We tried to measure the Boltzman constant, conducting the Johnson experiment.
We ploted V RMS vs the resistance and we get that for the higher resistances we lose the linear conection. We don't know how to explain it, and thought that it might be connected to this paragraph from Johnson's article (1928):
"If the input element and the amplifier are so chosen that R(omega) does not change much over the frequency range of the amplifier, then it is permissible to use the mean value of R(omega), obtained to a sufficient degree of approximation by replacing co by coo, the frequency of maximum amplification."
But he doesn't specify the considerations.
If anyone knows the reason to this phenomena it would be a great help.