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

Re: Solid-state tesla circuit

On Thu, 26 Sep 1996, Tesla List wrote:
> I'm using a pot to apply a varying voltage to the 3825, controling the
> pulse width - what I find is that instead of getting a nice increase in
> output power I get two distinct peaks one at about 1/2 way and the other
> at about 80%. Needs a bit of twiddling. This happens regardless of which
> output transformer I'm using.

That must be some sort of resonance or something. I definately think
you should wind a current transformer for investigating the problem
(with a scope, you'd definately need one). Current transformers
are also pretty nice becouse of very low losses.

> In the Mark 2, I was going to use a resistor on the high side for current
> sensing, with a 1:1 transformer and a bridge rectifier (as in Rob's circuit)
> to monitor the current and use current feedback. Or I may use an opto-coupler.

OK. On low side you gotta be very careful with resistors. They raise
the source level above the ground! They also must be of very low
inductance (no usual 4W wire-wound ones). Metal foil resistors work
well.

> Or I may dispense with the 3825 and use a handful of logic to give a square wave
> and an over current limitter.

..making an oscillator was one thing I had thought of too. However,
3825 is way more compact than oscillator, flip-flop, low-voltage
shutdown, drivers, ...

> My main problem with the transformers was internal arcing and saturation.
> Bathroom sealent (clear silicone) between windings (as well as tape) has solved
> the arcing, careful calculation and more primary turns has solved the

Yeah, I know. I had the same problem at first too :) I just wondered
how my driver did not blow up with an arcing load or at the time
I made some other funny things. (No current limiting!)

> The next output transformer will use 3 of the present E cores side by side. And
> it will
> be right next to the driver board, reducing wiring length.

I wonder if many cores are needed. At coiling frequencys (say 250kHz)
a supprizingly small core should do for 1-2kW. However, make sure
all the 3 coils share the current equally if you make a 3-core system!
Have either primaries or secundaries in series. That is a fool-proof
method for having the same current in each one.

> I did consider getting a bunch of ferrite radio rods - say 7 in a hexagon, or
..
> For very cheap - hard to saturate - lots of inductance per turn - should be
> OK at the higher frequency (made for the medium wave) - Coils on the outside of 

Those rods can hve pretty low permeability too! They can also be made
of materials ok for low flux (like a MW receiver receiver) but way
too lossy for high power at higher flux.

> the core for easy cooling - infinite winding window - easy winding

Cooling can be solved. That's why people use resins with good thermal
conducting properities in transformers. Silicone is not quite ideal.
Using a bunch of wires (or litz) can also reduse losses. At high
frequency one should also not keep the flux too high to keep the
core losses reasonable.

Infinite winding window is nice. However, that means it must be a case
where you can take the huge leakage inductance resulting from large
windings. Or you have to interleave primary and secundary. Leakage
inductance can also be made use of in resonant type SMPSU's. However,
that is not too easy in tesla drivers because the load resonant point
and driver resonant had to be same. That's why I'm more into zero
voltage switched resonant converters in my design. ZCS in not too easy
to impelement in coil driver IMO.

BTW, if you use say an E-core and buy the bobbin too, it is easy
to wind as well.. ;)

> Against - power losses (must be better than an air core say 70% efficent) -
> rf transmission - screened box.

Air-core seems too lossy or big. It seems a very difficult task to
design properly. Very high turns count and big size increase easily
losses. 

> I gather from your comments Hari that I can anticipate more problems 
> at higher powers :(

I'm not at high powers yet. However, I have driven a coil with about
20A peak sine wave current with a single pair of IRF740's. That did
result in sever heat problems and fets blowing up. Resonances in
transformers was also a severe problem with my poorly designed
step-up transformers.


--
"99% of the pain in the world is due to misunderstanding" -M. Jarvelainen

Harri.Suomalainen-at-hut.fi - PGP key available by fingering haba-at-alpha.hut.fi