This is most interesting in light of the totality of past experiments in LENR which are "believable" going back twenty years.
There seems to be excellent evidence for long-term COP of over one but less than two, often written off as measurement error; but far less reliability for experiments which have COP greater than two. Yeah, I know: "believability" or "reliability" is too subjective of a criterion to be meaningful - but what you are saying has a surprising ring of truth to me, which is not easy to verbalize. From: David Roberson I have run many simulations of an ECAT type device where temperature is the controlling parameter and find that the only way to get a decent COP (>2) is to use positive feedback. In this mode the device is thermally running away to generate effective gain and must be reversed just before total control is lost on each cycle. If a truly linear heat output versus material temperature is available one can have constant COP gain, but this is not true for any other function.
