Tomography at Injection in the PSB

The PSB was conceived as an intensity booster for fixed-target physics. Consequently, no attempt was ever made to synchronize the turns injected into each ring with the rf in that ring because as many as a dozen such turns were expected to be superposed longitudinally [1]. When only a small, non-integer number of turns is injected, this asynchronism results in a particle distribution whose initial phase fluctuates wildly from shot to shot with respect to the rf bucket into which it is subsequently captured. This has long been suspected to be an ingredient in the intensity non-reproducibility observed for low-intensity beams, such as pilot beams for the LHC [2]. An MD cycle has been built to pursue this suspicion (in one ring, at least) by introducing, in a fully ppm fashion, distributor timings that are first resynchronized to the rf train of Ring3 and that then count 40 MHz clock ticks to reduce any jitter to just 25 ns (cf., the bucket duration at Booster injection is 1.67 μs). Leaving aside the original motivation, this MD note summarizes a couple of remarkable by-products of the work. Most notably it was realized that the particular nature of capture on the fly at the Booster causes the longitudinal phase space motion early in the cycle to be radically different from that in a normal accelerating bucket and, further, that this could readily be taken into account by the Tomoscope to produce detailed images of the longitudinal distribution delivered by Linac2 and its debuncher.