1. RT Operations RF Power Sources Linac Systems System Operation
Ivan Konoplev, JAI Oxford University
Peter McIntosh, STFC ASTeC

2. Introduction & Goals Accelerator technologies: Quality Reliability
RF Sources
Linacs
Operational Experiences
Quality
Reliability
Size/Modularity
Servicing/Maintenance
Costs
Questions issued (non-exhaustive), collate feedback, try to converge on an optimum RT approach.
What are technology development opportunities:
efficiency, robustness, simplicity, stability

3. Session Programme 13:45 – 14:00 Introduction & Goals
14:00 – 15:45 RF Sources & Linacs (I Konoplev, P McIntosh)
Affordable Linacs (U Amaldi)
Quality
Reliability
Size/Modularity
15:45 – 16:15 Coffee
16:15 – 17:45 Operational Experience (J Van Dyk)
Service/Maintenance
Operability
Costs

4. Klystron – based on inertial bunching mechanism of RF generation
Efficiency - up to 85% (research) and 60% (commercial) –running cost savings
Commercial availability in broad frequency range operating range from 0.1GHz up to 15GHz
Versatile, reliable they are operating at 100s (A) beam current and 10s (kV)
Can be high average power up to 1MW
However they are more expensive
Basic schematic of two cavity klystron
Technical drawings of a multi-cavity cavity klystron

5. Example of typical operation data

6. Example of absolute rating

7. What to expect from vendors
The tube price is about £100k and can be negotiated with up to 15% reduction if one buys 10 of them.
Spec. for life time is 5000 hours (sure can run > but not specified as such in official offers).
Tube can be repaired (new cathode) for 30% of initial price.
New tube as compared to existing tubes with similar peak power should save >30% of modulator cost (low voltage, no oil tank) and it is very compact (1/3 of standard modulator in volume).

8. Magnetron – M-type tube generates RF if electrons are losing “potential” energy
It is an Oscillator and cannot be used as an amplifier
Not expensive (1/3rd of klystron price) but not as reliable, life time is an issue for HP devices
Power up to 3MW (peak power) good for low energy LINCAS (up to 6MeV)
Operating frequency range 0.1GHz to high 10s GHz
Versatile but not stable and operating frequency
may walk around
The cathode in a magnetron is harder to cool than
in a klystron and gets additional heating from back
bombardment so the maximum output is far less
than a klystron

9. Magnetron
Cathode is heated to generate the electrons but it is also heated by the electrons coming back – erosion of the cathode
Magnetrons are cheap to manufacture but it cannot be controlled to the level of the klystron
Schematic view of the magnetron and its operation

10. Example of magnetron controls

11. Magnetron and Klystron
Cost

12. Solid state RF power supply
Efficiency - comparable with klystron but less
Water cooling is required
Modular
High power (1MW level) will be available at 750MHz soon
At frequencies above 1 GHz the high power will be available in the next 3-5 years
Expensive but relatively easy to run
Typical table from one of the vendors
Klystron
IOT
SSA-1
SSA-N

13. Example of SSA RF power supplies
One of the most complicated parts are power combiner and cooling

14. Vacuum Tubes vs SSA

15. Just to make it more complex the following subsystems are needed
Regardless of the power supply chosen to drive the accelerator

16. Klystron vs Magnetron vs SSA
Efficiency
50%-70%
Up to 90%
Frequency (GHz)
0.1 – 15
up to 1
Peak Power (MW)
Up to 10
Up to 3
<1
Stability
high
low
Life expectancy
Cooling
yes
Maintenance
complex
simple
Size
compact
large
Mobility
Modularity
partial
Capital Cost
intermediate
Run Cost
SSA is a good solution for the future but klystron is the answer to have reasonable cost/quality ratio

17. Accelerator technology standing wave and traveling wave LINACs
Standing wave LINAC
Traveling wave LINAC
Variabilities
Frequency, gradient, phase velocity, geometry, shunt impedance, Q Factor

18. Standing wave LINAC
More compact c.f. TW structures (particularly when optimally coupled)
Typically utilised for heavy ion, proton or short pulse electron linacs

19. Traveling wave LINAC Short fill time structures c.f SW structures
Electron bunch propagates with the wave in accelerating phase
Short fill time structures c.f SW structures
Normally utilised for short beam pulse machines, linear colliders or SR injectors

20. TW vs SW
Incr. loss
Incr. Trans-time TW SW
Dominated by RF source costs
SW Side-Coupled optimised to reach >100M/m with small beam-pipe (6mm)
Resonant coupling – maximise group velocity
Comparison of Standing-Wave and Travelling-Wave Structures, R Miller (SLAC), Linac86

21. S-Band SW Structures (SAMEER)
Parameter
Value
Frequency (MHz)
2998
# Cells 24
Coupling Mode
Side /2
Effective Shunt Impedance (M/m) 87
Q Factor
15000
Length (m)
1.12
RF Input Power (MW)
5.5
Energy (MeV) 15
Compton X-Ray Source

22. S-Band TW Structures (TTX)
Parameters
Units
Length
1.5m
Mode
3/4π
Cell length
39.3mm
Number of cells
38(with coupler) f
~2856MHz a
10.6mm-8.1mm
vg/c
0.83%~0.3% Rs
64.6MΩ/m~71.8MΩ/m
Filling Time
960ns
Eacc(Pin=30MW)
30MV/m

23. C-Band SW Structures (Tsinghua U/Nuctech)
Parameter
Value
Frequency (MHz)
5712
# Cells 12
Coupling Mode
On-axis
Effective Shunt Impedance (M/m)
130
Q Factor
10500
Length (m)
0.29
Eacc (MV/m) 21
RF Input Power (MW)
2.16
Energy (MeV) 6

24. C-Band TW Structures (SwissFEL)
Structures are machined “on tune”, no provisions for dimple tuning!
Cup manufacturing with micron precision at VDL ETG Switzerland
Coupler manufacturing at VDL ETG
Stacked by robot at PSI
Vacuum-brazed at PSI
Production rate: 1-2 / week
Production finished August 2016
High power results for first structure:
Conditioned to 52 MV/m
Break-down rate at 52 MV/m ≈ 2 x 10-6
At nominal 28MV/m, break-down rate negligible (well below the specified threshold of 10-8)
R. Zennaro et al.,
“Measurement and High Power Test of the First C-Band Accelerating Structure for SwissFEL”, Proceedings of LINAC2014, Geneva, Switzerland

25. X-Band SW Structures (SLAC & BIEVT)
Performance of two of them was limited to gradients lower than 55 MV/m by breakdowns in the couplers.
Breakdowns may be related to the high RF magnetic and moderate RF electric fields on the sharp edges at the waveguide-to-coupler cell opening.

26. X-Band TW Structures (CLIC)
GHz, X-band
100 MV/m accelerating gradient
Input power ≈50 MW
Pulse length ≈200 ns
Repetition rate Hz
25 cm
Micron–precision disk

27. Linac Comparison S-Band C-Band X-Band SW (SAMEER) TW (TTX)
(TU/Nuctech)
(SwissFEL)
SW (SLAC)
(BIEVT)
TW (CLIC)
Frequency (MHz)
2998
2856
5712
11400
9300
11994
Operating Mode
Side /2
3/4
On-axis /2
2/3
On-Axis 
On-Axis /2
Eff. Shunt Impedance (M/m) 87
71.8
130 80
81.9
133
112
# Cells 24 38 12
113 15 11
Q Factor
15000
16566
10500
10000
8600
7100
6265
Length (m)
1.12
1.5
0.29 2
0.2
0.15
0.23
RF Input Power (MW)
5.5 31
2.16 28 6 1
63.8
Eacc (MV/m) 17 30 21 50 14
100
Energy (MeV) 45 56 10 23

28. Acknowledgements RF Sources. Linacs I Syratchev (CERN)
Acknowledgements RF Sources Linacs I Syratchev (CERN) W Wuensch (CERN) G Burt (Lanc U) G Burt (Lanc U) A Dexter (Lanc U) A Wheelhouse (STFC) Affordable Linacs, U Amaldi (TERA Foundation) RF Questions

29. Accelerator Technologies
RF Sources
Linacs can potentially be driven by Magnetron, Klystron or Solid state devices - do you have any preferences? Why?
Which accelerator technology is most commonly used?
Linacs
Which accelerator technology is most commonly used
Pro’s/Con’s of the accelerators used for the treatments, what is the most annoying part of working with the accelerator
List the electron beam energies needed for a cancer treatment
List electron beam current needed for a cancer treatment
Can current generation of LINAC available in hospitals deliver the required parameters
Linacs based on Travel Wave/Standing waves do you have any preferences? Why?

30. Quality Electron beam quality i.e. beam dimensions and dispersion?
Accuracy of beam delivery to the target?
Electron beam stability over the treatment time?
Which accelerator technical parameters are the most important for a cancer treatment?

31. Reliability
Accelerator stability i.e. how stable it should be and what are the acceptable boundaries of parameters variations?
Change of the environment and reliability: how stable environment conditions can be provided i.e. cost of running environment stable conditions against the cost of environment robust accelerator?

32. Size Favourable preferable dimensions for the accelerator?
What is more preferable compact but more expensive accelerator or capital cost to upgrade the hospital available rooms to accommodate the facility?

33. Second Session – 16:15 – 17:45
16:15 RT Technology Operation Experience, J Van Dyk (Western Univ, London, Canada) Servicing/Maintenance Operability Costs

34. Services/Maintenance
Availability of technical personal?
Availability of mechanical and electrical workshops to fix things locally?
Stable water and electricity supplies?
What are servicing requirements, capability available, how modular does system need to be?

35. Operability
What is important efficiency criteria – initial costs or operating costs, (is it important to reduce energy consumption by 10% or it would be important to reduce the initial cost of the equipment by 10%)?
What cooling configurations are deemed appropriate (additional water cooling or well-conditioned room)?
What level of controls needed – cost implications? Manual but cost reduced system or fully automated but more expensive?
What are stability requirements – which components need tight control?

36. Costs Capital vs Operating? Technology drivers?
Manufacturing techniques?

37. Combined Session Summary
Capture of key points raised
Energy
Frequency
Size
Source technology
Staffing – expertise and training needed
Operability constraints – local infrastruture
Opportunities for development of RT concepts:
Scope?
Market research?
Technology R&D?