DOE OSTI2013
The magnet alignment of the 90 th and final multipole girder assembly was completed January 17 th , 2013. Our procedure of “rough” pre-alignment by laser trackers and final alignment by vibrating wire & “Differential Variable Reluctance Transducers” (DVRT) has been quite successful, achieving magnet to magnet alignment well under the 30 micron specification. The pre-alignment accuracy is verified by the initial scan from the vibrating wire. The alignment residuals can be calculated from the offsets of a best fit line. The results are impressive, especially considering the environment of the Building 902 Annex (which has a large bay door that would open during pre-alignment, without fail), the sloppiness of the magnets’ fiducials, and especially the fact that the girders were pre-aligned to their mechanical centers, but the vibrating wire reports the magnetic centers which may be considerably different. This result hints at the potential feasibility of alignment with laser trackers for replacement magnets of NSLS-II and maybe even all magnets of future accelerators. But, it is difficult to predict the achievable accuracy until the main three error sources (a.k.a. Survey group excuses) are eliminated: measuring environment, bad magnet fiducials, and mechanical/magnetic center differences. The test performed on a large aperture quadrupole on girder assembly C19G4 was designed to answer the following questions with the aforementioned error sources either eliminated or simulated: 1) How accurately can a NSLS-II multipole be realigned after repair or other disturbance given the existing fiducials? 2) How accurately can a replacement NSLS-II multipole be installed in situ if its magnetic center is known relative to good fiducials? 3) If the current NSLS-II spares are never magnetically fiducialized, how well can they be installed using only mechanical data? 4) Could the magnets of a future accelerator (with similar alignment specifications as NSLS-II) be aligned using commercially available metrology instruments? This paper will answer these questions.