Crab Cavities “ from virtual reality to real reality” - CERN

juicebottleAI and Robotics

Nov 14, 2013 (3 years and 10 months ago)

109 views



LHC
Crab Cavities

from virtual reality to real reality”
R. Calaga, BE-RF, LHC-PW, Chamonix 2012
On behalf of the LHC-CC
collaboration


The Real “Problem”
Nominal → 4 IRs, 120(+) parasitic encounters
Sufficiently large crossing angle inevitable (8-12

sep)
Beam-Beam Team
CERN-ATS-2011-217
8 to 16 LR
encounters
No collisions or LR
2011 MD: 36 bunches
50 ns, 2 Collisions
Reducing crossing angle


2

Consequence
Φ
=
σ
z
σ
x
ϕ
c
σ
eff
=

σ
x
2
+
σ
z
2
ϕ
c
2
Piwinski angle
Upgrade: reduce

* (by factor 2-4)
Consequence → approx double the crossing angle (10

sep)
Ineffective Overlap
Note: don't forget hour-glass effect (~15% loss for

*/

z
)


Some Numbers
2011
2012
after LS1
after LS3
Energy
3.5 TeV
4 TeV
7 TeV
7 TeV

* [cm]
100
60
55
15
2

[

rad]
260
313
247
473
R

(

z
=7.55cm)
0.94
0.85
0.82
0.37
R

(

z
=10.1cm)
0.76
0.74
0.28
2
ϕ

d.

ϵ
/
β
ip
Assume
:

N
= 2.5

m, d=10
very inefficient


For the Upgrade
12

separation
10

separation
N
b
= 2 x 10
11
p/b

N
= 2.5

m

* = 15 cm
S. White, LHC-CC11
L
pk
< 7 x 10
34
(12

sep), little margin for leveling
Note: don't forget synchro-betatron resonances

~2-4
BBLRs might alleviate partially
Nominal


To Recover
Δ
p
x
=
qV
E
.
sin
(
ϕ
s
+
ω
t
)

c


- Bump”
RF Deflector
RF Deflector
V
crab
=
cE
tan
(
ϕ
c
)
ω
R
12
.
2
sin
(
π
Q
)
cos
(
ϕ
cc

ip

π
Q
)


Cavity Voltage
~6MV/ IP-side (2 cavities)
add a cavity


Why 400 MHz
LHC bunches are long
RF non-linearity (longitudinal)
Higher frequency (for example 800 MHz)

Smaller” cavities
Less voltage (V
T


1/

) →
Not really
Easier phase noise control ? (see later)
800 MHz Cavity, K. Ohmi
L

N
b
2
σ
2
R
Φ
F
RF
Φ
GUINEA-PIG simulation, Y. Sun
F
RF
~ 10-25%
Form factor ~1 (

* 10-55 cm)
=1


Pillbox Cavity
beam
Transverse Cross Section, squash
beam in/out of the plane
TM
010
TE
111
TM
110
TE
011
freq spectrum
TM
011
TM
210
TM
110Y
f
res

1
R
R
(independent of length)
crabbing mode (HOM)
R: 400 MHz ~ 610mm

800 MHz ~ 305mm
Too big for
IR regions


Lengler et al., NIM 164 (1979)
Karlsruhe-CERN RF Separator

1
st

SRF Deflector
Assembly into cryostat
RF separator for 10-40 GeV/c from the
SPS
Unknown heavy particles, baryonic states/exchange, K
±
& p-bar
F = 2.865 GHz
V
T
= 2 MV/m (104 cells)
Still in use at U-70 setup at IHEP


KEK Freq: 508.9 MHz
Power: 50-120 kW
(Qext: 2x10
5
, BW: 2.55 kHz)

1
st”

e
±
Crab Cavity
Feb 2007
LONG R&D, but short lifetime
(2007-2010)
Complex HOM Damping Scheme


THEY WORK!
The real question: will the technology be
efficient/transparent for the HL-LHC operation
Real answer: you may have to wait a little while


The LHC Pillbox
Conceptually simple, but practically difficult (KEKB experience)
Main Constraints:
Frequency


800 MHz
Damping LOM/SOM/HOM remains a challenge
Complexity of multiple frequencies in LHC
Only vertical crossing at both IPs
Surface field to kick gradient ratio is poor
1-cell version, CERN, L. Ficcadenti et al.
2-cell version, USLARP, L. Xiao et al.


Pillboxes → TEM Cavities
~4yr of design evolution
Exciting development of new concepts
(BNL, CERN, CI-DL-LU, FNAL, KEK, ODU/JLAB, SLAC)


Short History
80yrs later
similar concepts to be applied
for LHC crab cavities
Concentric Conducting System
short for coax
Leading to the telephone etc..



Its strongly reentrant form makes the field pattern
at the outer radius
predominately TEM
with the
consequence of only moderate current flow”
E. Haebel
Freq = 100 MHz
Gap Voltage = 0.5 MV
P
beam
= 200 kW
(1.6 MW @400 MHZ, NC Cavities)

/4
More History



/4 TEM Resonator
I
0
V
0
a
b
~

/4 = 187.4 mm
V
0
a
b
~

/4

~

/4

gap
Z
0
=
V
0
/
I
0
Frequency


resonator length
and “not” the gap or radii of the concentric cylinders
194 mm
194 mm
142.5 mm
122 mm
BNL: I. Ben-Zvi et al.


400 MHz LHC Cavity, quasi

/4
Z
0
tan
(
β
l
)
=
1
ω
C
gap

/4 Resonator, HOMs
Note, due to large aperture & residual Ez the
LHC cavity will only a quasi

/4 resonator
For a pure

/4 resonator, next HOM is x3
the fundamental mode
Therefore, damping is a LOT more easier
(for example use a high-pass filter)
56 MHz RHIC Prototype
Pedestal to cancel E
z


V
0
I
0
-I
0
~

/2

/2 TEM Resonator
Two

/4 resonators →

/2

Use HOM (TE
11
like) for deflection

More elegant is to use two

/2 resonators
Single

/2
Two

/2
SLAC, Z. Li
ODU, J. Delayen

Height of the cavity is symmetric about beam pipe

Only compact in dimension, LHC needs both x-y compactness
~

/2



Joint SLAC-ODU Effort
2010
2011
Fill these regions
Full design change
SLAC, Z. Li
ODU, J. Delayen

/2 TEM Resonator
Symmetric Ridges
Also, Initially proposed by
F. Caspers (Crab WS 2008)



/4 = 187.5 mm
Courtesy G. Burt, B. Hall
4R (LU-DI-JLAB)
Four co-linear

/4 resonators
500 MHz CEBAF Separator
Conical resonators for mechanical
stability
Downside is that the deflecting mode
is NOT the lowest order mode
4 eigenmodes, mode 2 is our crab mode


Performance Chart
Double Ridge
(ODU-SLAC)
4-Rod
(UK)
¼ Wave
(BNL)
Cavity Radius [mm]
147.5
143/118
142/122
Cavity length [mm]
597
500
380
Beam Pipe [mm]
84
84
84
Peak E-Field [MV/m]
33
32
47
Peak B-Field [mT]
56
60.5
71
R
T
/Q [

]
287
915
318
Nearest Mode [MHz]
584
371-378
575
Kick Voltage: 3 MV, 400 MHz
Geome
trical
R
F
194 mm
B1
B2
< 60 MV/m
< 100 mT
damping more
complicated


Impedance Thresholds
Longitudinal impedance
2.4 M

total (7 TeV)
Strongest monopole mode:
R/Q=200

→ Qe
<
1x10
3
Damping → Qe < 100-500
Transverse
Courtesy: Burov, Shaposhnikova
H
OM
HOM
H
OM
H
OM
Crab
Strongest dipole mode:
Z < 0.6 M

/m (0.58 GHz)
(Qext = 500)
Longitudinal


HOM probe
Input
HOM
Broadband
LOM
3-5 stage Chebyshev
High pass filter
HOM Damping
56 MHz Prototype
(placement not fixed yet)
4 Symmetric couplers
on the end caps
(notch/high-pass ?)
4 asymmetric couplers on
cavity body


ODUCAV
SRHW
KEKCAV
UKCAV
QWAVER
FRSCAV
Vz(x=0) [kV]
0.0
-2.1 - 2.5i
-4 +
1378
i
0.0
0 +85.7i
-0.1 -0.2i
Vx [MV]
5
5
5
5
5
5
B(2) [mTm/m]
0
0 -0.04i
-32.7 - 0.1i
0.02 + 0i
25 + 0i
0 +
108
i
B(3) [mTm/m2 ]

1250 + 0i

229 + 0i
250 - 0i
2452
- 0.5i
464 + 0i
-233 +1i
B(4) [mTm/m3]
0
0
266 - 5i
0
540 +0i
-189 -
14209
i
RF “Multipoles”
Courtesy: A. Grudiev, R. deMaria, J. Barranco
Linear tune shifts ~ 0.0 -10
-3
Non-linear effects (b3, b4) → Negligible
See slide A5 for mitigation


Cavity Tuning
Thoughts
Up/down motion
±
2mm → 1 kHz
Push/pull on
cavity body
Scissor jack type
mechanism
CEBAF Tuner
Double lever
(Saclay type)
SM
SM
Modified screw/nut
(SOLEIL type)
SM


Low gradient (weak or moderate)
High Field (weak)
Multipacting
Courtesy G. Burt, J. Delayan, Z. Li
Medium gradient (strong)
beam-pipe region (similar to KEKB)
Not a serious worry, will require RF processing


RF Power
V
b

Q
L
I
b
R
T
Q
0
(
k
Δ
x
)
50 kW
RF Power ~8kW (
V
T
=3 MV)
Margin
R/Q = 300

I
b
= 0.55 A
For Comparison,
Main RF 300kW (
V=2 MV)


RF Power Options
Courtesy E. Montesinos
2.0m
2.5m
2.0m
Solid State Amplifiers
190 kW, 352 MHz
Single tower < 3m
IOTs
(TV Transmitter)
Light Sources
Tetrode
(SPS)
400 MHz, ~50kW
Electrosys
2.5m
50 kW/cavity, moderate power
Simplified (modified) LHC coupler
Common platform for 3 cavities designs
Three available choices
For SPS tests, reuse Tetrodes used in SPS tests


Crab Cryomodule
Graphic Courtesy: S. Weisz
(Space in bypass extremely limited)
RF Distribution
~300m
LLRF (Coupled feedback)
P. Baudrenghien
Need ~20-25 m space for
amplifiers on each IP-side
Waveguides/Coax

Preliminary thoughts”


RF Noise
Δ
x
IP
=
θ
c
k
RF
δ
ϕ
Δ
V
T
V
T

1
tan
(
θ
/
2
)
σ
x
*
σ
z
For example:

c=570

rad;

V/V=0.4%

x
*=7

m,

x
*=7.55cm

err
=1.2

rad
Amplitude
jitter
Phase jitter
For example:


= 0.005
0
,

c=570

rad

x
IP
= 0.3

m
(5% of

x
*
)
LHC Main RF,

= 0.005
0
at 400 MHz (Philippe)
(summing noise at all betatron bands from DC→300kHz)
Note: IOTs & SSAs are
less noisy
+ betatron comb (


0.001)


Planning Overview
M2: Compact Validation
& Selection (2012-13)
M2: Beam Tests
(2015-16)
Prototype Cryomodule
Final Implementation
(2022-23?)
Production of Cryomodules
Detailed planning, see E. Jensen (LHC-CC11)
LS1
LS2
LS3
Cavity Testing


Sheet metal (deep drawing, spinning, hydro-forming)
Multiple dies, electron-beam welding
Solid Niobium & machining
Material costs & leak tightness
Fabrication Options
{Total 16 cavities (2 IPs, B1 & B2)

With sheet metal (4mm thick)
We need approx 500-600 kg Niobium (RRR>300)}


0
100
200
300
400
500
600
0.00E+00
1.00E+02
2.00E+02
3.00E+02
4.00E+02
5.00E+02
6.00E+02
Position [cm]
Ez [V/
m
]
4R Al-Prototype
Courtesy G. Burt, B. Hall
Bead-Pull
Niobium cavity to be delivered in
March 2012
Nb Cavity from
solid Ingot
Al-prototype for field measurements


Double Ridge Fabrication
Courtesy:J. Delayan, Niowave
Nov 2011
Jan 2012
Niowave
STTR, Phase I/II
Testing April 2012


Real Reality ?

If it is real, we believe in it”
The Church of Reality


100%
90%
80%
70%
Leveling with crossing angle
A1: Leveling, X-Angle
Courtesy Beam-Beam Team
CERN-ATS-2011-217
Demonstrated in 2011 w/o affecting other IPs and emittance
w/o crabs range is extremely limited
To fully exploit leveling with x-angle, an RF cavity is ideal


A2: Why SC-Cavity
Q
0
=
G
R
s
Geometrical factor
~ 200

Microwave resistance
Copper ~ m

Niobium-SC ~ n

With ~6MV/module, NC-RF is not a viable choice
R
s
=
1
σ
δ
G
=

E
3
dV

H
2
dA
Maximize aperture & minimize # of cavities (reduced impedance)
A choice of 2K cryogenic system optimum for crabs (LHC-CC11)


A3: SPS As a Testbed
Long. Position: 4009 m +/- 5m
Total length: 10.72 m

x,

y: 30.3m, 76.8m
Cavity validation with beam (field, ramping, RF controls, impedance)
Collimation, machine protection, cavity transparency
RF noise, emittance growth, non-linearities,
Instrumentation & interlocks
Present COLDEX


4 LHC Cavities in SPS (1998)
RF Power Setup (~50kW, Tetrode)
A4: SPS, BA4 Setup
Y-Chamber like, similar to present COLDEX
Courtesy E. Montesinos


5 dbm/div
500 kHz
500 kHz
A5: RF Noise, LHC
with 1-T feedback
P. Baudrenghien

Selective reduction at all f
rev
lines (V=1.5MV, Q
L
=60k)

Using a betatron comb, we can expect ~16dB reduction
at selective frequencies


Courtesy G. Burt, J. Delayan
A6: RF Non-Linearity
Voltage deviation over 5mm:
Horizontal: 20% → 5%
Vertical: x2 → 10%
Tuning (shaping) to suppress multipoles


A7: Other Applications
Emittance exchange x-z (P. Emma & others)
Φ
=
σ
z
σ
x
ϕ
c
HE-LHC (16.5 TeV)
= 0.6, similar to nominal
(

z
= 6.5cm,

x
= 9

m,

c
= 160mrad)
R

= -12% wr.t. to head-on
Compensate offset collisions due to beam loading for LHeC (Zimmermann)
May not be needed if phase modulation removes the phase-slip
Momentum cleaning: Qacc = (fcc/f0)

(S. Fartoukh)
For effective Qacc ~ 0.3 → 8GHz, too high freq (Y. Sun)

x




z



SRF Deflector
10 MV, 366-447 MHz
3 GeV LINAC
Mode l
TE113
Freq
447 MHz
R/Q
500

Epk
34 MV/m
Bpk
74 mT
Aperture
75 mm
A8: ProjectX Synergy
LHC Type Concept(s)
Courtesy M. Champion, Y. Yakovlev


A. Facco, SRF09
A9: TEM Resonators
Saclay
IPNO
Argonne
New Delhi
INFN LNL-MSU
TRIUMF
INFN LNL
Sputtered
INFN LNL
Right here at CERN
(HIE-ISOLDE)
Cavity reached (ANL 72 MHz)
Ep=70 MV/m, Bp=100 mT
Q0 = 1 x 10
9
at 4.6 K (IPAC10)