Re: SSTC - Q's

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

Original poster: biomed-at-miseri.winnipeg.mb.ca 


Well I just built one of these too and I did some research regarding design
of this type of circuit to learn more about it.

See comments dispersed below,

Shaun Epp


 >List,

 >I am in the middle of building my first SSTC. It will be very similar
 >to the PlasmaSonic II of Dan McCauley (Dan, thank you very much for
 >providing all the details and especially for your step-by-step-set-up
 >manual).

 >Scoping the waveform at the gate of the power MOSFETs confused me
 >somehow. Trying to understand it, I read AN-937 of International
 >Rectifier. On page 7 it is stated: "....to short out the dc bus. For
 >this reason it may be necessary to slow down the turn-on of the
 >HEXFET while leaving the turn-off as fast as practical.....".

The MOSFET's that are used on the plasmasonic have a "fast body diode"
meaning that the short out condition that can occure in this type of
circuit are not likely to happen.  The problem with most MOSFETS is the Trr
- reverse recovery time, which means that the internal diode that is part
of the MOSFET by it's internal construction takes a long time to stop
conducting when the voltage on the gate decreases.  What happens is the
other MOSFET turns on while this body diode is still conducting and then
you have a short.

 >IMO any slow down, whether on the turn-on or on the turn-off slope,
 >will increase the losses. Fast turn-off will create huge ringing if
 >the load is inductive. These very large spikes at the drain could
 >cause a dv/dt induced turn-on.

The turn ons are delayed abit the the resistor and parallel shottky
combination.  The turn off you want to happen asap, because of the Trr I
mentioned above.   there not needed for this on the plasmasonic board, I
tried it in hopes to cool down the driver chip but it didn't help and it
just delays the turn on.

I thought of the voltage spikes too, so I made four snubbers to go across
the Source to Drain of each MOSFET.  51ohms resistor in series with 1nF
cap.  I ran it with the snubbers, but I don't have any proof yet that they
help protect the MOSFET's better.

 >Q: Why it is not recommended to slow down turn-off instead of turn-
 >on?

See above

 >Another means to avoid shot-through could be sensing the current in
 >the source of the low-side MOSFETs. After pulse shaping this signal
 >with a Schmitt trigger it could be used to enable the driver of the
 >opposite bridge branch.

 >Q: Did somebody try this?

 >Pulse transformers are said to be used for duty cycles between 35 and
 >65 % best, because of the Volt-seconds problem. Low duty cycles will
 >cause problems therefore.

 >Q: Has somebody used high-side drivers e.g. optocouplers?

 >BTW: I have wound two different GDTs using the same type of toroid
 >material. One was made of a very close twisted trifilar winding of
 >AWG 27 and the other was made of a trifilar winding of PVC insulated
 >stranded AWG 23 wires. Connecting each of the secondary windings to a
 >dummy load of a 1 kOhm resistor in parallel with a 3300 pF capacitor
 >I didn`t find any distinct difference in waveshape or peak voltage.
 >Because of the large voltage difference between the windings I am
 >using the PVC insulated wire now.

The plasmasonic MOSFETS are 10 nF capacitance each

 >Lastly my driver chips TC 4422 are getting hot quickly and I solved
 >this problem by using a heatsink.

Use a cooling fan, the drivers need it.

 >Thank you for your attention.

 >Herwig

Shaun Epp