Taken at face value, the large claimed gains should easily allow a self-running device, even if resorting to a TEG to convert the heat to electricity at 5% efficiency or less.
A self-running device should remove all skepticism and doubt (except for the possibility of criminal fraud) so why not make the self-runner your immediate goal? The most obvious answer is that the claimed gain, which may be real on occasion, is extremely unreliable - and the average gain is far less. Jones From: Peter Gluck A quantitative evaluation- see please the claims in the Focardi Rossi paper- is foolproof I think.
