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

Re: Mosfets survive

On Mon, 25 Mar 1996 tesla-at-grendel.objinc-dot-com wrote:

> caps into service for the moment. The electronics are now being driven
> by a CA1524 PWM chip supplying 170kHz to 240kHz on a 5% to 95%

Beware. That chip will not give you nice and fast risi/fall times. Add
a fet driver for more current. Low current can mean fet parasitic 
capasitance can induce high gate voltage at turn-off and fet will then
turn on unintentionally. I noticed some problems with that phenomena with
the same driver (3524)

> I need to used a grounded electrode though to pull a 1.5" spark out of the
> toroid. Next step is to up the voltage to 150v, and add 2 or 4 more FETs.

Should be no ned to. I remind you I did drive around 20A peak with IRF740's!
They were getting pretty hot but it was just manigable.

> was simply to hand when I grew frustrated with transformers and bottom
> feeding. Pulsed DC means that transformers are not an option with this
> circuit - they would just saturate.

Not if properly designed. You can even use air-cored transformers if
you will make it just resonate with a proper cap at the primary. I'll
investigate that one some day more.

> Am I correct in saying that my coils should be closely coupled. And that

If you drive it with rf at the resonance you should go for maxinum
coupling.

> I should construct a primary which has a reduced width, more hieght

..and severe primary-secundary flash-over problems most likely..

> (I dont anticipate sparks bigger than 6" to 10"), but increase the number
> of turns to keep the induction about the same. (too low an induction
> would give fast rising currents overwhelming the current proctection).

If the inductance is low it will load the fets. The power will then just
be circulated at the primary and never go to the tesla coil. Resonate it
with a cap or make it large enough in inductance. Large here means that
it should be many times larger reactance than the coil looks like.

> Snubbers still seem too hot, I'm using 2   5W, 3.3K resistors in // plus a
> 1000pf 2kV cap (in series) . time constant about 1uS, I figured that 

Lower the resistor for one thing. Something like 470 ohms in *series* with
say 1nF cap is much better.

> I've just recieved Duane Bylands book, "modern tesla coil theory".
> good material, though the print quality is poor. He includes the
> suggestion of building 6 identical fet driven coils in a hexagonal
> arrangement. Driving each one 60 degrees out of phase with
> respect to its neigbours, giving a field which rotates at the coil
> frequency, has anyone tried this. It should create a very attractive
> display.

That is a good idea! I guess I'll have to construct a 6-phase driver
when I have the time.. :)
--
See "Malin Tunic: Chemistry of structure - function references in cheese" for
further details!
Harri.Suomalainen-at-hut.fi - PGP key available by fingering haba-at-alpha.hut.fi