A MultiMission 8 GeV Superconducting Injector Linac

1
A Multi-Mission8 GeV SuperconductingInjector
Linac

• G. William Foster
• SNS Seminar
• May 6, 2004

2
Two years ago now
3
Fermilabs Accelerator Chain
4
Fermilabs Accelerator Chain
Cockroft -Walton
Tevatron
LINAC
BOOSTER
Main Injector
5
Plan 1 more neutrinos
New Booster needed for big increase in neutrino
beam intensity
6
WHY REPLACE THE BOOSTER?

• Neutrino Physics
• Beam Power on Target
• It is older than dirt.
• reliability problems
• The Bottleneck in Accelerating Protons at
  Fermilab
• Space-charge at Booster Injection
• If you do it right, it opens up many new
  opportunities for the experimental physics
  program.

7
Space Charge 101

• Beams are many parallel current /line charges
• Parallel Line Charges Blow Themselves Apart
• (electrostatic space charge repulsion)
• Parallel Line Currents Attract
• (Lorentz force right-hand rule)
• Cancellation is Perfect for Relativistic Beams
• The amount of beam that can be circulated
  without space charge blowing it apart goes as
  ??2.

8
Three Possible Approaches

• Adiabatic Improvements of Existing Booster, Linac
  MI
• A New Booster Synchrotron
• Superconducting 8 GeV Proton Linac

9
Demand Schedule for 8 GeV Protons
8 GeV Linac gt1E18/hr
R. Webber
10
The Linac and Booster in 1971
11
Excavation for the Booster 196970
Sometimes these photos are the best documentation
of what is buried underground at Fermilab !
12
R. Billinge in 1970 installation of the Booster
ComponentsBooster Design was copy of Cornell
10 GeV Synchrotron
13
The Booster ismy Favorite Machine

• it is like an old VW Beetle where you can
• look at the engine and understand every part of
  it.

14

• Drift-tube Linac module arrives 1970

15
The BossInspects
16
Eventually
17
Sometimes you had to crawl inside to fix
it(Drift Tube Linac)
18
Original Proton Source Electronics
19
Proton Source Linac Front End

• Original FNAL Cockroft-Walton

MODERN REPLACEMENT
AccSys PL-7 RFQ with one DTL tank
20
The DTL is also used for Neutron Therapy

• ? THERAPY means it is not experimentalit works.

21
BOOSTERTUNNELCONSTRUCTION(196970)A
Training Manual for Spotting Safety
Violations.not enough shielding for
high-intensity operations
22
Losses During Typical Booster Cycle
At Various Injected Intensities (E. Prebys)
Transition
Intensity (E12)
Energy Lost (KJ)
Protons check in, but they dont check out.
Time (s)
23
Winging it on a dirty floor

• Booster RF Cavities

24
Superconducting RF Today

• Preparation of RF cavities for TESLA linear
  collider

25
Adiabatic Booster/Linac Improvements(partial
list spanning 30 years)

• Linac Energy Upgrade (again)
• Larger Aperture RF Cavities
• Linac Beam Chopper
• Beam Tube/Perforated Liner
• Aggressive Collimation of Losses
• Magnet End Shims (improved field shape)
• More, Better Dampers
• Modernize LLRF System
• Upper 8 GeV Line (Injection Accumulator)

26
ACCELERATOR CONSTRUCTION IN FULL SWING
27
Fermilab Long-Range Planning Exercise

• Leaks Started
• Strong Support for Neutrinos and Proton Driver.
• (especially SCRF Linac)

28
RECOMMENDATIONS

• We recommend that Fermilab prepare a case
  sufficient to achieve a statement of mission need
  (CD-0) for a 2 MW proton source (Proton Driver).
  We envision this project to be a
  coordinated combination of upgrades to existing
  machines and new construction.

29
RECOMMENDATIONS

• We recommend that Fermilab elaborate the physics
  case for a Proton Driver and develop the design
  for a superconducting linear accelerator to
  replace the existing Linac-Booster system.
  Fermilab should prepare project management
  documentation including cost schedule estimates
  and a plan for the required RD. Cost schedule
  estimates for Proton Driver based on a new
  booster synchrotron and new linac should be
  produced for comparison. A Technical Design
  Report should be prepared for the chosen
  technology.

30
Proton Driver Charge Letter
31
Proton Driver I Synchrotron

• 16 GeV Synchrotron
• Proton Driver for 2001 Neutrino Factory
  Design Study
• Exceeded Minimum Specs for MI Injector
• 242M 30Tax
• 315 M

32
Proton Driver II Synchrotron

• 8 GeV Synchrotron
• Same size as Booster
• Optimized for 5x more protons in MI
• 200M 30Tax
• 260 M

33
Proton Driver II Synchrotron

• Proton Driver Study II is for an 8 GeV, 0.4 MW
  synchrotron, upgradeable to 2 MW. It is smaller,
  but also cheaper, than Proton Driver I.
• Design features
• Same size as the present Booster (474.2 m).
• Racetrack shape in a new enclosure.
• Transition-free lattice with zero-dispersion long
  straights.
• Reuse of the existing 400 MeV linac, addition of
  another 200 MeV rf ? Total linac energy 600 MeV.
• 3x1014 protons per second at 8 GeV (380 KW )

34
Proton Driver II Synchrotron Parameter Table
() Although originally designed for 15 Hz
operations, the present Booster has never
delivered beam at 15 Hz continuously. In the past
it used to run at 2.5 Hz. In the near future it
will run at 7.5 Hz for the MiniBooNE experiment
35
Main Injector Intensity Upgrade

•        RF Major upgrade. Need a second power
  amplifier for each cavity (and 4 more cavities
  for synch).
•        Power supply moderate upgrade.
•         Magnet Ok.
•         Cooling capacity Ok for magnet, but
  need to be doubled for rf.
•         Gamma-t jump system New.
•         Large aperture quad New.
•         Collimation system New.
•         Passive damper and active feedback New.
•         Stop band correction New.
•         Shielding Ok.
•         NuMI and other 120 GeV Beamlines Under
  study.

http//www-bd.fnal.gov/pdriver/
36
Conclusions on Synchrotron Proton Drivers

• Two Excellent Design Studies Completed
• http//www-bd.fnal.gov/pdriver/
• Feasible and Affordable 250M
• Will Feed 2 MW upgrade of Main Injector
• 5x current Proton intensity
• Great Benefits to Neutrino and FT Programs

37
8 GeV Superconducting LinacPossible Sitings for
MI Injection
38
8 GeV Superconducting Linac

• New idea incorporating concepts from both SNS and
  TESLA.
• Copy SNS Linac design up to 1.3 GeV
• Use TESLA Cryomodules from 1.3 ? 8 GeV
• H- Injection at 8 GeV in Main Injector
• ? Super-Beams in Fermilab Main Injector
• 2 MW Beam power at BOTH 8 GeV and 120 GeV
• Small emittances ? Small losses in Main Injector
• Minimum (1.5 sec) cycle time
• MI Beam Power Independent of E (flexible ?
  program)

39
120 GeV Main Injector Cycle with 8 GeV
Synchrotron
40
120 GeV Main Injector Cycle with 8 GeV Linac, e-
and P
41
8 GeV Linac Allows Reduced MI Beam Energy
without Compromising Beam Power

• MI cycles to 40 GeV at 2Hz, retains 2 MW MI beam
  power

42
Running at Reduced Proton Energy Produces a
Cleaner Neutrino Spectrum

• Running at 40 GeV reduces tail at higher neutrino
  energies.
• Same number of events for same beam power.
• (Plot courtesy Fritz Debbie)

43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
A. PARA
48
(No Transcript)
49
(No Transcript)
50
(No Transcript)
51
K. McFarland
52
Superconducting LinacParameters
Project Info tdserver1.fnal.gov/project/8gevlinac
53
Why an H- Linac? (part 1)

• To use linac efficiently, we want a long pulse to
  fill the MI over many turns
• With a normal (e.g. proton) beam, the kicker
  magnet used to place the injected beam on the
  circulating orbit will eject the circulating beam
  (Liouvilles theorem)
• Budker Dimov (1963) invented foil stripping
  injection to cheat Liouville.

54
Multi-turn H- Ion Injection
Circulating Beam
4 pulsed ORBUMP magnets
DC Septum
Beam at injection
400 MeV H- beam from LINAC
Stripping foil
(Eric Prebys slide)

• At injection, the Linac H- beam is injected into
  the Booster over many turns.
• The orbit is bumped out, so that both the
  injected beam and the circulating beam pass
  through a stripping foil, after which they
  circulate together.

55
Why an H- Linac? (part 2)

• The foil stripping method can also be used to
  perform essentially perfect collimation of the H-
  beam downstream of the linac.
• Normal beam collimation (e.g. a copper block with
  a hole in it) makes a big mess downstream.
• A FOIL with a hole in it will strip the H- beam
  halo outside a certain radius.
• The stripped halo and unstripped beam can be
  magnetically separated and the halo sent cleanly
  to a dump.

56
8 GeV S.C. Linac Can Accelerate electrons,
positrons, H- and protons

• The linac pulses at 10 Hz, but the MI only uses
  0.6 Hz.
• The last 7 GeV of the linac can accelerate e- or
  P
• Requires fast ferrite phase shifters (SNS RD)
• Other possible missions for unused linac cycles
• 8 GeV electrons can drive XFEL
• 8 GeV ? program, Spallation Neutron or Muon
  sources, etc.
• 8 GeV Linac could eventually become e
  pre-accelerator for TESLA _at_FNAL

57
8 GeV Superconducting LinacWith X-Ray FEL, 8 GeV
Spallation Neutrino Sources and Neutrino Factory
Anti- Proton
58
8 GeV Superconducting LinacWith X-Ray FEL and 8
GeV Spallation Neutrino Source
59
Multi-Mission 8 GeV Injector Linac
60
Benefits of 8 GeV Injector Linac

• Benefits to ? and Fixed-Target program
• solves proton economics problem gt 5E18
  Protons/hr at 8 GeV
• operate MI with small emittances, high currents,
  and low losses
• Benefits to Linear Collider RD
• 1.5 scale demonstration of TESLA economics
• Evades the Linear Collider R D funding cap
• Simplifies the Linear Collider technology choice
• Establishes stronger US position in LC technology
• Benefits to Muon Collider / ?-Factory RD
• Establishes cost basis for P-driver and muon
  acceleration
• Benefits to VLHC small emittances, high
  Luminosity
• 4x lower beam current reduces stored energy in
  beam
• Stage 1 reduces instabilities, allows small beam
  pipes magnets
• Stage 2 injection at final synchrotron-damped
  emittances

61
8 GeV Superconducting LinacTECHNICAL SUBSYSTEM
DESIGNS EXIST AND WORK
FNAL/TTF Modulators
SNS Cavites
TTF Style Cryomodules
Civil Const. Based on FMI
RF Distribution
62
Linac 0 - 87 MeV

• Direct Copy of SNS Design
• Ion Source
• RF Quadrupole (RFQ)
• Medium Energy Beam Transport buncher (chopper?)
• Drift Tube Linac (402.5 MHz Normal Conducting)
• SNS work provides technical existence proof

63
At Reduced RF Duty Cycle of 1, the Front End
is a Commercial Product
CUSTOM LINAC SYSTEMS AccSys proprietary and
patented linac technology can provide a wide
range of ion beams and energies for specialized
applications in research and industry. AccSys
experts will design a system to customer
specifications consisting of a carefully selected
combination of our standard modular subsystems
radiofrequency quadrupole (RFQ) linacs, drift
tube linacs (DTL), rf power systems and/or other
components such as high energy beam transport
(HEBT) systems and buncher cavities. Radio
Frequency Quadrupole Linacs
AccSys patented Univane (US Patent No.
5,315,120) design provides a robust,
cost-effective solution for low-velocity ion
beams. This unique geometry incorporates four
captured rf seals, is easy to machine, assemble
and tune, and is inexpensive to fabricate. The
extruded structure, which is available in lengths
up to three meters, can accelerate ions injected
at 20 to 50 keV up to 4 MeV per nucleon. Cooling
passages in the structure permit operation at
duty factors up to 25.
Drift Tube Linacs
Drift Tube Linacs provide a cost-effective
solution for ion beam energies above a few MeV
per nucleon. Designed to accelerate ions from an
RFQ, the DTLs permanent magnet focusing and high
rf efficiency result in a minimum cost per MV.
AccSys patented drift tube mounting scheme (US
Patent No. 5,179,350), which is integral to the
twin-beam welded vacuum tank, provides excellent
mechanical stability and low beam loss.
D. Youngs work...
64
AccSys Source/RFQ/DTL

• AccSys PL-7 RFQ with one DTL tank

• Appears to have shorter length and lower price
  than cloning the SNS Linac, for 10 Hz operation

D. Youngs work...
65
Superconducting Cavity Gradients

• 8 GeV design
• assumes peak
• field in cavities
• of 45 MV/m.
• SNS
• 37.5 MV/m
• TESLA(500)
• 47 MV/m
• TESLA(800)
• 70 MV/m

66
SNS b 0.81 Tests at TJNAF
from N. Holtkamp Nov 01 SNS Review
Eacc gt 20 MV/m for protons is now reasonable
design goal
67
9 Cell Beta1 Cavities, 1207.5 MHz
68
Self-Consistent Accelerator Physics Design(Jim
MacLachlan)

• Had to increase number of quads to get to good
  design...

69
CRYOMODULES

• BIG Differences between SNS TESLA
• Key Specification
• SNS Cryomodules can be swapped out in 1 shift
• TESLA cryomodule replacement takes 25 days
• comes from having 2.5 km section of linac
• 8 GeV LINAC 2 day repair time specified
• possible because linac sector is much shorter
  300 m

http//tesla.desy.de/new_pages/TESLA_Reports/200
1/pdf_files/tesla2001-37.pdf
70
SNS/CEBAF Cryomodules

• Warm-to-cold beam pipe transition in each module
• 2K Coldbox, J-T HTX in each Cryomodule
• Bayonet disconnects at each coldbox
• Only 3-4 cavities per cryomodule 50 fill
  factor
• Expensive Design forced by fast-swap requirement

71
TESLA-Style Cryomodules for 8 GeV( T. Nicol )

• Design conceptually similar to TESLA
• No warm-cold beam pipe transitions
• No need for large cold gas return pipe
• Cryostat diameter smaller than TESLA ( same
  as LHC)
• RF Couplers are KEK/SNS design, conductively
  cooled

72
8 GeV H- Stripping in Magnets

• B 0.06 Tesla strips only 1E-6 of Beam in 10m
  length
• 500m Bend Radius is OK
• Stripped Beam Power is lt1 Watt

73
8 GeV Injection Line Optics with Betatron and
Momentum Collimation
Similar to BNL Design for SNS
Ensures that no beam halo is delivered to ring
74
H- Injection Layout in MI

• Foil Stripping Injection at 8 GeV
• Slow orbit bump disappears as beam accelerates
• (fast, smaller orbit bump also required to escape
  foil)
• Injected beam misses nearest quad in MI straight
  section

75
H- Injection Painting(A. Drozhdin)

• Painting from 2pi into 40pi with 90-turn
  injection seems feasible (2-3 traversals/P)
• Peak foil temperature 2200 degC (tolerable)

76
8 GeV Linac Baseline 2 MW
77
RF - Klystrons

• 402.5 MHz / 2.5 MW (7 total)
• 805 MHz / 5.0 MW (10 total)
• 1207.5 MHz / 10 MW (36 total)
• Also the possibility of standardizing on TESLA
  frequencies and eliminating 805 MHz
  entirely

SNS Actual
TESLA Design Scaled by 7 from 1.3GHz
78
RF Power Budget Coupler Power
79
RF Fan-out for 8 GeV LinacA. Moretti, D. Wildman
80
RF Fanout at Each Cavity
81
805 MHz RF Distribution in Tunnel
No Active Electronics in Tunnel only Ferrite
Bias Coils.
82
RF Distribution Technical Parameters
83
Modulators for Klystrons

• Biggest single component in RF costs
• Pfeffer, Wolff, Co. (FNAL BD) have been making
  TESLA spec modulators for years
• FNAL Bouncer design in service at TTF since 1994

84
Modulator Circuit

• IGBT / Capacitor Discharge circuit
• Bouncer to maintain flat top
• Redundant Switch with Ignitron Crowbar
• Pulse Transformer 10kV to 130 kV (typ.)

H. Pfeffer, D. Wolff, sons.
85
(No Transcript)
86
8 GeV Klystron Gallery Floor Plan

• Linear array of equipment - simpler than SNS
• Much more room allocated for electronics than
  TESLA design

87
FRONT END BUILDING
88
Tunnel Cross Sections
89
MAIN INJECTOR CONNECTION
90
LCW System from Fess
8 GeV Linac LCW System
91
TESLA Tunnel Klystrons
92
TESLA RF DISTRIBUTION SYSTEM
93
SNS March 2000 Design, 12 Cavities /
KlystronIndividual Collective Cavity Control

• SNS pursued dropped due to lack of RD time

94
SNS Final Baseline RF Design 1 Klystron Per
Cavity, Individual Control

• Conceptually simpler but 10x more Klystrons

95
Fast Ferrite Phase Shifter RD

• Provides fast, flexible drive to individual
  cavites of a proton linac, when
  one is using a
  TESLA-style RF fanout. (1 klystron feeds 12-36
  cavities)
• Also needed if Linac alternates between e and P.
• This RD was started by SNS but dropped due to
  lack of time. They went to one-klystron-per-cavit
  y which cost them a lot of money (20M / GeV).
• Making this technology work is key to the
  financial feasibility of the 8 GeV Linac.

96
ELECTRONICALLY ADJUSTABLEE-H TUNER
Attractive Price Quote from AFT (ltlt Klystron)
FERRITE LOADED SHORTED STUBS CHANGE ELECTRICAL
LENGTH DEPENDING ON DC MAGNETIC BIAS.
TWO COILS PROVIDE INDEPENDENT PHASE AND
AMPLITUDE CONTROL OF CAVITIES
97
(No Transcript)
98
(No Transcript)
99
Ferrite Phase Shifter High-Power Test Stand
A. Moretti, D. Wildman

• 805 MHz Klystron
• 12 MW x 100usec
• (need 0.5 MW x 1 msec)
• First goal
• See if existing YIG tuner functions at 500kW.
  (yes!)
• Ultimate Goal
• 0.2 dB loss for
• 360 deg. phase shift
• in 100500usec.

Dry Load
12 MW Klystron
YIG Ferrite Phase Shifter
Door-knob Transition
Hybrid Tee
Ferrite Bias Supply
100
Tuners and Cavity Resonance Control

• TESLA RD has shown that piezoelectric tuners can
  correct for Lorentz detuning and cavity
  microphonics in a 1 ms pulsed SC linac.
• SNS Cavity assemblies used for cost basis include
  both Mechanical and Piezoelectric tuners
• RF phase and amplitude control provided for
  individual cavities via fast ferrite phase
  shifters.
• Simulation ( SNS experience) needed to determine
  bandwidth requirement of shifters.

101
(No Transcript)
102
At What Energy and Klystron Fanout does Vector
Sum Regulation Work?

• If TESLA-style Vector-Sum Regulation works
• No need for FAST phase shifters
• No need for Ferrite Phase Shifters at all if we
  start with H-/protons only.
• Design Study assumes Vector Sum works in
  Beta1.00 section.
• Marcus Huening embarking on detailed simulation

103
(No Transcript)
104
RF Phasing in Linac for Protons vs. Electrons

• Cavity cell length changes as proton accelerates
• not all cavities can be same design
• lose some gradient by running off design ?
• Protons are non-relativistic
• energy error ? downstream phase error
• Protons run off-crest
• only get 85 of accelerating gradient at crest
• more sensitive to phase errors
• Must change cavity phases to accelerate electrons
  and protons on alternate cycles

105
(No Transcript)
106
Summary of Phase Shifter Specs805 MHz and 1207.5
MHz
107
Phase and Amplitude Tuner Specs
108
Staging Cost Reductions

• Preserve 2 MW beam power Main Injector
• Need 1.5E14 25uC per linac pulse.
• Reduce stand-alone power of 8 GeV Linac from 2
  MW? 0.5 MW
• Beam current 26mA ? 8.7mA
• Beam Pulse Length 1msec ? 3 msec
• Rep Rate 10Hz ? 2.5 Hz
• Klystron Count 41 ? 12-17
• Option to install the rest as later 2 MW upgrade
• Investigate SCRF front end instead of DTL.

Same Charge Per Pulse
109
8 GeV Linac Baseline 2 MW
110
0.5 MW with SNS Frequencies
111
0.5 MW with SCRF Front End
112
0.5 MW with TESLA Frequencies SCRF F.E.
113
SCRF Front End Option

• Replace DTL with independent, low-power SCRF
  resonators
• RF system 1/5 the size (total watts)
• 300 MHz - 400 MHz
• Independent power tubes,
• or Fan out power from one big tube to many
  resonators?

114
(No Transcript)
115
(No Transcript)
116
Radial Combiner (or Fanout)
SNS 402.5 MHz Klystron
117
Advantages of using the TESLA-compatible
frequencies

• 1)    There are only two klystron types (433
  1300 MHz) in the final machine, rather than three
  klystron types (402,805, 1207 MHz) in the
  SNS-compatible design.
• 2)    No Klystron RD program is needed. For the
  main 1300 MHz Multi-Beam Klystrons, TESLA has
  working prototypes and is in the process of
  qualifying three vendors (Thales, CPI, Hitachi).
  For the front-end 433 MHz linac, several MW
  pulsed klystrons (e.g. Thales TH2120) also exist.
  Development cost for a new Klystron design is
  1M/vendor and takes 1-2 years. Cavity
  development is cheaper.
• 3)    There are fewer overall klystrons, since
  one 10 MW TESLA Klystron replaces two 5 MW SNS
  klystrons. This is a savings of 5 klystrons in
  the 2 MW design, or 2 klystrons in the 0.5 MW
  staged design.
• 4)    The long-term availability of the 5 MW SNS
  klystron is not clear, since the SNS upgrade plan
  is to replace the warm-copper CCL which uses the
  5 MW klystrons with a SCRF section that does not
  use them.
• 5)    We don't have an immediate need for an RF
  coupler RD program, since the conductively
  cooled TESLA (TTF3) coupler is adequate in both
  price and performance. For 805 MHz, we would
  have to design and test a conductively cooled
  variant of the (vapor cooled) SNS coupler, as
  well as a separate 1207 MHz coupler. The TTF3
  coupler is not perfect, and the longer term, we
  may wish to continue RF coupler development.
• 6)    No immediate cryostat RD is needed, since
  the betalt1 cavities will fit into the same TESLA
  cryomodules as the beta1 cavities. For the
  larger 805 MHz cavities, a new cryostat design is
  necessary.
• 7)    Ceiling height on the Klystron Gallery will
  be about 5 ft shorter. The TESLA Klystron height
  is 8.2 ft vs. 13 ft for SNS 805 MHz klystron.
  For an underground gallery the cost difference
  will be significant.
• 8)    Waveguide components, circulators, loads,
  ferrite tuners etc.are smaller cheaper at
  1300 MHz than at 805 MHz.
• 9)    There is some evidence that the larger 805
  MHz cavities are harder to rinse than smaller
  TESLA cavities, and that this may be responsible
  for recent SNS cavity yield problems.
• 10)           Frequency compatibility with A0
  infrastructure may be an advantage for early
  tests

118
Advantages of the SNS-Compatible 402.5 MHz / 805
MHz / 1207 MHz solution

• 1) Successful models of three out of four cavity
  designs exist the SNS beta0.83,0.61, and the
  RIA beta0.47. Only one new cavity design
  (1207.5 MHz beta1) would need to be created.
• 2) Only three beta ranges (rather than four at
  1300MHz) are required to cover the betalt1 region,
  since the 6-cell 805 MHz cavities have a wider
  velocity acceptance than the 9 cell 1300 MHz
  cavities. One fewer cryomodule type (plus
  spares) will be required.
• 3) Cost data exist for the 805 MHz components,
  althought they would still have to be
  extrapolated to 1207MHz in the back 7/8 of the
  linac.
• 4) Waveguide losses are lower for the larger 805
  MHz waveguide.
• 5) Compatibility and parts interchange with SNS
  is possible for the first 1/8 of linac but no
  parts exchange with anyone for the last 7/8 of
  the linac.
• 6) Frequency compatibility with existing FNAL
  linac front end. This would be useful only if
  the warm front end gets replaced off-project, or
  if we want to do early beam tests in the linac
  gallery or muon area.
• 7) Accelerator Physics are well documented for
  the 805 MHz SNS design. The 1300 MHz design will
  have a smaller RF bucket area and a smaller beam
  iris. Petr Ostromov's quick look at a point
  design did not uncover any problems, however.
• tc.are smaller cheaper at 1300 MHz than at
  805 MHz.
• 9)    There is some evidence that the larger 805
  MHz cavities are harder to rinse than smaller
  TESLA cavities, and that this may be responsible
  for recent SNS cavity yield problems.
• 10)           Frequency compatibility with A0
  infrastructure may be an advantage for early
  tests

119
Project Information

• 125 Page Design Study
• Cost Estimate Spread Sheet w/ BoE
• http//tdserver1.fnal.gov/project/8GeVlinac

120
COST ESTIMATE (online)

• 284 M 30 Tax 369 M

121
Conclusions

• Either a Synchrotron or SC Linac Based Proton
  Driver upgrade is technically feasible
• The SC Linac Option is somewhat more expensive
  but has a number of advantages and possible
  secondary missions.
• It looks like Fermilab has (finally) identified
  its next accelerator project. Were looking for
  collaborators.