Re: Capacitor test results

Tesla List wrote:
> >From Esondrmn-at-aol-dot-comMon Jun 10 20:35:03 1996
> Date: Mon, 10 Jun 1996 12:01:59 -0400
> From: Esondrmn-at-aol-dot-com
> To: tesla-at-pupman-dot-com
> Subject: Re: Capacitor test results
> In a message dated 96-06-08 15:26:40 EDT, you write:
> >
> >As a rule of thumb, when using neons, I like to size the neon at
> >least 30% higher in current sourcing ability than the reactance of
> >the RF capacitor demands.  This ensures more firings of the gap per
> >cycle and better overall streamer activity.
> >
> >Unless you can come up with 120 milliamps worth of neon power for
> >this test, you may have to go back onto your pole pig.
> >
> >Good luck, rwstephens
> >
> rwstephens,
> Good point.  Thanks for the reply.  Yes, it is a .025mfd capacitor.  I
> remember when I first fired this coil I was using one of these same neon sign
> transformers and a home made rolled poly cap of .019 mfds.  The gaps did fire
> with several gaps in series and sounded much hotter than this one does now.
>  I was reluctant to fire it with the pole pig because it is somewhat
> difficult to control at low current.  At the lowest welder setting I think I
> get about 10 amps which is 2.4kva.  With no primary, I thought this might be
> quite a jolt for the capacitor - and of course I didn't really want to ruin
> it if it is any good.   I also was reluctant to pull the rotary gap out of
> the current setup, but maybe I will have to.
> Ed Sonderman


I've been monitoring the messages on your capacitor for a while now. A 
few thoughts I had: 

If you still have your 0.019 uF Rolled poly cap, try retuning and 
substituting it into your tank circuit (at lower voltage/power levels) to 
see if all of the other components seem to be operating properly. If so, 
it rather strongly points to the cap as the problem. 

If the Condenser Products cap appears to be the only component which does 
NOT work, a couple of measurements might be made to try to identify the 
failure mode. Since the capacitor seems to give only rather weak sparks 
at the gap, potential failure modes could include internally open 
sections, leaky, shorted, or high voltage breakdown of the dielectric, or 
a high equivalent series resistance (ESR) internal bus connection. 

First of all, fully discharge the cap. Then use a capacitance meter to 
measure the capacitance, or if you don't have a C Meter, use your tank 
circuit (with no secondary) and a signal generator to measure the 
anti-resonant frequency. Since you can compute the inductance of your 
primary, this should allow you to compute the low voltage capacitance 
versus expected value. 

If this test is successful, you next need to verify that the capacitive 
reactance is appropriate under high voltage conditions. With the pole-pig 
secondary open-circuited, apply a moderate amount of voltage (and record 
its value: Vapplied) to the low voltage side. Measure (and record) the 
no-load current drawn by the pig (Ip(Open Circuit) under this condition. 
Turn off power and hook your cap directly across the pig's high voltage 
output and record the primary current at the same Vapplied as the 
previous measurement (Ip(with cap). The difference between these two 
current levels is the current attributable to the secondary reactance 
"reflected" back into the primary. 

If the capacitor is not breaking down internally, the current should 
remain at a steady AC level. Assuming you are using a 7200 Volt Pig 
driven off a 120 Volt winding, the pig's turns ratio would be about 60:1. 
The expected secondary capacitive reactance would be 
1/(2*Pi*60*0.025x1e-6) or about 106,000 Ohms. The resulting input 
impedance "looking into" the pig's low voltage side will be 106,000 
divided by the turns ratio squared (3600). Thus, the resulting reactance 
"seen" by the primary should be 29.4 Ohms. 

The reflected reactance should be about Vapplied/(Ip(with cap) -Ip(open 
circuit)), and sould be about 29.4 Ohms. A markedly higher value, or 
erratically changing current levels may indicate excessive leakage 
current or internal breakdown under AC voltage stress. A markedly lower 
value may indicate that an open section or bus connection is resulting in 
substantially lower capacitance. 
If the above tests are OK, it is possible that an internal bus connection 
has come loose, and is introducing a relatively high resistance path in 
the tank circuit when the sparkgap fires. While this would not markedly 
impact capacitive measurements or reactive current measurements above, it 
would have a dramatic impact on the capability to deliver a crisp high 
current pulse when the gap fires. The "acid test" recommended by Richard 
Hull (connecting a rotary directly across the parallel combination of the 
pig and capacitor) if done briefly at low to moderate power should 
confirm (Hotly and LOUDLY) whether the cap can indeed deliver repetitive 
high-current pulses. It may be that only part of the capacitor is 
delivering high current while the remainder has a high ESR. 

It sure sounds like you have a flakey cap...

Best of luck!!]

-- Bert --