Test results of HD1b, an upgraded 16 tesla Nb3Sn dipole magnet

The Superconducting Magnet Group at Lawrence Berkeley National Laboratory has been developing high-field, brittle-superconductor, accelerator magnet technology, in which the conductor\´s support system can significantly impact conductor performance (as well as magnet training). A recent H-dipole coil test (HD1) achieved a peak bore-field of 16 Tesla, using two, flat-racetrack, double-layer Nb₃Sn coils. However, its 4.5 K training was slow, with an erratic plateau at ∼92% of its un-degraded "short-sample" expectation (∼16.6 T). Quench-origins correlated with regions where low conductor pre-stress had been expected (3-D FEM predictions and variations in 300 K coil-size). The coils were re-assembled with minor coil-support changes and re-tested as "HD1b", with a 185 MPa average pre-stress (30 MPa higher than HD1, with a 15-20 MPa pole-turn margin expected at 17 T). Training started higher (15.1 T), and quickly reached a stable, negligibly higher plateau at 16 T. After a thermal cycle, training started at 15.4 T, but peaked at 15.8 T, on the third attempt, before degrading to a 15.7 T plateau. The temperature dependence of this plateau was explored in a sub-atmospheric LHe bath to 3.0 K. Magnet performance data for both thermal cycles is presented and discussed, along with issues for future high-field accelerator magnet development.