J. Kissel, J. Driggers

While Darkhan was improving the SNR of the CAL lines (see LHO aLOG 30476), I noticed that the reported value for the darm coupled cavity pole frequency, f_{cc} -- newly calculated in the front-end -- was at ~345 Hz, where for the past two nights it's been measured by both calibration line front-end calculation and sweeps to be ~325 Hz (see LHO aLOGs 30391 and 30431). Indeed, the result was repeatable over several lock stretches tonight. 

I spoke with Jenne, and she mentioned that they've changed the POP A offsets (H1:ASC-POP_A_PIT_OFFSET and H1:ASC-POP_A_YAW_OFFSET, from 0.34 to 0.44 and 0.23 to 0.47, respectively) in order to push the improvement this has had on the power recycling gain even further. It worked -- and it apparently also had side benefits of increasing f_{cc}.

To confirm (so that we finally can just trust the front end calculation and not have to take the sweep every time), I took one more set of calibration sweeps; a PCAL2DARM transfer function to obtain C / (1 + G), and a DARM open loop gain transfer function to obtain IN1/IN2 = G and EXC/IN2 = (1+G). The ratio of these measurements ( C/(1+G) * (1+G)) gets us direct measurement of C alone.

Excellently, the sweep reveals that f_{c} has increased to 343 [Hz], the DARM optical gain went up, and there's even less SRC detuning. Sweet!

Unfortunately, I naturally inquired if we can go further with this POP OFFSET knob, but Jenne says further pushing doesn't seem to have any more effect. Looks like we've spun as much out of this knob as we can.

However, this does give me much greater confidence in the front-end calculations of f_{cc} and relative optical gain change, kappa_{C}. 

Lots more details and changes to come on the model comparison (2016-10-13_H1DARM_SensingFunction_Comparison.png) but I show it just so you can get a good feel for the changes. Note that the reference model (based on September measurements) are compensated for by the kappa_{C} and f_{cc} value take just after each sweep as measured by the calibration line. Also note that the comparisons use time delays from ER9, which -- while they make the sensing function and DARM open loop gain TF measurements match the model -- don't make sense with other timing measurements (LHO aLOG 29259).

Details:
Templates for sweeps line here:
/ligo/svncommon/CalSVN/aligocalibration/trunk/Runs/PreER10/H1/Measurements
DARMOLGTFs/2016-10-13_H1_DARM_OLGTF_4to1200Hz_fasttemplate.xml
PCAL/2016-10-13_H1_PCAL2DARMTF_4to1200Hz_fasttemplate.xml
Note that these templates measure significantly faster because I've reduced the number of points, and reduced the cycle/measurement time per frequency point. Delightfully, I can still get great coherence down to 10 Hz for both templates. However, if we want to stay well-informed of SRC detuning, we should probably re-up the DARM OLG TF measurement/cycle time to regain great coherence.