- zCorr which goes to RM only
- zGc which is splitted to RM and marionette
- zM which goes to marionette only
We measured the TF to the dark fringe, which should be given by the closed loop transfer function of the DARM loop (equal to 1 at this frequencies) times the mechanical response of the used actuators, times the cavity pole.
The first three attached plots show the results for zCorr noise, zGc noise and zM noise. The north end is behaving well, while the strange feature around 730 Hz is well visible in WE.
This is clearly visible as a very large resonance in the WE marionette transfer function. Indeed, similar but much smaller resonances are visible also for the NE marionette (the noise level needed was almost 10 times larger).
The last three plots compare the dark fringe and the two marionette correction with and without noise. Comparing the noise levels, it seems unlikely that the normal noise floor is the origin of the excitation of this resonance.
The Q of the marionette resonance can be estimated from the width at half height. It is of the order of 400.
1 899.42 2 arms along z
2 915.71 2 arms along z
3 1048.5 4 arms
4 1192.2 2 arms along x
5 1217.0 arms
6 1293.5 2 arms along z
7 1406.9 2 arms along x
8 1445.2 arms
9 2109.6 body
10 2526.7 body
the details can be found on the link:
http://www.roma1.infn.it/people/puppo/Virgo/FEM/FEM_marionetta.htm
a) Assuming that they are exactly equal, we should have seen a similar peak for the NE marionetta (and even NI and WI, which have been used for marionetta reallocation).
b) The peak shows large drifts and jumps in frequency (see Irene‘s entry). The jumps correspond to earthquakes and Gc crashes, both of which give a mechanical shock that could move something around.