CERN Accelerating science

Magnetic Field Prediction for the LHC Control System

31st May 2005

Present: Luca Bottura, Mike Lamont, Laurent Deniau, Delphine Jacquet, Nicholas Sammut

The aim of this meeting was to clarify how the Magnetic Field Prediction system (AT-MTM) will be applied to the LHC control system (AB-OP).

The meeting started with Laurent’s introduction of what is being prepared by AT-MTM for the MAD program that is used for particle tracking. A sketch of the query, information flow and output of the system can be seen below:

This was followed by a discussion concerning the form in which the transfer function should be transferred to the control system. Two possibilities were discussed:
1) The integrated transfer function over one sector is transferred in Tm/kA,
2) The AB-OP magnetic length is transferred to AT-MTM to compute the transfer function in T/kA. This is turn is transferred to AB-OP.

Luca stated that it would be better to opt for option 1) so that AB-OP apply their own magnetic length to the transfer function. The prediction system database should however also have the measured magnetic length of the magnets in the sector available.

It was decided that the static reproducible contributions to the field errors be supplied to the controls in a table format that is embedded inside the control database system. Mike will give Laurent the code that accesses this table so as to establish its format.

Mike explained how the trim system is used to adjust the settings of the power supplies on the fly.

Since the dynamic error contributions (decay and snapback) vary as a function of time, the control system must handle error variations as a function of current as well as a function of time. This is particularly true at injection and at the beginning of the ramp. Mike verified that this dual dependence can be handled by the control system.

Due to the current history effect on the decay amplitude, the dynamic effects cannot be uploaded as a simple table but must be programmed as a function inside the feed foreword prediction system. The exact time in which this calculation will be done during the LHC cycle still has to be determined.

A proposed date for the next meeting is 21st June 2005 at 16:00 in room 30-02-05.

Matters arising for the next meeting:
1) Laurent’s feedback on the table format used for the static prediction
2) Further discussion on how to implement the dynamic corrections during the LHC cycle