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

Capacitor Problems


 * Original msg to: Ed-at-alumni.caltech.edu

 Ed> I'm still stubbornly trying to find more ceramic capacitors,
 Ed> but have now decided that I am really committed to building  
 Ed> some caps of my own, when and as time appears.  I will at    
 Ed> least buy the materials and get the stuff in storage along   
 Ed> with all the other projects...

I think that if I were in your shoes I would go ahead and order
the correct capacitor from a firm such as Condenser Products. You
can pick up a NICE Tesla rated cap for around $200.00. In the
long run it will be time and cost effective.

 Ed> Main problem with Tesla coil work here is potential problems 
 Ed> with the neighbors if I get too ambitious.  Other hobbies... 
 Ed> are at least quieter and lest conspicuous.  I don't see how
 Ed> guys like Bill Wysock get away with it, but guess Bill's     
 Ed> tactic of inviting the neighbors in to watch (thus keeping   
 Ed> them away from their TV's) is a brilliant one.  

Humm, I have found that by proper RF grounding of the coil system
followed by Faraday shielding the spark gaps, that considerable
power can be pumped through a coil system without excessive RFI
and interference problems. I can fire at 2-3 kVA with televisions
running on the same circuit with only the slightest bit of snow.
The neighbors are not affected at all even with power levels up
to 10 kVA.

 Ed> One final thought or question.  I see lots of mention in the 
 Ed> capacitor notes about the voltage across the (60 cycle)      
 Ed> transformer rising to peak values way above the name plate   
 Ed> rating. This can clearly be true with a capacitor load, but  
 Ed> my question is this:  If you set the main spark gap to fire  
 Ed> at the secondary output of the transformer without a cap-    
 Ed> acitor,  would you expect any trouble from over-voltaging    
 Ed> the transformer????? I can't see how it could happen, but    
 Ed> interested in your thoughts...

There are several things going on that when factored together 
mean a hard life for a capacitor in the spark excited Tesla tank
circuit: RF dissipation factors in the dielectric, resonate rise
in the tank circuit, peak values as opposed to RMS values in the
line voltage, and shock excitation of the circuit.

The first factor, RF dissipation in the capacitor dielectric, is
perhaps one of the most critical aspects to consider in a cap-
acitor for these circuits. Even the G-Series mica transmitting
caps get hot in spark gap Tesla applications. True, they are
rated for the temperature, but they are converting a lot of RF to
heat through dielectric losses, and big mica transmitters are a
rare as hen's teeth these days. Barium titanate capacitors (HV
ceramic "doorknobs") are not RF rated. These ceramics are
typically DC rated and get hot very fast in Tesla tank circuits
due to large RF dissipation factors. This means that glass,
ceramic, and mica are not the best choice for capacitor
dielectrics. They can be used, but the duty cycle must be very
small (less than 1%) to allow internal heat to migrate and
radiate. Even watching the duty cycle carefully to reduce the
incidents of heat related failures does not alter the fact that
the dielectric efficiency is poor. When these dielectrics are
used the spark from the top of the coil tend to be violet, thin
and spindly.

Resonate rise in the tank circuit is a factor that many people
have a hard time getting a good handle on. I like the often
abused "water" or "swing" model of an oscillator to point this
one out. The water model is a half full aquarium or bathtub.
Using your hand to move the water it is possible to create a
resonate wave in the tank or tub: just like timed pushes on a
swing causes the swing to climb higher and higher, timed pushes
on the water creates a wave that will reach over the top of the
tank, or out of the tub, even though the container is only half
full. The ends of the tub (or tank) represent the capacitor
dielectric. Peak voltages from resonate rise in the oscillator
appear only on the capacitor dielectric, nowhere else in the
circuit. Just like the resonate wave in the tub only reaches it
peak at the ends, the voltage rise in the oscillator only reaches
it peak across the capacitor dielectric. The only way to protect
the cap from resonate rise in the tank circuit is to place a
safety gap directly across the capacitor terminals. Safety gaps
placed behind RF chokes, bypass caps, etc., will not help a cap
survive resonate rise over-voltages.

Voltage peaks on the 60 cycle AC sine wave creates another source
of dielectric stress. You are correct in stating that a properly
set spark gap should protect the capacitor from the line voltage
peaks. In many cases it does. The problem arises when you have
synergistic factors that come into play at the same time: the
capacitor dielectric gets too hot and a 60 cycle voltage peak 
reenforces the resonate rise in the tank circuit.

Shock excitation of the circuit, that is pulse discharging the
capacitor through the breakdown of the main system spark gap, is
another source of stress on the capacitor design. The amperage
flow across the gap can easily run into the hundreds of amps.
Lead wires and connectors that are too small for the job of pulse
discharging will cause resistance heating that in turn creates
hot spots on the dielectric. Heat is not allowed to flow out of
the capacitor through the conductor as the conductor becomes an
additional heat source. This is common on high-voltage rated DC
filter caps that are pressed into service pulse discharging in
Tesla tank circuits (where they don't belong). 

The solutions to these problems point to specialized capacitor
materials and designs. Plastic dielectrics such as polyethylene,
polypropylene, or polystyrene have extraordinarily low RF
dissipation factors. These dielectrics have low melting points,
but if air and moisture are removed from the construction, and
the capacitor is covered in high quality oil, they don't get hot. 
In fact they rarely get warm. This also means more energy is
processed by the capacitor with smaller losses.

Resonate rise and AC sine wave voltage peaks are simply dealt
with by making the construction as heavy duty as possible. Pulse
discharging is facilitated by using wide, flat, and smooth
conductors suitable for high current RF. This is a perfect
application for litz wire.


Richard Quick


... If all else fails... Throw another megavolt across it!


___ Blue Wave/QWK v2.12