Today I spent some time on alignment of PRMI and the SRC, as well as some guardian things that I hope will make these alignments a little easier.  After the alignment DRMI locked easily on 1F (I haven't tried to engage alignment loops).

I did a walk of the PRMI alignment using PRM, PR2 and BS.  The build ups in PRMI are now POP90 norm 73.5 and POP18 48.3, which is 40% better than earlier this week for POP90 and 60% better for POP18.  These builds ups are in the same ball park as the buildups that we had Sept 10th 2015 at 5:26 UTC when POP 90 norm was 84 and pop 18 was 56.  I think that we have had multiple modulation depth changes since Sept 2015.  It would be good to search for a more recent time when PRMI was locked without arms, and also to calibrate these into sideband gains. 

After aligning PRMI, I had a look at the alignment in HAM4/5/6.  

• First, leaving SR3 alone I closed the AS centering loops in single bounce and moved SR2 to see when the power on the AS WFS started to drop.  I found that for the current SR3 alignment (P:521.5, Y 489.3), the beam is roughly in the middle of the aperture (from the Faraday and baffle) with SR2 at P 192 Y 3619.5, in each direction the aperture is about 250 urad wide in SR2 slider values.
• I turned off all 3 whitening stages on AS_C and redid the dark offsets, because we were saturating something which made searching for AS_C confusing.  We should check if there is something wrong with AS_C, as Jenne had suspected earlier. 
• Another confusing feature was that after we changed the sign on the SRM alignment (test) biases Friday there was a ghost beam on the AS QPDs from SRM, so I moved the test offsets again.   
• After centering SR2 in this way I couldn't see any light on AS_C, so I rastered SR2 in single bounce to find the beam and used the AS_C pico to keep it on the diode as I moved SR2 back to the position where I think the beam is roughly centered in the aperture.  
• SRM alignment was very far away, so I put a large raster on SRM to looks for some interference from SRY.
• I locked SRY (needed to lower the trigger compared to settings in the guardian) and manually aligned. 
• SRM T2 coil output is now at -58700 counts, which means that this shift in SRM alignment won't solve our problem with the T2 coil being too close to open.  

After that DRMI locked easily.  The MICH and SRCL gains were fine with UGFs of 6.3Hz and 35Hz, the PRCL gain seemed low, I increased it by 86% for a ugf of 65 Hz, and also added a gate_valve_flag to the guardian that can be used to identify that the arms are not controlled when ETMX is aligned.  

I also spent some time today on things that I think could help make our initial alignment process better (both in the short term while we are missing cameras and once we get back to the full ifo):

• Fixed some bugs in MICH bright states, (made it a replica of MICH_DARK states).  I also lowered the threshold for MICH_BRIGHT in the is_locked(dof) function.
• I added a state which centers the AS WFS for MICH_BRIGHT.  Since we have switched to using ASC_AS_A for locking MICH, we get different LSC gains if we are falling off the diode. Since we don't have the AS camera yet, the easiest way to align the Michelson seems to me to be to lock on the bright fringe and maximize the power.  Minimizing can be confusing because we don't have centering loops and falling off the diode seems a lot like aligning the michelson to a dark fringe.   We could implement a dither loop to maximize the bright MICH and eliminate one of the last manual steps that the operators do in our inital alignment procedure. 
• PRXY will now lower the laser power to 2W if it is set higher than this.  (We use high power for MICH and SRX/SRY, but PRX can't lock at 5W).
• I added a WFS centering state for PRMI.  I used this to look at which WFS might be useful for PRMI while I was aligning by hand, it seems like perhaps we could use AS 45Q for the beam splitter, REFL B 9 for PRM, and REFL A 9 for PR2. 
• I moved the integrators for the DC centering loops from the ASC filter modules into the tip tilts, and added a line to the DRMI guardian down state to clear these in DOWN.  The reason I did this is that it will allow us to add the DC centering loops to the generic alignment offloading guardian state (although I haven't tried to do this yet). 

For now, I think that the best way to align PRMI is to use PRX and manually walk PRM and PR2 to maximize the build ups on POP, then choose the WFS_CENTERING_MICH_BRIGHT state and adjust the beam splitter to maximize the power on AS QPDs.