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

Re: More RFI filter testing...

Hi Terry,
          I think you are up against the quantum nature of matter 
here:

> Original Poster: Terry Fritz <twftesla-at-uswest-dot-net>
> 
> Hi All,
> 
>     I was trying different RFI filters tonight...  Nothing worked...  didn't
> even effect it at all...
> 
>     I tested the scope and probe equipment to be sure something screwy
was not
> going on there.  Aside from pure RFI pickup on the case of the fiber-optic
> transducer, the signals appear to be very real.  I could switch transducers
> off and disconnect optic cables and such and the signal will stop just like
> it should.  the signal was definitely being feed to the scope through the
> measurement equipment.
> 
>     Now get this...  Frustrated, I stuck a 1K ohm resistor in series with the
> primary circuit.  The pulse did NOT change at all!!!  Even stranger, I can
> get the pulse by just turning the variac up to before the point it will
> arc.  There will be an occasional high power pulse with no normal arc!
> 
>     As far as I can tell, the real meat of the pulse lasts about 20nS and is
> composed of very high frequency and very high power signals.
> 
> My latest theory of the minute is....
> 
>     The pulse seems completely unrelated to the primary coil or other primary
> parts.  It appears that the pulse is caused by the actual arc at the spark
> gap.  The heavy wiring in that area, just serves as an antenna to transmit
> these 1GHz+ signals.  I suspect the gap stores energy as capacitance across
> the gap.  When conduction starts, the arc becomes a super high power high
> frequency transmitter for about 20nS.  Apparently, this initial arc can
> occur by itself without starting the primary circuit into conduction.  Ie.
> it can occur so fast the primary circuit will be unaffected by the fast
> local arc of the gap.  I must assume this is common to any spark gap system
> and not just Tesla coils.  This is good in that it may have more data about
> it somewhere.  Unfortunately, the power, speed, frequency, and connected
> metal parts will make this thing bazaarly difficult to stop or even shield
> against!

Exactly so. You (and every other coiler) have a parasitic relaxation 
oscillator in the primary. How to get rid of it: provide an unlimited 
current once the gap discharge begins :)

>     So to make a long story short.  It looks like the initial gap
> capacitance's stored energy is going into the initial arc at the gap and
> feeding a tremendous amount of power into the arc for about 20nS.  That
> power is being converted to very powerful, high-frequency RF.  Since the
> arc size is about 1/4 inch, I assume the frequency extends well into the
> low number of GHz region...
> 
> Cheers,
> 
>     Terry

BTW, I think you associated an inductance figure of 20uH with the gap 
in an earlier piece (correct me if I'm wrong). That figure is highly 
suspect. I have a coil with about 12uH of real inductor in its 
primary and it runs as if no (or minimal) other inductance is present.

    Another BTW - I use a Sencore LC35 LC analyser for my 
measurements and it comfortably goes down to 1uH (below if one 
ignores the inherent inaccuracies). The only thing I have so got 
bad answers on this machine on was measuring my MMC strings. 
**Aside: I have strings of seven caps. Each cap is bridged by a 10M 
Ohm resistor. The instrument appears to measure correctly up to five 
in series but then goes wildly inaccurate. I finally resorted to 
putting the strings in a tuned circuit to determine their real value. 
I don't think the leakage current due to resistors played a part but 
will try again without them. Otherwise, the machine has worked 
perfectly. 
**End Aside.

The Sencore even detects a single shorted turn in transformer 
windings with ease with its "Q" measurement.

Malcolm