Re: Ground strike currents and time...

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>

At 11:17 PM 6/3/2006, you wrote:
...........

> > http://drsstc.com/~sisg/files/scantesla/DanQ-08.gif
> >
> > 40V at 10nS rise times!!!  That is the limit of the scope 
> > now!!!  And the real signal is obviously "faster still"...
>
> And that might actually be more the ringing in the wire and/or 
> transmisson line than the actual current waveform.  How long is the wire?

6 feet.  We are probably just seeing the "ring" of the cable.  Like a 
bell, we hear the ring, but not the strike...

Paul said:

> That's a very nice recording of the discharge.  Notice how it
> decays away and then comes back a little after 100nS or so.
> It's like there's a little echo at about 550nS - a reflection
> from something 50 - 80 metres away?  Seems you might have made a
> TDR to send pulses into the ground system.  The 'echo', if it is
> one, is not inverted, thus indicating a bit of an open circuit
> at the discharge frequency.

There is a big problem of course, the waveform has very little DC 
offset.  Instead of being a nice RC discharge such as 108000 e^ 
(-t/(128 x 31e-12), it is an oscillation.  The bell is ringing from 
the strike, but the actual strike is far too fast to see.  Jim 
suggest the actual strike might be on the order of 10's of pS!!

This was all done to suggest how fast the models should discharge the 
system in a strike, - That has been answered.  Just pick a time, and 
zero all the variables "instantly" ;-))  I transfer the energy to the 
load energy.  I imagine a fair amount is actually radiated....

If the time is say 100pS.  Resistance is gone and only inductance of 
the path matters as Jim says.  If we arc a few inches to another 
large sphere of considerable capacitance then the inductance of the 
arc path might be say 100nH.  Ipeak = Vp x SQRT (L / C)  == 108000 x 
SQRT(100nH / 31pF)  ==  6.1 million amps!!!  However, the actual 
current flow might be governed by how fast the electrons can actually 
"move".  Light only travels about in "inch" in 100pS so the electrons 
in the arc would be governed mostly be the propagation speed of the 
electrons then.  For a say 12 inch sphere, the electrons on the far 
side need to travel pi x r  or 19 inches.  That takes 1.6nS.  So the 
current might be a 1/2 sine pulse based on integral of the surface 
area being discharged over time.

So we can forget about LRC factors and consider it as only a problem 
in moving charge from the surface of sphere A to sphere B limited 
only be the speed of light and perhaps electrostatic propagation 
factors as the electrons travel over the surface of the sphere.

Since C = Q / V  We have 31pF = Q / 108000  or Q = 3.35 
uCoulomb.  The peak current is 2Q/t = 6.7e-6 / 1.6e-9 
coulomb/second  = 4188 coulomb/second or 4188 amps peak.  The current 
pulse would be:

i(t) = 4188 SIN(1.96e9 x t)  from 0 to 1.6nS.

Forgive simplifications, math errors, and general errors :o)))  But 
that does make sense and it shows I can stop trying to see it on the 
scope :D  It does also suggest that for good big close conductors, 
the "speed of light" is the limiting factor.  If we add long wires 
and long arc paths things slow down and the equations may change if 
other factors slow the pulse down more than the speed of light.

For a beat up oxidized corrugated aluminum dryer duct toroid held 
together with duct tape.  A real current pulse reading would me 
mostly very fast "noise".  But for electrically and mathematically 
nice top terminals, the pulse might be fairly easy to predict 
accurately.  There are probably eddy current and inductance of the 
sphere surface effect to worry about but stuff like that makes my 
brain hurt %:-)

That's my thinking of the moment...


Paul also Writes:

> I wrote:

http://drsstc.com/~sisg/files/scantesla/DanQ-08.gif


> > there's a little echo at about 550nS - a reflection
> > from something 50 - 80 metres away?

No.  Near field dirt, concrete, heavy house on top...


> Correction, that's 550nS into the trace, only 350nS after the
> pulse, so 30 - 50 metres.

That pulse was surprisingly repeatable.  Possibly a cable reflection 
even if they were all terminated.  But the loop was 50 ohm in 
parallel not series so the termination was not really "right" on the 
source.  I did add a 50 ohm tiny resistor in series with the source 
for proper termination later, but it was all monster noise then...


> Terry wrote:
> > The 20MHz 110 sees something, but it is pretty garbled.
>
> That 110 signal looks quite interesting.  I shouldn't write it
> off. What's that long low frequency ringing?  Have you got the
> 110 terminated correctly?    And there seems to be a change to
> the character of the signal at about 110nS and again at 350nS
> after the pulse.

The CTs were terminated correctly.  The waveform was zero after the 
shown point.


> Wonder what point in the waveform the discharge cuts off?
> At 10nS, 350nS, or much later - after a uS or so?  Can you
> capture a light curve?

What is a "light curve"?


> Lots going on in that little snap of a discharge.

Mostly noisy ringing I think.  The magic "strike" is totally 
invisible.  I could move the probe away and the signal went away too, 
so the scope and instrument side where clean.



Cheers,

        Terry