Abd ul-Rahman Lomax <a...@lomaxdesign.com> wrote:
> Cart before the horse, Peter. The first issues are scientific, and > exploring the parameter space is *more difficult* if, at the same time, > high "signal" is required. "Pushing noise down" by careful experimental > design can save a lot of money and time. > I agree. I think the NRL in Washington goes overboard with this approach, but generally speaking, I agree. Sometimes you can measure a lower level of heat with more confidence than higher level. Small-scale calorimeters work better. See: http://lenr-canr.org/acrobat/RothwellJbutterside.pdf This could be why STMicroelectronics (a.k.a. "Big International Company") is using a smaller amount of wire than Celani, with a different cell and calorimeter configuration (according to Celani). > If the exact conditions for heat production are not known, if they depend > on very difficult-to-control conditions, such as the exact size and number > of cracks in palladium deuteride, as appears to be the case with the FPHE, > then your scaled-up experiment might unexpectedly produce a lot more heat > than you expected. It's dangerous. Pons and Fleischmann scaled *down* for > exactly this reason. > Exactly right. > In a similar way, "reliability" is certainly desirable. However, if we > don't have "reliability," if, say, half our experiments shown nothing while > the other half, seemingly the same, show significant heat, we are not > stopped and we need not -- and should not -- demand reliability before > proceeding. Heat/helium was conclusively demonstrated with not-reliable > experiments, that is the power of correlation. The "dead cells" serve as > controls, such that the hidden variable is all that is varying, plus, of > course, the output. > Right. And important. Work with what you have, don't hold out for something better but unobtainable. - Jed
