Studies of Performance and Field Reproducibility of a Precision 40KG Superconducting Dipole Magnet

An 8° bending magnet system for 30 GeV primary protons from the Brookhaven AGS is to be installed in the summer of 1973. This bend requires operating two magnet modules, each 6-ft. long, at an excitation up to 37kG. The cold bore I.D. is 3.375-in. A 20-in. long full-scale cross section model has been extensively tested. Multipole field components due to the magnet circuit, produce field error ~1 × 10-4 parts below 20kG out to the measurement radius at 82% of the radius to the superconductor. The "good field" (~1×10-4 parts) shrinks to a radius ~50% of the radius to the superconductor by 40kG. Diamagnetic effects in the superconductor are completely negligible by 5kG, with the exception of the (r4,5¿) multipole. This contributes 6.5 x 10-4 parts to the field at 82% of radius. End effects are sufficiently small that for 6-ft. long magnets little or no modifications of end contours are required for even very exacting applications. The magnet has operated to 45kG with no evidence of significant training. The magnet has been thermally cycled several times, and the field aberrations allowed by magnet design symmetries reproduce exactly from the first to last measurement runs. At 19. lkG, all allowed multipoles, i. e., 5¿,7¿,9¿,11¿, with 3¿ tuned to zero agree to 1×10-4 parts error over the duration of experimental measurements. However, ferromagnetic effects due to martensitic transformations in the stainless steel inner form on which the coils are wound have grown with repeated thermal and magnetic cycling.