A few rules of thumb for the solid state crew.
Please, exercise extreem caution if you choose to follow these rules of
I am looking for anyone who is interested in designing a tesla coil that
follows a series of design rules that I have found that in simulation
produces up to several Megawatts done right. I know the calculations, and
the basic design rules that are used to produce such a circuit. This all
starts with this basic tuned circuit as an example of exactly how a primary
winding should be resonant. Now, with a solid state oscillator this is
possible but, if you insist on using spark gaps then the means isn't
effecient based on certian basic principals of signal inputs, and type of
signal as to whether the proper circumstances exist for the circuit to
Now, just a capacitor and a coil produce the results that you see on the
breadboard when the circuit is at resonance. Now, I looked at my past
example, and it lacked a bit of information, and details that on the
schematic shown the wrong parts values for the same circumstances to exist
on the primary as this single tank circuit.
There is a way to do this right, and in small scale versions I have seen
the same results but, I cannot afford wire that can handle a Megawatt, let
alone several. The only way I have found to control this is the fundimental
oscillator frequency, and the tuned gang. Basically, what I am trying
relate to you is that I don't know how you can control all of the energy
that I can get to pump through a circuit with not more than the 5 volts that
are driving the 1 to 100 step up arrangement in the link I have provided.
The turns ratio is providing the voltage but, the current provided by the
oscillator is a very small fraction of the power actually used by the
circuit. So, when I take 3.488 amperes, and divide it by 8.743 mA a power
gain ratio factor of 398.948.
I have spent months studying this and only a sine wave can produce a
factor that high, and it must come from the oscillator which drives the
tuned circuit. The oscillator must meet some very specific criteria in
order to even start a low impedance tank circuit, and bring it to a point of
a high effeciency oscillation. In my experiments, a low impedance tank
circuit could rest at only providing a root of the square as an output when
the output impedance of the oscillator is higher than the independant coil,
or capacitor in the circuit as inductive reactance, or capactive reactance.
The first 10 cycles of the oscillators output clear to the first few hundred
may demand as much current from the oscillator as is found in the high
energy resonant state. Which would be 3.488 amperes for the circuit in the
link I provided. This means that at 5 volts you would have a impedance that
when the circuit started was only 1.433 ohms. With each successive cycle
from the oscillator the impedance of the tank circuit goes up, and
eventually reaches 571.886 ohms.
Now, in order to generate more energy than that, you need to step up the
voltage, and step it back down. If the inductance in the example of the
primary were 1 uH, and the secondary were .1 mHs then the next coil to step
it back down would have to be .1 mHs, and its step down secondary would need
to equal 1 uH, and match the step up transformer to bring the power back
down. Transformers do not have a trend for a high energy loss, they tend to
only convert amperes to volts, or visa versa with very little power loss.
Now, steping the voltage back down includes another parallel tank circuit.
The value of this capacitor can only be one of two values, and aid the
present situation found in the primary of the oscillator driven side of the
circuit. The first even harmonic up, or twice the oscillators frequency, or
at the same frequency. In my studies of this specific circuit arrangement.
I found that the average power output is more reliable at this stage of the
circuit at choosing the capacitor that combined with the step down coil is
at 2 times the oscillator frequency. Now, if you measure the inductance of
the primary of the step down coil, then calculate the value of capcitance
that is require for the resonant frequency, I use the inductive reactance
formula to turn an use the capacitive reactance formula and find the equal
value of reactance at the resonant frequency. Then I simply take that
precise value of capacitance, and divide it by to, place what ever
capacitors in parallel of a smaller group of capacitance values to
accumulate the precision value of capacitance I require. I measure the
resulting value of capacitance because, exacting it, centers the frequency
exactly. So, when building the actual circuit, the oscillator is a Wein
bridge, Armstrong, Colpits, or other variable oscillator that produces a
strong sine wave. The wave shape is everything here as to whether this
circuit will produce for you or not. There is nothing wrong with
hybredizing high frequency responce Op amps with Armstrong, or Colpits style
inductive/capacitive networks, and in fact they tend to operate more
consistantly than a single transistor arrangement. From there you need a
good amplifier to match impedance with the first stage tank circuit at half
of its calculated inductive reactance at its resonant frequency or it will
not start to resonate effeciently.
What is wrong with a square wave is that it is actually established as a
series of odd harmonics in relation to the actual resonant frequency.
Resonance is a result of the time constants of the inductor, and the
capacitor. Eventually a capacitor charges completely but, it inherently
slows down as it reaches its peak voltage, and an inductor will not reach
it's maximum rate of current flow until a specific span of time has passed.
But, at that point it's resistance to an AC current is at it's lowest, and
the charge capacitor can then be discharged by the reversal of current flow
through the inductor, which places the inductor in the position of first
opposing current flow, and eventually allowing for it. With a square wave
the instance is different, and the capacitor suddenly charges expending the
energy as a high frequency bandpass event, as the square wave is allowed to
remain on for its time period it then near it's end becomes a inductive
short, and a low pass event. So, at the very start of a square wave, and
the very end of a square waves half cycle some significant part of the
energy is bypassed to ground, and the tuned circuit is acting more like a
filter removing the square edges of the input. This just rules out chopper
circuits as oscillator drivers of maximum effecency at any frequency for
this style of circuit..
Going back to the original values of inductance, that are the same in
this single transformer arrangement at the link I have provided. At an
inductance of .1 mH to center the frequency on the first harmonic would
require a capacitor of 500 pf to step down a twice the oscillator frequency.
The energy output will be much higher than the input but, the signal will no
longer be sine but, chaotic, and the average power stable over 5 milli
second period, or less at higher frequencies. If you attempt to tune both
circuits exactly to the same frequency you will only find yourself
approaching an AM type of modulation that can end up at 1Hz. The average
power then is 1/2 for at every 1/2 second. It doesn't make sense to have a
low frequency of that nature hanging there, when a chaotic will maintain a
maximum output power factor at the same average power in only 1 millisecond.
If the resonant frequency were 159 KHz for the primary of the oscillator
driven tank circuit then the second stage would be at 318 KHz. Now, if you
were building a Tesla coil the output of the second transformer would be fed
to the primary of the Tesla, and it's resonant frequency would need be 318
KHz, or 636 KHz which would be experimental beyond the scope of the circuit
I have found appropriate.
Please, exercise extreem caution if you choose to follow these rules of