reading about the thermal camera: http://www.optris.com/thermal-imager-pi160?file=tl_files/pdf/Downloads/Infrared%20Cameras/PI_Brochure.pdf http://www.optris.com/thermal-imager-pi160?file=tl_files/pdf/Downloads/IR-Basics.pdf
Has stated accuracy of +/- 2% which at typical 1000K of Rossi's demonstration would be +/- 20K. So not too bad, and supports claims but nowhere near as good as a thermocouple. Rossi has at least done 1 thing right - painting the surface black to improve accuracy, but there is apparently some choice in spectral emission range and also it is only rated for 900°C, so pretty good chance for errors in setup and trying to operate outside measurement range.
