Re: SSTC Q's

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

Original poster: "Jan Wagner by way of Terry Fritz <twftesla-at-qwest-dot-net>" <jwagner-at-cc.hut.fi>


> 	I'm considering building a SSTC.  I know just about nothing on the
> subject, so I need some help.  

Well well... welcome to an highly expensive hobby!! ;o) 
And keep that large bag of spare mosfets well in reach...
 	

> 	First, what kind of voltage/currents are typically used in the primary
> circuit?  I just bought some MOSFETs, rated 500VPk, 12 amps for $6.50 a
> piece; will they work?  Did I get ripped off?? (I doubt it :-))  I know
> I'll need to parallel some even if they are suitable.  	

Rectified mains voltage - 'cause it is easily available! ;)

110vac => 110*1,414 => about 150VDC
230vac => 230*1,414 => about 320VDC

RMS currents depend on the power you want I=P/U - which typically is 
"much power", so maybe 5A to start with.

If you use an interruptor you can have large peak currents in the driver
but still normal mains current, but the mosfets have to be rated for those
peak currents. <=> high peak power but low average power
 I_average_from_mains = I_peak * T_on/(T_on+T_off)
(Hopefully that formula went right...) T_on and T_off are the interruptor
times. Interruptor == 555 timer circuit with adjustable freq and duty,
freq <1 kHz


The IRFP460's (18A 500V) should go for about $4 per piece, so I don't know
about that ripoff thing you mentioned...  ;-)


With a 500V mosfet and 110Vac mains you could do a simple forward
converter style SSTC,
  http://www.angelfire-dot-com/electronic/cwillis/sstc.html
but you absolutely must have a proper snubber network for the mosfet,
otherwise it fries - even at low input powers and voltages.

The better route to go, if you want power and no heavy voltage spikes
on the fets, is a half-bridge,

  http://www.angelfire-dot-com/electronic/cwillis/fetcoil.html
  http://www.angelfire-dot-com/electronic/cwillis/fetcoil2.html

a bit more complicated but much easier to handle than push-pull or
flyback/forward converters. The mosfets need only ne rated for rectified
mains plus a 100V-or-so safety margin, unlike the >>800V mosfets for power
forward converters and push-pull.


The simple "forward converter" schematic looks a bit like
  http://www.personal.u-net-dot-com/~hills2/electron/smps.htm#flycnv
but without the diode, and the transformer is just the primary coil with
secondary standing inside as usual. 
That's probably already the style how you built your air-core HV xfmr?


Btw, mind you, that the pic is just a schematic! For actually building it,
use as short as possible wires/PCB traces, and as thick traces as possible
or Litz enameled wire (braided wire??).


> 	What sort of drive circuits are used?  I could probably use a 555 timer
> thingy to fire the drive transistors, but there are better ways, right? 

As the signal generator, yes, 555 is nice. For example a 555 wired up as
50% duty, or a CD4046 which is 50% duty "by default".

There are some schematics at:
http://www.angelfire-dot-com/tx4/ionwind/sstc2.html
http://www.richieburnett.co.uk/sstate.html
http://www.angelfire-dot-com/electronic/cwillis/fetcoil2.html

Personally I don't see any need to go with pulse width modulation
(PWM) ICs like UC3846/LM3524/TL494 and the like, because for SSTCs a
fixed 50% duty cycle is just simply "best". ;o)

A cheap 555 or CD4046 or other 50% duty square wave generator will do very
well.

My sig generator versions (555 and CD4046) at:
  http://www.hut.fi/~jwagner/tesla/SSTC/half-bridge.htm
There are plenty of other schematics on the net as well.


The output of the signal generator must be beefed up a bit, amps wise, so
a good mosfet driver IC or a some other setup will be required. Especially
when you have to charge the gates of many mosfets, paralleled etc.


> Do you guys use square or sinusoidal wave drives?

The mosfets are working as switches here, so use a square wave drive
signal. With negative bias, this means a square wave/gate signal -15V in
off state and +15V in on state.

 
> 	What is the tank circuit arrangement?  Is a cap used, or is the primary
> inductor driven straight from the MOSFETs, without a cap.

No tank circuit.
You just drive the primary coil directly from your circuit output. 

There are also other ways to drive the secondary, but this is really the
simplest one.


> 	I've been playing with my new FETs today, and they sure are easier to
> deal with than bipolars! :-)).  I understand that MOSFETs can be
> paralleled with ease (no thermal balancing resistors).  Is this true?

Yup, easy to parallel. Not easy as in "a piece of cake", but easy anyway.

Current doesn't share 100% equally between paralleled mosfets. But that's
not really a big problem. The major problem is _how_ to parallel them -
how to connect the gate drive. For two paralleled mosfets very close
together a single gate drive (shared ground) may work. For more mosfets,
you are likely to end up with separate gate drives (each has a _separate_
ground i.e. directly connected to each mosfet source pin, and the signal
to the respective gate). That's the only way to keep the SSTC driver from
screwing up at large currents and when slightly out of tune and switching 
mosfets fast.

Ok I hope that all this didn't sound too cryptic, and maybe even made a
little bit of  sense sometimes... ;o)

You can read more on solid-state sites. I've listed up some at 
 http://www.hut.fi/~jwagner/tesla/links.htm
and there's also the solid state webring at
 http://q.webring-dot-com/hub?ring=solidstateteslac&id=2&hub


Good luck!

cheers,

 - Jan

--
*************************************************
 high voltage at http://www.hut.fi/~jwagner/tesla