A much needed test to locate the deviation of the Ch1 mean. And the harmonic sweep is very obvious in these as well.
I have the corrections of the test configurations documented for tests 1-7 and for 8 and beyond. Hopefully I can get that posted up sometime this week after I finish on the inductive reactance analysis and get it presentable. I also intend to implement a cross check against your estimated resistor inductance and test data to compare to measured instantaneous current. You may recall also that I mentioned the ring form on the Ch1 data regarding the slope decay of each successive wave - I knew I had seen it before; Hewlett Packard published a chart of Laplace Transforms and that matches number 4 in their chart for the time function: f(t) =0,t<0,e^-αt(sin βt, t>0. So it is a familiar progression decay of standard ringing with no special characteristics and easily duplicated mathematically. The timing of the negative I still have not figured out and that little spike off in the middle of nowhere with no explainable cause also has me puzzled as well. There must be an event occurring that we are not monitoring...a high to low transition through a capacitor to ground somewhere, I just haven't checked it out well yet.
I don't anticipate any issues with any of the inductive wiring although the 80MHz on that 6" piece between the CSR and the grounding star would normally raise my eyebrows a bit. The fact is, those signals have such a low amplitude, and they are ringing above and below the mean value such that they entirely cancel so the real frequency is well down in the kHz where it just becomes moot. What is even more intriguing is the timing of that impulse. If we use the slope of the load resistor collapse as our determining frequency component for the inductive reactance calculations and apply that to the CSR leads, then we are admitting that negative current is flowing through a closed MOSFET the wrong way when the other side of that MOSFET is at 600V against its flow - can it really be? That body diode is surely reverse biased to the max at that instant. How is that current flowing? Source to gate leakage?; I wouldn't think there is a potential there either. The Spatial test you did ruled out any 'boot strap' effect that I thought maybe happening. So I find myself entertaining some sort of capacitive push occurring in time 90° before the event even occurs. This would be the 100V bump capacitively pushing through the battery, out the B(-) and through the shunt before the body diode has a chance to open and would be likened to charging up the source pin from the back side in anticipation of the expected negative spike on the drain that never occurs. That drain spike drops just as fast as it builds and that 600 volts is being spent somewhere. Since it is not going through the MOSFET it must be moving through the load resistor back to the B(+) terminal. This produces a ringing at the B(+) which can be seen in the scope shots. So all of that energy is either spent in the resistor or put back into the battery. The puzzling thing is that it seems to be just fine without a closed circuit and our CSR just isn't monitoring the current flow from the drain to the B(+) with the MOSFET closed and the drain always above zero so as to prevent the body diode from conducting.
So what other possibilities are there? Quantum Tunneling, Negative Resistance, Electromagnetic return path, Bidirectional Vortex currents, what else?
Even if we were upside down on our current by 2W we still would have to account for 200% more heat than it should be doing.
Leave a comment: