After a lot of discussions, visual inspections and transferfunction measurements we think the odd MCO transfer functions are due to 3 independent effects:
- There is a coil actuation strength imbalance on maybe 30% between left and right. We set the new coil gains to [1.22,-1.22,0.75,-0.81]. Old: [1,-1,1,-1].
- The optical lever sensing seems to have a 2.8 degree rotation between PIT and YAW. This manifests itself as a feature in the YAW transfer function at the PIT resonance frequency of ~12Hz, as well as a ratty PIT transfer function at lower frequencies. We updated the rotation matrix, see plot 4.
- Finally the optical lever seems to be sensitive to length motion of MCO (at least more than MCI). This results in the YAW transfer function having a feature near the main pendulum mode. In principle this could be fixed by changing the gouy phase of the optical lever photo diode, but I don't think we care enough.
- Plot 1 shows the as-measured transfer functions from each coil to PIT and YAW.
- Plot 2 show the same transfer functions after adjusting the coil gains and the optical lever rotation.
- Plot 3 is a screen shot of the new coil gains.
- Plot 4 is a screen shot of the new optical lever rotation matrix.
[Stefan-san, Matteo-san, Kohei Yamamoto-san, Kirii-san, Ohishi, Izumi-san, Kokeyama-san, Enomoto-san]
On 30th May:
In the morning, some paramters in digital system have been changed in order to
*adjust the left (H1, H2) and right (H3, H4) imbalance
and to
*adjust little rotation between optical lever axis and mechanical axis
(perhaps due to roll rotation of the mirror).
Then we measured TF again but it seemed not to be improved as expected(see attached).
Izumi-san and Enomoto-san will adjust the parameters again.
Naoko, Yutaro, Kiwamu,
Following Ohiishi-san's discovery of the coils being not well balanced after the IMC alignment work, we re-adjusted the coil output gains.
We excited each coil at a time at 7 Hz and looked at the optical lever yaw spectrum in order to compare the peak height driven by differrent coils. Initially the coils seemed to be balanced to a 50% level which is consistent with the plots that Ohishi-san posted. We then tuned up the output coil gain so that the peak heights of all the coils match.
The attached shows the current best gains. The imbalance seems to be as small as 5-ish% at 7 Hz.
Also, we left an auto measurement running to evaluate the effect at other frequencies. We will have a close look tomorrow at these measurements.
[EDIT]
The measurements are done and they are stored at:
-
/users/VIS/180530/MCO/MCO_H1_N_180530_1527677795.xml -
/users/VIS/180530/MCO/MCO_H2_N_180530_1527677978.xml -
/users/VIS/180530/MCO/MCO_H3_N_180530_1527678162.xml -
/users/VIS/180530/MCO/MCO_H4_N_180530_1527678346.xml
