LHC Real-time requirements

1
LHC Real-time requirements
What is real-time?
The violation of one or more temporal constraints
causes a loss of functionality. If it takes too
long it dont work. It aint necessarily speed.
What is latency?
The time between asking for something to be done
and it being done
2
Implications
Real time
To allow real time control/access of

• Front-ends
• Field bus
• Gateway
• Network
• Workstation

• Power converters
• Beam Instrumentation
• Reference magnets

• Widely available industrial standards
• ATM, WorldFIP, LynxOS, PowerPC which would
  satisfy LHC requirements (and in fact are already
  in use, planned to be used, or being used in
  prototypes.)

E.g Full ATM for LHC 2M CHF
But note things are moving quickly out there
3
Why bother?
1. Multipole errors
Energy during injection plateau
Orbit
150 units of Q
Large but reproducible, will clearly have to be
measured
4
Control assumptions

• Snap-back and low energy provide the challenge
• Nominal is the worse case...
• The multipole factory takes care of at least
  80 of the effect i.e.
• magnetic measurements
• non-linear models
• on-line reference magnets
• Characteristic frequency of snap-back slow not
  changed by feed-forward from multipole factory

Corrections to horiz. orbit correctors (b1)
tuning quads. (b2) sextupole (b3) and decapole
(b5) spool pieces
70 s
Characteristic frequency is low
5
Beam parameter changes
6
Beam based feedback
Tune orbit we can do but b3 caveat...

• max rate db3/dt ? 0.067 units/second, 3 - 5 units
  Q/second
• ? 1 unit per second given the 80 from the magnet
  factory (NB out of tolerance in 1 second)
• Chromaticity is NOT an easy measurement and is
  not likely to be available for the physics beams
  (emittance growth). Available for pilots ?
  reproducibility (again)

7

• So we have large multipole swings during the
  injection plateau and snapback...
• Beam stability will depend totally on
• feed-forward from the reference magnets
• feedback on the key beam parameters
• tune, orbit, energy, chromaticity
• Have to make a phase jump in our expected
  performance of a control system
• at present it is in the stone-age
• have a look at synchrotron light sources to get
  an idea of what is possible...
• We will need similar to deal with multipoles,
  tight physics constraints the low tolerance to
  beam loss.

8
LEP Orbit
The stone age...
Maximum sampling frequency 0.067 Hz
LHC Orbit
Proposed sampling frequency 10 Hz
9
RT Requirements

• Reference magnets
• feed-forward of corrections to machine
• Feedback-loops
• Global Orbit
• Local Orbit - e.g. orbit stability at collimators
• Tune
• Chromaticity
• Real-time knobs
• typically tune, orbit bumps
• extension of feedback, operator close loop,
  need...
• Display
• Orbit
• Beam loss (not for abort, but display possible
  feedback)
• Luminosity, Beam sizes, Lifetimes
• Timing like actions (possibly)
• e.g post-mortem trigger

10
RT requirements II
Response limited by PC/magnet
11
Work in progress

• Power converters
• WordlFip 2.5 Mb/s PowerPC/LynxOS gateway
• already planning RT capabilities up to 100 Hz
• BPM/BLM
• work in progress but use of RT functionality
  assume
• ATM prototyping
• SPS Q-loop
• LHC orbit correction simulation
• Requirements elucidated
• LHC report 221 Requirements for real time
  correction
• CO/OP forum
• LHC-CP workshop
• here...

12
SL-BI

• Two classes of instrumentation global local
• The global systems e.g. orbit and beam loss will
  provide the main control challenges-
• Orbit 10 Hz RT connection to PCR
• BLMs 10 Hz RT connection to PCR (NOT to decide
  on beam abort)
• Local systems such as tune chromaticity will be
  involved in RT feedback
• Proposed Architecture
• VME
• World FIP for RT
• Ethernet for off-line retrieval Post-mortem

Under discussion
13
Power Converters

• Major system - all magnets - set, ramp, trim etc.
  etc. Offering RT control at 100 Hz (whether you
  need it or not!)
• Digital controllers- CERN design, manufactured
  by industry. Prototypes for String II.
• Gateways- standard SL-CO - Power PC running
  LynxOS
• Fieldbus WorldFIP from CEGELEC
• Timing via an IRIG-B receiver card

MCCS Server
RT LHC Network
Gateway
Gateway
Gateway
Gateway
80 Gateways
Up to 30 Digital Controllers Per WorldFIP fieldbus
1700 Digital Controllers
14
Reference magnets
SM18 Magnet Test Benches WorldFIP fieldbus
Instrumented Magnet
C
FB
Power Converter
I
MCCS Gateway
Multipoles Factory
DB
3-10Hz
Real-time LHC controls network
Real-Time LHC Control System
15
SL-CO-FE - Real Time

• Investigating RT architectures
• Front-ends, Networks
• Prototype
• WorldFIP, FIP running LynxOS, ATM
• SPS as test-bed
• Q-loop
• LHC orbit correction simulation

16
Conclusions

• Demonstrable need for real-time control at the
  LHC
• Essential requirements established.
• Foresee dual system
• real-time for purposes outlined above
• non-deterministic but high bandwidth for
  non-critical monitoring, command flow,
  post-mortem etc. etc.
• Accept RT is necessary, form WG under auspices of
  LHC-CP, ask them to investigate possible
  architectures
• Much work already done in prototyping possible
  system, however technology choice can wait. Final
  architecture will clearly take into account such
  things as
• cost
• reliability
• maintenance

17
Acknowledgements

• Michel Jonker
• Pedro Ribeiro
• Thijs Wijnands
• Quentin King
• Alan Burns
• Oliver Bruning
• Luca Bottura
• Robin Lauckner