The RF output sprayed into the air will light up a fluorescent tube several feet away, set fire to one of MSC's potted plants placed on top of the secondary L2, or engulf an entire door in a continuous wave of bluish-purple RF power arcs.
Basically, it is a resonant high-frequency transformer capable of generating
RF on the order of kilovolts or megavolts, which can both transmit power
and create big sparks!
Here is a simple theory of operation,
followed by the story of an actual implementation.
Figure 1 is a rough diagram of how the coil
works. These are the keys to getting a really, really, high voltage
at the output terminal:
The seond point is also pretty obvious. L1 is typically on the order of 10 turns. L2 may be 3 feet of closely wound 26 guage enamel magnet wire. This is on the order of 20 tpi * 36 inches or 720 turns. The turns ratio would be 72 in this case, giving a classical transformer output of 15 kV * 72 = 1 MV.
The way the circuit really works is that
The RF output sprayed into the air will light up a fluorescent tube
several feet away, set fire to one of MSC's potted plants placed on top
of the secondary L2, or engulf an entire door in a continuous wave of bluish-purple
RF power arcs.
Although I don't have a photo of the aforementioned pottled plant, here is an arc drawn from the top of L2 to a ground wire held in the hand of the author. The arc here is about 6 inches long.
And this is the corona discharge from the top of L2.
In 1986 through 1989, I embarked on a technology pilgrimage, a quest
for the holiest of Grails: the million Volt, ear-destroying, ozone-spewing,
EMI-radiating Tesla Coil power arc. I came close, if not in the voltage,
at least in the bonus side effects! I estimated about 600 kV, based
on a 15-inch (40 cm) open-air coil to ground arc length, a 3 MV/metre
breakdown voltage for air at STP, and a 50% derating for increased field
strength due to radius of curvature effects (OK, pointy parts, for you
artsies).
This model wass known as The Electrocutioner, or der Scharffrichter.
In the beginning, natural scientists called voltage "EMF", or electro-motive
force. An apt, name, for this is the Prime Mover. In the case of this pilgrimage,
the prime mover consisted of 115 VAC provided on a standard 15 A circuit.
And He looked at the EMF, and He spake, saying, "Hmmm, we're still a
bit short."
The next piece of technology was a pair of 30 mA, 15 kV RMS surplus
neon sign transformers, purchased for about $50 apiece, from commercial
establishments in the Kitchener Waterloo area. These provide a nominal
450 VA each, supplying me with almost a kilowatt of drive power.
I have heard estimates that the final arc length (output voltage) can be
estimated in an efficient design simply as a function of input power (can't
remember the scale factor, though).
This is component T1 in the schematic above.
The transformers are physically wires as a centre-tapped output winding,
with the centre to ground (the case). This keeps the potential between
each terminal and ground symmetric, so you don't have one hot end and one
cold end. This prevents us from wiring the two transformers in series
without REALLY stressing the insulation (read: very bad thing to do), so
I connected them in parallel, for 15 kV at 60 mA.
Now that I have a hundred bucks invested in my sign transformers,
I don't want to blow them up by letting a kilowatt of RF go back into them
and melt that nice black tar in which the windings are potted. (Well, there's
another story about that, and all I can say is many thanks, to Fred
in the Engineering Student Shop, for milling the case off one of those
puppies so I could repair the primary after it shorted and spewed molten
steel all over my nice polyester residence carpet...)
The protection consisted of as many turns of about 22 or 24 gauge stranded insulated wire as I could wrap around a nice ferrite core to make a choke about the size of a golf ball. You might say to yourself, "That doesn't seem like much of a choke". Well, let me tell you about the day I accidentally wired the chokes in the primary discharge loop! After powering up, the output seemed abnormally low, so I shut it down after about 10 seconds. After unplugging and probing around the primary capacitor, I almost burned my finger off when I touched the chokes! They had been disapating most of the kW power into the hysteresis loss of the ferrite!
Now for the main high current part: the main capacitor, the spark
gap, and the primary coil winding.
At 15 kV RMS, we're looking at about 22 kV peak across the main capacitor. For some reason, memory tells me I actually had double that across the caps, possibly because we need to worry about the sum of the RF in the primary discharge loop and the low frequency "bias" supplied by T1. Also, the 15 kV was a loaded voltage. under open-circuit output it would be more (probably by a factor of root two).
For maximum power transfer, I needed a capcaitor with an impedance at 60 Hz that matched the source impedance of T1.
Z = E/I, or 15 kV/60 mA.
Zc = 1/(2*pi*F*c)
C = 1/(2*pi*F*Zc)
C = 1/(2*pi*F*E/I)
C = 1/(2*pi*60*15000/0.06) = 10.6 * 10^-9 F, or 10.6 nF
Now a 10.6 nF cap rated for 44 kV peak is not something you pick up at Radio Shack! This puppy can store 10 Joules, which can do serious damage to your heart, not to mention those annoying little holes through the surface of your skin where the spark enters. But enough about those incidents.
TheARRL Handbook (amateur radio relay league) provided assistance here, in the form of tables of dielectric constants for printed circuit board matrials. I was able to determine that a choice piece of surplus store double-sided blank copper clad board approximately 3ft x 3ft would be more than adequate for my needs. As experiments later showed, it didn't quite have the dielectric punchthrough strength I had hoped for, and I was forever drilling blowouts and patching them with epoxy and Syncone, a cheap and pleasantly aromatic substance available even today at Canadian Tire stores.
(Side note: do NOT, under any circumstances, attempt to build a nitrogen pulse laser out of shards of rough cut glass, attempt to glue the pieces together using Syncone to fill the gaps between the ill-fitting pieces, and then go to sleep in a summertime 30 degree roasting oven of a tiny residence room containing freshly squeezed Syncone. Remember that comment about being "pleasantly aromatic"? Well, "toxic and nearly fatal" might be a better phrase.)
The primary coil was pretty simple. I was too cheap to buy 1/4 inch
copper pipe like most of the books recommended, so I used roughly-woven
18 guage solid house wire. This gave an effective 14- or 16-guage wire.
HF skin effects may have nulled much of this out, but the multi-stranding
was an attempt to model Tesla's recommended Litz wire.
The next challenge was the spark gap. Apart from the constant
blowouts of the main capacitor, the gap gave me the most problems
of all. There were a number of issues around the spark gap:
The gap should have flat electrodes or spherical electrodes to keep
the field low for repeatability of performance.
I wound up trying metal furniture casters about 1 inch in diameter,
but they were hard to keep parallel. Ithen tries using 3/4 copper
water pipe caps, with similar results. My final solution was to use
two pillars of 1 1/2" ABS sewer pipe (MSC's old standby...) with holes
bored in for 1/4 inch threaded rod. Copper water pipe caps were soldered
(and I use the term loosely:-) to the ends of the rod, so that by turning
the threaded rod you could adjust the gap. Now that I actually have useful
tools in the garage, I think one could do a much better job without too
much work.
The picture above shows the rudimentary ABS pipe and copper cap gap I used for a time.
The initial secondary consisted of 24 AWG vinyl covered stranded
wire wrapped around a garbage pail! I still remember the day I bought that
pail. The hardware clerk took a look at the young engineering student,
passed me a knowing smile, and asked how much beer I thought I would be
able to brew in it!
As things turned out, without a reasonable attempt to design the coil based on classical cyclindrical inductor equations, the resonant frequencies were off and I really didn't get much output. I then tried much finer enamel covered magnet wire on a piece of 4" diamtere ABS sewer pipe. I made a section about 3 feet long and it still did not have anough output. After analysing the frequencies I believed I was still off.
I then discovered that the catenation of the pail with half of my ABS
pipe topping would give me results that were pretty good. I was able
to demonstrate the tuning of primary and secondary using an HP signal generator
and an old Dumont scope. As a first approximation, you could even see that
a neon bulb would light up on the secondary when the primary was driven
by the generator at the right frequency!
This photo above shows the pail and pipe secondary. The black wire near the base of the pail is the primary coil. The shiny rectangle in the foreground is the foil and glass cap.
Alas, all good things must come to an end. And thus it was
with der Scharffrichter, the Electrocutioner.
Amidst untrue allegations of a number of infractions of residence rules, I was paid a visit by the campus radiation safety officer and a special Hydo inspector. The end result was that der Scharfrichter was hidden under a friend's bed (which later turned out to be structurally unstable; that's another story) while the visit occurred. The residual equipment in my room was enough to convince the inspectors to "strongly recommend that the equipment be relocated to a designated laboratory".
The photo above shows that sad day; the Electrocutioner, the trusty scope and HP generator, the Nuclear Data multichannel analysers, and a few other tidbits. All packed up and ready for transport to a "designated laboratory"... heh, heh, heh.
To be continued.....