1. ISIS Accelerator Division
ISIS Upgrade Options
John Thomason
ISIS Accelerator Division

2. 0.24 MW ISIS • Assumes an optimised 2RF system giving 300 µA in the
synchrotron
• 4/5 pulse pairs to TS-1
(192 kW) and 1/5 pulse pairs
to TS-2 (48 kW)
• Must keep beam to TS-2 for
the foreseeable future

3. Upgrade Options
Option
Comments
Beam Power
(MW)
Neutron Yield
1(a)
Add 180 MeV Linac
Technical Issues
~ 0.4
1.7
1(b)
Add 800 MeV RCS
Operational Issues
~ 0.5
2.0
1(c)
Upgrades 1(a) + 1(b)
Technical/Operational Issues
~ 0.9
3.8 2
Add ~ 3 GeV RCS
Recommended 1st Upgrade 1
3.2 3
Add ~ 6 GeV RCS
Technical/Cost Issues
5.6 4
Upgrades 1+2 or 1+3
~ 2 – 6
~ 6.4 – 16.8 5
400 – 800 MeV Linac + 3 GeV RCS
Recommended 2nd Upgrade
2 – 5
6.4 – 16.0 6
1.334 GeV Linac + Accumulator Ring
Good “Green Field” Option
18.8
• Upgrade routes for ISIS are summarised in the table above. All designs are to be developed primarily for an optimised neutron facility, and should provide provision of an appropriate proton beam to the newly built TS-2. The list here is not exhaustive, but presents the main, reasonable routes that would provide a major boost in beam power. Primary considerations are the cost relative to a new facility and the impact on ISIS operations.

4. Linac Upgrade • Replaces oldest and probably most
vulnerable part of ISIS accelerators
• Increased energy (up to 180 MeV)
gives increased intensity in
synchrotron and more beam to
target
• Similar designs are already
available e.g. CERN Linac4, J-PARC
• Would require a new building and
reconfiguration of the TS-1
neutron and muon targets

5. (Steve Jago, Stuart Birch, Adrian McFarland)
180 MeV Injection
(Steve Jago, Stuart Birch, Adrian McFarland)
HEDS Line
- DC Dipoles and
Quadrupoles
Vertical “Sweeper”
- Pulsed Dipole
Injection Septum
- High Current DC Dipole
Injection Dipoles
- 4 Pulsed Dipoles
- Closed Orbit Bump
Stripping Foil
- Beam rigidity
- Aperture
- Increased field/current/voltage
- Stripping efficiency
- Lifetime
Single Turn / Ferrite Core
155 mm Aperture
Peak Field ~ 0.11 T for 70 MeV
Peak Current ~ A
Rise / Fall Time ~ 80 µs
Flat-top ~ 400 µs

6. 70 MeV Magnetic Model Calculated current - 12195 A
Actual current – A
Calculated voltage V
Actual voltage ~ 250 – 300 V

7. 180 MeV Magnetic Model Peak field ~ 0.18 T for 180 MeV
Calculated current A
Calculated voltage V

8. Power Supply I > 21000 A Rise/fall ~ 80 µs Flat-top ~ 400 µs
Possibility of tuneable
gradient on “flat-top”
Current waveform
Voltage waveform

9. 180 - 800 MeV Acceleration (Chris Prior)
Initial calculations (done in 2003) show that 180 MeV injected
beam should be cleanly accelerated to 800 MeV in the ISIS ring at
intensity levels giving 0.5 MW on target
There should already be enough RF acceleration available in the
ISIS ring to accommodate this (calculations were better
optimised for fundamental only rather than dual harmonic
acceleration)
Did not include transverse effects
Requires beam chopping

10. Recommended MW Upgrades
• Based on a ~ 3 GeV RCS fed by bucket-to-bucket transfer from ISIS 800 MeV synchrotron (1MW)
• RCS design also accommodates multi-turn charge exchange injection to facilitate a further upgrade path where the RCS is fed directly from a 800 MeV linac
(2 – 5 MW)

11. RCS Rings (Chris Warsop, Grahame Rees, Dean Adams,
Ben Pine, Bryan Jones, Rob Williamson)

12. Outline Work Plan Main topics to cover Pull work together: comparisons
Longitudinal Calculations/Simulations for candidate FI Rings, then MTI
Build up MTI simulation 2D to 3D, then acceleration
Model Injection Magnet and Region (CST based, as profile monitors)
Transverse Space Charge and Working Point, Correction Schemes
Lattice Optimisation, Analysis, Correction
Instabilities: standard checks, e-p work
Collimation: outline designs, full simulations with activation
Review all main systems
Pull work together: comparisons
Best lattice and ring size; RF and MMPS combinations
Estimated losses and activation
Started
Next
Uncertain Cover

13. Simulation Hardware SCARF: Linux only: Red Hat Enterprise 4.4
GNU, Intel, PGI compilers for Fortran, C, java
AMD, Intel, ATLAS, Goto BLAS, BLACS,
ScaLAPACK, FFTW libraries
MPI is the parallel framework
3 main sections of cluster
Can choose section or just submit
Access by grid certificate
Obtained from e-Science locally
ORBIT MPI installed as a module
Synchrotron Physics and Intense Beams Groups have purchased 17 nodes
= 136 processors
Will be available soon after 16th February
Larger jobs possible using normal SCARF queues
Will still maintain access to 16 processor development rig

14. Simulation Codes ORBIT MPI: Strip-foil injection, including
painting and foil scattering
RF focusing and acceleration
Transport through various
magnetic elements
Longitudinal and transverse
impedances
Longitudinal, transverse, and
three-dimensional space charge
forces
Collimation and limiting apertures
Calculation of many useful
diagnostic quantities
Parallel version of ORBIT available
on SCARF
Set Development (Ben Pine):
Parallel version of Set running
on SCARF
Currently 2D with space charge
Work advancing on adding
2.5D
Main goals this year:
- Benchmarking against
standards
- Adding new 2D boundary
conditions
- Starting to think about full 3D
And:
Add other software
Add other accelerator codes

15. Visualization
e-Science Application Development Group
Parallel visualization cluster
Expertise in data mining, computational steering
Could use existing tools if data put into nexus data format
Simulation hardware, code and visualization resources on site to do
the world-class accelerator physics required for ISIS upgrades

16. (Grahame Rees, Ciprian Plostinar)
800 MeV H- Linac
(Grahame Rees, Ciprian Plostinar)
Design Parameters:
Beam power for 2 MW, 30 Hz, 3.2 GeV RCS: MW
Beam pulse current before MEBT chopping: mA
Beam pulse current after MEBT chopping: mA
Number of injected turns for 370 m RCS: ~500 turns
Beam pulse duration at the 30 Hz rep rate; ~700.0 μs
Duty cycle for the extent of the beam pulse; ~2.1 %
MEBT(in) normalised rms emittances: , (π) mm mr
MEBT(out) normalised rms emittances: , 0.42 (π) mm mr
Cell equipartition transv/long phase shift ratios:

17. All options have the same 324 MHz, 74.8 MeV stage 1:
Design Options
All options have the same 324 MHz, 74.8 MeV stage 1:
IEBT
DTL
MEBT
RFQ
74.8 MeV
Options F (MHz) Stage Stage Stage 4
ScL ScL ScL 3
CCL ScL ScL 3
ScL d ScL e ScL c
? ScL a ScL b ScL c
~200 MeV MeV

18. Beam Envelopes for Option 2
DTL CCL (193 MeV) ScL (800 MeV)

19. 800 MeV H- Linac Studies Still Needed
MEBT comparisons
Selection of option
ScLa/b spoke cavities?
Refine stage matchings
Linac error effect study
Effect of failed cavities
Ring collimation line