RE: SSTC - experiments

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

Original poster: Jan Wagner <jwagner-at-cc.hut.fi> 


On Sun, 11 Jan 2004, Tesla list wrote:
 > Original poster: "David Trimmell" <humanb-at-chaoticuniverse-dot-com>
 >
 > Steve, why wouldn't a "self-resonant" system work for a audio modulated
 > SSTC?

"Traditionally" audio modulation is done by pulse-width-modulating the
drive signal, that is, making the pulse duty cycle smaller than 50%,
proportionally to the audio input voltage.

You can add PWM to the self-resonant setup too, but then it will work
properly only in a narrow freq range and you can't just plug in different
secondaries any longer.

Here's a self-resonant SSTC audio PWM how-to: integrate the
squared-up feedback signal minus half of Vdd/Vcc. This gives you a
positive triangle wave. Next use an ultrafast comparator on the triangle
wave and the audio reference (centered around half triangle wave peak
voltage), and let the comp output control the gate drive chip /ENABLE or
INHIBIT or similar pin. It actually works 8-)

The integrator time constant and audio volume is what limits the freq
range where this PWM scheme will work.

 > I think that a resonance "correction" circuit could be quite
 > useful. But I am not sure how to do it. The TL494 seems to be a less
 > then perfect IC for modulation, but Dan has proven it to be more than
 > capable! I wonder if it would be possible to do modulation with the
 > IR2153D IC that Dave Sharpe has been looking into? BTW, where to find
 > this IC?

You can buy it from www.irf-dot-com. Digi-key corp, Future Electronics
and Newark stock it too.

(Btw, has anyone actually tried this IR2153 design idea yet? Dave?
Someone else? Would be interesting..)

I've never tried the IR2153 but it appears to use a simple RC
oscillator. If you use an R-C combo you could add the audio as a "DC"
offset from ground e.g. don't connect the timing cap to ground, but to a
buffered audio signal. Just an idea. This gives you frequency modulation
/ FM. Although, with the nonlinear impedance of the TC secondary, this
direct FM is guaranteed to sound crappy, ranging from "flat" to
highly distorted sound. :| However, correcting comments welcome! ;-)

It would sound better if a low-pass filtered feedback signal were compared
to the audio, and the frequency corrected up/down according to the error
voltage. But low-passed feedback vs audio input signal phase shift
(filter phase shift, transistor switching delays, TC coil phase shift) is
going to cause problems there...

 > I have checked the usual suspects, but haven't found... What
 > are folk using to design their boards, Eagle? But finding the right
 > component list is troublesome!

If it's not in the libraries, just add it! :) Eagle has a library editor -
it's easiest to open some other library that already contains 8-pin
DIP/SOIC components. Then you can reuse these packages. Add and draw your
own symbol for the chip. Then connect package pins to symbol I/O pins, and
that's it. IIRC the Eagle PDF manual (www.cadsoft.de, downloads section)
has some help on how to edit the libraries.

regards,

  - Jan

--
****************************************************
  Helsinki University of Technology
  Dept. of Electrical and Communications Engineering