nutsinee.kijbunchoo@LIGO.ORG - posted 00:27, Saturday 18 June 2016 - last comment - 02:50, Monday 20 June 2016(27830)
Violin mode damping guardian is up-to-date
As of June 18, 7:22 UTC. See Violin mode table for details.
Added ETMX, mode 7 narrowband damping filter for 507.194Hz. Setting: FM1, FM3, FM4, g=+150 (or g=+450).
Thois damps by itself, as well as together with the (still 2-mode damping) ETMX mode 6. Setting: FM1, FM3, FM4, FM5, g=-100 (or g=-300).
Both are in the guardian. this damps both 509.194Hz and 509.159Hz.
Fond an error in the VIOLIN_MOIDE_DAMPING state. In short, all gain settings for ETMX were actually applied to ETMY.
After that, the modes on ETMY were the biggest - in particular EY mode 5 ( 508.220Hz).
Turns out all the band-pass filters of EY mode 3,4 and 5 (508.146Hz, 508.206Hz, 508.220Hz) overlapped.
I changed all three bandpass filters to narrow-band filters, with zeros on the other modes. The following settings damp all 3 modes:
EY mode 3 ( 508.146Hz): FM1, FM4, Gain=-20
EY mode 4 ( 508.206Hz): FM1, FM3, FM4, Gain=-20
EY mode 5 ( 508.220Hz): FM1, FM3, FM4, Gain=-40
All these settings are in the Guardian.
Some notes:
- EY mode 5 ( 508.220Hz) is by far the hardest to damp (presumably rung up very high, but only a small couling to DARM). The daming setting above is the best among the 6 possible phase combinations.
- The filter for the 3 EY modes 3,4 and 5 (508.146Hz, 508.206Hz, 508.220Hz) are now narrowband, meaning that their impulse response has significant support out to ~3min.
- Additionally for all three modes the dampling drive slightly couples directly back into lenght, resulting in an inital response that looks like driving the modes up. Together with the filter response time, the actual damping only becomes apparent after ~4min.
- I then also turned on IX mode 7 (502.621Hz): FM1, FM4, gain=50; and IX mode 8 (502.744Hz): FM1, FM3, FM4 gain=10; both worked just fine.
- The only modes that still need attention (i.e. they are not shrinking any further) are
EX mode 2 (505.707Hz) / EX mode 3 (505.710Hz)
===============================================
As reference, here are the 3 band-pass filters:
EY mode 3 ( 508.146Hz): FM1, FM4, Gain=-20 :
zpk([-0;-0;-0;-0;-0;-0;0+i*508.206;0-i*508.206;0+i*508.086;0-i*508.086],
[0.00392661+i*508.126;0.00392661-i*508.126;0.011182+i*508.129;0.011182-i*508.129;
0.0167351+i*508.135;0.0167351-i*508.135;0.0197404+i*508.142;0.0197404-i*508.142;
0.0197404+i*508.15;0.0197404-i*508.15;0.0167351+i*508.157;0.0167351-i*508.157;0.011182+i*508.163;
0.011182-i*508.163;0.00392661+i*508.166;0.00392661-i*508.166],1e-36,"n")
gain(1.67)
EY mode 4 ( 508.206Hz): FM1, FM3, FM4, Gain=-20:
zpk([0+i*508.192;0-i*508.192;0+i*508.22;0-i*508.22],
[0.00284641+i*508.202;0.00284641-i*508.202;0.00284641+i*508.210;0.00284641-i*508.210],1,"n")
butter("BandPass",4,508.192,508.22)gain(1.06229e+11)
gain(1.15269e-06)
EY mode 5 ( 508.220Hz): FM1, FM3, FM4, Gain=-40:
zpk([0;0;0+i*508.206;0-i*508.206;0+i*508.234;0-i*508.234],
[0.00208372+i*508.212;0.00208372-i*508.212;0.00569283+i*508.214;0.00569283-i*508.214;
0.00777655+i*508.218;0.00777655-i*508.218;0.00777655+i*508.222;0.00777655-i*508.222;
0.00569283+i*508.226;0.00569283-i*508.226;0.00208372+i*508.228;0.00208372-i*508.228],-1,"n")
gain(-356.902,"dB")
Added the following violin mode dampers to the guardian:
ITMY mode 4 (501.811Hz) FM1,FM4,FM5, gain=-100
ITMY mode 7 (504.804Hz) FM1,FM3,FM4. gain=+500
ETMY mode 2 (508.010Hz) FM1,FM4, gain=-40 (gain critical: x-talk with ETMY mode 3)
ETMY mode 8 (508.661Hz) FM1,FM3,FM4, gain=-500
ETMX mode 4 (505.805Hz) FM1,FM3,FM4, gain=-500
ETMY mode 6 (508.289Hz) FM1,FM3,FM4,FM5, gain = 500 (Note: gain oscillations kick in between gain=1000 and gain=1500 - as a result, this mode is probably the slowest daming one. )
This completes all fundamental violin modes - the violin monitor screen is now clean.