Today I check the beam shaking at another place, as shown in attached Fig. 1. I compared it with the calculation I did with ABCD matrix.(Fig. 2 The result is 0.00059) However, the result doesn't match the measurment(Fig. 3 this value should ccrrespond to 0.000142). I will try to find why my code is wrong.

I also measured the noise spectrum in low frequency, as shown in Fig. 4 and 5. As told by Matteo, our loop can sense the beam shaking as a length shaking. So we can see the noise level is increased.

So I decided to take the video of this shaking on the first iris in the vacuum tube between IM and EM.

(Video is here)https://drive.google.com/open?id=1jNG-MD1ATfqfCkNuEixQux5IXnzmwJTk