TSSP: List Archives

From: "Terrell W. Fritz"
Date: Sat, 23 Jun 2001 21:45:45 -0600
Subject: [TSSP] Direct Voltage Measurement of Topload Voltage for TSSP

Hi All,

I sketched up some ideas for a 500kV capacitive voltage divider at:

http://hot-streamer.com/temp/HighVoltageProbe.gif

The top diagram shows the circuit.  The 5pF and 250,000pF (oops, says
50,000pF :-)) capacitors form a 50,000:1 divider.  This takes a 500kV input
and divides it to 10Volts.  That signal is buffered and gain adjusted to a
nice 100000:1 output signal that can be feed to a fiber optic coupling
system.  The voltage could actually be taken off the lower cap and feed
directly to a scope with two 50ohm matching resistor and 50 ohm coax if one
was "feeling lucky".  That is how most commercial dividers do it.

The 5pF cap can be a coaxial capacitor made from a 12 inch sphere with an
~12 inch long 1/4 inch rod protruding from the bottom.  The rod leads
inside an oil bath.  There is a 5 inch diameter "can" inside the bath into
which the rod goes in about 5 inches.  This basically forms a 5pF cap that
can stand off 500kV (hope hope!).  There is another grounded can around the
first which basically acts as a shield for the inside can and adds some to
the divider capacitance.  All the parts are oil insulated and it would
stand about 30-36 inches high and ~12 inches in diameter.  It would need
6-10 gallons of oil.  The top sphere maybe could be smaller for less
capacitive load.  There would also need to be a "cable" from the coil to
this probe that would add capacitance and maybe add other problems?  There
really aren't probes like this around.  The commercial ones are
high-frequency but to catch "spikes" rather than to continually run.  The
"real" high frequency probes don't go this high in voltage...  I bet there
are reasons for that :-))

One problem may be that the amplifiers can't take a 20V p-p signal given
their unity gain bandwidth and all.  However, that is not hard to fix.  I
could easily add the direct 50 ohm stuff for safer lower power tests.

This is just my quick sketch of the thing.  If anyone has suggestions or
sees any flaws in my thinking here, I am all ears.  The commercial dividers
use gas (probably SF6) but that is hard to get.  The fiber optics add a lot
of complexity but until one is sure it works or one can develop really good
proven protection circuits, the fiber optics provide absolute protection no
matter what.  I expect this stuff as full of "problems" :-))

Cheers,

	Terry


Maintainer Paul Nicholson, paul@abelian.demon.co.uk.