I compared the results of the shaker injection for the MCF legs.

In the accelerometer signal, we can see excess at 60-140 Hz for the bellows legs (1,3,5), but not for the straight legs (2,4,6).

No significant excess was found in the MCi/MCo oplev signals.

I compared the results of the shaker injection for the MCE legs.

In the accelerometer signal, we can see excess at 40-210 Hz (in X,Z) and 40-120 Hz (in Y) for the bellows legs (1,3,5), but not for the straight legs (2,4,6).

No significant excess was found in the MCe oplev signals.

I compared the results of the shaker injection for the viewports or the floor for the oplevs. 

No significant excess was found.

I tried to estimate the projection from the ACC to the Oplev signal, using table tapping test data.

But it looks not working well because the projected result is larger than the reference.

Washimi-kun,

I have several questions on the measurements and the results.

1. What are the frequency regions you injected white noise?
2. Are the excess noises linearly coupled (Do the excess noises have a coherence with injected signals)? Or, nonlinear coupling?
3. Does the reference signals see some signals (real vibrations)? Or just see the noise?
4. How about the relation between injected vibration and vibration on the table? Smaller, comparable, or even larger?

Answers for Ushiba-san:

1. For the shaker, the Injected signal (digital) was 20-5000Hz white noise. However, due to the shaker's specs, it was high-pass at ~60Hz. 
2. I checked the coherence in xml files, but somehow it was not recorded. (Only in the template file, coherence was recorded. that's a mystery). I will try it from frame files.
3. The sensor and DAC noise level is about 2e-6 m/s2/√Hz for the ACC. So about over 30 Hz is actual vibration.
4. The injected vibration should be much larger than the vibration on the table because the shaking vibration was very large and I could feel it by hand.

Washimi-kun

Thank you for your replies.
I have several comments and additional questions.

1. You mentioned injection signals have high pass at 60 Hz, and it means the injected vibration is no frequency response (white noise in frequency region above 60 Hz). However, excess noise around 60 Hz is much smaller than those around 80 Hz. Does it imply vibration inside the chamber around 80 Hz is enhanced somehow from the ground vibration?
2.You said sensor noise and DAC noise are about 2e-6 but I cannot understand what the DAC noise level is. Does it mean DAC noise (~1e-6V/rtHz) is equivalent to 2e-6 m/s^2/rtHz?
3. What I would like to know is not the injected vibration is large enough with respect to the ground motion but the vibration on the optical table is enhanced from the tje injected vibration. If so, the ground vibration is somehow enhanced by the stack (or chamber) and it is so problematic. So, I would like to know the relations between injected signal and vibration on the stack.  

Sorry, "DAC noise" was typo of "ADC noise".

Today I checked how our portable shaker could generate vibration using the 3-axial accelerometer, near the MCF chamber.

The output signal was the same as in the last week's work.

Because the shaker moves vertically, the Z signal was the largest, but an excess of around 70 Hs was found in X&Y.

Even though the shaker was OFF (blue), the peak at 60 Hz and 58.8 Hz was much larger than last week. Due to vacuum evacuation?

After this work, I moved this accelerometer to the OMC area.