Trsstc - Not so good, apparently

["Tesla list" <tesla@xxxxxxxxxx>]

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

Hi:

I was looking at the behavior of a "Triple resonance SSTC", that would
have the structure of a "magnifier", but using external excitation by a
voltage source insted of capacitor discharge.

.            k12
.   +---C1--+   +------+--L3--+
.  +|       |   |      |      |
.  vin      L1  L2     C2     C3
.  -|       |   |      |      |
.   +-------+   +------+------+

I was hoping to obtain something similar to what happens in the magnifier,
where the splitting of the secondary coil of a conventional Tesla coil,
and the inclusion of the capacitance C2 results in interesting solutions
with very fast energy transfer.
http://www.coe.ufrj.br/~acmq/tesla/magnifier.html

I tried a design similar to what I had done for a double resonance SSTC,
where excitation between the two resonances of the system, with proper
dimensioning of the elements, results in fast energy transfer with high
voltage gain, with all the energy in the system being concentrated at
the output capacitance after a given number of cycles, and the input
current in phase with the excitation voltage during most of the time.
http://www.coe.ufrj.br/~acmq/tesla/drsstc.html

I was hoping that designs would exist where the gradual inclusion of
C2 in the circuit would gradually transform a DRSSTC into a TRSSTC.
But what I found is that these designs -do not exist-.
C2 would have to be negative if the condition of complete energy transfer
is to be retained.

There are sets of designs with complete energy transfer, however.
The system resonates at three frequencies, f1, f2, and f3, and
there are two possibilities, one with excitation between f2 and f3
and another with excitation above f3. The design mentioned above would
be with excitation between f1 and f2, but it is impossible.

In the possible designs, the voltage gain is always smaller than in
a DRSSTC using the same C1 and C3, the voltage gain grows slowly with
the increase in the number of cycles for energy transfer, and the input
current is never in phase with the excitation voltage.

Primary current feedback doesn't improve much the situation, and
can even result in excitation at a too low frequency, with large
increase in the primary current. It remains to be seem if there is
an adequate way to design these systems for excitation at the
resonances.

I wrote a simulator: http://www.coe.ufrj.br/~acmq/programs/trsstcd.zip

It can design the system and simulate what happens, with forced
excitation or with primary current feedback. For comparison,
I have also the designer/simulator for drsstc systems:
http://www.coe.ufrj.br/~acmq/programs/sstcd.zip

I finally figured out a reliable way to simulate the stopping of the
driver after a prescribed time. The programs can simulate what happens
if the driver is open, short-circuited, or opens with free-wheeling
diodes. I had to consider an open-circuit driver resistance, otherwise
the simulator becomes unstable. Artifact of the trapezoidal integration
method.

Antonio Carlos M. de Queiroz