Logbook entry
https://logbook.cern.ch/elogbook-server/GET/showEventInLogbook/4344164
EPA Dedicated MD Summary Compared to previous MDs, we slightly decrease the decrease the chromaticity at the SFT injection, and play SFTPRO - MD1 - SFTPRO1 - MD1 - SFTPRO1 - MD3, where MD3 is an MD1 with only dipoles flattened.
We record first turn and and orbit on the SFT after MD3, with and without dipoles. Without cycling the dipoles the first turn is centered, and orbit is flat, thanks to the momentum ramp at the injection plateau C200 - C1460. When keeping the MD3 at idle current, we find that the first turn jumps to 4.6-5.2 mm, and the RF kicks the orbit out to 8 mm. When applying a 3-exponential decay to with parameters with time constants 111 ms, 670 ms, 3.0s, we adjust the first turn by 0.6 G, and move the orbit to just below 3 mm, and find that the orbit is slightly overcompensated with the fast decaying component, but then stays within 1 mm delta. To fix the offset we add 7e-5 T, which centers the first turn, and consequently we fix the orbit (besides the overcompensation). With the injection corrected the beam passes well to extraction, and we proceed with flat top and full cycle compensation. In addition, the feed-forward corrections apply the correct corrections to each SFT based on cycle (dI/dt) history for the eddy current decay.
According to the B-train, there is no difference in injection field between SFT with and without MD1 precycle.
We excite the conditions by switching to an LHC-type SSC with an LHCPILOT (then INDIV). The first turn and orbit differences are similar to the previous SSC, and we correct the eddy current decay by 0.6 G, and then add a constant offset (which is visible on the B-Train, albeit with a large drift gap when the marker is applied) to correct the orbit. We don’t find the hysteresis any different on flat bottom between SFT-only and SFT-LHC SSC. When applying the same offset as before we center the beam.
Due to bugs in the trimming, we are unable to trim across beam processes. We make a trim on the flat top (with LHC cycle in SSC) to compensate the flat top with or without MD1. We find that the compensation overcompensates by 0.2 G, well within the accuracy limitations of the model. This is the first time we manage to compensate flat top without MD1!. The spill is already screwed up, so we don’t see any improvement.
Important findings The MD1 precycle on the dipoles seems to to have little effect on hysteresis across different supercycles, although we see a 5e-5 T - 7e-5 T shift, which is not always visible on the B-Train, so we conclude that we are not able to consistently compensate the main dipole field at injection. We are however able to model and correct the eddy current decay to a sufficient degree of accuracy. Finally, it seems the MD1 precycle is primarily important for the sextupolar contribution in the dipole remanent field, which causes the beam to go unstable due to wrong chromaticity, and it is crucial to characterize the b3 component of the dipole at injection (with operational cycle sequences) in the lab.
Further MDs will focus on quadrupole compensation QF/QD.