Henning - 15 September 2010 08:38 AM

I think they don’t say anything about a current rise time of 100ns from 0 to 1 MA with a resonating circuit. That’s the problem.

Somehow I like the idea of getting rid of the switches a lot.
Would there really be an issue with rise time?
Basically as I see it is all about the current - how fast you discharge your capacitors into the induction coil.
Still I think you would need some kind of switches to connect the resonating circuit with the DPF, but probably you could choose between voltage and current.

I’m not sure how you would get a 100ns rise time for 50kV/2MA with an RLC circuit without loosing most of the power to resistors. Or choose LC / RC / RL circuits if you like, which are in reality RLC circuits, just with one of the factors being very small. Then I’m not an electric engineer, so prove me wrong.

Remember: You don’t get a perfect signal from the induction coil + photovoltaic system + steam turbine. At least the input of turbine is mostly independent of the fireing rate. And even if you only take into account coil + photovoltaics, you’re getting in only a relatively shallow input with a rise time of maybe a millisecond.

Additionally to that you’ll need different resonating circuits for different firing speeds. Here you need the switches again to connect the different circuits to the DPF, but with less strict specifications.

remember that the parameters R, L, and C are a natural part of the circuit. R comes from the resistance of the feed wire, etc. The whole point of ZCT techniques would be to use multi-phase resonance, and switch at precisely the right time, so there is zero stress on the switches.

if you need a 24-phase (or more) circuit to accomplish that, i think we’re okay.

if you can put up with a 300V transition on a 50kV feed, you can still select your firing rate.
but even that transition could be eliminated with additional small-resonance circuitry.

bringing this back to the topic of history, here’s a 1989 U.S. patent for an alpha-particle capture coil:
—http://www.freepatentsonline.com/4835433.pdf

Putting the rise time issue aside one interesting solution to high power problem could be having multiple resonance circuits and connecting them in serial/parallel manner to increase voltage/current. You could take 40 circuits with 1000 volts to get 40kv voltage. The multi-million amperage seems more of an issue.
The question is how do you control such a large number of circuits precisely? This project seems worth of ITER challenge, but the small testing circuits connected a mini-dpf could be worked on by students on an university level.

vansig - 13 September 2010 04:35 PM

...

that depends on the size of the wire. if 50 ns is the needed rise time (= 1/2 wave, ie: 10MHz), and you can tolerate a .006 ohm resistance, the skin effect limits the wire dimensions (in copper, ρ = R·Area / length = 1.68×10^−8 Ω·m; skin depth= 20.5 μm @10MHz) to a length no more than 7.2 x the circumference.
...

So, per the formula, a 1” wire connector could be only a bit less than 2’ long. Tight quarters!  The caps would be in a tight ring around the vacuum chamber, I think.

Brian H - 16 September 2010 09:48 AM

vansig - 13 September 2010 04:35 PM

...

that depends on the size of the wire. if 50 ns is the needed rise time (= 1/2 wave, ie: 10MHz), and you can tolerate a .006 ohm resistance, the skin effect limits the wire dimensions (in copper, ρ = R·Area / length = 1.68×10^−8 Ω·m; skin depth= 20.5 μm @10MHz) to a length no more than 7.2 x the circumference.
...

So, per the formula, a 1” wire connector could be only a bit less than 2’ long. Tight quarters!  The caps would be in a tight ring around the vacuum chamber, I think.

I’d build the ring to minimize distance to the anode, where it exits the base plate, which is at the high voltage system’s ground potential. The extreme is to build with a single cap.