My s.s. primary configuration

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Kennan C Herrick by way of Terry Fritz <twftesla-at-uswest-dot-net>" <kcha1-at-juno-dot-com>

Some of you will recall that my s.s. coil (Herrick's Patented Solid-State
Alternating-Current Thunder and Lightning Entertainment Machine to be
precise) utilizes an arrangement of power MOSFETs and electrolytic
capacitors in the (patented) "current ring" configuration that I have
previously commented upon.  Presently I incorporate 4 modules of 6 pairs
of MOSFETs and 4 modules of capacitors driving 3 equivalent turns of
primary conductor.
I post the following from an explanatory paper I'm preparing to go along
with some drawings I will offer to those who might be interested.  For
now, anyone so inclined might care to try figuring out what is the
optimum configuration of MOSFETs, capacitors and quantity of primary
turns in the general configuration that I use--from the following
information:
I designed the MOSFET/capacitor circuit-boards so that up to 6 could be
emplaced within a 24"-square configuration in a horizontal plane around
the coil of primary cables. Various arrangements of up to 6 board
assemblies and up-to-several primary-turn quantities can be configured
using the existing hardware.
It appears that, with the present configuration of 4 MOSFET modules, 4
capacitor modules and 3 primary turns, I could do with fewer, or smaller,
MOSFETs. I have made the following measurements:
1. Arbitrarily chosen spark duty cycle: 6.5 ms on & 75 ms period yielding
0.087 duty cycle.
2. Using a DVM, ac voltage measured across a 100mv/50A shunt inserted in
series with the 115 V line input = 25 mV. This yields a measured line
current of 12.5 A--essentially all going into the storage capacitors.
Caveat: the DVM, a Fluke Model 77, did not measure true RMS and so the
figure is only approximate since the current waveform necessarily
reflects the switching characteristics of the 4 unsynchronized) power
supplies.  Nevertheless...
3. Since storage-capacitor current out = current in, the output current
during the pulse bursts is 12.5 divided by the duty cycle or 144 A.
4. But that current is from all the capacitors. Since there are 4 groups
of them in series within the current loops, total loop (primary) current
= 144/4 = 36 A.
5. The 36 A is shared amongst 12 MOSFETs per MOSFET group; thus each
MOSFET carries, on average, 3 A, or 6 A peak during each half cycle of
excitation.
It appears that the same quantity of, say, "garden variety" IR IRF840
MOSFETs could well be used, at that current level, rather than the
IRFP460LCs of the present machine. Alternatively, fewer of the IRFP460
type could be used. Note that, at this writing, the IRFP460LC is no
longer recommended by IR; instead, they suggest the IRFP460A (of which I
just bought 25 at $3 each).
It would seem that, using this general scheme, it is best to maximize the
primary’s ampere-turns (the determinant of the magnetic flux that is to
excite the secondary) by arrangements that minimize the amperes per
resistive component: amperes mean IR drops that diminish efficiency. Thus
one would want to maximize the quantity of MOSFET and capacitor groups
and at the same time choose the quantity of turns such that MOSFET-
and/or input line-current limits are not exceeded. One can roughly
extrapolate from the 6 A peak figure, in the "4 + 4 + 3T" configuration
of the present machine (yielding a "benchmark" figure of 6 x 3 = 18
partial ampere-turns), to other configurations:
1. 5 + 5 + 3T: 5/4 of the current or 7.5 A. This yields 3 x 7.5 = 22.5
partial ampere-turns not considering the greater IR drops.
2. 6 + 6 + 3T: 6/4 of the current or 9 A. This yields 3 x 9 = 27 partial
ampere-turns not considering the greater IR drops.
3. 6 + 6 + 2T: 6/4 x (3/2)^2 of the current or 11.4 A (since inductor
impedance varies as the square of the turns). However, this yields only 2
x 11.4 = 22.8 partial ampere-turns--and at the expense of considerably
greater IR drops in the MOSFETs and capacitors.
4. 6 + 6 + 4T: 6/4 x (3/4)^2 of the current or 5 A. This yields 5 x 4 =
20 partial ampere-turns but the IR drops are cut almost in half as
compared to the 6 + 6 + 3T configuration. However, with only 9 A peak
through each transistor, the transistor drops should be moderate, anyway,
in a 6 + 6 + 3T configuration.
5. Finally, one can extrapolate still further: Reduce the quantity of
MOSFETs per module to 4 pairs, say, rather than 6. Then, in the benchmark
4 + 4 + 3T configuration, each MOSFET would conduct 6 x 6/4 = 9 A peak.
Then provide 8 such modules, and 8 capacitor modules, rather than the
six. 8 + 8 + 3T: 8/6 of 9 A or 12 A peak MOSFET current. This yields 12 x
3 = 36 partial ampere-turns: a significant improvement over 27 and
utilizing 8 x 4 = 32 pairs of MOSFETs rather than the 36 of the 6 + 6 +
3T configuration. But that is at the expense of 12 A MOSFET current
rather than 9 so the IR drops are proportionally greater.
It need be noted that the primary wiring scheme for some of the above
configurations, or others, might be difficult or awkward to realize. In
the present system, each MOSFET-to-capacitor loop-circuit travels 90°
around the primary circle in the counter-clockwise direction but 180° in
the clockwise direction (this is because 1 MOSFET group and 1 capacitor
group are incorporated on each of the circuit boards: physically at the
same angular position). Such asymmetry could increase with other, more
complex configurations.
So...that's a lot of detail.  Just thought I'd throw it out.
Ken Herrick
________________________________________________________________
GET INTERNET ACCESS FROM JUNO!
Juno offers FREE or PREMIUM Internet access for less!
Join Juno today!  For your FREE software, visit:
http://dl.www.juno-dot-com/get/tagj.