[Prev][Next][Index][Thread]

8-9 RFI noise thoughts.

Hi All,

	I think I have identified the source of the waveforms I see on the scope
from the primary circuit firing.  The waveform is composed of basically
three separate mechanisms:

1.	The first wave form is simply our old friend the few hundred ampere
primary decrement wave that powers our coils.  It is simply the ringdown of
the primary capacitor and inductor.  No surprise there...

2.	There is a fairly powerful 5-30 MHz burst right after the gap fires that
last around 100uS.  I think this is caused by the self-capacitance of the
primary and the inductance of the primary wiring.  My coil seems to have
about 39pF of self-capacitance on the primary and around 890nH of wiring
inductance.  When this system is hit by the gap firing, it rings for a
short time.  This is especially important in that it hits the lower
television frequencies.  However, since this oscillation travels through
the whole primary loop, there should be ways of filtering it.  Perhaps a
small inductor or two in line with the primary would lower this frequency
to where it would damp out before oscillating too much or simply be so low
as not to be a problem.  I think the zero crossing bursts are basically
this mechanism too.
There may also be about a 5MHz signal too that may be formed by the primary
inductor and it's self capacitance.  This signal gets a little lost in all
the other fun but it should be there, but of fairly low power...


3.	The third noise source is the gap.  When the natural capacitance between
the electrodes is suddenly shorted as conduction begins, wild things
happen.  Although this is not a true circuit theory phenomenon, I can
reproduce the noise in computer models with a cap and inductor across the
gap.  Judging from my measurements and making some guesses,  the gap
capacitance seems to be on the order of 20pF.  The inductance is around
80nH with a series resistance of a fraction of an ohm.  If the gap is
charged to 21kV the burst power is enormous.  3GHz at around 2000 amps for
a 10nS burst!!  Note that 2000 x 21000 = 42 megawatts!!  The RMS power may
be in the tens of watts.  I don't think this signal travels too far down
the interconnect wiring but is rather an intense point source of RF.  It
can easily spray into the surrounding area and affect nearby electronics
(like fiber optic probes).  However, I would think such high frequency
would not travel too far into the surrounding neighborhood (I really don't
know that??).
	I will try ferrites at the gap tips when they come.  The frequency would
suggest that the electrodes may carry some of this current and the ferrites
may present a good load to this signal...  Otherwise, one may have to put
the gap in on old microwave or something to "try" and stop this burst from
getting out.  I think trying to contain this burst will be tricky at a
wavelength of a few inches.  The normal shielding may just act like and
antenna when hit with this RF.  May need some microwave expert advice on
this one...
	I assume the real mechanism is some fancy plasma thing...  Probably
nuclear bomb science... ;-)  Hopefully, it could pre prevented all
together.  Shielded, at least...


Soooo.  The ~27MHz stuff should be filterable without too much trouble.
Hopefully this would be the majority of RF that could irritate the
neighbors.  The 3GHz pulse may be really nasty to stop...  I have no doubt
of its power. 

BTW - I messed up on yesterday's post.  It is not 3000000 amps, more like
~5000...  Was "thinking" too fast there :-)  Of course, the field probe
would react differently to such a source, so the real level can only said
to be "high"...

Cheers,

	Terry