[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Vortex gap loss measurements
Gary, Dan, John, Jim and all.. :^)
It turns out that introducing virtually any spark gap into a single
(uncoupled) RLC circuit will induce linearly-decrementing behavior
instead of the exponential decay seen with a classic RLC circuit. It's
interesting to rediscover this, since it was originally reported by Dr.
J. Zenneck almost 100 years ago (1904!). Because of the linear
decrement, tank circuit energy goes to zero in a sharply-defined period
of time. The waveform on Gary's site clearly shows this happening in 295
uSec for the Vortex Gap. An exponential decline will asymptotically
approach zero but will never quite get there (at least in theory).
The spark gap's nonlinear, arc-like behavior is characterized by a
relatively constant voltage drop (only tens of volts) that's pretty much
independent of gap current. The actual voltage drop is mostly a function
of the geometry, materials used in the gap, gas pressure, and the type
of gas(es) and metal(s) that form the arc. The instantaneous energy
that's lost in the gap is almost directly proportional to tank circuit
current (E = V*I), and it's this characteristic that actually drives the
John H. Morecroft devotes a fair section of his book to the behavior of
spark gaps in RLC and coupled circuits in "Principles of Radio
Communication" (in all three editions, but the 1921 1st edition is the
most thorough). Of particular interest is an equation, originally
developed by Zenneck and Stone, which describes the linearly
decrementing tank circuit current as a function of the Effective Gap
Section A: Section B: Section C:
---------- ---------- ----------
I = -E*Sqrt(C/L) * (1-(R*t)/(2*L)) * sin(t/(Sqrt(LC))
While Sections A and C of the above equation simply describe the peak
current and sinusoidal oscillations, section B represents the linear
decrementing term that forms the "envelope" of the waveform. If the
spark gap behaved as a pure resistor, section B would instead be an
exponential function of time. The current envelope declines to zero when
(R*t)/(2*L) = 1.
This means that, if we know L and measure t, we can solve for the
effective gap resistance R and use it to compare different types of
gaps! Morecroft suggests that gap resistance was actually governed by
the magnitude of the FIRST current maximum. Apparently, the higher the
first current peak, the larger the initial plasma channel, and the lower
the effective gap resistance for the remainder of the decrement. This
also implied that there was relatively little modulation of gap
resistance by the oscillating RF current inside the envelope.
Let's re-look at Gary's experiment now in the light of the above
relationship. Gary's primary circuit resonated at 138 kHz, and per his
coil specs, he was using a 0.021 uF tank cap. This implies a tank
inductance of about 63 uH. The Vortex gap decremented to zero in about
295 uSec, while the vacuum gap did so about 17.5% quicker, or in about
243 uSec. This implies that the vacuum gap lost energy more quickly, and
thus had a higher effective gap resistance. Solving for the effective
gap resistances of the two gaps styles:
RVortex = 2*L/t = (2*63e-6)/(295e-6) = 0.43 ohms (lower = GOOD)
RVacuum = (2*63e-6)/(243e-6) = 0.52 ohms
The Vortex gap has significantly lower gap resistance, possibly because
of the greater number of charge carriers available at the higher
operating pressure. However, gap resistance is only one parameter making
for an efficient gap. Another is its quenching capability. Higher
pressure gaps often take longer to quench. However, some of Gary's
earlier measurements seemed to indicate that the vacuum gap was not
quenching very well, so it remains for another set of experiments to
determine if the Vortex gap has better quenching ability.
Great job, Gary!
Safe coilin' to you all!
-- Bert --
Web Site: http://www.teslamania-dot-com
Tesla list wrote:
> Original poster: "Lau, Gary" <Gary.Lau-at-compaq-dot-com>
> Today I found some time and performed a comparison between the gap losses of
> my single vacuum gap, and my new single vortex gap. To do so, I scoped the
> primary ringdown with no secondary in place. I used a Terry Fritz fiber
> optic voltage probe across the primary coil and a digital storage scope to
> record the results. I have not yet accurately calibrated the voltage
> readout, so for now, the results are just relative to each other.
> With no secondary in place, the ringdown is a linearly decrementing
> waveform, not logarithmic. As such, the slope of the ringdown indicates the
> losses in the circuit and is independent of the gap firing voltage. I
> performed ringdown slope measurements at a variety of gap widths to vary the
> initial voltage, but the ringdown slope is a constant, independent of Vgap.
> The power to the blower motor is varied through a lamp dimmer and I tried
> varying the motor speed to see what effect that had. At very low speed, the
> linearly decrementing waveform became slightly logarithmic-looking, but
> still predominantly linear. The gap breakdown voltage appeared to change
> slightly at low speed, but this was hard to measure as it was slight and the
> bang-to-bang gap breakdown voltage is not as consistent as one might hope.
> The slope decrement figures are assuming that my probe is accurately
> calibrated for voltage, though I suspect it may not be, so the figures are
> useful only for relative comparison purposes.
> The pressurized vortex gap decremented at 200V/usec.
> The vacuum gap decremented at 235V/usec (17.5% faster).
> The vortex gap breakdown voltage is about 20% higher than the vacuum gap at
> the same gap distance.
> Vortex gap web page:
> Vacuum gap web page:
> Regards, Gary Lau
> Waltham, MA USA