J. Kissel

I've installed IPC broadcasters into the h1isiham4 and h1isiham5 front-end models in order to broadcast the HAM4 & HAM5 ST1 GS13s. Subsequently, I've installed IPC receivers into the h1sussrm, h1sussr2, and h1sussr3 models. With this installation, it completes the signal chain such that each SUS can now compute their own suspension point motion. All that computational infrastructure had already been in place (calibration filters, transformation matrices, etc), but they were just not receiving any input signals. I attach a pdf demonstrating the power of this fully operational battle station for SR3. The pictures attached show trends of displacement SRM, SR2, and SR3 indicating that their positions have been restored to what they were before I started.

Detailed, roughly chronological notes:
---------------------------------------------------
Alignment Offsets before getting started:
        P [urad]    Y [urad]
SRM   -934.049    430.123
SR2   2963.7     2728.0
SR3    430.3      142.6

- made and saved mods to h1isiham4.mdl, h1isiham5.mdl, h1sussrm.mdl, h1sussr2.mdl, h1sussr3.mdl by adding 6 cdsIPCx_PCIE (i.e. corner station dolphin network IPCs) sender parts to each HAM ISI top-level model, with the following "channel names" H1:ISI-HAM[4,5]_2_SUS_GS13_[X,Y,RZ,Z,RX,RY], and created a corresponding 6 cdsIPCx_PCIE receiver parts in the SUS models.
- paused HAM4 and HAM5 SEI manager guardians
- changed ISI guardians to "EXEC", brought ISIs to "READY", turned off master switch (and left HEPIs alone)
- brought SUS guradians to "SAFE"
- captured new safe.snaps for SUS (forgot to do ISI, but J. Warner / H. Radkins are confident that their safe.snaps are up-to-date. I've confirmed that all local to cartesian matrices still obey the conventions defined in T1000388)
- logged onto h1build as controls, used cdsCode alias to get to build directory, /opt/rtcds/lho/h1/release/
- compiled models, e.g. make h1isiham4, making sure to compile the sender models first (i.e. the HAM ISI models)
- installed models, e.g. make install-h1isiham4, again making sure to install sender models first
- logged on to each model's front end, i.e. for the ISIs h1isih45, and for the SUSs, h1sush34 and h1sush56
- restarted models, e.g. starth1isiham4, again restarting sender models (the ISIs) first. Out of habit, I made sure to have the GDS_TP screens open and hit the BURT button as soon as it became available. I don't thin it's necessary, but for big models with tons of EPICs channels, I can click the button faster than the BURT restore can install the right value.
- Brought SUS back to "ALIGNED", confirmed that all SUS had the same alignments restored as mentioned above.
- Brought ISIs back up to "HIGH_ISOLATED,"
- changed ISI back to managed, SEI manager guardian back to exec
- DAQ restart, because I found nonsense (and its needed when one adds new *receiver* parts)
- Went through each model's GPD_TP screen and hit "diag reset" to clear initial, start-up transient, errors from both the timing and IPC- Started trending alignments of SUS to confirm they've gone back to the same place.
- SR3 optical lever damping came up with both stages (M2 and M3) on -- maybe turning this on is hard coded in an ISC guardian somewhere? 
- Turned the SR3 M3 optical lever damping OFF, now both filter modules' SWSTATs report what they were before I got started.
- Now SR3 optical lever reports the P and Y position has been restored to within 1 [urad].
- Found SRM in the same place (according to M3_WIT_P and Y OSEM sensor channels) with larger noise on trend, found with getting constantly pulsed by ISC L from. 
- Changed ISC_DRMI guard state to DOWN (it had been in INIT, and its background RED indicating an error; probably because it had lots communication with one of its subordinates, i.e. SRM). Noise in trend went away. Now restored to with 0.5 [urad] in P and T.
- Not sure about exponential shape of SR2's M3_WIT_P and Y signals, but they appear to be asymptotic toward the previous original position. Spot on AS camera appears centered, and as of this entry, Dan has been able to re-lock the OMC using an IX single-bounce.
- Committed modified top-level isi and sus models and safe.snaps to userapps repo.

This completes work permit Work Permit #4944. Note that this installation is simply completing what had been installed for the rest of the SUS in H1 back in October 2012 - Jan 2013, see e.g. LHO aLOG 4553. The theory behind the matrix transforms (that were already installed) can be found in T1100617.