30MHz Spark Gap Testing - Is this real??

From:  Jim Fosse [SMTP:jim.fosse-at-bjt-dot-net]
Sent:  Tuesday, April 07, 1998 2:07 AM
To:  Tesla List
Subject:  Re: 30MHz Spark Gap Testing - Is this real??

Hi Terry,

>	When equipment capable of much higher bandwidth is employed this picture
>seems to change dramatically.  When one terminates a quality antenna into
>the proper 50 ohm impedance the fundamental signal seems overwhelmed by
>heavy noise spikes.  This explains why a simple wire is often used as a
>scope probe to receive primary waveforms as opposed to a higher quality
>antenna system.  The simple wire and its very poor impedance matching to the
>input of an oscilloscope attenuate these noise signals so that only a nice
>clean signal is left.   
>	If one uses a high bandwidth properly terminated antenna, a series of noise
>spikes are seen (my testing was done without the secondary inductor in
>place).  Careful examination will show these spikes appear as a series of
>noise bursts that occur at the peaks of the fundamental frequency.  The
>power of these noise bursts is vastly higher than the fundamental waveform.
>In my testing, I have found that these burst consist of ~50MHz signal bursts
>that persist for about 100nS and then fall off to a much lower level.  The
>power of these bursts is remarkable. 

First, you are hitting any LC circuit (distributed elements or not)
with what amounts to a Dirac delta function. If it can ring, it will.
I've seen "RFI" (using a 3' wire antennae) into the 100s of MHz when
just running my rotary spark gap from my PIG with 6' of connecting
wire. I did not have the primary or secondary of my TC connected in
this test. I was just testing out my rotary gap.

Secondly, the power levels are high enough that unintentional pickup
becomes a real problem. 

Put some opaque tape over your fiber optic cable and then run your
system. Do you still see the signal? If so, it's from a parasitic
antennae. Don't forget that black plastic IS conductive. I've measured
1 megohm per foot on black 1/4" vacuum hose.

I've included 3 private posts to Malcolm Watts regarding my
experiments with a photodiode looking at my static spark gap. I had
some very read problems with unintentional pickup.

------------- include 1 --------------------------
To: MALCOLM-at-directorate.wnp.ac.nz
Subject: Spark Gap Meterology
From: jim.fosse-at-bjt-dot-net (Jim Fosse)
Date: Sun, 11 May 1997 03:22:11 GMT

On Sun, 27 Apr 1997 20:30:28 GMT, jim.fosse-at-bjt-dot-net (Jim Fosse), you

>I'm waiting for some fiber optic cable to arrive (monday or tuesday)
>to start scoping my gap. Meanwhile I've been building a repulsion
>coil. I'm now levitating a 1" Cu ring about 1' up the core. One can
>really see the effects of K on the primary current by forcing the Cu
>ring lower.  {4/6/98, I eventually started putting divits into my 
>garage ceeling with this;)  }
Hi Malcolm,
	I finally connected the fiber optic cable to those pin
photodiode and led that I talked about last month.

I sanded the top of the led flat and sanded off the lens of the
photodiode. (400 grit then 600, then 1000 grit sandpapers) I used
thick supper glue to hold the plastic fiber optic cable to the diodes.
While the glue was still liquid I hand adjusted the position of the
fiber for max illumination of the die. A squirt of hardner and it was

In spite of the LED's greater output when it was illuminated with
sunlight (through the fiber pointed at the sun), at lower light
levels, the photodiode was much more sensitive. I was using both the
LED and the photodiode in photovoltaic mode.

I am picking up to much RFI from the spark gap using the diodes in the
photovoltaic mode. Tomorrow I'll build a current to voltage amp and
run the diodes in current mode.

	More tomorrow,

------------------------include 2 -----------------------------
To: "Malcolm Watts" <MALCOLM-at-directorate.wnp.ac.nz>
Subject: Re: Spark Gap Meterology
From: jim.fosse-at-bjt-dot-net (Jim Fosse)
Date: Mon, 12 May 1997 00:10:07 GMT

On Mon, 12 May 1997 08:26:43 +1200, "Malcolm Watts"
<MALCOLM-at-directorate.wnp.ac.nz>, you wrote:

>Hi Jim,
>        Please keep me posted on this one. I can see some interesting 
>avenues of measurement coming out of it.
	Here is the circuit I'm using. I didn't draw the 10uF tant.
across the 9V battery. I'm using a LMC660 quad opamp. It's GBW is only
0.5mHz. I didn't have a faster amp at home.  The two unused amps are
connected as unity gain followers with their + inputs connected to
virtual gnd. (The output should be 1 volt per 10uA. but until I get
the RFI under control, I'll not increase the gain)

	+ 9v batt
     ___|                                     _	
     |	|				      ^	pin photo diode	
     |	/ 470K				      |	 100K
     -	\				      |--/\/\--|
 0.1 -	/     |--------|		      |	|\     |
     |	\     |	|\     |		      |-|- \___|____o
     |	|     |-|- \___|__virtual gnd-----------|+ /
     |--|-------|+ /		       |        |/	    output
     |	|	|/		       |
     |	/			       ---------------------o
     |	\
     -	/ 470K
 0.1 -	\
     |	|
	- batt ret

note: this circuit limits when I place the fibre optic cable in

I'm getting 1.5V peak RFI into my scope when I fire the 15kV 30ma neon
into a spark gap 7' away! (without the light pipe connected) It looks
like 6 or 7 sawtooth cycles on each half mains cycle. I've connected
my neon directly to 5 gaps of my RQ spark gap with ~1' of neon sign
wire. With just this circuit you can see the leakage inductance
secondary-self-C charge cycle.  If I just disconnect my scope probe -
no RFI, so it's not coming in the mains.

Connected the scope ground to the scope probe next to the circuit
looking for antenna loops - no RFI.

I put the unit inside of a grounded tin can. The RFI was back. Put the
scope in differential mode - I still have RFI. I started to build a
shielded box for it, stopped for lack of parts: no box top, no full
braid cable, dozens of BNCs with no center pin. ARG!

I'm stumped, It's been 25 years since I've done any RF work and it
looks like I forgotten it all!




---------------------- include 3 ----------------------
To: "Malcolm Watts" <MALCOLM-at-directorate.wnp.ac.nz>
Subject: Re: Spark Gap Meterology
From: jim.fosse-at-bjt-dot-net (Jim Fosse)
Date: Tue, 13 May 1997 05:34:05 GMT

>           The circuit came out right this time :)  I'm glad the 
>interference has gone down a lot. The box sounds like a good plan. 
>Please keep me posted.
The box really helped!  I was able to reduce the Oscope scale factor
to 5mV before I started seeing RFI again. (remember, I'm at 1V/uA
scale factor here; 3 divisions -> ~15nA current injection) As I
reached to the scope to change a switch, I touched the bottom of the
box and the RFI jumped! If I touched the grounded "case" of the box -
no change.

The box is painted. The bottom was not making contact with case. As
soon as I scraped off the paint and put internal star lock washers on
the screws, the RFI disappeared.

	Shielding 101 revisited;) Like I said yesterday, it's been 25
years since I've done any RF work. Had I been building a receiver, I
would have expected RF pickup. To much digital makes the neurons go
soft - fuzzy even;)  But, I'd never expect the body of a 1/4 watt
resistor to pick up 15nA 2" from a metal shield that itself was 6'
from the spark gap!

I only had time for 2 quick peeks at the gap. With just the neon
running the gap I was seeing 0.5 - 1.0 volts per spark.  I added my
10nF cap across the gap. The voltage went to 4.2 volts per spark and I
may be saturating the op amp. More measurements tomorrow. I didn't
have time (7 year old attention span at the time) to see more than
just a number of pulses per half cycle. I expect to see some RF in the
output of the photodiode when the gap is connected to my TC primary.


---------------- end includes ------------------------

I've become sidetracked since then and have not done any measurements
with my primary or secondary connected.