Original poster: "B2" <bensonbd@xxxxxxx>
Hi Marco,
High and low bandwidth high voltage probes have been in some of my
experiences. Your answers to the questions have helped me to gain a
better perspective of what you setup is. I will share some of my
experiences. I cannot explain a lot of the theory. Maybe you can help
me out with that?
T = (RC)/6 risetime formula from Standard Techniques for
Dielectric Tests, July 12, 1968, IEEE
C = 2.7 meters (your probe length) x 111pF/meter = 300 pF Approximately
R = "20 Mohm" as stated in a previous post (Is this correct?).
T = (20,000,000 x 0.0000000003) / 6 = .001 very roughly
If you believe this formula, then the response time of your probe
should be approximately 1,000 Hz. I believe that you want 1,000,000 Hz
or better for accurate Tesla Coil work.
For lightning work within about the same frequency range, I have used
20,000 Ohms and less. For a Tesla coil this might be difficult. This
sounds close to what Terry measured for the output impedance of a Tesla
coil? I don't know how to make this work for measuring the output
voltage of such a low impedance source. Matching the impedance to that
of your coil could provide some interesting data such as Imax into its
characteristic impedance.
I would try the toroid without the resistor. Run a copper wire or
tube through a sliding feed-through such as a SwageLok. The end of the
wire or tube is terminated with a small plate. The height of the plate
is adjusted to intercept the least amount of flux that will provide a
good signal.
You could also try using the wall as a vertical ground plane. Stick a
wire with a disk on the end through one of the bolt holes. Put your
oscilloscope on the outside of the building, also. This would be like
putting the Tesla coil in a screen room.
I was once told by an engineer that the input of an oscilloscope
should always have a 50 Ohm feed-thru terminator on it. I asked why. He
said that the terminator was cheaper to replace than the oscilloscope.
The output of a liquid resistor probe should be around 100 volts to
overcome distortion from electrolytic effects associated with the metal
electrolyte interface (experience only). I am not sure how to account
for these effects other than to perform experiments with a sinusoidal
waveform (without the Tesla coil connected or shorted to ground).
> 6. Why are the voltage divider plates so wide? (In my
> experience 1.4"
> (3.5 cm-approx.) worked down to 1 ns.)
The basic idea was to have a self-compensated design (see the principle
drawing on the top on my web page). I just thought the top, bottom and
tap plate should be about the same size. Why a smaller tap plate would
be better?
> 7. Why wouldn"t the capacitance limit the bandwidth?
We had the following preliminary thought with a friend of mine. The
divider is compensated if R x C of the lower arm equals the one of the
upper arm. Or, better, if every section of the divider satisfies R x C =
constant. In a homogenous field that is ensured by the basic design (one
pipe, one diameter). In the non-homogeneous field between the probe
toroid and the TC toroid the same result is achieved by varying the
probe section diameter together with the field strength. That's what I
did.
Do you mean that the capacity of the upper arm to gnd still affects the
bandwidth and spoils my design?
Make the plates infinitely wide and it seems that the capacitance
would short out the signal that you are trying to measure to ground. I
don't know how the size affects the signal when smaller. I always try to
make it as small as possible with the afore mentioned in mind.
I will wait and see how you receive what has been said so far before
saying any more
B2.