I moved another optical table in the cleanroom because I didn't have enough space to properly test the laser noise reduction.

 

I removed the laser fiber from its collimator without moving the collimator, so that, when I will put it back, I will not loose the alignment.

 

I set the laser at 200mA which is half of the typical operation, this current gives 60mW in output. I put an OD2 filter after the collimator and a beam dumper to collect its reflection. Then a non-polarizing cube beam splitter 50:50 and a 50mm lens for each beam to focus the beam on 2 identical photodiodes (DET10N and DET10N/M). On one beam I measure the beam size at different positions. To do that, I put a blade moved by the old translation stage and controlled by the old software that makes scans of the DC. On the DC channel, I sent the analog output of the PM100D power meter .

 

I found a waist of 20micron at 54mm from the lens and I put the PD at about that position. I didn't repeat the waist measurement for the other lens but I put the other PD at the same distance from the lens. Each PD is mounted on a XY micrometric translation stage, so, by watching the signal at the oscilloscope I centered the beams in the PDs (moving the PDs instead of the beam).

 

I put a T with a load resistance of 7.5kOhm for both the PDs. This gives 2.5V DC signal for one PD and 2.8V for the other. The two signals, coupled in AC, overlap on the oscilloscope and show the "jumps" that are probably the noise reason.

 

I closed the loop with a lowpass at 3Hz and gain 1000. See the spectra.

 

The noise in the out-of-loop PD reduces by a factor of 10 (20dB)

 

I also connected the PD at the lockin and after demodulation at 380Hz (with the internal oscillator) the noise significantly reduces (by about 10 times by eye).

 

Next thing to do is to place the PDs in the absorption setup and check if the absorption signal noise reduces in the same way