Controls Overview
October 27, 2005
Outline
ES&H
Goals
Commissioning Schedule through the LTU
FAC Responses
Status update
Next 6 months
Conclusions
October 27&28, 2005
FAC - Controls Overview
1
LCLS Control Group
Lcls-controls@slac.stanford.edu
LCLS Control System ES&H
PPS for the Laser and Injector is designed.
Internal design review of wiring, racks, and
AC distribution will be done for existing and
new designs to determine safety and noise
issues.
All personnel has up to date safety training
for required work.
All equipment being purchased meets SLAC
electrical safety standards.
October 27&28, 2005
FAC - Controls Overview
2
LCLS Control Group
Lcls-controls@slac.stanford.edu
LCLS Control System Goals
Provide a fully integrated control system to support the
construction, test, installation, integration, operation and
automation of the LCLS Accelerator
Standardize on all devices and components across all
subsystems.
Identify all data either by pulse id, beam pulse related time
stamp, or 500 msec rough time stamp.
Full integration with the SLC – timing, use of LCLS data in
SLC high level applications, and use of SLC data in LCL
Provide an upgrade path for the SLC
October 27&28, 2005
FAC - Controls Overview
3
LCLS Control Group
Lcls-controls@slac.stanford.edu
Integration with the SLC Control System
SLC
Alpha
All High
Level
Apps
EPICS
EPICS
EPICS
W/S
EPICSWS
W/S
EPICS
W/S
Distributed
W/S
Distributed
Distributed
Distributed
Applications
Distributed
Applications
High Level
Applications
Applications
Applications
Xterm
Xterm
Xterm
Xterm
SLC Net (Data Communication)
KISSNet (fast closed loop control data)
PNet (Pulse ID / User ID)
MPG
Ethernet (EPICS Protocol)
micro
Camac
I/O
October 27&28, 2005
FAC - Controls Overview
Micro
P
E emulator
I/OC
N
V
E
G (SLC-aware)
T
RF reference clock
4
LCLS Control Group
Lcls-controls@slac.stanford.edu
System Design
SLC
Alpha
Apps
EPICS
EPICS
EPICS
W/S
EPICSWS
W/S
EPICS
W/S
Distributed
W/S
Distributed
Distributed
Distributed
Applications
Distributed
Applications
High Level
Applications
Applications
Applications
Xterm
Xterm
Xterm
Xterm
Fast Feedback over Ethernet?
SLC-Net over Ethernet
Channel Access
IOC
C
P
U
E E
HPRF
V V LLRF
I/O
G R
Boards
IOC
C
P
U
E
V
R
Diag
C
P
U
E Pwr
V Supply
R Ctrl
Single
Bunch
Beam
Dumper
Machine Protection
Beam Code + EPICS Time + EPICS Events
October 27&28, 2005
FAC - Controls Overview
C
P Vacuum
U Ctrl
16 triggers
16 triggers
Drive
Laser
Off
IOC
5
LCLS Control Group
Lcls-controls@slac.stanford.edu
LCLS Installation Schedule
Laser Commissioning May 31, 2006
Injector Commissioning Nov 2, 2006
BC1 / L1 Commissioning Nov 2, 2006
Inj Spectro. Commissioning Nov 2, 2006
L2 / BC2 / L3 Commissioning Oct 31, 2007
LTU Commissioning Jan 31, 2008
E Dump Commissioning Jan 31, 2008
Laser Heater to be commissioned with Undulator
October 27&28, 2005
FAC - Controls Overview
6
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (1)
Global standards are required
Most of our standards are in place. Power supplies between the
Undulator and the LINAC are not standardized.
You desperately need a central database for all project data
Andrea Chan has joined the LCLS and we are using their group for
further support. There were two positions open, one in management
and one in controls. The one is controls was removed. We share the
one person with some additional support.
We have selected the PEP II database. The SNS database had too
much in it that was SNS specific.
Sergei Chevtsov in the LCLS Controls Group has been looking at
the relational database needs for the control system, in particular
the RDB requirements for online modeling for the accelerator..
BPM location and anchoring must be decided
Resolved.
October 27&28, 2005
FAC - Controls Overview
7
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (2)
SLC-aware IOC good progress prototyping, still lack
functional requirements and top level design decisions on
Applications and which should be EPICs vs SLC.
The SLC aware IOC is complete. The SLC applications are provided
as a backup to the EPICS based applications. We have a team of
three people starting on the EPICS based high level applications.
The high level applications needed for each phase of
commissioning have been identified and a plan is being executed.
http://www.slac.stanford.edu/grp/lcls/controls/global/facilities/slc_ioc/
index.html
X-ray beam line controls requirements must be defined
Steve Lewis has identified all of the requirements, standardized all
I/O, vacuum control, and motion control and is in the process or
reorganizing the WBS to reflect these changes.
October 27&28, 2005
FAC - Controls Overview
8
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (3)
Undulator BPM performance requirements should be articulated including max
working range
Done See Patrick’s talk
next week Patrick, Till, Bob and Lill meet to discuss the engineering specifications
The engineering process seems to be proceeding without explicit requirements
and reviews. An example of this is the PNET module
Work groups for all subsystems are in place
These groups jointly develop requirements and design
High level ESDs should be complete this month.
All detailed ESDs are near completion, but have not been reviewed or published.
Where is global controls on the org chart?
All design and integration effort is now in 1.1.3.5
Subsystem equipment and implementation is in x.x.2
Signature authority on x.x.2 is global controls and x.x system managers
October 27&28, 2005
FAC - Controls Overview
9
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (4)
Can we get x-band klystrons and waveguides from the NLCTA?
We can get klystrons, elbows, converters – but we need to make the
waveguide runs. (In the LINAC scope of work)
Injector laser controls interface needs close communication between
vendor, controls group and laser experts/operators
Dalesio met with the selected vendor to review the design and verify that it
more than meets the requirement to time stamp data within 500 msec.
The Thales control system will be used for control of the drive laser for the
first year to maintain the warranty. The integration with EPICS is scheduled
to be done after this first year.
Thales will make all source code available as well as the communication
protocol to their embedded controller.
All 120 Hz control and shut down will be done external to the Thales drive
laser.
Sheng Peng is working with the LCLS laser group to control all devices
outside of the Thales laser.
October 27&28, 2005
FAC - Controls Overview
10
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (5)
Follow new code requirements for low-smoke
fireproof cables
Done
It is unclear how to design feedback to control Xband phase- explore "dithering".
Addressed in Patrick’s talk – compensating drifts in the X
band system with beam based feedback in the S Band.
October 27&28, 2005
FAC - Controls Overview
11
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (6)
The PPS system will be implemented with PLCs.
This is the first time this will be done at SLAC.
We are glad you are helping bring SLAC PPS
into the present century. There is a relatively
new standard for safety certified PLCs.
Consider using only PLCs that meet this new
standard.
We are working together with SPEAR3 to develop a
common approach based on the work at J-Lab and
SNS using the SIL3 Cat4 safety certified Pilz PLCs.
The LCS injector PPS plans are scheduled for
presentation to the SLAC safety committee this
November.
October 27&28, 2005
FAC - Controls Overview
12
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (7)
The magnet power supplies have only one current read-back. This is used both in the
feedback loop of the power supply and in the read-out by the control system. While
many accelerators operate like this, it has the problem that if the current readout
(shunt or transductor) breaks such that it gives the wrong reading by a few percent,
there will be no indication in the control system, but the beam will misbehave. While
such failures may not happen too often, they can take many days (suffering with poor
beam) to track down. The use of redundant current read-backs would solve this
problem. One is used in the supply feedback circuit while the second is used for the
control system read-back. If either fails, the control system will show that the current is
out of tolerance. Please consider using redundant read-backs.
The orders for the powers supplies for the injector have now been placed. The small, corrector
supplies use the COTS Bira™ MCOR supplies with only one current read back. The
intermediate supplies for the injector solenoids and bend magnets will use a SLAC design
Ethernet controller, used at SPEAR3, which does have the capability for redundant read back.
We are also further investigating the power supply controllers for the rest of the project with
redundant read back and better integration into the control system in general.
October 27&28, 2005
FAC - Controls Overview
13
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (8)
The MPS system needs to be defined soon. It is unknown if it is a big job or
a little job. Until it is defined well enough to know, it will remain a concern.
Reasons to be worried that it might be a big job is that its actions are mode
dependent, it must stop the beam before the next pulse, and it may need to
interface with the master pattern generator of the SLC control system. It is
complicated enough that it may need to be programmable, yet as an MPS
system, great care must be taken to make sure it keeps things safe. A
similar project on the SLC took several man years to implement. At least the
functional requirements and top level design should be done and presented
at the next FAC so it will be know how big a job it is.
The functional requirements for the MPS have now been defined including both
the range of fault sensing devices and the beam shut off systems to be used,
along with the necessary response time.
It will interface to the master pattern generator for LCLS. One of the mitigation
actions is to rate limit the beam.
The earliest need date for MPS is 10/31/2007 with the commissioning of BC2/
L2/BC3.
October 27&28, 2005
FAC - Controls Overview
14
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (8)
While SLAC is using a CVS system at SLAC as a
controlled software repository, the other labs are doing
their own things. There needs to be a common repository
that all the developers use.
It is a good idea and we need a system manager to set such a
thing up. Until then, we plan to receive updates from the other
labs (i.e., in tar format), import into the SLAC CVS repository,
build, and release to production as needed. We also provide
SLAC AFS accounts to external engineers and allow those
accounts to update the SLAC repository.
As a result of budget concerns, we will not fill a full time system
manager. We will use the management services around the
laboratory.
October 27&28, 2005
FAC - Controls Overview
15
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (9)
The present plan is to move the Undulator in and
out of the beamline by using two stepping motors
(one at each end). This may cause problems of
keeping them synchronized to avoid the jamming
that would occur if the Undulator gets crooked.
Consider using a single motor with gears and
chains connecting the two ends or air cylinders to
move them instead of the two-motor solution.
We will test a fail-safe motor controller software system
in the next year. This would sense the position of both
motors to avoid jamming. We believe this is preferable
to the additional mechanical complexity of gears and
chains linking each end of the Undulator.
October 27&28, 2005
FAC - Controls Overview
16
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (10)
The very tight temperature tolerances in the Undulator tunnel (+/- 0.2 C) have severe
implications on controls. There are plans to put electronics in the ceiling air return duct
where it will be difficult to maintain and concerns that the stepping motors will give off
more heat than allowed. The air conditioning system necessary to maintain that
temperature stability is also very expensive. The accelerator physicists should have a
hard look to see if there is a way to increase this tolerance.
The temperature stability tolerances for the Undulator tunnel have been re-examined both with
respect to their influences on the Undulator magnetic field as well as to the positional stability
of the quadrapoles and BPMs.
GENESIS simulations of the effects of errors of the average K values for each Undulator
segment, both random and systematic, show that temperature errors from a uniform
distribution with a width of ±1 degree F (±0.56 degrees C) are consistent with a total overall
error budget for a 25% reduction in FEL power (but not taking credit for simple Undulator xposition adjustments to compensate temperature variations).
In parallel, a thermal expansion study was carried out at the APS with the result that for
temperature changes of ±0.5 degree C the critical components will stay with in the position
tolerances (±5 microns over 24 hours).
Based on these analyses, which will be presented during the next FAC meeting, the
temperature tolerances for the Undulator tunnel have been relaxed. The requirement
specification says now: “The absolute temperature along the Undulator will stay within a range
of 20±0.6 °C at all times.”
October 27&28, 2005
FAC - Controls Overview
17
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (11)
The project is making good use of commercial hardware
solutions. This will help keep the engineering costs
down.
We are continuing with this policy.
We are now exploring COTS small processors that can run
RTEMS and EPICS to support small, standalone controls
applications without the use of a full VME crate.
These could potential save significant hardware and wiring costs.
However, in order to expedite the injector installation we have
opted to use a small number of SLAC built power supply
controllers for the medium sized magnets while we further explore
other commercial solutions. There are 5 of these at $2K each that
we can replace if we implement an intelligent controller on
schedule.
October 27&28, 2005
FAC - Controls Overview
18
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (12)
The power distribution system is being planned
with extra capacity. This will allow upgrades to
the Linac to be conveniently done by
implementing the new hardware using the new
power distribution system. The old hardware can
then just be turned off when it comes time to
switch over.
Electrical power load tests have been scheduled for
the Linac klystron gallery and after evaluation the
Conventional Facilities group will begin planning any
upgrades necessary to support the new equipment.
October 27&28, 2005
FAC - Controls Overview
19
LCLS Control Group
Lcls-controls@slac.stanford.edu
FAC Responses (13)
A fiber solution for distributing the RF phase reference
was presented as a future upgrade to the planned
distribution via cable. This looks very promising. We
encourage actively pursuing this R&D with the hope that
the fiber system could be used from day one.
We are keenly following the progress of two groups at present,
who are pursuing fiber distribution of RF. The MIT/DESY
collaboration uses interferometry at the RF wavelength to stabilize
the fiber, while LBL uses a method developed for astronomy
based on interferometry at optical wavelengths.
We have not yet built a test stand of our own for LCLS R&D.
October 27&28, 2005
FAC - Controls Overview
20
LCLS Control Group
Lcls-controls@slac.stanford.edu
Update April 2005 – October 2005 (1)
Finish hiring – team in place
Laser Control – Sheng Peng
Power Supplies – Kristi Luchini, Paul Bellermo, Anthony DeLira, Dave MacNair
Vacuum – Stephen Schuh, Steve Lewis, Tom Porter
Diagnostics – Patrick Krejcik, Till Strauman, Doug Murray, Sheng Peng, Michael Cecere, Steve Smith, Tom
Porter, Ron Akre, (Bob Lill APS).
Personnel Protection – Patrick Bong and group
RF Control – Dayle Koturri, Arturo Alarcon, Ron Akre
Wiring, Racks, Conventional Facilities – Mario Ortega and company.
High Level Applications – Diane Fairley, Debbie Rogind, Sergei Chevstov, and Stephanie Allison
Fast Feedback – Stephanie Allison working with diagnostics and controls
Network and workstations – Doug Murray and lab network support
Undulator Control – Josh Stein and company
XTOD Control – Steve Lewis and company
Physics requirements and diagnostic design support – Patrick Krejcik
RDB for project documentation and beamline control – Andrea Chan and company.
Machine Protection System – Stephen Norum
Group Leader interviews have begun.
October 27&28, 2005
FAC - Controls Overview
21
LCLS Control Group
Lcls-controls@slac.stanford.edu
Update April 2005 – October 2005 (2)
SLC-Aware IOC is complete.
Prototypes in test:
PNet – Tests ongoing. Looks good.
Interface to facility controls ALC system tested
Timing – EVG/EVR100s operational – need the 200 series to test the passing of the PNET data onto the EVG.
Prototype operational for the Newport motors for laser control and wire scanner control.
Power Supply – using the SLC power supplies. VME I/O ordered to set up the test stand. Ethernet controller for
more stable power supplies is running at SSRL. We are looking for an embedded controller to replace this one as
there is some concern over the reliability of the UDP communication.
Video – cameras is operational but needs the EPICS driver
LLRF test is using an analog front-end is operational. A new embedded controller is prototyped and nearly ready
to test.
Parts for the vacuum control I/O have been ordered.
A BPM prototype from Steve Smith’s group is nearly ready for test.
Parts for the wire scanner test are in hand, but a change back to previous wire scanner hardware may require a
redesign.
The RTEMS BSPs for the 5500 and 6100 cpus are operational.
SLC aware IOC test stand is operational using multiple platforms. Integration with timing is required.
October 27&28, 2005
FAC - Controls Overview
22
LCLS Control Group
Lcls-controls@slac.stanford.edu
Update April 2005 – October 2005 (3)
Prototypes needed
120 Hz fast feedback
Machine Protection
Micro IOC for use in MPS, Power Supply Control, Vacuum Control,
LLRF Control, and video control.
WBS LINAC being updated to reflect staged
installation
No contingency available in the budget due to
extreme overruns in the conventional facilities.
Work is being done to handle this and provide
possible contingency. Until prototypes are
operational our budget cannot be refined further.
October 27&28, 2005
FAC - Controls Overview
23
LCLS Control Group
Lcls-controls@slac.stanford.edu
Update April 2005 – October 2005 (4)
Laser Installation Requirements
Thales laser controlled in Labview for first year
Newport motor controls for laser table prototype is operational and all motors identified
Video diagnostic prototype is operational – needs integration with EPICS
Laser safety system is designed.
Injector / L1 / BC1 Installation Requirements
PNET interface is operational
SLC-aware IOC is operational
Power supplies ordered
Vacuum hardware is identified and first articles ordered for the test stand
LLRF design with a prototype ready for test
Timing System prototype needs evg200/evr200 which are ordered
BPM prototype from Steve Smith’s group is ready for test
Wire scanner hardware may not interface to new wire scanner design – needs some rethinking
Questions remain about reading the wire directly – or using pickups further downstream
Personnel protection is designed and SIL3 compliant hardware ordered for prototype
October 27&28, 2005
FAC - Controls Overview
24
LCLS Control Group
Lcls-controls@slac.stanford.edu
Next 6 months
Identify micro computers to use for our
applications in the next three months
Complete subsystems needed for Laser and
Injector
Complete detailed designs
Complete prototypes with EPICS integration
Complete reviews
Revamp costs
October 27&28, 2005
FAC - Controls Overview
25
LCLS Control Group
Lcls-controls@slac.stanford.edu
CONTROLS Breakout
Breakout Session 3: Controls
Location: Redwood B
1:30 Controls Overview B. Dalesio
2:00 Integration with SLC S. Allison
2:30 RDB A. Chan
2:45 High Level Apps D. Fairley
3:00 Controls plans for the 2006 installation B. Dalesio
3:30 Break
3:45 Laser Safety and Personnel Protection Systems P. Bong
4:15 Laser Control S. Peng
4:30 BPM Control T. Strauman
4:45 Undulator Controls J. Stein
5:00 X-Ray Transport/Optics/Diagnostics Controls S. Lewis
October 27&28, 2005
FAC - Controls Overview
26
LCLS Control Group
Lcls-controls@slac.stanford.edu
Conclusions
An excellent team is in place.
A group leader is to be hired in the near future.
The WBS is being changed to reflect the staged
installation of the LINAC.
The timing hardware must go into test soon to give
our SLC integration a conservative amount of test
time.
MPS and fast feedback have teams working on
them and are not needed in the next year.
Formal design reviews are needed on all
subsystems to be used in the next year.
October 27&28, 2005
FAC - Controls Overview
27
LCLS Control Group
Lcls-controls@slac.stanford.edu