I wrote:
> 3. It is VERY difficult to measure 4 mW. > It is *much, much* harder than measuring 100 mW or 1 W. You may not appreciate this until you have spent a good deal of time looking at raw data from calorimeters, and looking at different calorimeters. For nearly all instruments, 4 mW is in the noise. > 4. Even when you have 1 mW precision, it does not apply across the full > range of power, and it is extremely unlikely to apply near zero. Look at > the data from Miles and others who have done high precision calorimetry. > The curves turn 90 degrees and goes off in the opposite direction as you > approach zero. Calorimetry is not linear or predictable at all scales, at > all power levels, to an arbitrary level of precision. > I refer to Fig. 5 in this paper: http://lenr-canr.org/acrobat/MilesMcalorimetr.pdf The cell constant is inconstant below 0.6 W with this particular instrument. With another instrument that works on another scale, you might be able to measure below 0.6, but at some point you will see a distortion like this. If you get the same response fitting the same data points each time you test at a certain power level, you might search for excess heat as a deviation from that curve, even though it is highly non-linear. However, that is a dicey way to do things. You really need a calibration curve that it not headed straight up to infinity in the area of interest. Miles discusses the reasons for this non-zero intercept problem, which he calls a "large error." He shows that this issue calls into question the results from Lewis et al. Remember: * There is a difference between accuracy and precision. * Precision is not the same at all power levels. * 5 mW is very low power, unless you happen to be Rob Duncan working with a picowatt calorimeter to measure collisions by individual cosmic rays, in which case 5 mW is a gigantic level of power, off the scale. That's a different kind of instrument altogether. - Jed
