HV Choke Design (fwd)

---------- Forwarded message ----------
Date: Sun, 18 Jan 1998 14:04:53 +0000
From: Gwyn Zucca <gwyn-at-tcbod.demon.co.uk>
To: tesla-at-pupman-dot-com
Subject: HV Choke Design

Hello All,

Regards HV chokes it seems to me that in order to meet the main criteria
a few design/construction problems always arise. For the choke to be
effective it needs to present a high reactance to high frequencies which
means a reasonable value of inductance, however the choke must not only
have this inductance but should have a low value of self capacitance as
once you get above the chokes resonant frequency its impedance becomes
capacitive in nature and so its effectiveness decreases, also if the
choke rings near to the TC operating frequency it is sure to be useless.

Due to the difficulty in obtaining large ferrite ring cores I have
devised a different approach. What I did was as follows.

I used small cores which were originally commercially wound common mode
mains suppression chokes, removed all the wire and bobbin to just leave
the ring core. These cores are of the right material for RFI suppression
use (Generally Mn/Zn ferrite of a high permeability). These type of
chokes can be obtained as surplus for very little money! 

The cores were insulated with heat shrink sleeving and wound with 17
turns of PTFE insulated aircraft wire again bought as surplus, in a
single layer wind as this gives the lowest self capacitance, the cores
must be insulated because not only does the ferrite have properties of
conductance but also tents to have a high permittivity so capacity to
the core is reduced if the wire is kept from being in contact with the
core material.

Nine of these cores were wound as a continuos string as below (If you
experiment with one core to find out the length of wire needed and then
multiply by the number of cores this makes the hand winding easy, BTW
the PTFE slides through the cores with no problem if you take your

   |---| |---| |---| |---| |---| |---| |---| |---| |---|
===|   |=|   |=|   |=|   |=|   |=|   |=|   |=|   |=|   |===
   |---| |---| |---| |---| |---| |---| |---| |---| |---|

The cores were threaded onto a 3/8" diameter Tufnel rod with 1/4"
between cores and the end wires soldered to 2BA brass threaded inserts
in the ends of the rod. The whole was the slid inside a length of
polypropylene tube which was filled with potting compound to insulate
and mechanically fix the cores in place.

The total voltage across the choke is thus split between the cores and
so the electrical stresses on the wires insulation are reduced according
to the number of cores. 

Because the individual self capacitances are is series the overall
capacity is low, whereas the inductance is high. In the case of my
chokes the inductance total is 26mH with self capacitance of only a few
pF. The self resonance being several hundred kHz, in other words several
harmonics above the TCs operating frequency.

Gwyn Zucca (UK)