Original poster: Jim Lux <jimlux@xxxxxxxxxxxxx> At 08:01 PM 6/3/2006, you wrote:
Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx> Hi All,I got to looking at currents involved in ground strikes. Apparently the current is real high and the time is real fast in order to drain the secondary so quickly. But, How high and how fast?<snip>
40V at 10nS rise times!!! That is the limit of the scope now!!! And the real signal is obviously "faster still"...That puts the current pulse in the range of 10 - 20 nS and the current in the 1000 to 2000 amp range!! For a 108kV 3.7 inch spark off a 31.1pF toroid!!! Goodness only knows what a 5 foot streamer to ground is =:O The streamer resistance is roughly 100 ohms in that case.The peak power for a few nS is over 100,000,000 watts!!!! For just a "little" spark...
With the longer streamer, the pulse will last longer (that 1uH/meter inductance will slow things down)
Well.. think about this... say the top load is 30 pF, and it's discharging through what is effectively a wire a meter or so long (call it 2 uH). As a resonant LC, that's 20 MHz. And, you can do a simple Li^2=Cv^2 kind of calculation and see you get kiloamp kinds of peak currents (with 10 ns rise times). Given that some of the energy is actually stored in the charge on the leader as it's being formed, I wouldn't be surprised that it's a lot faster (and higher current). Consider this scenario... The streamer is reaching out towards the ground (but hasn't yet touched).. The charge on the 30 pF topload has moved almost entirely into the streamer (considering the streamer as a wire, the charge would take 5-8 ns just to move down the length of the wire). Then, when it touches, you've basically got a charged wire that's a couple meters long. This isn't much different from firing a transmission line pulse generator.. the equivalent inductance is very low as the traveling wave propagates down the arc. Rather than a nice LC, what you really have is something more like a step function with very high di/dt propagating along a transmission line.
At a guess, the step might last a few tens or hundreds of pS (how ever long it takes the current to start flowing.. comparable to spark initiation time in a gap, which is well under a nanosecond). di/dt might be in the MV/ns range, which is really quite impressive.
It's not the air streamers or the 100 kHz RF from the TC that kills electronics nearby. It's the magnetic transient from the ground strikes. As you note, very high currents, very fast rise times make for a great EMP generator. I did some calculations a few years back on this (they're probably in the archives) and after doing it, I was amazed at how much voltage you could induce in a relatively small victim loop a fair ways away.
Those TCers who have fooled with fast marx banks or exploding wire rigs have already encountered these sorts of problems. Kevin Ottalini killed a bunch of power supplies when building his little table top Marx, before realizing what was going on.