During the power up, there is some variation of the DRMI optical gains. This can lead to loop instability and lockloss. This could be mitigated by:
- perfect TCS
- digitally compensating the optical gain
- or having a loop w/ more margin.
I have made a new low pass filter to give some more margin. Its more stable with no noise penalty, so it is now loaded by default.
Plots:
- open loop gain: (green) old loop w/ 4 dB gain margin and 25 deg phase margin. New loop has more gain since the old LP filter had a DC gain of -2 dB(?).
- comparison of old (CYAN) with new (Purple); ~ 10 deg more phase @100 Hz
- comparison of spectra: old (right) and new (left). Basically none of the SUS DAC RMS is due to the high frequency stuff, so its OK to relax the low pass filter.
Looking through the elogs, the most exciting LSC input matrix element seems to be POP_9_I -> SRCL. It gets the most attention.
Over the past couple of weeks, I've been measuring it every couple days. It moves around with the thermal state, sometimes changing sign.
Today I measured and minimized the coupling from a PRM excitation to SRCL_IN1 at 85 and 250 Hz. The matrix coefficient I found is the same as Gabriele and Sheila found on Oct29, but not what Stefan set it to on Nov5. I guess Stefan had a semi-warm measurement.
Putting the element back to 0.038 from 0.024 reduces the PRCL 2 SRCL coupling by a factor of 10. With the element set to 0.024, at PRM motion actually shows up 2x stronger in SRCL (in terms of meters) than in PRCL. This was measured at thermalized 23 W, so could change at 30 W.
