[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] predicting phase shift

DC Cox skrev:
Zapp's phase shift correct will not work.  Yes, like smoke and mirrors it
works in the circuit he has, but when he hooks it up to a real TC load it is
not going to work properly.
Since I am the only person having posted anything related to this subject, not just recently, but in all the 10+ years I have monitored and contributed to this list, I feel compelled to assume that you are referring to the contents in my paper: "The Hammertone Predikter".

Let me first point out, that to you, sir, I am Mr. Finn Hammer, of "Hammertone" fame.

I also strongly object to your change of subject heading.

It might also be a good idea to read the paper twice, and discover the finer points, like the link to this video. It shows the circuit driving what would appear to be a tesla coil: delivering 8+ feet of spark, with no snubber or any other overvoltage protection of the IGBT`s, despite the 800V buss voltage.


Zero crossing under a load can be achieved as long as there is not change
--- no change in potential or no change in current.  What may look like
perfect zero crossing at full load may cause very high current switching
before you get to the full power level.  This is not shown on the video
because no load is connected.  When a real load is connected the
currents will definitely dynamically change.
Others, and perhaps even you, will be interested in this video, showing the relationship between the predicting signal, the tank current, and the voltage across the load:

What you see is a remarkably spikeless Voltage square wave, voltage being reversed quite close to the current zero crossing, and the feedback predicting signal, which does not change. Neither "rapidly, nor dynamically", as you so eloquently put it. This is with the primary current being varied between 50A and 500A.

Those of us that sit down in front of a scope and find resolve to challenges in the nanosecond range find it hard to take terms like "Rapid change of signals" seriously, I might add.
With no load applied and just adjusting the firing angle of the gate drives,
you can absolutely set perfect zero crossing.  When the load is dynamically
applied there is a slight shift in phase angle because the inductance
changes due to the current changes --- which are very rapid and dynamic.

The overshoot he is trying to null out is not a zero cross over problem.
Overshoot, which is very high in his waveforms, and which he is trying to
null out, is caused by improper turns ratios on the gate drive
transformers.  This is where the overshoot problem is occuring, not from the
phase angle.

If you watch the you tube video you will see the amplitude change and there
is no change in phase shift.  You need to adjust and perfectly match your
gate drive transformers to correct this problem.

The only way to make it work is dynamically follow the rapidly changing load
perhaps with a microprocessor.  This would be a step in the right direction
to solving this problem which tends to amplify itself as coils get larger
and the power levels increase.  There is no simple, easy solution.  It can
only be solved in a dynamic fashion accurately tracking the currents.

I see no point in responding to the rest of your complaints, as, to me, they are poppycock.

But then, I asked for "peer review". Do you _really_ consider yourself as my peer?

Cheers, Finn Hammer
Tesla mailing list