Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

Guys,

If you want to output high power sinusoidal waveforms, you basically have 3 options:

1) Sine waveform signal generator + analog power amplifier (PA)
2) Rectangular waveform signal generator + digital power amplifier (PA) + analog filter (LP or odd harmonics rejecting)
3) Power oscillator, e.g. the Mazilli circuit

For frequency critical applications option #3 is unsuitable because it provides poor frequency control and load dependent harmonic distortions.  This option is good for simple inductive heaters where purity and frequency are not critical.

Option #1 is universal but very expensive at high frequencies.

The digital amplifier is a cheap alternative when driven by a suitable signal generator (lab SinGen, TL494, SG3525, CD4046, etc...)
Below please find some implementations of a digital power amplifier.  Follow it by an LC circuit or a Low Pass Filter (or odd harmonics rejecting) to obtain sine waveform at its output.

[Tomtech29]

Thanks because you can end up like this:
Upss.
time:1:04
https://www.youtube.com/watch?v=yypJQKDoYXQ
also you feel it?

[Hoppy]

Thanks because you can end up like this:
Upss.
time:1:04
https://www.youtube.com/watch?v=yypJQKDoYXQ
also you feel it?

Blimey! Feel it - I can smell it from here.

[Dog-One]

1) Sine waveform signal generator + analog power amplifier (PA)
2) Rectangular waveform signal generator + digital power amplifier (PA) + analog filter (LP or odd harmonics rejecting)

These are to me the correct approaches and what I'm focusing on.  I use a crystal controlled frequency generator driving a "WaveDAC" component to get my sine wave.  From there I dump the pseudo sine wave into a Class-D power amplifier and on to the toroid transformer.  There is a little redundancy with this design but it allows me to use off-the-shelf components.  The clear advantage of this is solid locked frequency and volume control allowing me to ease up the power transfer.  The current hurdle I'm faced with is the reaction of the toroid transformer when connected to the grenade and induction heater coils.  If you have the resonating capacitors incorrectly set, you get huge spikes that destroy the Class-D amplifier.  Designing a proper snubber network has been a challenging task thus far.  My goal is to come up with something that absorbs this reaction and bleeds it off as heat while protecting the amplifier.  Then it becomes a simple matter of tuning where power transfer is optimal and dissipation in the snubber network is minimal.

Verpies, if you have some ideas I can try, I'll get them implemented and see how they perform.

[Hoppy]

@ Dog-One

You wrote an interesting post about transients. I have experienced the same problems with both push-pull PWM and Mazilli oscillators, where the transients from the yoke are very prone to destroying the switching mosfets. Snubbers are effective to some extent but dissipate a lot of energy in heat. As you say, its a case of finding a resonant condition before the damage is incurred. The higher the supply voltage, the harder it is to tame the effects of the switching. Using a current clamped supply is important but as Nick has found, switching PSU's are prone to clamping or frying at the first sign of nasty transient spikes. I've used the PWM that Nick recently posted a link to and although spec'd as having every type of adjustment and protection, its very easy to fry the switching IGBT, if driven into a coil with too high inductance. I'm just in the process of replacing the IGBT. Its a minefield even for the experienced hobbyist, let alone the novice.