Craig, Varun, Elenna Today we went out to the PSL racks and injected into the error point of the input mode cleaner and intensity stabilization servo second loop to check the cross couplings between the loops. To do this we first locked the IMC, and locked the DC ISS second loop at 2W (we boosted the gain the same way as alog 61721). First, I measured the coherence between the IMC error point (H1:IMC-I_OUT_DQ) and the ISS second loop error point (H1:PSL-ISS_SECONDLOOP_ERR2), and found it to be unity at around 3 kHz in the digital world. We later verified this coherence in analog world by measuring the coherence between IMC CMB OUT2 (connected to IN1) and the ISS outer loop ERR2 port on the back of the chassis. This analog coherence is reported in the first plot. It appears to be real. Next we searched for the source this coherence by making injections. We unplugged IN2 from the IMC common mode board, and injected into IN2 while measuring [ISS 2nd loop ERR2 / IMC CMB IN1]. ISS 2nd loop ERR2 / IMC CMB IN1 is the second plot. We got good coherence with this measurement, implying that true frequency noise appears in the ISS second loop. This is not surprising, since the IMC is bound to act as a FM-to-AM converter, and the ISS second loop error signal is measured on IMC transmission. We just don't know the level at which this coupling is happening (yet). We injected a line by itself into the IMC loop, and saw strong bilinear coupling through the IMC to the ISS second loop as well (a second line at twice the injected frequency appearing above the noise of the ISS ERR2). Again, not surprising. We have even less of an idea about the bilinear coupling here because we didn't do an excess power injection, just a swept sine. Finally, we injected into the Analog EXC on the back of the ISS outer loop chassis, and measured [IMC CMB IN1 / ISS 2nd loop ERR2]. For this measurement, we injected up to 1V before we decided that we weren't measuring anything and it wasn't worth trying to inject more. So for now, no noticeable ISS 2nd loop to frequency noise coupling. Of course that could change when CARM enters the loop. For now we have plenty of modeling to do for calibrating the IMC into Hz and ISS into RIN so the second plot is useful.