Extra 1/c in test mass noise

There seems to be an extraneous factor of 1/c in the definition of either the test mass noise or the coupling into the measurement. The test mass noise is generated as a fractional frequency fluctuation, and then scaled by nu/c in instrument.py, which is not correct.

We should either generate the test mass noise as velocity, or remove the scaling factor from the noise coupling in instrument.py.

@j2b.bayle, what do you think is better? I think you wanted the glitch to be in velocity, not ffd? The TM noise should probably have the same units (so velocity for both, meaning the factor 1/c should be removed in the noise generator).