After developing my latest computer thermal feedback model of Rossi's ECAT, I realized that it would be a good exercise to look back at some of the earlier demonstrations that he conducted to see if the data presented made sense from my model's perspective. After a lot of staring and playing with numbers, I found that there appears to be reasonable correlation.
To save time and words I assume that anyone who is following this document has access to the data generated during that test. If you need assistance locating it send me a note since it is currently online. In October 2011 test Rossi has his ECAT attached to a heat exchanger and Mats Lewan was present to help gather data and publish it on his nyteknik.se site. The data collected is quite extensive and temperature readings are taken on approximately every 2 seconds. This is a large file and contains a great deal of useful information to analyze. The first section of the temperature curve for the ECAT box mainly begins at 12:32 where full power is applied to the device, which is this case is approximately 2700 watts. This level of drive is more than enough to cause the core of the ECAT to reach the positive feedback loop gain of greater than unity operational region. If left at this drive level for too long, the ECAT will reach meltdown and that is why the input power is dropped to 0 watts at 14:00. Rossi then proceeds to drive the input with a pulse width modulated waveform until one last final large drive period is engaged. A sample of these drives are as follows:0 watts at 14:00,Full Power at 14:10,0 watts again at 14:20,Full at 14:37, and etc. I performed a quick analysis of these drive pulses and see that the average power is roughly 1/2 the maximum level. The pulses occur rapidly enough to be smoothed out by the thermal time constants of the ECAT. The level at which Rossi began to apply this 1/2 power waveform appears to be slightly below the temperature which would result in a continuous rise in core temperature and generated power. This is evidenced by the slow fall off in the box temperature data during the time that the input function is applied. If you look at the temperature curve from 14:36 to 15:09 you will observe that the temperature is falling off at a moderate rate. My model suggests that he could have actually chosen a point in time to begin the PWM waveform that would have generated a truly flat curve throughout this region, but the system positive feedback gain would have made it next to impossible to hold that state over a large time period. To obtain flat performance, Rossi would have had to apply the initial full power curve for slightly longer to allow the core temperature and thus generated power an opportunity to increase. In the case we are studying, Rossi proceeded to apply full power at 14:59 and again, this is sufficient to allow the core to reach a state of thermal run away if continued for long enough. But a stable demonstration was his goal so he cut off the drive to the device at 15:53 and allowed the system to continue operation with virtually no input drive power until the end of the demonstration. When the high drive level was immediately reduced to zero there is a moderate delay during which the box temperature rises and then begins to fall off rapidly. The shape of this temperature falloff mimics that expected according to my model. It suggests that the slope of the falling curve has an inflection point at the temperature where the positive feedback loop gain exactly reaches unity. This is also the temperature at which the negative resistance is extinguished for those of us who like to view the world in that manner. Once the box temperature falls below this value, the ECAT has only one way to go and that is to cut off. The core will continue to generate excess power during all of the time that it is held at an elevated temperature by heat storage. This will contribute to the total output energy measured during the test, but there is no chance for it to reach another state of power increase unless the input power is reapplied. I am not able to calculate the amount of excess energy that originates from the core without much more sophicated information and that appears to be impossible to obtain at this time due to trade secrets, etc. The good news is that I can understand what is transpiring during the test by consulting my computer model. It must reflect the operation of the ECAT to a reasonable degree so I feel confident that the model is useful. Dave