In the SUS group** currently, we have 7 chains worth of TOP masses about the aLIGO project hooked up to electronics, with their damping loops closed:
H2 SUS ITMY M0 (Installed in H2 BSC8, production electronics)
H2 SUS ITMY R0 (Installed in H2 BSC8, production electronics)
X1 SUS QUAD03 M0 (Read out on LHO X1 QUAD Test stand, AA/AI filters at Rev 1)
X1 SUS QUAD03 R0 (Read out on LHO X1 QUAD Test stand, AA/AI filters at Rev 1)
H2 SUS FMY M1 (Installed in H2 BSC8, production electronics)
X2 SUS FMY M1 (Read out on LLO X2 QUAD Test stand, AA/AI filters at Rev 1)
X2 SUS PR3 M1 (Read out on LLO X2 Triple Test stand, production electronics)
These are examples of all of the four different OSEM/magnet arrangement types:
QUAD M0 (See E1000617)
QUAD R0 (See E1000617)
BSFM M1 (See E1100108)
HXTS/OMCS M1 (See E1100109)
In trying to understand what the heck is going on with FMY, and noticing some inconsistency in H2 SUS ITMY M0, I've put together a document that uses information from the above linked arrangement posters that explicitly spells out every single sign flip in the signal chains of all of these types of suspensions, see T1200015. In this document (which is for now just pictures, I'll fill in the writing later), it is assumed that the following is true:
(1) Retracting the flag from the OSEM (i.e. current on PD increases, because more light allowed to pass) creates a positive (+) signal.
(2) (a) A positive (+) current through the OSEM coil, paired with a North (N) magnet, forces the magnet away from the coil ("Push")
(b) A negative (-) current through the OSEM coil, paired with a South (S) magnet, forces the magnet away from the coil ("Push")
(Note that these are known truths, not really assumptions.)
With these two conventions, in addition to the pre-defined coordinate system, T1200015 defines a completely consistent sign convention across all suspension types.
---------------------------------------
Using T1200015 self consistent convention, I've then compared it with what is currently in place on the 7 suspensions across the project in order to conclude the current status, given that all of their damping loops can be closed, and are stable. See attachement.
From this assessment, I have concluded the following:
H2 SUS ITMY M0 -- F2 and F3 magnet polarities are flipped (N mistaken for S, or F2 mistaken for F3; currently compensated in the COILOUTF bank)
H2 SUS ITMY R0 -- all clear, as ideal
X1 SUS QUAD03 M0 -- All clear, as ideal (note 2 extra minus signs to account for electronics (AA/AI) gain; currently compensated for in OSEMINF and COILOUTF banks)
X1 SUS QUAD03 R0 -- All clear, as ideal (note 2 extra minus signs to account for electronics (AA/AI) gain; currently compensated for in OSEMINF and COILOUTF banks)
H2 SUS FMY M1 -- L, P, and Y signs are exactly incorrect in EUL2OSEM / OSEM2EUL basis (these matrices have the same sign, so they cancel in the conversion to-and-from the EULER basis -- we're just controlling -L, +T, +V, +R, -P, and -Y instead of the expected +L, +T, +V, +R, +P, +Y), AND overall minus sign on polarity of magnets (N mistaken for S, currently compensated in the COILOUTF bank)
X2 SUS BSFM06 M1 -- L, P, and Y signs are exactly incorrect in EUL2OSEM / OSEM2EUL basis (these matrices have the same sign, so they cancel, we're just controlling -L, -P and -Y instead of the opposite), AND F2 and F3 magnet polarities are flipped (currently compensated in the COILOUTF bank)
X2 SUS PR3 M1 -- Over all minus sign on magnets (currently compensated in the COILOUTF bank)
---------------------------------------
Why haven't we caught this before now?
(1) T1200015 is the first time (at least I've seen) the entire control loop (both analog and digital) signs have been written down, for all suspensions at once.
(2) Up until now, we have not had all of these types of SUS, covering the entire range of of possible arrangements with their damping loops closed.
(3) In our rush to get the answer -- are these SUS acceptable? -- given the plethora of things that might go wrong, and the abundance of data to absorb to determine that information, we have neglected tests that would confirm these signs, e.g.
(a) Looking at the phase of the Top2Top Euler basis transfer functions
(b) Taking Top2Top OSEM basis transfer functions, and looking at their phase
(4) In the confusion of old/incorrect BURT captures/restores, and hard coded scripts, copying-and-pasting between suspensions that are not the same, the range of possible electronics (Test Stand vs. Production), and the "fifty-fifty shot" problem, it has become habit to just flip the sign digitally to "whatever works at making the damping loops stable."
(5) Commissioning and assembly procedures were written before this slow, careful, big-picture assessment was made.
---------------------------------------
What do we do about it?
H2 SUS ITMY -- Leave as is for OAT. Switch magnet polarities when we take out the SUS to replace ITMY test mass.
H2 SUS FMY -- Fix flaws in matrices, but leave magnets as is for OAT. Switch magnet polarities when we take out the SUS to put in glass optic.
X2 SUS BSFM06 -- Fix flaws in matrices, and fix magnet polarity when remaining work on SUS is being done during phase 1 (nowish).
X2 SUS PR3 -- Fix polarity now, while still in phase one.
** This aLOG DOES NOT include a TMTS TOP mass, which has YET ANOTHER OSEM/magnet arrangement.
