Luminosity Variations Along Bunch Trains in PEP-II

In the spring of 2005 after a long shut-down, the luminosity of the B-Factory PEP-II decreased along the bunch trains by about 25-30%. There were many reasons studied which could have caused this performance degradation, like a bigger phase transient due to an additional RF station in the Low-Energy-Ring (LER), bad initial vacuum, electron cloud, chromaticity, steering, dispersion in cavities, beam optics, etc. The initial specific luminosity of 4.2 sloped down to 3.2 and even 2.8 for a long train (typical: 130 of 144), later in the run with higher currents and shorter trains (65 of 72) the numbers were more like 3.2 down to 2.6. Finally after steering the interaction region for an unrelated reason (overheated BPM buttons) and the consequential lower luminosity for two weeks, the luminosity slope problem was mysteriously gone. Several parameters got changed and there is still some discussion about which one finally fixed the problem. Among others, likely candidates are: the LER betatron function in x at the interaction point got reduced, making the LER x stronger, dispersion reduction in the cavities, and finding and fixing a partially shorted magnet