I retuned the PRCL subtraction from the SRCL element of the LSC input matrix, since the IFO has been at 430 kW circulating for several hours. 

Note that I did not first check the POP phasing (although I ought to have).

I generally followed the prescription from Anamaria's alog 66151.

• PRCL in NomLowNoise at this power could not handle the Not88 filter (ate too much phase, caused significant gain peaking). For PRCL, I created a new notch at 91 Hz.  Since the not88 was fine for MICH and SRCL, I just used those filters for those loops.
• I did quick measurements of MICH, PRCL, SRCL open loop gains, just to make sure none of them changed gain dramatically while I changed the LSC input matrix.
• Injected 91 Hz into H1:SUS-PRM_M3_LOCK_L_EXC.  I ended up using an amplitude of 10 counts to get a nice visible line in SRCL, although 3 counts was probably already enough.
• I changed the LSC input matrix in steps of 0.005 in the _5_1 element of POP 9 I -> SRCL.  It started at 0.049, and I'm leaving it at 0.095.  This is a pretty big change, but it makes the PRC length line disappear in SRCL. 
  • If this causes DRMI locking troubles, then maybe we can instead change this element once we're at high power.  We'll see how locking goes.
• Periodically after a few steps, I would turn off the line, and spot check the open loop gains.  None of them ever changed though.
• There wasn't a big difference between vals of 0.090, 0.095, and 0.100, so I left it at 0.095.

In the attachment, the blue reference might be from after I'd already started tuning the POP9I->SRCL element.  So, PRCL was contaminating SRCL at least this much, but could have been a bit more.  Red is now after the input matrix adjustment.