Re: OLTC II Initial calculations

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Stephen Conner by way of Terry Fritz <teslalist-at-qwest-dot-net>" <steve-at-scopeboy-dot-com>

OK, so I have John Freau telling me that 60" sparks from a 24" tall 
secondary is fine, and Malcolm Watts telling me that a safe maximum voltage 
gradient is 8ft per megavolt. I wonder what the voltage gradient associated 
with 60" output from a 24" secondary is?

60" sparks at 120bps = 1250W

1250W -at- 120bps = 10.4J bang energy

Capacitance of 6" x 26" toroid and secondary ~ 25pF

Bang energy=0.5*Ctop*Vtop^2

Vtop=sqrt(10.4/(0.5*25p))=900kV. Probably nearer 750-800 due to losses.

Voltage gradient=800kV/24"= 30" per megavolt.

I believe the Tesla coil secondary can stand a higher voltage gradient than 
a plain insulator of the same length, because the winding forces the 
electric field to be uniform. It's always "X" volts per turn.

Anyway, I redid my design, I now have a 10" x 30" 2000 turn secondary that 
resonates at 58kHz, a 17uf tank cap, and a theoretical output of 55". The 
peak primary current is a whopping 3150A per brick. This is well within the 
overdrive capabilities that our intrepid experimenters discovered for 
discrete IGBTs. However we don't yet know if bricks behave the same.

The main worry is that the inductance in the emitter wiring inside the 
brick will put bad ringing onto the gate signal. Sure there is a "Kelvin 
connection" for the emitter but it's not going to be perfect. If there are 
several dies (Dice?), it can only be a true Kelvin connection for one of them.

The manufacturer probably allowed for this but overdriving the gate voltage 
will eat into the safety margin at the same time as the high emitter 
current is making the ringing worse :( This could lead to uneven current 
sharing as some of the dies turn partly off on current peaks, or Game Over 
due to destruction of the gate insulation :6 The easiest way to find out is 
probably to build the primary circuit and see- which is what I'm going to do.

BTW I solved the charge circuit problem, using a voltage doubler and a 
variac off 240V I can get 0-650V on my DC bus, and with a suitable meter 
circuit (basically a small measuring capacitor that charges via a diode 
from the main cap) I can read out the peak charging voltage. All nice and 
simple and it still qualifies as Off-Line :D

Steve C.