We got a few noise budget injections done in this lock before we lost lock. We will post a more complete noise budget once we have some more of the excitations done, a partial noise budget for the ASC noise contributions to DARM are attached.
The other interesting thing that we found was that MICH is only about a factor of 3 below DARM from 60-80Hz.
After we lost lock we needed to redo initial alignment, this might require realignment on ISCT1 since Jenne updated the initial alignment references for our new spot positions. 47168
Redid the MICH FF and added another filter, Feb27c, to MICHFF FM8. Filter shown in the plot.Feel free to try it in the morning.Feb27c takes advantage of Danny's beamsplitter suspension resonance measurement. So far the best known feedforward is Feb27, in FM6. Feb27b gave worse noise than Feb27. Feb27c also gave worse noise than Feb27. The problem is the ~1 degree phase mismatch around 70 Hz in the MICHFF TF fit. Also some Feb27d was made and is sort of better at 70 Hz but is worse at 100 Hz. Leaving this one on for now.
Here's a fit to the MICH FF which might have a better residual.
zpk([1030.2733+i*950.7986;1030.2733-i*950.7986;-94.6683+i*1386.0676;-94.6683-i*1386.0676;-38.9617+i*1095.7877;-38.9617-i*1095.7877;-47.1472+i*199.6596;-47.1472-i*199.6596;-15.1585;-0.025039+i*111.0189;-0.025039-i*111.0189;8.3412;-10.6324+i*71.093;-10.6324-i*71.093;-0.91023+i*62.1857;-0.91023-i*62.1857],[-5.0157+i*52.8173;-5.0157-i*52.8173;-1.2463+i*63.023;-1.2463-i*63.023;-13.5628+i*70.3543;-13.5628-i*70.3543;-0.0019152+i*111.1789;-0.0019152-i*111.1789;-48.6805+i*199.3746;-48.6805-i*199.3746;-335.845+i*587.6419;-335.845-i*587.6419;-37.2718+i*1094.6672;-37.2718-i*1094.6672;-88.2415+i*1288.8722;-88.2415-i*1288.8722],1)
Gain at 100 Hz: 1.24488138312429 (complex = -0.625528556960358 - 1.07631021665528i)
Tried Gabriele's filter in FM1 (copied below, scaled by -0.21415 gain, the one above does not agree with foton), as well as a new FF from me, called Feb28 in FM3 and plotted below.
Reran the injection test from last night. Both Gabriele's FF and Feb28 are much better than yesterday's MICH FF. These tests were done at the beginning of a lock, prior to complete thermalization taking place. Unclear if this matters to the FF quality.
Gabriele's FF
zpk([1030.2733+i*950.7986;1030.2733-i*950.7986;-94.6683+i*1386.0676;-94.6683-i*1386.0676;
-38.9617+i*1095.7877;-38.9617-i*1095.7877;-47.1472+i*199.6596;-47.1472-i*199.6596;-15.1585;
-0.025039+i*111.0189;-0.025039-i*111.0189;8.3412;-10.6324+i*71.093;-10.6324-i*71.093;
-0.91023+i*62.1857;-0.91023-i*62.1857],
[-5.0157+i*52.8173;-5.0157-i*52.8173;-1.2463+i*63.023;-1.2463-i*63.023;-13.5628+i*70.3543;
-13.5628-i*70.3543;-0.0019152+i*111.1789;-0.0019152-i*111.1789;-48.6805+i*199.3746;
-48.6805-i*199.3746;-335.845+i*587.6419;-335.845-i*587.6419;-37.2718+i*1094.6672;
-37.2718-i*1094.6672;-88.2415+i*1288.8722;-88.2415-i*1288.8722],
-0.21415
)
Measured MICH with the different feedforward filters after 3 hours into a lock. Both filters are better than Feb27. I switched out Feb27 with GabFF and used it to update MICH in the the noise budget.
We re-reran the MICH coupling later in the same lock. Danny said his test above had a slightly different excitation for each filter. Both GabFF and Feb28 are much worse. GabFF seems to be the better filter now. The overall frequency coupling is changed as well. It seems that further thermalization changed the feedforward even further, spoiling the linear noise cancellation. MICH to DARM length noise is projected to be only a factor of 3 below DARM with these levels of feedforward. We may need to make early/late lock MICHFF filters.
