Re: Terry's DRSSTC -"different" H-drive functions...

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Antonio Carlos M. de Queiroz" <acmdq@xxxxxxxxxx>

Tesla list wrote:

> Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

> After the primary voltage rings up to say 10kV, we turn off the IGBTS.  At
> that point, the primary capacitor voltage appears across the now open
> H-bridge.  But the reverse diodes will catch that voltage and limit it to
> +- the buss voltage:
>
> http://hot-streamer.com/temp/DRSSTC-H-Bridge_sch.gif
>
> http://hot-streamer.com/temp/DRSSTC-H-Bridge_Vdiff.gif
>
> In the second picture, you can see to voltage across the bridge when the
> gate drives are turned off (yellow circle).  We get about +- 300V at say 50
> amps average for roughly 100uS.  That is ~~1.5 watts into the reverse diode
> heating.  At say 200BPS that is 300W!!  That is what I am trying to avoid
> by keeping the lower IGBTS conducting during the turnoff and keeping that
> power in the primary LC system.

With some delay, I was looking at this problem, using my tuning
method. There are three situations:

The first is if it were possible to really open the driver. If this is
done when there is no energy in the primary circuit, the voltage that
appears is determined by the voltage of the secondary oscillation and
the transformer, now with the primary in open circuit. The voltage that
appears is a sinusoid with 8/pi (2.55) times the amplitude of the driver
voltage. Looks difficult to implement...

The second situation is considering the free-wheeling diodes in the
bridge. With no breakout at the secondary, all the energy is just
returned to the power supply through the diodes. If the driver is
interrupted before complete energy transfer, some energy remains in
the secondary. The primary voltages and currents remain within the
normal limits.

The third situation is if the driver is short-circuited. The system
from this time on behaves as a regular Tesla coil, when all the energy
is in the secondary (or split between primary and secondary if the
driver is short-circuited before complete energy transfer). The two
tanks are slightly mistuned, but the waveforms show the beats seen
in the energy transfer of a regular Tesla coil. The problem is that
in a drsstc with my tuning method the primary capacitor (or the primary
tank) stores at most about 1/4 of the bang energy during the energy
transfer. With the driver short-circuited, almost all the energy in
the secondary system, assuming no breakout, returns to the primary
circuit, generating capacitor voltages and currents with about twice
the value that appears in the forward energy transfer with the driver
in operation.

So, my conclusion is that it's best to operate in the second case,
letting the free-wheeling diodes return to the power supply the
unused energy.

Antonio Carlos M. de Queiroz