Reports until 11:26, Sunday 11 December 2016
H1 General
cheryl.vorvick@LIGO.ORG - posted 11:26, Sunday 11 December 2016 - last comment - 13:17, Sunday 11 December 2016(32443)
out of Observe to run a2l

After reading Keita's alog 32025, I used the template to passively look at coherance over this lock, and pitch is at 0.9 and has been and this was not in the previous lock, interestingly enough, though yaw was worse at the end of the last lock.

Going out of Observe now to reduce the coherance made sense in light of the long lock we just had.

Comments related to this report
cheryl.vorvick@LIGO.ORG - 11:36, Sunday 11 December 2016 (32444)
Link

paasive a2l checks:

  • first file: 18:59:54UTC - before running a2l
    • pitch coh = 0.9, which is significantly higher than the referances
    • yaw coh is below the referances in the 15-21Hz range
  • second file: 19:23:25UTC - after running a2l
    • pitch coh is now at or below referances
    • yaw coh is now higher than the first measurement (first file), and is at, above, and below the referances in the 15-21Hz range
Images attached to this comment
cheryl.vorvick@LIGO.ORG - 12:41, Sunday 11 December 2016 (32445)
Link

Here's what's odd - pitch and yaw coherances are better an hour after running a2l, but the entire region 10Hz to 40Hz is elevated, and the range has gone down.

Plot start time is 20:32:06UTC.

Images attached to this comment
cheryl.vorvick@LIGO.ORG - 12:55, Sunday 11 December 2016 (32446)
Link

New plot - start time 20:43:21 - coherence is higher but 10-40Hz is not elevated.

Images attached to this comment
keita.kawabe@LIGO.ORG - 13:17, Sunday 11 December 2016 (32448)
Link

Smaller coherence with larger DARM noise at 20Hz means that the excess noise is not due to a2l coupling. When something other than a2l coupling dominates the noise at 20Hz, running or not running a2l is irrelevant, so please don't run it.

Again I'll say this, it's a perfect project for a fellow to write a script to plot the coherence and DARM BRMS for 15-25Hz or so VS time. With such a script it would be easy for operators to determine (and show to others) if running a2l would be beneficial or not.

Anyway, going out of OBSERVE to run a2l script is OK when all of below are satisfied.

  • Measurements show that coherence has been constantly large for hours (not just one or two) in a single lock stretch.
  • L1 is out of OBSERVE.
  • We're not chopping up otherwise good science segment into shorter than 2 hours chunks (2 hours is just my preference).

Do it sparingly, single IFO data is still valuable for CW.

(Added later at 13:30 local time: If the coherence is super bad like Cheryl's "before" plot for PIT, there's not much point in waiting hours to run a2l, but make sure that the alignment itself is not suspect.)