AGK update procedure

To adjust the Abort Gap Keeper to a given train length the following 3 systems must be updated:

  • The MKI and AGK settings,
  • The IQC (kicker module) settings,
  • The SIS AGK settings.

The procedure can be found here. A xls sheet with the parameters is available here. All green background parameters are inputs, all other values are outputs or constants. Values with yellow background are settings to be updated.

A BT expert with the appropriate roles (see procedure) must adapt:

  • The MKI kick length limits (min & max, +- 100 ns around reference value),
  • The MKI limits in the IQC analysis module.

Once this is done, the MKI pulse length can be set using the MKI application with the lhcop role.

The SIS expert with role MCS-LHC-SIS-EXPERT has to update the SIS settings in the non-multiplexed BP. The max number of bunches and the total train length must be updated.

Once all parameters are updated, a filling scheme must be created or modified (for example Single_8b_8_0_8_0_AGK) for the test. The filling scheme contains standard buckets around 1 (and before) as well as 4 buckets around the new AGK limit:

  • The last valid bucket for the new AGK setting NV,
  • The buckets NV+10, NV+20 and NV+30.

See the example below for NV = 32811 (BCMS 3 x 48b with 250 ns SPS batch spacing).

Probe bunches must then be injected to check that:

  • The last valid bucket NV is always OK for injection,
  • The bucket NV+30 is always blocked.
  • The the intermediate buckets maybe either OK or blocked.

For this test the following SIS interlocks will have to be masked (first test that they block !):

  • SIS INJECTION_BUCKET_B1(B2),
  • INJECTION_REQUEST_BUCKET_NO_BUNCHES,
  • SPS_BQM

An example of test may be found in the afternoon shift of 02.05.2018 (OP and BT elogs, http://elogbook.cern.ch/eLogbook/event_viewer.jsp?eventId=2563789).

You are here