[Stefan, Kiwamu]
We are now in the process of checking/adjusting the suspensions as our length actuators.
-
We adjusted the L2P balance in the M1 stage of PRM and PR2.
-
This was done by applying a big offset in the LOCK L filter of the M1 stages and corrected the L2P by putting a non-zero gain in their DRIVEALIGN L2P.
-
As for PRM, we simply looked at the REFL camera and adjusted the L2P such that the beam doesn't move with the length offset.
-
As for PR2, because we didn't have an intuitive sensor, we adjusted it by looking at REFL_DC when PRX is locked.
-
This eventually allowed us to offload low frequency feedback to the M1 stages. Because the M1 stages take care of low frequencies below 1 Hz, we don't have to an L2P adjustment at high frequencies.
-
H1:SUS-PR2_M1_DRIVEALIGN_L2P_GAIN = 0.0045
-
H1:SUS-PRM_M1_DRIVEALIGN_L2P_GAIN = 0.006
-
PRM crossover frequency was measured to be approximately 10 Hz (see the attached).
-
We then locked PRY and actuated on BS to see how the length transfer function looks like.
-
The UGF was set to be about 10 Hz by lowering the gain so that the control doesn't suppress BS's length signal at high frequencies.
-
Only FM6 of M2 LOCK FILTER in BS was enabled to make the entire transfer function look more like a pendulum, namely 1/f^2.
-
Indeed the length transfer function was 1/f^2 above 10 Hz and looked good (see the attached).
-
Also, we copied the filters for the BS oplev loops over from the L1 to H1 directory.
-
The pitch loop looked helpful in the sense that it reduced the RMS of the motion.
-
The yaw loop may not be necessary at this point because the dominant disturbance seems mainly at 0.2-ish Hz which is out of the oplev control band.
-
For now, we leave the yaw loop off as it doesn't look doing something useful at this point. (Though we probably will need it when we use BS to lock the Michelson.)
One thing I am confused at this moment is the length transfer function of PRM and PR2. I did the same TF measurement as for the BS length drive for comparison, and found that the TF were 1/f above 10 Hz rather than 1/f^2 (see the attached). There must be some stupid reason because otherwise we wouldn't be able to lock PRX or PRY. I will revisit this issue tomorrow.