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

Re: Question about capacitors for coil



Original poster: "resonance" <resonance@xxxxxxxxxxxx>



Yes, your program provided good verification of this equation.

Thanks for the info on the losses though --- I wasn't aware of that data. I hope we haven't confused the new coil builder too much! A larger sec coilform would be in his best interest.

Dr. Resonance

Although spark length is really directly related to power, not potential, it does require a high potential to begin the ionization process, so the basic idea is to apply the equation for potential in an air coil inductor:

It seems to require "both" power "and" potential for the optimal sparks. There are very high voltage coils that perform poorly and there are very powerful coils that perform poorly. But high voltage powerful coils do very well. This problem shows in ScanTesla's streamer length predictions if one uses parameters that are not "typical". The program does do it fairly well now in most cases by combining power and potential together to come up with leader length and strike distance. It is not perfect yet or anything, but there are no giant obvious errors now. I also have plenty of fudge factors to fix the remaining small errors :o))

.......

The rate change of current, dI/dt really depends on a number of efficiency factors such as spark gap resistance, primary connection resistance, etc. We use 4 AWG welding cable with copper battery lugs for our primary connections (inside tygon tubing) and use 3/4 - 1 inch dia. copper pipes for our sparkgap thus keeping the switching (dI/dt) value high as possible in current and fast as possible in quenching time of the sparkgap). High current (the dI) and fast speed (the dt) is important.

From the old post at:

http://www.pupman.com/listarchives/2006/Aug/msg00201.html

At t~0:

dI(t)/dt = Vfire x SQRT(Cp / Lp) x 1 / SQRT(Cp x Lp) x COS(t / SQRT( Cp x LP))

If t ~~= 0 and with some simplification we get:

dI(t)/dt = Vfire / Lp

dI/dt is really almost totally "independent" of losses!!! What losses do is "eat up all the power" before it turns into streamers by destroying the primary system's "Q".

Qp = SQRT(Lp / Cp) / R

R (losses) kill primary "Q", not primary dI/dt.


It is very interesting to note that the Q numbers for the SISG PIRANHA are:

Qp = SQRT(15e-6 / 165e-9) / 0.16 = 59.6

However, if it used a conventional gap where R = 3.0 ohms:

Qp = SQRT(15e-6 / 165e-9) / 3 = 3.17

It would not even run "at all" with conventional spark gaps!! Thus my warning to "don't even try" to use that topology without SISG spark gaps!!

http://drsstc.com/~piranha/PIRANHA/PIRANHA.pdf
....

Assuming the average coil's efficiency around 70%, then the sec voltage equation becomes:

Vsec  =  Vpri x 70% x SQR (Lsec/Lpri)

I've used antenna current x field measuring systems developed by Terry Fritz (our super-duper moderator) and this equation is very close ---
usually within a few percent so it really does work.

Glad to know that number works for someone else too!!!! It is hard to get independent verification of such things!

...

Cheers,

        Terry



Happy coiling,

Dr. Resonance