All of the test data and lots of pictures can be found here: https://www.dropbox.com/personal/albiston fusion/test data/Reactor 6 - Conduction 1/Test1
The test used a stainless steel fuel tube which contained 3g AH50 Ni, 0.45g LiAlH4, and 0.3g Nanoshel Li. The Ni was heated in air and Hydrogen reduced using the procedures in the MFMP "signal". The SS fuel container was contained in an 18" Alumina tube with a thermocouple butted up against the SS fuel container to measure the core temperature. The 18" alumina tube was wound with Kanthal A1 heating coils and placed inside two concentric Alumina tubes with a gap filled with lead (see the dropbox pictures). The lead jacket was then heated to lead melting and additional lead was added to completely fill the gap between tubes. A thermocouple was placed inside the lead jacket and a second thermocouple was placed on the exterior of the outer Alumina tube. These two thermocouples acted as a conduction calorimeter. A conduction calorimeter works by first having a good conductor (molten lead) accumulate and spread out the heat and then measuring the temperature differential across a second less good conductor (Alumina tube). Ideally the temperature difference will be linear versus heat generated inside. Mine turned out to be close to linear but not quite likely due to radiation from the outer layer.
Calibration:
https://plot.ly/803/~fear_nuts/
https://plot.ly/813/~fear_nuts/
Calibration only proceeded to 1000C and 540W because my power supply ran out of juice. I will run additional calibrations covering full range for future tests.
The test was started Saturday night with additional reducing/vacuum cycles. Test continued until this morning when I had to end it because of test controller problems. I'm not sure what the failure was yet, but it seems like either the core thermocouple or the thermocouple embedded in the molten lead jacket failed.
I think this chart shows the results the best:
https://dashboards.ly/ua-ocR57qTE73RUFWTYA2FkUB
The calorimeter temperature differential appears to be significantly higher on the active run, but uncertainties are large. Comparing active and calibration curve fit temperature differentials versus power input the COP can be estimated to be up to 1.2. Further testing will need to be completed to reduce uncertainties, but I consider this promising that the MFMP "signal" procedures will result in more consistent positive results.