Reports 1-1 of 1 Clear search Modify search
IOO (IMC)
kenta.tanaka - 1:37 Friday 02 July 2021 (17355) Print this report
Sensing matrix measurements and Trial of closing MCE pit/yaw loops

C. Hirose, K. Tanaka

We measured the response at each QPD (REFL WFS1,2, MCE TRANS DCQPD1, 2, IMMT1 TRANS DCQPD1)when we actuated each actuator (MCI, MCO, MCE, COMMON, DIFFERENTIAL between MCI and MCO) and summarized these results in form of the matrices. 

PIT

cnts/rad MCi MCo MCe COM DIF
REFL WFS1/oplev -377 100 355 -746 -1018
REFL WFS2/oplev 373 31.7 502 1234 1159

 

1/rad MCi MCo MCe COM DIF
MCE TRANS DCQPD1/oplev -0.234 -0.0122 -0.0263 -0.216 -0.193
MCE TRANS DCQPD2/oplev 0.0665 0.00316 0.0776 0.0611 0.0553
IMMT1 TRANS DCQPD1/oplev -0.238 -0.00224 -0.0871 -0.204 -0.199

YAW

cnts/rad MCi MCo MCe COM DIF
REFL WFS1/oplev -92.6 32.2 323 -242 -257
REFL WFS2/oplev -65.5 206 51.4 55.6 -267

 

1/rad MCi MCo MCe COM DIF
MCE TRANS DCQPD1/oplev -0.00780 0.0134 0.0252 -0.0200 0.00536
MCE TRANS DCQPD2/oplev 0.00806 -0.0159 0.0198 0.0236 -0.00736
IMMT1 TRANS DCQPD1/oplev -0.0130 0.0428 0.0138 -0.0527 0.0281

Based on these matrices, we obtained the combination of sensors with the highest sensitivity to the motion of each actuator. To start with, we created sensors that are sensitive to the each angular fluctuation of the MCE. Looking at the MCE TRANS QPD row of the above matrices, we can see that the motion of the MCE in the YAW direction can be obtained by adding MCE TRANS DCQPD1 and 2 together at 1:0.85. Similarly, for MCE movement in the PIT direction, MCE TRANS DCQPD1 and 2 should be added together at a ratio of 0.3 to 1.

We created control loops to control the each angular fluctuation of the MCE using these sensors. Figs. 1 and 2 show the open-loop transfer functions of the control loop for the MCE in the PIT and yaw directions, respectively. The bottom right of these figures show the spectra of the in-loop sensor and oplev during control on/off. As far as we can see, in the pitch case the noise on the high frequency side needs to be reduced a little more, and in the yaw case, the noise on the high frequency side is returning the noise and making the yaw sway a little more, so we need to review the filter design.

Images attached to this report
Comments to this report:
tomotada.akutsu - 12:00 Friday 02 July 2021 (17364) Print this report

It is a nice progress for you as you finally start locking the ASC loops!!

By the way, I got a little bit confused on the logic how you created the "sensors" used for these control loops. Could you please summaize it more clearly? so that we (I?) can follow the flow of your thought. For example, is my understanding as follows correct?

  • For yaw, in MCE_DCQPD1, as COM (MCi+MCo) has a comparable signal with MCe, you would like to erase it by using MCE_DCQPD2: 0.0200/0.0236 = 0.85.
  • For pitch, in MCE_DCQPD2, as COM has aa comparable signal with MCe, you would like to erase it by using MCE_DCQPD1: 0.0611/0.216 = 0.3.

Or, maybe you have already "diagonalized" the matrices with a more systematic scheme than I put in the above??

It would be nice for us to have a habit of keep traceability of our work. Concretely speaking, could you please point where the raw data sources (i.e. your measurements; spectra, transfer functions, whatever...) from which these numbers in the tables came?? For example, diaggui figures might be able to easily show which channels were used)

It might be nice to (re-)make wind shields for the MCe oplev :)

Search Help
×

Warning

×