The spectra taken since the M2 stage AOSEM bracket swap last Friday (15th) show slightly anomolous peaks (not glaringly wrong, but not great).  This morning we attempted to remedy the mis-centering of the M1 UL AOSEM which we thought might be brushing the magnet.  After deeming it ok, Arnaud's spectra revealed it still to be bad.  This time I went in and played with both the UL and LL attempting to center the AOEM with it buried over the magnet blocking all of the light.  This old standby technique was not very helpful though because as you draw the AOSEM back out to 50% open light, the AOSEM walks all over in the bracket nulling the alignment.  The stackup of brackets at this stage are not very nice (putting it nicely).

The picture below shows how all 4 AOSEM brackets are crooked relative to the structure and the suspended mass with the "modified" brackets.  We needed to make new brackets here.  Too late.  (Although I have Tyler making a fresh set for SR3, now that I am aware of the problem.)

I also found that one of the 4 top mass EQ stops was brushing the top mass.  This could have been the culprit the whole time, but nevertheless, the M2 stage brackets don't look right.