CERN Accelerating science

LHC-CP sub project on RT controls for the LHC

Meeting no.7 : Friday 27-04-2001 - reported by Mike Lamont/Thijs Wijnands

Present : M. Jonker, A. Butterworth, P. Ribeiro, M. Lamont, O. Berrig, Q. King, T. Wijnands

Agenda :

Prototyping in the SPS

Minutes :

No comments on the minutes of the previous meeting

Training. Contact has been made with Mick Storr of the training division. Essentially it's fine with them, they will give us all support required. However, it is up to us to find a suitable candidate to give the course.

Enrique Blanco who has recently give a talk on cryogenics closed loop control is a potential contact as is Bala Srinivasan from EPFL. EB: "We currently have a collaboration with Valladolid University for control systems, but probably we will establish another one with them also. The location is very convenient to follow up different projects. In fact, we, in the LHC/ACR group, were already working with them in different projects applying some basic advanced control techniques."

Prototyping in the SPS

Pedro gave an overview of the history and motivation for the installation of ATM in the SPS.

At present the only thing that it has been used for is the SPS Q-loop. It is at present switched off.

SPS would provide a nice prototyping environment. Could attempt 10 Hz acquisition, check response of SPS correctors, model plant dynamics and compare with reality. Should be feasible for BI and would give them a chance to test: transmit, jitter, calculation and response on the real SPS rather than in the lab.

Use of ATM for data transfer

The overhead for the equipment groups is zero (PR). JJG questioned this and asked, regarding the orbit "will the application run in exactly the same way?". The changes will be transparent for the clients.

Why ATM v. FDI/ethernet?

(SCI - dead duck). difficult DMA to get MOPOS, For 10 Hz acquisition do we really need to bother with ATM? Might be useful to use public network, if it bottlenecks we can always say "see, told you so..."

MOPOS

MOPOS at present does a turn-by-turn acquisition, these data are stored locally and read out at the end of the cycle.

It would not be recommended to change the operational version of MOPOS. Could modify MOPOS and launch test version of the software during MD (with coasting beam for example). Plus a test bench is available.

The ROCS system will accept RT corrections and AD is apparently waiting for RT input.

Discussion:

• LHC electronics is present on 1 pickup in BA4, and we could start by playing with this... start from scratch...
• Was suggested to perform simulation instead of trying to do it for real
• Close to what Olav Berrig is doing with ROSY - check for overlap
• Acquisition server would run in PCR: VME, RT OP system on a powerPC.
• GPS required.
• Some discussion on timing: millisecond & SC number from TG8. This in coast as well.
• Classical local orbit correction algorithms are not very fast. JW has developed a fast local orbit correction algorithm and is prototyping next generation orbit control with JAVA.
• Local fault tolerance to bad pickups etc. will be important.
• (IR3 - cleaning collimators will have 12 pickups and 4 correctors - i.e. enormous redundancy

QK commented on the timing of the project effort & whether this was worth while... however it was felt that things are going to be very busy in 2 or 3 years time and if we don't gain some experience now the opportunity will be missed and we'll have to do it on the real thing.

Need to check: effort required to use ATM & how much they really want to sell. Work required to develop and install new version of MOPOS. What Olav is doing?