New Energy Research Laboratory Device and Process Testing Update
by Ken Rauen
Sonofusion
Research, true research, is full of unfulfilled predictions, unanticipated results, and mistakes. A researcher cannot be fully prepared when facing the unknown. A researcher goes where no one has gone before, otherwise it would not be research. We have had our share of those experiences. Unfortunately, this report announces one of those mistakes, coming from unanticipated and unknown operating conditions of our ultrasonic reactor. We must scale back the magnitude of the excess heat reported in our last issue. The 8.5 watts of excess heat reported in Issue No. 40 was incorrect, despite all of the calibrations performed on the system before, during, and after the experiment. Retesting of the same ultrasonic transducers with our new understanding of what is going on still found excess heat, but only as high as 1.5 watts. This was with 10 watts of ultrasonic power input.

The error was found by serendipity. Weeks after the results were written up, I accidentally connected a pair of ultrasonic transducers out of phase. This resulted in excessive voltage in the resonant system of the piezoelectric transducers and their series-connected inductors. The acoustic load on the piezos was reduced, being out of phase, and the "Q" of the resonant circuit rose, thereby increasing the voltage across the piezos. Our custom-made wattmeters which were inserted into the resonant circuit did not like the overvoltage, and some components were "smoked." The Crest oscillator was also damaged. While testing to find the problem, I measured test points with an oscilloscope, which normally were not measured or were not measured frequently.

Another pair of titanium-tipped transducers was tested. Only 1.5 watts of excess heat was found. Operating conditions were changed in an effort to find out what increases or decreases the excess heat. The oscilloscope was kept on the piezo voltages; we were seeing higher voltages than we had ever anticipated. Originally, the wattmeters were scaled to handle 2kV, peak; our hardest-driven signals into Roger Stringham's reactor and original transducers only reached 1.6kVp when we boosted the oscillator to 140 VAC from a variac. Now we were easily seeing over 2kVp even after the out-of-phase condition was corrected. Chris Eddy, the designer of the wattmeters, said that the 2kVp design point was for calibration purposes and that the instrument could take much more before "clipping" occurred. Clipping is an electronic engineer's term for a specific form of distortion. It is the operating point where the electronic signal entering or leaving a transistor or an operational amplifier integrated circuit chip (called an opamp) reaches the component's power supply voltage, and the voltage of the signal above that point gets "clipped off," as if a sine wave drawn on paper were clipped with scissors at the top and bottom of its curves.

Continued probing of the electronics with an oscilloscope during sonofusion runs resulted in the discovery of the wattmeter clipping. One day, my 100x scope probe died in a puff of smoke as I tried to see what a piezo voltage was. The scope registered 3.6kVp before it was fried by the oscillator. At this point, Chris Eddy's reassurance was inappropriate consolation that the system could handle the higher voltages; I was no longer content that the electronics were safe and accurate. My 100x probe was gone. All I had left was a 10x probe, so I went inside a wattmeter to monitor the lower voltages across the sense resistor of the voltage divider network (the solid state components cannot sense several thousand volts directly) and across the current sense resistor. Both points registered voltages way above normal opamp inputs and even their power supply voltages, clearly indicating trouble to me for the first time. Inspecting the signal stages of the wattmeters showed me that the wattmeters were clipping. The DC outputs were not off-scale, so I never suspected that the inputs and intermediate stages could be clipped. We knew the resonant system had a nearly 90-degree phase shift between voltage and current, indicating a predominantly reactive load. We expected large voltages and currents for a tiny power dissipated. We did not anticipate the voltages produced at the wattmeter inputs to ever exceed the overload values.

Testing of many titanium-faced piezo transducers since then has shown us that great variability exists in electrical loading characteristics with the Crest oscillator, far greater than we anticipated. Since the transducer assemblies are of our own design, not even Crest, our contracted manufacturer of the transducer stacks, could know of their performance. We had overdesigned the wattmeters, but it was not enough. So goes research.

We have sought and have seen several runs where "live zeroes" with heavy water were observed. Nominally +0.4 and +0.5 watts of excess heat were observed with heavy water by reducing oscillator supply voltage, changing the frequency sweep window, and changing the argon pressure. Turning off the oscillator produced +0.32 watts of nominal excess, an offset error, showing the live zeroes to be fairly good. We are just now running a deuterium-depleted water sonofusion run with fresh titanium stacks, which has maintained about +0.35 watts of excess, that same offset error seen before. Despite the calibration offset, the system is capable of detecting an energy balance, so we believe that the excess heats measured do exist.

We are not as close to commercializing a sonofusion demonstrator reactor as we had thought. We still think we can do it, though we are not sure if what we have at present will be decisive enough to convince the skeptical and disinterested side of the scientific community that cold fusion is real. We will explore improvements that may increase the excess power.