Gabriele, Sheila
Summary:
We have higher noise in SRCL than we did in O2, which is part of the reason why our SRCL noise contribution to the DARM noise is large. We changed the input matrix element which should cancel the PRCL signal in SRCL using POP9I, and now the SRCL control signal is similar to what it was like during O2 at low frequencies.
Details:
POP9I to SRCL input matrix:
We made an injection at 132 Hz into PRCL with notches engaged in PRCL and MICH, and measured the resonse of POP9I and POP45 I. We found that for the POP45I to SRCL input matrix element of 1.856 we should use a POP9I to SRCL element of 0.038 to cancel the PRCL contribution to SRCL, while the element in place was -0.01
The first attached screenshot shows the improvement in the SRCL noise from 10 to 50 Hz when we retuned this input element. The second attached screenshot shows a similar comparison for the control signal. You can see that at high frequencies the SRCL control signal is still noisier than it was during O2, but that the noise from 40 Hz and below is now comparable. The third attached screenshot shows measurements of the SRCL OLG that Gabriele took before and after this change in the input matrix, there is some impact of the cross coupling from 20-50 Hz.
Whitening gain:
A few weeks ago Keita found that one of our locklosses were due to ADC saturations in POP45, 44420, and detchar followed up 44521. We reduced the POP45 whitening gain to avoid those locklosses. Then Patrick Godwin and I found that the MICH residual motion was large because of a boost missing in MICH 44693, and Hang put the boost back in 44703 I think that Hang may have also added a boost to SRCL.
I looked at the RMS of the ADC inputs for POP45 I was 71 ADC counts while Q was 41 counts. I have increased the whitening gain by 21dB and compensated with a -21dB filter in the digital path. We have locked once since then but quickly lost lock because of the earthquake in the pacific near El Salvador.
Note: I didn't reset the dark offsets for POP45, which I should have before we relocked. Edi: dark offsets updated this morning. Increasing the whitening gain for POP45 didn't change the sensing noise much.
