I was reviewing the latest analysis of Rossi's Ecat by Mr. Ruer. Over the last few years I have posted many descriptions of the thermal behavior of a system of this type with positve feedback being a consequence of the design.
The curves generated by Ruer match my first models fairly well and that is a good beginning. Thus far he has not pursued time domain analysis of the system. Had he done so, he would have discovered that the device would have began to self heat at a temperature of around 625 C according to his graphs. You can predict this behavior by viewing the graph he published, since that is the temperature at which the input power requirement begins to display a negative slope. One interesting consequence of breaching the negative resistance region is that the output temperature will begin to rise exponentially even thought the input heating power remains constant. This type of behavior matches what I have referred to as a type 2 or type 3 thermal feedback system within my earlier posts. It is not possible to stabilize the output at any temperature that exists within the negative slope region unless significant negative feedback is used to adjust the input drive. For instance, the input must be reduced by a greater number of watts than the output rises due to an incremental increase in temperature. The same criteria must be applied to an incremental decrease in the operational temperature or the device will cool back past the negative resistance knee where it stabilizes. Ruer has defined a critical temperature level that I have likewise chosen as the point of no return. This is where the input heating power shown on the curve has been reduced to zero watts. As he must realize, if you exceed that operating temperature you will not be able to force the ECAT to cool down without resorting to some extra technique such as cool water spray. Incidentally, Ruer's calculation of the device COP versus temperture does not take into consideration the thermal run away that begins once the temperature enters the negative slope region. If a PWM drive signal is used similar to what he describes, it is possible to keep the device stable as its temperature oscillates between a couple of different levels, but the COP will not be greater than approximately 10 without great difficulty. If normal quasi linear negative feedback is used for control, I suppose that higher values of COP can be obtained as you approach the point of no return. The feedback sign must be adjusted to handle the fact that the slope of the power input versus temperature curve begins positive and changes to a negative direction once you begin to operate at a temperature past the knee. It is good to see that others have replicated my conclusions to the first order and hopefully they will extend their models to be more inclusive of device operation. This will happen once they begin to simulate the time domain behavior of their models. Only then does it become evident that operation within the negative slope region has interesting consequences. Dave
