H1 SR3 in chamber (HAM5 locked) undamped transfer functions for the three stages were measured yesterday night and are attached below. Overall they look good. Damped transfer functions are currently running.

Attachments in order :

1) M1-M1 undamped tfs compared with the model
Clean tfs, excellent match with the model.

2) M2-M2 undamped tfs compared with the model
Data is noisy, which is probably as good as we can get for now.
Pitch seems to couple with the first top mass vertical mode at 1.06Hz, yaw seems to couple with the second top mass roll mode @1.97Hz. Those coupling should be reduced with damping. 

3) M3-M3 undamped tfs compared with the model

Data is noisy, but gives a good idea of the tf shape. The only odd thing in the results was the factor of 4ish discrepancy between the model and the measurement for the pitch dof (meas=4ish*model). After spending some time trying to understand, it appeared to be a wrong euler to osem basis matrix. The pitch to osems factor (8.3) was 4ish times higher than the nominal value (2.4) (from make_sushlts_projections.m under HLTS/Common/MatlabTools/). Screenshot attached shows the corrected matrix.

I doubled check for PR3, which was also wrong. The new values were saved under a safe.snap for both suspensions and commited under the svn.