Sheila, Kiwamu, Evan, Matt, Lisa, Jenne, Corey
Tonight the locking has been stable enough that we were able to try several low noise steps. We ended up with a range of about 20 Mpc, and were locked for 3.5 hours.
I used the 332 Hz and 1 kHz pcal lines to update the calibration front-end values for the DARM gain and pole. With these new values, we see that the DARM sensitivity is below the O1 sensitivitiy in the few kilohertz region (depending on the behavior of the intensity noise).
I did not change the value for the antispring or the actuation strengths.
J. Kissel, E. Hall Providing some more quantitative details of Evan's calibration change: The reference optical gain (newly installed in FM8, called "ER10gain") is now 8.40e-7 [m] / DARM_ERR [ct], changed from 1.102e-6 (installed in FM4, called "ER9 gain"), a change of 24%. This new optical gain, in physical units is 5.08 ([mA] of DCPD SUM) / ([pm] of DARM displacement), the expanded version of the Evan-loved units he simply calls "[mA/pm]." The reference cavity pole frequency is now 342.0 [Hz], (newly installed in FM7, called "SRCD-2N") where it used to be 328.7 [Hz] (installed in FM3, dreadfully also named "SRCD-2N"). Both the optical gain and cavity pole were determined in rough fashion by taking the magnitude of the (PCALY's TX PD / DARM IN1) transfer function at the calibration line frequencies (331.9 and 1083.7 [Hz]), and solving the following system of equations for K (the optical gain) and fc (the cavity pole frequency): ( |H|_{@331.9 Hz} )^2 = K^2 / (1 + (331.9 / fc)) ( |H|_{@ 1083.7 Hz} )^2 = K^2 / (1 + (1083.7 / fc)) i.e. the naive single-pole approximation. Solving such a system is a one-liner in Mathetmatica (which is what Evan did): NSolve[{3.64^2 == (K^2 / (1 + (331.9 / fc)^2)), 1.52^2 == (K^2 / (1 + (1083.7 / fc)^2))},{K,fc}], where 3.64 and 1.52 where the respective TX PD / DARM IN1 transfer function magnitudes at 2016-09-17 07:15 UTC, or Sep 17 2016 00:15:00 PDT). Evan chose PCALY's TX PD instead of PCALY's RX PD, simply because he wasn't aware that RX PD was the standard. It should be mentioned that in trying to quantify from where he gathered these numbers for this aLOG comment, we grabbed the transfer function all through out the lock stretch of the above summarized night; highlighted in the screencap of the summary page. The transfer function values over the course of the lock (spot-checked every 15 minutes or so) yielded an optical gain and cavity that varied wildly, from optical gains from ~3 to 5 [mA/pm] and cavity pole frequencies from 250 to 370 [Hz]. We do not expect either physical parameter to varying that much. Thus, this method -- albiet delightfully simple and quick, we assume it has a very large uncertainty . Also, though it was assumed that the optical spring frequency from SRC cavity detuning did not change (i.e. Evan did not change it), we're not confident that it has not changed. Regardless, for the time being, the reference optical spring frequency remains 9.81 [Hz]. In the words of Evan "I just wanted to do something quickly for the purposes of, and at the accuracy needed for, noise hunting. I knew the calibration group would do this much better within a week or two, so it wasn't worth the time to go through the whole shebang." I agree. I've saved the template for the attached transfer function in /ligo/home/jeffrey.kissel/Templates/H1DARM_Calibration_mA_per_pm.xml which we can use for future quick-studies like this.