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

RE: The 1500t secondary myth (long)

Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi,

At 10:08 AM 12/1/2004, you wrote: 
I thought I better put in my 0.006 GBP on this one ;)

>It's not a myth --- it's physics, pure and simple.
>The potential developed across an inductor is V = -L di/dt.
>It is very apparent that the larger the inductance, the large the output
>potential

This is true. But also, the larger the inductance, the lower the operating
frequency. Hence you don't really push any more current through a given
streamer capacitance than you did before (since I=C dv/dt)

If everything else is the same, the lower frequencies also tend to need a higher inductance primary coil which tends to lower primary currents which reduces primary losses significantly. Higher inductance primary systems tend to have lower losses in the coil, wire, gap, etc.


By this line of argument I wouldn't expect massive differences between
higher and lower frequency coils. And indeed in practice this seems to be
the case. There does seem to be a sweet spot around the 1000-1500 turn range
but it's nothing drastic.

DR R. also tends to make very nice coils with low loss primary systems. His results may not be completely true for those of use whose coils are not as well built. But the 1200 - 1500 turn idea seems to be a nice area to work in. You can't really go too wrong with it.



The new theory that I have been working on (with a lot of input from Terry,
Malcolm Watts, and others) suggests that the sweet spot is not any
particular number of turns, or even any particular inductance. Rather, it's
when the resonator has a characteristic impedance (Zo) of around 36,000
ohms. Zo is a function of both secondary inductance and secondary/toroid
capacitance:

Zo=sqrt(L/C) or Zo=2*pi*f*L or Zo=1/(2*pi*f*C) where f is your coil's
resonant frequency

To check my theory, I have calculated Zo for various coils with good
performance documented on the net, and they all come out around 40 to 50k.
See my earlier posts for details. John Freau's TT-42, which is still one of
the most efficient coils ever, has Zo=44k. John found that adding a bigger
toroid to his TT-42 (which would lower the Zo towards 36k) increased the
spark length.
And, Richie Burnett found that a coil with Zo=22k performed poorly, but
changing to a resonator with Zo=~50k gave much bigger sparks with the same
bang energy.

Hence I think there is fairly good evidence for my theory. If anyone knows
of any good coils with Zo greatly different from 36k, please let us know.

That Zo range does seem to be a winner! We don't know enough about streamer dynamics to know "why", but secondary impedance matching should be "real important" according to the known data and models. With just simple static models, playing with loading and impedance has dramatic effects. You say 36K is the sweet spot, I'll carve that number into my desk ;-))



So my recipe for an optimised coil is-

1) Decide what length of sparks you want

Using john Freau's formula, you can come up with that based on input power too which is usually a good starting point given what type of NST one found ;-)


2) Work at around 100 bps 

120 in USA  ;-)


3) Use Freau's efficiency equation to get the required bang energy

You usually have to start with the power source one finds and then get the best streamers out of that.


4) Size the toroid so it will just break out at this bang energy

I tend to like to over size them. I have had many too small, but none that were "too" big. You can always force breakout of the toroid were too big and too smooth, which never seems to be a problem.


5) Choose the secondary height big enough to avoid flashovers and racing
arcs (ie more than 1 meter of height per megavolt- you can get a more
accurate figure from Paul Nicholson's TSSP voltage gradient plots)


Yes!  Don't make it too small or too short.  You just can't "fix" that.


6) Now choose the secondary diameter and number of turns so that Zo is 36k.
If this is inconvenient (maybe it gives too high an operating frequency for
your *SSTC) you can increase the inductance provided that the "Q-limited
streamer length" works out bigger than your target streamer length. I think
I already posted how to calculate this.


CW coils may have different Zo matches than tube, disruptive, etc.

http://hot-streamer.com/TeslaCoils/MyCoils/CWCoil/CWImpedance.txt


7) Finally choose the primary and tank capacitor to match your power supply
voltage/current and secondary frequency, and deliver the required bang
energy. The procedure for this is well known already (at least for spark-gap
coils and OLTCs)

One will have to match the primary cap for the LTR NST match. But higher inductance primaries reduce losses if that can be worked in too. However, in the case of the DRSSTC, very low loss primary systems might not be good as Dan found. A little resistance there can dramatically reduce IGBT currents without hurting performance much.


PS. For what it's worth, I favour using a "fat" secondary like Dr. Resonance
does. A fat secondary is better at "catching" the magnetic field from the
primary than a thin one, so you can achieve high coupling without getting
the primary dangerously close. And it makes the coil look mean. It also has
a higher unloaded Q than a long skinny one, but in the light of the new
theory I don't think this is much of an issue (as the loaded Q of our
optimal coil is only 6)

I also like larger diameter secondaries as long as they "look" right. I note that most coils all have pretty much the same general proportions that have been carved out over the years. Making things that look "odd" compared to other coils is usually not good unless one knows exactly what they are doing.

Cheers,

        Terry


Steve C.