Re: Who needs a quenching gap ?

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Bert Hickman" <bert.hickman-at-aquila-dot-net> 

Tesla list wrote:
> 
> Original poster: "Finn Hammer" <f-hammer-at-post5.tele.dk>
> 
> Gang!
> 
> Today I finally managed to get a scope trace off the new RAT-coil under
> streamer loading.
> 
> This was particularly satisfying, since it confirms the Microsim based
> preliminary design effort, which predicted a clean 2nd. notch quench.
> This was recorded at 100BPS, about 1 meter streamer length.
> 
> http://home5.inet.tele.dk/f-hammer/ringdown.jpg
> 
> Horizontal=50µS/div
> Vertical:5kV/div
> 
> Since the gap is a SRSG with onely one gap in operation, and should
> therefore be a poor quencher (but possibly a "low losser") this picture
> tells, that this coil is quenching the gap very well.
> 
> I believe it is possible to determine the coubling from this trace, how
> is that done?
> 
> The coil is really just thrown together along the proven R.Quich
> guidelines 1000 turns, 1:3 dia/length (although it is a fine spacewind
> on threaded form). So the impedance match btwn streamer and coil is not
> precise. Zcoil/Zstreamer= 1.6. I have tried to find a way to meet this
> match, and if anyone is interested, we could talk more about this
> subject. According to this work, the perfect impedance match is met,
> when the toroid is arounf 10pF. I can hardly imagine that this would
> make the coil perform any better than with the present 32pF . I have
> never managed to create a model of a coil, that quenches at the first
> notch, so perhaps 2nd. notsh quenching is natural, in some sort of way.
> 
> Comments?
> 
> Cheers, Finn Hammer

Hi Finn and all,

Based upon the above waveform, it appears that your system's coupling
coefficient is relatively loose, being somewhere around 0.1 - 0.125. The
approach that Terry mentioned (estimating "k" by taking the reciprical
of the number of peaks during the first primary ringdown) is accurate to
within 5-10% for the typical range of "k's" used in 2-coil systems.
Quenching becomes easier as you decrease "k". The fact that you "lost" a
bit of your lower secondary winding is undoubtedly why "k" is lower than
originally desired... :^). With a low "k" your system's output (and
energy transfer efficiency) is likely not as good as it could be -more
energy is getting burned up in the gap before each "bang" is fully
transferred between the primary and secondary. 

As long as we have sufficient topload C to supply "instant" energy for
streamer propagation, and the surge impedance of the resonator is low
(relative to the streamer impedance) so that it can efficiently deliver
sufficient displacement current to the streamer to keep it "hot", the
actual value of secondary impedance is likely NOT very critical. 

Some further speculation on quenching:
=====================================
There is empirical evidence that there's relatively little to be gained
(at least in terms of streamer length) by using heroic efforts to
achieve 1st notch quenching versus 2nd notch. This appears to be
related to the way energy residing in the secondary:toroid system
couples to streamers. Because of the physics underlying streamer
propagation, streamers will propagate only while the secondary voltage
envelope is RISING - a pre-established streamer will NOT propagate any 
further if the secondary voltage envelope is falling. 

A "perfect" (1st notch) quench will result in only ONE rising secondary
voltage envelope, followed by a more slowly declining envelope as the
secondary rings down (i.e., the main gap goes open-circuit). It takes
significant energy to initally establish a streamer, or to rebuild one
from a previous bang. Once formed, and during ringdown, significant
displacement currents will flow between the toroid and the established
streamer. This displacement current will help to keep the channel hot,
but it will NOT support further propagation during secondary ringdown.
So we'll get a streamer of average length (L1) by the time we hit
maximum secondary voltage, but that's it.

Suppose instead that we quench on the 2nd notch. We now have another
opportunity to further extend the streamer during the rising voltage
created by the second ringup of the secondary. In well-built efficient
coils, the amount of energy available on the second ringup will still be
as much as 75% that of the first ringup. Further streamer propagation,
building off the 1st streamer, may result in average streamer lengths
(L1 plus delta) significantly longer than streamers from a 1st notch
quench alone! 

Well, if 2nd notch quenching is better, is 3rd or 4th notch quenching
be even better? Apparently not! It takes more energy to keep a longer
channel hot via displacement current joule heating. By the time we get
to the third notch, we may no longer have sufficient output voltage to
support further streamer propagation. Reduced secondary voltage also
reduces the magnitude of displacement current heating essential for
maintaining the conductive channels. As the outer ends of the streamer
cool down, they become less conductive, reducing streamer capacitance
and further reducing displacement current heating in the rest of the
channel. This sequence of events may lead to a death spiral for the
streamer itself. With less energy per bang available for streamers
because of 3rd notch (or higher) quenching, more energy will be required
to re-establish previous streamers during subsequent bangs, so we'll
tend to see less streamer growth from bang to bang. The observed result
would be reduced average streamer length, and weaker looking (lower
current) streamers. 

Perhaps an ideal system would use maximum coupling consistent with 2nd
notch quenching??

-- Bert --
-- 
Bert Hickman
Stoneridge Engineering
Email:    bert.hickman-at-aquila-dot-net
Web Site: http://www.teslamania-dot-com

-- 
Bert Hickman
Stoneridge Engineering
Email:    bert.hickman-at-aquila-dot-net
Web Site: http://www.teslamania-dot-com