E-log entry for Friday - will attach more plots to this entry on wednesday after i am back to the 40.
Started working at some 1030 hrs and recording the Open Loop Tfs for all 6 loops.
The control was not so good, and I lost the lock quite a number of times while measureing
WFS did not converge when the spot was aligned to the center. But there was convergence to a non-center point. So if the control system was switched on near those points, it was converging to that point.
Autolocker : switches WFS control on directly, whereas the best way is to gradually increase the gain to 1. Also, the autolocker code now needs to be changed to incoporate the switing off the MC2 oplev in down and switch it on in the up script.
After Koji locked the Reference Cavity in the evening, I resumed measurements for the Open Loop TFs.
Measurement of the Open Loop Transfer Functions :
noise waveform was generated using arbitrary wf generator and injected into each loop.
An LPF was applied to have max co-relation at minimum disturabnce. (thanks to Rana)
The Transfer functions, Co-relations and Power Spectra were then measured using the DTT.
Power Spectrum of the IN1, IN2 and EXC shows clearly the suppression of the noise, and OLTF shows the phase margins.
- Courtesy Rana again for suggesting the idea of plotting power spectra of all signals in the same graph.
Later in the night , Koji worked with me and we reflected upon all TFs and changed gains whereevr required according to the phase margin considerations from the Open Loop TFs.
We used the same output matrix given in the previous e-log.
Final gains -
Alignment Gain in the WFS Master - 1.000
| Loop |
Gain |
| WFS1 P |
0.27 |
| WFS1 Y |
0.7 |
| WFS2 P |
0.15 |
| WFS2 Y |
0.110 |
| MC 2OPLEV P |
-0.1 |
| MC2 OPLEV Y |
-0.1 |
this measurement invloved locking the MC to the correct position, with the spot centered at both the WFS and the QPD. invloved some cheating (offsets) after we tried centering w/o offsets.
demod signal was also centered while alignment.
credits to Koji for getting the correct lock position and also staying with me till late night in the lab
Important Points to be noted
1. All loops' histories have to be cleared while swtiching them on.
2. turn the loop output before the loop input so that there is no remnant history in the loop.
2. Alignment gain was gradually increased to 1. and tehn the oplevs turned on.
Later measured teh PSD of 6 error signals under 3 conditions -
New Control ON
New Control OFF
Old Control ON
Also measured the time series for the MC_trans and MC_refl for the 3 conditions.
|