I did the same measurenet as klog 30416. The measurements were done with ALS CARM locked and ALS DARM open. I measured the transfer functions from {TM, IM, MN}_LOCK_L to DARM_OUT, and the results are shown in Figure1 (TM), Figure2 (IM), and Figure3 (MN). From these measurement I found

• Overall gain of TM control is different from the past measurement
• A sharp peak at 1.67 Hz

I noticed that in the previous measurement gain of TM_LOCK_L is 1.255, but now it's 1. I also found that FM4 of TM_LOCK_L is named "gain" and just multiplies 1.255, therefore it seems to be designed to implement the gain factor. I turned on this filter, and now I got almost the same results with the previous measurement as shown in Figure4. I also modified LSC_LOCK guardian not to turn off this filter.

After that, I tried to lock ALS DARM, but still 2.4 Hz oscillaiton grew up and lock was lost. Then, I measured the transfer function from TM_ISCINF_L to DARM_OUT as shown in Figure5 and Figure6. They looked similar, but I found that around 2.17 Hz and 2.45 Hz there are a narrow dip and thay may cause the instability.

We may need to tune the relative gain of the hierarchical control to reduce these dips.

Note

I remember that on 7/25 when we had the power outage at Mozumi office some setting related to ETMX was not recorded properly in SDF and reverted. Perhaps we are still missing other settings.