In 1996 the LBNL D20 Nb 3 Sn dipole magnet reached 12.8 T at 4.4 K and 13.5 T at 1.8 K, and caught the accelerator magnet community by surprise. Not only did it achieve a record field for an accelerator dipole magnet but it also demonstrated a breakthrough in Nb 3 Sn conductor technology. This chapter summarizes the technical aspects of D20 development and test, addresses lessons learned, and includes suggestions relevant to developing high-field magnets for future accelerators.The design approach towards D20 (Dell'Orco et al. 1993a) was "think-outside-thebox," a concept that followed a similar approach previously tried with a unique Nb-Ti magnet (Dell'Orco et al. 1993b). D19, a 50 mm aperture Nb-Ti dipole magnet, was built and tested as a candidate for the US Superconducting Super Collider (SSC) program. With a record field of 7.6 T at 4.4 K and 10 T at 1.9 K, that magnet specifically addressed issues of high fields and Lorentz forces. By separating the traditional concept of using self-supporting collars that combine assembly and pre-stress, D19 used thin non-supporting collars for assembly alone, leaving the functionality of pre-stress to be applied during the final structural assembly. Thin elliptical collars (with a width of 3 mm on the mid-plane) were used for coil assembly and initial alignment. Pre-stress was split between a "rings and collets" loading system and a cool-down shrinkage of aluminum shell over iron. The close proximity of the iron to the coils contributed to the dipole field, and saturation harmonics were handled by the elliptically shaped collars. Aluminum bars between the yokes controlled the final assembly.The successful test of D19, performing without training at 4.4 K, became a model for the D20. The D20 design Scanlan et al. 1995) had a considerable number of additional steps with a complexity that required an integrated design approach. A multi-phased heat treatment, thermal expansion of materials, protection heaters, epoxy impregnation, assembly, and stress pre-loading were mostly still to be understood. This chapter describes the D20 design as it was done in the early 1990s: a time when many presently available tools and programs, especially 3D, were not available or had just started to emerge. Therefore, certain design aspects, routinely used today, were missing and were not implemented in D20.
Magnetic Design OptimizationD20 was designed to extend accelerator magnet technology to high fields by using Nb 3 Sn cables with sufficient current density to deliver fields over 13 T. The graded design used two double layers with different wire and cable sizes, placing the larger strand size in the inner layer. The magnetic design underwent three successive iterations:1. Infinite permeability, sector coils, no wedges; 6 LBNL Cos-theta Nb 3 Sn Dipole Magnet D20 135