Since the LSC FF injections went quicker than expected, I used some extra time to take a look at the status of the HAM1 FF. Gabriele noted larger than expected coherence of CHARD P and Y with DARM in his bruco. Up until this morning we had still been running with HAM1 FF tuned on 76W data on April 23. The FF was set for CHARD, PRC2 and INP1 both pitch and yaw loops. Comparing on and off times of the feedforward, it appeared that the HAM1 feedforward was actually injecting noise into the yaw loops in some cases, or having no effect. I immediately turned off all feedforward for the yaw loops (this does line up with memory that we had no need for feedforward in the yaw loops at 60W). It also appeared that the 76W HAM1 FF was worsening the noise in the pitch loops. I also reverted the pitch feedforward to the filters tuned on April 6th, during a thermalized 60W lock.

The plot attached shows all 6 loops (p/y of CHARD, INP1, and PRC2). The gold traces show the spectra with feedforward OFF. The blue traces show the spectra with the 76W feedforward ON. The red traces show the spectra after I reverted to the April 6 feedforward for the pitch loops, and turned the yaw loops feedforward off. This plot shows that yaw feedforward is unnecessary, and made the noise worse. The plot also shows that the April 6 pitch feedforward is decent, and does improve the noise.

It is difficult to see any change in DARM right now. I attribute this to higher LSC coherence. I think with the LSC feedforward retuned, we might actually see the improvement.


Quick edit add: these have been SDFed in both safe and observe.