Keeping the high-power laser current at 0.8A we have 420mW right after the fiber output.
Today the best I could get from the IPC was a power maximum and minimum (as a function of the rotation angle of the halfwaveplate) of:

Right after the IPC
407mW (max)
15mW (min)

On the imaging unit breadboard (after all the optics on the path, including the chopper)
190mW max
3mW min (comparable with the power-meter sensitivity)

In order to get 30mW (the power we usually use for calibration), I increased the laser current up to 1.3A

In order to notice a waist difference (in the absorption signal) between low and high current, we have to be able to have 30mW with high current. I think we have to get the IPC better optimized, or, if not possible, we need to add another IPC