RE: Parallel resonant DRSSTC (long)

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Steve Conner" <steve.conner@xxxxxxxxxxx>

Just my $.20 worth on this-

1) Parallel resonant DRSSTCs do exist. I have seen two made by Vladimiro
Mazzilli and Liviu Vasiliu. The secret is to make the inverter current-fed
rather than voltage fed. They used the so-called "Royer oscillator"
push-pull circuit with two IGBTs. IMO this SSTC design is very simple and
elegant and deserves more attention than it gets. But it's not so easy to
use in pulsed operation because of the choke in the DC link.


2) Myself and a few others (mostly on the 4hv.org forum) have spent some
time analysing and simulating various triple and quadruple resonant SSTCs,
including the ones based on the LCL induction heater concept. (I think we
all wanted to be remembered as the inventors of a new *RSSTC.) However none
of us has yet found a network that when simulated matches the performance of
the plain DRSSTC. The reason (IMO) is that the more reactive elements you
have, the more ways the energy input from the inverter is split. Ie. instead
of ringing up two resonant circuits it's ringing up three or four.

Of course I'm sure Antonio can derive a formula for optimum tuning of a
multiple resonant SSTC (MRSSTC?) in the same way as he did for
multiple-resonant spark-gap coils, that will ensure that all the resonant
elements ring up, then suddenly snap together and lash all that energy into
the topload. But this level of theory is quite beyond me.


3) Dave Sharpe said
>Having <100A versus 1kA + going
>through your power devices might improve reliability just a slight
>bit...

This is true. But remember that in the DRSSTC we need a very fast and
violent delivery of power to a resonator with a Q of about 10. Typically the
"peak" power is from ~25kW for a tabletop coil to ~500kW for a Steve Ward
sized monster. So high currents in the power devices are unavoidable if you
want good performance from a pulsed SSTC. In fact they are the cause of the
good performance!

It's really not the same as a LCL induction heater where you're delivering
say 5kW of real power into a resonant circuit with ~250kVA of reactive power
circulating in it (ie a Q of 50) I would say stop thinking induction heater
and start thinking Chain Home radar transmitter ;)


>25kVA+. I don't think (at last check) anyone in the WORLD has a SSTC / SS
Converter
>that big on the "amateur" level...

Well if you're talking about peak power ;) any kid who ever played with
TO-247 IGBTs has one. In terms of average power you are probably right. The
reason is that (extrapolating from the current state of the art) a DRSSTC
with 25kW average power would put out around 28ft of streamers and I don't
know any DRSSTC builder with enough lab space to take that on. It would also
have a couple of megawatts peak power which is daunting.


>1.  The fewer power devices in the power path, the more reliable the
>converter.
>3.  Placing a DC buck regulator in series with the half bridge as a high
>speed current limiting/controlling means will also "dramatically" improve
>converter robustness

So, the fewer power devices, the more reliable, and yet the more power
devices, the more reliable. I prefer a full-bridge because it gives twice
the output voltage so the primary components can be higher impedance. DRSSTC
bursts are so short that there's not really much time for a buck regulator
to act. I would control the primary current by cutting off the gate drive to
the bridge when the current gets above the desired level and cutting it back
in when it goes below that level. This is done with flip flops so it always
cuts out whole numbers of cycles beginning and ending at current zeros.

I plan to use this method for a CW induction heating inverter too since I
don't want to be bothered redesigning the driver. But it would probably be
unacceptable in high powered commercial heaters because of acoustic noise
output.


Steve C.