Re: A SSTC oscillator

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

Original poster: dhmccauley-at-spacecatlighting-dot-com 


It depends on how you want to use the SSTC.

First off, the TL494 can operate perfectly with 0% deadtime when running in
single-ended mode.  You are probably referring to the PlasmaSonic board
which is using
push-pull mode.  In push-pull mode, the minimum deadtime is 3%.  However,
even the PlasmaSonic can be hardwired re-configured to operate at perfect
50/50 duty
cycle (0% deadtime) single-ended mode.

You also need to determine what you want out of your SSTC.

1.  If you are looking for the longest arcs possible and run the highest
power possibly, you want to run 50% duty cycle single-ended with the TL494.
This will allow for maximum
output power and arc length when using half-rectified AC input to the
full-bridge.  This mode of operation is obtainable from the PlasmaSonic
boards with some very minor hardwire
changes.  The new PlasmaSonic boards out next month will have jumper
selectable operation for this.

2.  If you are looking for audio modulation, you obivously will have lots of
deadtime because you are PWM modulating the signal.  Also, audio modulation
works best with
push-pull mode with the TL494 as there is a minimum of 50% deadtime which
allows you to run more efficiently and reliably than when running
single-ended at 50% duty cycle.
Push-pull mode with the TL494 reduces average power by 50% therefore
providing a much cooler running and reliable system than single-ended mode.
Audio quality is also
greatly enhanced (through my experience)

The TL494 is a proven chip and the workhorse of the majority of commercial
switching supplies in the past 20 years.  It also works extremely well and
is very readily available which is why its
used for SSTC work.

The Captain


 > Gang,
 >
 > I am searching for an oscillator for a SSTC. From my limited experience,
 > it appears, that there are 2 importand design criteria that must be met,
 > if the H-bridge is going to be able to switch quietly and cool:
 >
 > It must be driven in tune with the load, and with as litle dead time as
 > possible.
 >
 > The issue of dead time rules out the TL494. Although it has been the
 > workhorse in the development of SSTC`s, it cannot go below 3% deadtime.
 > If there is deadtime, there is hotswiching, and therefore noise.
 >
 > The other issue is tuning. I would like to use the method suggested by
 > Ken
 > Herrick, where the voltage across a cap in the ground lead is derived
 > into a square wave which the oscillator locks onto.
 >
 > The MAX038 function generator fits the bill, if it was not for the
 > prize, so the question is this:
 >
 > Anybody know of a chip that can deliver a 50%/50% square wave, and has
 > the ability to lock on to an external square wave?
 >
 > Cheers, Finn Hammer
 >
 >
 >
 >
 >