We've been seeing a ~28Hz line come and go in DARM. I ran a bruco on a time yesterday when the line was high, results here, and saw the line was coherent with PRCL. I checked the PRCL UGF in NLN and it was very low, ~26Hz with a couple of unity gain crossings (first attachment).

I increased the gain from 6.4 to 11.6, putting the UGF closer to 40Hz (second attachment). I put this gain in the guardian - it's set in LOWNOISE_LENGTH_CONTROL.

It might be that we are losing gain as our plant changes over the course of a lock. This measurement was taken at the ~20 hour mark. If PRCL looks like it is gain peaking at start of the next lock, we might want to revisit it.

Edit: this did not remove the 28Hz line from DARM :'(