Re: SSTC-Starting Out

["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 embarking upon a journey that will lead to the demise of many an
> innocent semiconductor switch ;-)).  It's taken me 3 days to build one
> of the three portions of the drive circuitry.  I wouldn't be surprised
> if you guys could do it in 5 minutes, given my idiocy.

naah... Myself it took a couple of _weeks_ to get one running. :-) Mainly
because of grossly overcomplicating things. (as usual ;o)

> 	Next, I need to make the pulse width controller.  Do any of you have
> schematics or ideas regarding single shot pulsers?  I need to generate a
> pulse that is SHORTER than the incoming trigger pulse, that is variable
> from 0.5 uS to 20 uS in width.  It can be either rising or falling edge
> triggered.

A pulse width controller/modulator isn't a necessity, IMHO. Much too
complicated.

Actually, you may be better off by simply doing driver frequency
modulation instead of power modulation. The end result is exactly the
same, but with less inductive spikes and interference etc...
Frequency modulation would be simple - with an existing 555
50%-duty output circuit, it's as simple as connecting the R and C timing
components to a voltage corresponding to a buffered low-voltage audio (or
other) input instead of ground.

Then again, if you don't want to do anything fancy (yet :), like
modulation with audio, you could just use just a beefy dimmer in front of
our SSTC.

OTOH if you really want to do PWM, the way it is "usually" done is to have
a linear ramp at the driver base frequency, then compare this ramp to
a "fixed" input voltage, and while the ramp is still below the reference,
the output pulse is kept ON, after that, OFF.

For this you only need one ramp generator (a 555 will do this, too) with
adjustable frequency, and a fast comparator. The comp output is your
square wave drive signal, i.e. you don't need any separate 555 to generate
the square wave. Neither flip-flops. But maybe a second comparator to
retrigger the 555. Just an idea. :)

> 	The final thing will be the duty cycle control (Stacato Circuit???),
> but that should be easy.

Yup - a second 555, one cap, a resistor, two diodes and pots, and you're
there. Values so that you get into the milliseconds to 10s of milliseconds
range, maybe. 100k pots 0.47uF cap.

> I intend to feed the outputs from all three
> controls to two 3-input AND gates (one for each set of two drive
> transistors).

Sounds reasonable!

> The output of each AND gate will drive the first stage of
> the amplifier, which of course ends with the coil's main drive
> transistors.  I'm not sure whether to use MOSFETs or IGBTs, but the
> driver should work for either one at this point.
>
> 	The coil will run between 75-100 kHz if I use IGBTs, or 180-250 kHz if
> MOSFETs are used.

With the driver running at f_res, the IGBTs could be pushed to higher freq
operation. Or so I've heared. ;) Haven't tried it out yet.
Maybe somebody could comment on this?

MOSFETs at <100kHz are absolutely fine too. At those freqs the main choice
criteria between IGBTs or MOSFETs is what power level => currents you
intend to run at. IGBTs have a constant, fixed V_ce voltage drop. MOSFETs
on the other hand have a fixed drain-source channel resistance, so the
voltage drop increases linearly with current.
Voltage drop times current causes transistor heating. If the drop is
larger than the ~5V or so V_ce of the IGBT, you'd probably choose the
IGBT. :o)

One additional thing for you to consider is that with IGBTs you can leave
out the schottky diodes required with mosfet operation. Sometimes the
IGBTs also already have an ultrafast reverse diode inside, so that'd be
additionally less components.


cheers,

 - Jan

--
*************************************************
 high voltage at http://www.hut.fi/~jwagner/tesla
 Jan OH2GHR