I think they're bad candidates.
Quote:
Originally Posted by Sean Williams
...what has me slightly baffled is the design of the secondary coil....nearly all of the design data I can find so far indicates the secondary coil resonating at around 1/10-1/20 of the primary frequency - am I missing something fundamental here?
|
I don't know what designs you've seen but yes, the tank circuit comprises rather few turns and hence the L component on its own has a relatively low inductance. But the parallel LC circuit together is resonant at approximately 75-80% of the secondary coil resonant frequency.
All except low-powered valve tesla coil designs I've seen over the years have a parallel LC circuit. The capacitor is typically strontium titanate, (one of those common RF doorknob types) and usually between 3-5nF. Or you can build what's called an MMC from tens or more pulse caps.
The streamers at the antinode have some (series, obviously) capacitance with the secondary coil. It might only be a few tens of pifs but it does detune the coil variably variably.
Usually the self-resonant frequency of valve tesla coil secondaries is lower by means of a large topload, typically toroidal, but spherical will do. A break-out point is essential.
It's recommended to build an interrupter to lower the duty cycle of the valves if you don't want to red-plate the anodes in seconds. The 'on'-time of the square pulses only needs to be 20mS for persistence of vision to kick-in. Typically you modulate the cathode via a big MOSFET thesedays.
The RF circuit needs to be good, but you're expert on this - very modest post, considering your background!
This design was John Freau's and is one of the best out there. I got 22 cm of streamers from a single Svetlana 572b after my first build in 2001, when I built the first of a total of five (to date) coils.
(I abandoned the last EHT one because of my orthopaedic injuries and being unable to safely shift a circuit that already weighed 35kg and wasn't finished! I think Albert's rig is twice that or something.)
Happy to help esp with typical glitches and I'm familiar with good resources and bad!
P.s - I am still working on my solid-state build, very slowly. I thought I'd give Albert plenty of elbow-room to exhibit his amibtious project! So there are now three known Tesla-coil activists on the forum!
Spark length in inches for a valve Tesla coil is estimated by 0.5 *Sqrt (power in)