40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  40m Log  Not logged in ELOG logo
Entry  Thu Sep 29 20:21:29 2016, Johannes, Update, General, YARM loss measurement 
    Reply  Mon Oct 3 21:24:02 2016, Johannes, Update, General, XARM loss measurement ReflectionLoss.pdf
       Reply  Tue Oct 4 22:18:24 2016, Johannes, Update, General, X/YARM loss measurement anomalousData.png
          Reply  Wed Oct 5 19:10:04 2016, gautam, Update, General, Arm loss measurement review 
             Reply  Mon Nov 14 19:15:57 2016, Johannes, Update, General, Achievable armloss measurement accuracy 
                Reply  Tue Nov 15 20:35:19 2016, Johannes, Update, General, Achievable armloss measurement accuracy 
                   Reply  Thu Nov 17 21:54:11 2016, Johannes, Update, General, Achievable armloss measurement accuracy 
                      Reply  Thu Jan 12 02:45:53 2017, Johannes, Update, General, Next armloss steps ass_illustration.pdf
                         Reply  Fri Jan 13 08:54:32 2017, Johannes, Update, General, DC PD installed ASDCPD_up.jpgASDCPD_down.jpgscrambled_osci.jpg
Message ID: 12704     Entry time: Thu Jan 12 02:45:53 2017     In reply to: 12624     Reply to this: 12710
Author: Johannes 
Type: Update 
Category: General 
Subject: Next armloss steps 

As stated in elog 12618, using an oscilloscope to average the reflected powers and thus circumventing all filtering yielded much better results than before:

XARM: 21 +/- 35 ppm
YARM: 69 +/- 45 ppm

We can probably decrease the measurement uncertainty further by using a larger photodiode that is more suited for DC measurements. It will be placed in the AS pathtemporarily. If we get below 10 ppm systematic errors will begin to matter. To get those under control I will have to re-determine the visibility in the arm cavities and the modulation indices. The numbers to match from an estimate via the power recycing gain are <= 50 ppm arm average from elog 12586. Once the measurement scheme is up and running, we can proceed to generate ETM lossmaps. ITM will still be tricky but let's see what we can do.

Following Yutaro's approach, we can move the beams on the optcs in a deterministic way by several mm on the ETMs. Moving the beam is achieved by introducing offsets into the ASS auto alignment. As an example, the Yaw dither for ETMY is shown:

Each of the 8 test mass rotational degrees of freedom is driven by a particular frequency, and 2 signals are digitally demodulated in the real-time system: The arm transmission ("T") and the LSC arm length feedback signal to the ETM (L). The T signal feeds back to the input pointing, aka Tip Tilts and BS. This maximizes the transmission for a given test mass orientation. The L feedback controls the beam position on the mirrors in the arms. It minimizes the coupling of the dither to the length feedback, which is achieved when the beam goes through the axis of the rotational motion. This is where we introduce the offset:

The signal C1:ASS-YARM_ETM_YAW_L_DEMOD_I_OFFSET (for this example) moves the locking point of the dither-to-length coupling and thus moves the beam around on the ETM. This is true for the PIT and YAW of all test masses except ITMX. In the current configuration the TTs optimize the alignment into the YARM, and for the X we only have the BS, which is why the beam spot on ITMX cannot be independently controlled as-is. We could, however, for the sake of this measurement, temporarily temporarily give TT authority to the XARM feedback to control the ITMX beam position. I imagine something like dither-aligning with ASS the normal way, and then run a customized script in which the XARM is treated as the YARM, feecback to the BS is cut, and the YAW signals are inverted due to the reflection on BS.

Knowing the angle of the offset gives us a way to calculate the beam spot displacement with the cavity geometry. For best results I want to make sure our OpLev calibration is still good (laser power decay, although last time this was done was only about a year ago), which would be analogous to elog 11831.

As for ITM beam position, this scheme only works partially, because it would require the beam to steer further off its axis than in the ETM case. This is problematic because of the spacing between tip tilts and ITMs. I summarize:

  1. Place larger DCPD in AS path
  2. Confirm mode-matching and mod-indices
  3. Assess loss in center with zero offsets
  4. Uncertainty low enough? If not get better.
  5. Calibrate OpLevs
  6. Introduce calibrated offsets in dither alignment
  7. Wander beam on test masses, recording arm losses
  8. ???
  9. Profit
Attachment 1: ass_illustration.pdf  29 kB  Uploaded Wed Jan 11 23:00:43 2017  | Show | Show all
ELOG V3.1.3-