Logbook entries
https://logbook.cern.ch/elogbook-server/GET/showEventInLogbook/4336879 https://logbook.cern.ch/elogbook-server/GET/showEventInLogbook/4336892 https://logbook.cern.ch/elogbook-server/GET/showEventInLogbook/4336901 https://logbook.cern.ch/elogbook-server/GET/showEventInLogbook/4336944
Log
Supercycle SFTPRO1 - ZERO - LHC3 - MD1
We switch off the MD1 (put it on idle current)
We see the orbit moves 8 mm, so we look also at the first turn
The first turn moves by 4.6 mm, equivalent to about 1.3 G. The eddy current model predicts the following trim w.r.t. to an SFTPRO1 with an MD1:
The orbit does not improve much, but the first turn looks like
So we move the beam by 2.2 mm, and the rest should be hysteresis.
Hysteresis compensation
We reset the trims on SFTPRO1 and MD1, and switch to hysteresis only compensation.
The trim applied is upwards -1.6 G, and the first turn moves from 4.6 mm to 6.9 mm, since the field is decreased.
Based on the field measurements, the main dipole field increases by about 1.5-2 G when MD1 is removed, however we find that orbit increases by 4.6 mm, which corresponds to a field decrease of 1.2 G.
Originally we believe this behaviour comes from drift in the B-Train, however we find that the field measurements are without gaps:
This would suggest that at first turn the field changes by +1.5 G on each the measurements after the reference was taken (SFTPRO1 with MD1).
We also tried keeping the MD1 at 26 GeV (ramping up to “injection”) so the B-Train could mark, but tripped the power converters every time.
This inconsistency in field measurement is not yet explained.
Autoregressive field prediction
We find that the predicted field is along the lines of what is measured. Green is before MD1 is removed, and orange and red after. Purple is what was corrected.