Here's the complete MR2.5 run with manually sampled 5-minute data:
I thought that compared to MR2.4, pressure increased differently, with each step giving a larger differential value than the previous except for the last one. This probably means that more lattice-bound deuterium was in the mesh. Whether this is a good thing or not remains to be seen (although according to Mizuno and Storms, D loading itself is not what one should be after), but employing the same calibration as used earlier (modified calibration 10 curve with 3rd degree polyfit after adding a 0W step), this time there was no apparent few watts of excess heat for the initial few steps and actually it tracked calibration values quite well except for the last input power step.
|Input (W)||Tcell2 (°C)||Output (W)||Excess (W)||Efficiency|
However, magicsound used a different calibration that gave lower values,
Putting this aside, there were a few odd temperature and pressure bumps which will possibly be easier to see with the full data.