Re: Holy Crap! WOW, that's the origin of quarter-shrinking!

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>

Hi,

> One real problem is the SISG. If fired perfectly "once" but I never 
> stopped to think what 800+ joules might do to it. All the IGBTs are 
> shorted. Seems 100X its pulse rating is "too high". Easy to fix, but 
> that leaves us with no reliable switch for the fireball thing... 
> Don't know how to solve that yet.

With more thought....

It fired twice:

1. 1800V 60e-6F into probably ~15ohms (97.2J). The timing on the SISG 
was set for 400uS and it turned off with about 800 volts still on the 
cap. The SISG survived this.

2. 5400V 60e-6F into probably ~15 ohms (875J). All the IGBTs shorted 
to zero ohms.

After some calculation, the instant thermal heating on the die in the 
first case was well below its rating. Breaking open the circuit at 
400uS was also fairly easy.

However, in the second case the die critically over heat in about 
50uS. The higher current (into about 8.22 ohms) dissipated 10X the 
power and the IGBTs never had a chance. Then they get hit with 
another 28 joules at 400uS to finish them off for sure.

So easily explained and no problem at all for Tesla coil stuff. But 
the 8.22 ohms is probably good for a conventional spark gap now in 
the case of the plasma ball experiment. The 800V at 400uS gives great 
knowledge there A very complex dynamic high current ohmmeter...  It 
also gave a nice test for the SISG.  If you charge a cap with two 
MOTS for 4 minutes, that is probably too much power for it 
:O))))  But the SISG can take one heck of a punch!!!  There are 
larger IGBTs that could do it but the next option is just easier now...

Steve Ward simply suggested a big ball electrode HV relay like early 
quarter crushers used.  That would be perfect for the ball lightning 
experiment thing.

Mcmaster sells unglazed standoffs but they are only an inch long.  I 
would think two inches long with a 1/4 inch hole and 1/2 inch 
diameter would be pretty close.

Cheers,

        Terry