New measure about Study of Noise With Accelerometers

We want to analyze the noise of the out of loop Photodiodes (PD 12595) in close loop about the frequency 1-50 Hz and to find a solution to mitigating this noise. First of all we search a corrispondence with the mechanical vibration.
With the New bench Layout (see entry of Alessio) I put on the bench 3 accelerometers:
1) On the Laser
2) On the Chiller
3) On the Optical Table
We want to study the Coherence of the noise of the accelerometers respect to the noise of the out of Loop Photodiode ( PD 12595) in close loop.
I measured the following parameters:
- Out of Loop PD (12595) NO Pb: RIN of 12595 PD (out of loop) in close loop (G=300, 30-300 Hz of band) when there isn't a brick of Lead (Pb) on the Laser and chiller and when
- Out of Loop PD (12595) With Pb: RIN of 12595 PD (out of loop) in close loop (G=300, 30-300 Hz of band) when there is two bricks of Lead (Pb) on the Laser and chiller
Then i measured the noise spectrum of accelerometers:
- Accelerometers NO Pb: noise of the accelerometers when NO lead bricks are positioned on the laser and chiller
- Accelerometers with Pb: noise of the accelerometers when lead bricks are positioned on the laser and chiller
Then i measured the Coherence of the Accelerometers respect to the noise of the out of loop PD 12595.
In the following the graphs:





The noise about the out of loop increase...

I noticed that when i put (and remove) the lead on the laser the feedback system goes to the oscillation.