Previous A2L measurement is based on the assumption that actuator efficiencies are identical for all 4 coils.
We thought that the unbelievable "tilt" may be caused by imbalance of the coils.
1. Setup an optical lever.
2. Dither the optic by one coil and demodulate oplev outputs(OL_PIT or OL_YAW) in that frequency.
3. Compare the demodulated amplitude. Ideally, the amplitude is proportional to the coil actuation efficiency.
What we did:
MC2 is the least important, but the easiest.
1. Placed a red laser pointer at MC2 trans table. During the installation, I moved the mirror just before QPD.
2. Made a python script that measures coil actuation efficiency using the above method. I set the driving frequency to 20Hz.
It is /cvs/cds/caltech/users/yuta/scripts/actuatorefficiency.py.
The measurement result is as follows. Errors are estimated from the repeated measurement. (Attachment #1)
MC2_URCOIL 0.953 ± 0.005
MC2_LRCOIL 1.011 ± 0.001
MC2_LLCOIL 0.939 ± 0.006
For MC1, we can use the main laser and WFS1 QPD as an oplev.
But we only have slow channels for QPD DC outputs(C1:IOO_WFS1_SEG#_DC).
So, we intentionally induce RF AM by EOM(see Kiwamu's elog #3888) and use demodulated RF outputs of the WFS1 QPD(C1:IOO_WFS1_I/Q#) to see the displacement.
1. Replaced HR mirror in the MCREFL path at AP table to BS so that we can use WFS1.(see Koji's elog #3878)
The one we had before was labeled 10% pick-off, but it was actually an 1% pick-off.
2. Checked LO going into WFS1 demodulator board(D980233 at 1X2).
power: 6.4dBm, freq: 29.485MHz
3. Turned on the hi-voltage(+100V) power supply going into the demodulator boards.
4. Noticed that no signal is coming into c1ioo fast channels.
It was because they were not connected to fast ADC board. We have to make a cable and put it in.
Is there any place to place an oplev?
- prepare c1ioo channels and connections
- I think we'd better start A2L again than do oplev and coil balancing.