[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: The 1500t secondary myth (long)

Original poster: Greg Leyh <lod@xxxxxxxxxxx>

Hi Steve,

Determining the best Zsec to drive the arc impedance seems to be a good design
approach. Malcolm's argument that the loaded Q can't be <<6 is intriguing! I'll have to think more on that one to understand it.

Having a good-sized, lowZ capacitance [large toroid] appears to be critical too,
perhaps to support the *fast* [10s to 100s of nsec] dart leader features that
Jim Lux mentioned earlier in this thread. Unfortunately, larger toroids tend to
pull Zsec downward, so a compromise must be struck here.

One potential limitation as coil scale increases is the inevitably decreasing Zsec.
Based on my noisy statistical data, I set the lower limit for an acceptable Zsec
around 20k, since the 120L50K coil at the shipyard operates with Zsec~20k,
and gives nominal output arcs for its secondary length. http://www.lod.org/Projects/120L50K/120L50K.html

Note the relatively short arcs of the first three photos, compared to the
fourth photo.  The coil features the larger toroid in the fourth photo.
The new
power level with the larger toroid is about 25kW, yielding a nominal arclength
per power of about 1 foot/kW.

As it turns out, other factors limit ALF's size before the Zsec drops too far.
At present, the Zsec for the ALF is 38.8kOhm per tower [1186mH at 5200Hz.]
Site power is the principal limitation, at just under 8MW. Hopefully the energy
recovery scheme will provide some extra margin here.

Zsec = 36k does seem to be well in the range for good performers, although
the crest of that curve seems quite broad, with a lot of statistical noise. It sounds
like you might have better statistics on coil performance than I do, however.

-GL


Original poster: "Steve Conner" <steve.conner@xxxxxxxxxxx>

>I haven't been following things enough to see where the 36k came from

I worked it out in the following way- According to Terry the impedance of
streamers is 220k (irrespective of the length) in series with 1pF per foot.
So as the length of a streamer tends to infinity, the impedance tends toward
a resistance of 220k. It sounds counter-intuitive but in my own experimental
work I've not found anything that would disprove this.

And Malcolm Watts argued that the loaded Q of a Tesla coil can never be less
than 6. He predicted that if you try and design one with Q<<6, the streamers
would just shrink to restore the Q to 6.

Again this agrees with experiment. It's particularly obvious with SSTCs: as
you drive them harder you reach a point where the system doesn't accept any
more power and the streamers don't get longer. At this point, the only thing
that will give more spark length is adding a bigger toroid (ie lowering Zo)

Note: this is not a quantitative proof because SSTC streamers probably have
a lower resistance and higher capacitance than Terry's 220k/1pF formula
which was for disruptive TCs. But qualitatively it seems to back up my
argument.

Putting these two constraints together means that the Zo of the resonator
should be about one-sixth of 220k. This will guarantee that streamer length
is never limited by running out of Q.


Steve C.