RATs

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

Original poster: "Finn Hammer by way of Terry Fritz <twftesla-at-uswest-dot-net>" <f-hammer-at-post5.tele.dk>

Some experiments are great: if they succeed, the by-product is a great
toy.
Other experiments suck: although it is good to know, how _not_ to do
things, the byproduct is a pile of junk.
The RAT coil that I talked at length about some time ago is like the
latter, and this is a report of why:

The adjustable electrodes of the SRSG should make it a breeze to dial
"equal bangsize" in, at different powerlevels, but reality proved the
opposite:
The only way to know where the electrode is positioned, once the coil is
running, is by watching the position of the ringdown on the scope.
However, on a 300 BPS system, the electrodes can easilly get too close,
so that no discharge occeurs, because the cap doesn`t have time to
recharge. So the electrode "disappears" on the scope, and after a few
minutes of running it is impossible to distinguish, which electrode is
producing what discharge where. Sounds confusing?: it is, and that is
also why I think that a remotely adjustable gap must have some sort of
indication of the actual electrode position at the controll box. This
could be done by using a servo motor drive at the SRSG, or by using
stepper motors, but if the operator cannot see where the electrode that
doesnot produse a discharge is at, then he doesn`t know which electrode
controll knob to adjust to get it to produce a discharge, and then the
whole thing gets very confusing. I will describe another solution to the
problem later.....

The rotary was also designed to have only one gap in the path. The
current is delivered to a center contact on the rotary disc, and
delivered onto the rotating electrodes. This center contact does not
work for extended periods of time. Fusing takes place, and the operation
gets noisy. It was silly of me to try to test this idea out in the
narrow confines of this particular coil. I have no space to add cooling
finns to a static electrode here, should I desice to add a small amount
of spark gap here.

Another thing that I found is that tin-bronze is a poor material choice
for primary`s: this is the first coil that displays severe heating of
the primary coil. The low q of the primary circuit that the high ESR of
the coil brings about is also the worst companion imagineable, of a
remote tuning arrangement: Although this mechanism works exeedingly
well, (at last something that does) the tuning is less than critical:
the same output is generated over a range of 4 to 6 primary turns.

Another aspect of the problems with the coil is attributable to the
ballast, not the MOT`s. At first I got some very strange charging
waweforms, which i thought were the result of the MOT`s. Not that they
were explainable by any known property of these transformers, but rather
because they have been given a bad rap at times. To investigate this
particular aspect, I wound a mini-pig on one of the already well prowen
cores from BLUM, the same core as the one I used for the MUSEUM coils
pćorkchop transformer, only this time I wound it to 8kV, so that it
would match the 4 MOT`s output voltage. Same charging waweform!
So the natural thing would be to try a slotted core on the CLI (Current
Limiting Inductor) and sure enough: the waweforms got into accordance
with what could be expected from Microsim simulations. We still don`t
have any unlinear core models for our Microsimulations, although I seem
to remember Tero from Finland mentioning something along those lines?
The current limiting shunts on a NEON effectively appear as a choke in
series with the _Secondary_ of the transformer: It is not affected by
the idle current in the primary winding, and this makes it natural for
this transformer to reach full voltage, even without load. With a MOT
supply, the idle current is so large, that the transformer doesn`t reach
full output voltage in the unloaded state.  However, when it is driving
a capacitive load, it does indeed reach full voltage, and in case of
around resonant capacitive loading, the voltage soars way above this
value. The unproportionately large idle current of the MOT`s is no doubt
part of the reason thet these transformers are seen as low value, but at
present I think they are allright, provided that the relatively low
output voltage can be tolerated.
The logical conclution of this experience is, that it is indeed right to
slot the current limiting variacs, and this comes as a big surprise to
me.

So how can I salvage this coil from ending in the scrap bin? I still
want the controll over the timing that is supplied by a rotary, and I
still want more than 100 BPS operation.
The recent discussion about triggered gaps, rekindled by Metlicka Marc,
got me thinking: I have tried it on a jacobs ladder which was powered by
mot`s, and therefore would not fire well, because of the aforementioned
current draw trough the external ballast, and it worked well. However
the problem of timing it with more than one pulse per 1/2 phase still
persists, and although there no doubt are pulse generators on the market
that could deliver the pulses into some fancy driver circuit, then this
path is not for me: I have too litle knowledge of electronix to do it.
However, another way could be running a very small sync motor with a
cam, breaking breakers, just as it is done in a distributor in an
ordinary (oldfasioned) car distributor would give me the things that I
want: A device that provides remote indication of where the discharge is
positioned on the timeline, relative to the incoming AC wawe, and a
device that can be repaired without posessing a EEdiploma.


Well, long rant, wee`l see.....

Cheers, Finn Hammer