RE: Tesla coil tuning

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "BJ" <firebee@xxxxxxxxxxxxxxx>

In principle, your first idea of sweeping has some merit. A squarewave
signal should still produce a global max (for primary resonator) and min
(for secondary resonator). A simple averaging peak detector will work. You
don't need to have accuracy in an absolute voltage/current sense. Just sweep
and search for the max or min across the range. A bigger issue will probably
be getting useful frequency resolution over a wide sweep range if you
planned to use the microcontroller as the signal source (or frequency
counter if using a separate sweep source).

Your second idea will work in theory, but the maths is not trivial and would
need floating point as well, especially if the test frequencies are
arbitrarily chosen you could be a long way off the resonant response and two
or more of the equations would be almost identical. This solution also does
not lend itself to measuring the primary/secondary coupling factor, which
can be determined from the two response peaks seen when the primary and
secondary are coupled together. For those who operate DRSSTCs, this is a
very useful parameter to measure as part of the tuning procedure.

A while back I experimented with an oscillator configuration that could
operate with either a series or parallel LC resonator (series mode when
connected to the secondary and parallel mode when connected to the primary).
The oscillator then automatically fired up at the resonant frequency, and
only a frequency counter is required. This suffered in two respects; first,
it didn't help with determining coupling factor. Second, my large secondary
acted as an antenna picking up local EMI and phase modulating the
oscillator, hence it was not stable enough for my liking. So the parts went
back into the junk box !

Regards
Bevan

-----Original Message-----
From: Tesla list [mailto:tesla@xxxxxxxxxx]
Sent: Wednesday, 18 April 2007 8:58 AM
To: tesla@xxxxxxxxxx
Subject: Tesla coil tuning

Original poster: "Breneman, Chris" <brenemanc@xxxxxxxxxxxxxx>

Hello,

I'm thinking about trying to make a digital Tesla coil tuner based on
(probably) an Atmel microcontroller, and was trying to think of
various ways to determine the resonant frequency digitally.
My first idea was to make something similar to Terry's TCT, but
digital.  ie. sweep a frequency range when connected to the
primary/secondary circuits and see which frequency passes the
highest/lowest current.  The problems with this are that it's not
trivial to accurately do this digitally because of the harmonics
inherent in any digital square wave and the difficulty in accurately
measuring RMS voltage/current on a waveform that's not perfectly
sinusoidal without carefully calibrated equipment.  This method would
also be relatively slow.
Then I had another idea ... but I'm not sure if it would work.  Would
it be possible to calculate the DC resistance, inductance, and
capacitance of the primary and secondary (and from that, the resonant
frequency) by applying three sine wave (generated by external
oscillators) of different frequencies, measuring the current for
each, and solving a systems of equations?  This method would be much
faster and would solve many of the problems above.  Any feedback?
Also, if anyone else has any other ideas for methods, they are very
welcome.  If I get it finished, I'll post the schematic, parts list,
microcontroller code, and hex dump of the code.  It should be a
pretty cheap way of easily and accurately measuring resonant
frequency (if it works).  Atmel microcontrollers are all <$10 a piece
(usually around $5 or less from retailers), and the other components
(like cheap LCD display, resistors, caps, etc.) would be pretty cheap as
well.

Anyway, thanks for any feedback,
Chris