I measured the TF between seismic noise and beat, then projected the measured seismic noise measured from the table to the noise budget.
Noise due to seismic is really high.
I used T240 to measure seismic noise on the table on z and y direction. z is vertical direction, y is horizontal along the table length(cavity beamline), and x is horizontal along the table width.
TF measurement setup
When I measured the TF, the PZT, driven by high V, excited the table along y direction.
Beat was measured with 100kHz input range setup.
The seismometer has a pole at 30Hz, so the data at high frequency is probably invalid, and mostly noise.
I tried 100 integration cycles for low frequency (1 - 50Hz). At high frequency I chose 500 cycles. There was no significant difference between both setups.
I made sure that the TF did not change with excitation's amplitude.
Above, the TF between seismic and beat measurement, peaks around 170 Hz are periscope's resonant frequency.
From the TF, we know how seismic adds into beat noise. The next step is measuring the noise of the table.
Then, multiplying the noise by the TF to see what is the frequency noise in the beat due to seismic.
The figure above contains 4 important plots. (Another figure with unchanged y-axis is plotted below as I promised not to change the y scale on the plot, but I want to zoom in on this one)
1) old beat measurement with previous spring suspension on RCAV
2) current beat measurement with new suspension on RCAV (spring removed)
3) seismic noise projection from z vertical direction
4) seismic noise projection from y (horizontal, beamline, table length) direction
Note that the measured TF is beat divided by seismic. Above T215's pole frequency(30Hz)the seismic signal is very low compared to the beat noise, so TF becomes large and very noisy. The projection from 200 Hz and above will be replaced by another set of data, from accelerometers. So the higher noise projection at f > 200Hz is not alarming.
The noise projection from vertical direction from 2 to 10 Hz that is higher than the measured beat noise is probably the result of
y direction driving force on the table. The table does not shake that much along vertical direction, but the beat experiences a lot of resonance peaks originated from y-direction movement. This causes the large TF on Z-direction which exaggerates seismic noise in z direction to the noise budget.
The noise projection from y direction lies nicely with the both beat measurements. This suggests that those peaks around 20 Hz are from y-direction seismic.
Since the springs we removed are not directly related to y direction movement, we might want to find a way to damp the suspension on this direction.
We have to win against the seismic by at least 4 orders of magnitude to see coating thermal noise at 1 Hz region.