Today we installed two accelerometers on the table, for vertical and horizontal(beamline) positions.
The TF between the PZT's driving V and the signal from the accelerometer are plotted below.
I did some calculation on frequency noise due to beam/cavity mismatching in translational and pitch which causes the natural axis to change from the designed value. For example, if the beam translate away from the center by a little, same angle, the cavity length as seen by the beam will be shorter because of the curve mirrors.
It looks comparable to what we see on the beat signal.
Frank found me 2 of Breul & kjaer 8318 accelerometer. I installed them on the table for vertical and beamline horizontal directions.
I chose vertical direction because the spring was removed from the cavity suspension and become susceptible to vertical seismic.
For beamline horizontal direction, I chose it because it is the same direction the PZT pushes the table, so I expected to see some strong signal.
The signal from 8318 is sent to SR560 preamp, ac couple(doesn't change between dc/ac), roll off at 1kHz, gain200.
The signal from SR560 is sent to response (B channel on SR785)
The source is sent to HV amplifier that drives the PZT and reference (A on SR785).
The integration cycle is 50, with 10 settle cycles. I tried 100 integration cycles (quickly cheking between 100Hz and 1kHz again), but there was no significant change.
No strong signal on the horizontal direction. The magnitude is even smaller than that of vertical one.
I'm not sure how valid the TFs are. I tried changing gain on SR560 and looked at high frequency, 100 - 1kHz,TF the magnitude changes correspodingly
with the gain, the shape looks the same.
The next plan will be measuring the TF between cavity motion (shadow sensing technique)and acceleration on the table. I'll also compare with the seismometer on the table.