10/25/2001 Cornell October 2001 1 LEP Operation and Performance Outline: 1) Brief History 2) Injection & TMCI 3) Beam-beam tune shift & Luminosity performance 4) Optimisation 5) Equipment 6) Operations, controls and instrumentation 6) Polarization 6) Other issues 7) Conclusion Mike Lamont, CERN Will try and concentrate on physics & lessons that might be relevant to future machines.
59
Embed
LEP Operation and Performance - Cornell University...65.0 0.050 249 91.5 0.055 89 94.5 0.075 81 98.0 0.083 73 101 0.073 66 102-104 0.055 63 Beam-beam limit not reached Observed in
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
• Beam current around 3 mA • Pretzel test• Lots of waist scans• BIG beam sizes…
8.6 pb-1
Conclusion from Chamonix 91
• a 70/76 team has been set up
• a dispersion team has been set up
• a dynamic aperture team has been set up
• a closed obit team has been set up
• an intensity limitation team has been set up
• a longitudinal oscillation team has been set up
• a crash pretzel team has been set up
• a beam-beam team already exists!
10/25/2001 Cornell October 2001 7
1999 - cruising
BORING!
253 pb-1
10/25/2001 Cornell October 2001 8
Performance
• Two distinct regimes:– 45.625 GeV characterised by working well into the soft beam-beam limit
and approaching the hard limit.– 80.5 GeV and above
• Staged installation of RF cavities• Maximum collision energy (c.m.) raised to 209 GeV• Accelerator physics regime of ultra-rapid damping• Not beam-beam limited
• 2000: Operational strategy to maximize discovery reach with operation in the regime of ultra-rapid damping
10/25/2001 Cornell October 2001 9
Injection
• A lot of effort in to pushing the bunch current in anticipation of high energy,
• Efficiency always variable, synchrotron injection used• In the end limited way below maximum by RF system (power
levels and stability)• Fundamental limit at LEP TMCI which was eventually reached
despite more practical problems – coherent tune shift & resonances (in particular synchro-betatron)– Increase injection energy– Removal of copper RF cavities– Increase of synchrotron tune– wigglers
• Some evidence that long-range beam-beam reduced TMCI limit
Influence from beam-beam: Lower TMCI threshold by ~ 12 %
Synchro-betatron resonances (SBR):
Longitudinal single-bunch instability: Not understood. Avoided with bunchlengthening.
∑ ⊥
=)(
2
s
srevth ke
QfEIσβ
π
+ 1.5 %
Raise Qs (also helps RF)Experimentallyfound 1998 to bearound ~ 1 mA
Qv = n · Qs with n = 1, 2, 3(coherent and incoherent)
AvoidSBR
Injection limits in 1998
10/25/2001 Cornell October 2001 11
(MD-results by P. Collier, G. Roy, R. Assmann and K. Cornelis, M. Lamont, M. Meddahi)
1998 standard workingpoint (SWP):
Qh = 0.28Qv = 0.23Qs = 0.132
780 µA per bunch reached with two beams…
Extended up to ~ 940 µA in single electron bunch MD
(chromaticities ~ 1-2)
LEP working Points:
10/25/2001 Cornell October 2001 12
Qh = 0.29Qv = 0.30Qs = 0.142
High Qv workingpoint:
1030 µA per bunch in 4 bunches (limited by TMCI).
Qs > 0.144 inconclusive. Qs = 0.16, 0.166, 0.174 with 850 µA per bunch.(low injection efficiency 20%, injection would require re-optimisation).
(single beam, separators off)
(lowered chromaticities by 0.5)
New working point (Cornelis, Lamont, Meddahi):
10/25/2001 Cornell October 2001 13
20001999
1998
Overview of Luminosity and Energy Performance
10/25/2001 Cornell October 2001 14
With the strong transverse damping (60 turns at 104 GeV)…
… second beam-beam limit (tails, resonances) is overcome… beam-beam limit is pushed upwards… we then profit from smaller IP spot size and higher currents… 1/3 resonance can be jumped… beams can be ramped in collision with collimator closed
… but also…
… no radiative spin polarization above 61 GeV (energy calibration)
Unique experience with ultra-strong damping at LEP
Larger emittances / energy spread (ε ~ E2, σE/E ~ E)• Less luminosity• Higher backgrounds
Solenoid coupling is weaker (θ ~ 1/E with B=const)• Residual coupling contributes less to vertical emittance
Strong transverse damping (τ ~ 1/E3, 60 turns at 104 GeV)• Second beam-beam limit (tails, resonances) is overcome• Higher beam-beam tune shifts with higher beam-beam limit• 1/3 resonance can be jumped• Beams can be ramped in collision
Luminosity Performance at High Energy
10/25/2001 Cornell October 2001 17
Horizontal beam size: / rmsxx x x x xJ D Eβ εσ β∝ ⋅= ⋅
Compensate increase with energy (smaller luminosity, larger background):
1) High Qx optics with smaller Dxrms (D. Brandt et al, PAC99)
2) Smaller βx* (2.0 m - 1.5 m - 1.25 m)
3) Increase damping partition number Jx via RF frequency
0102030405060708090
100110
00:40 01:00 01:20 01:40 02:00 02:20 02:40 03:00
∆ R
F fr
eque
ncy
[Hz]
Time
Safety settingJx = 1.04(larger σx)
Jx = 1.6 (smaller σx)
Jx = 1.4
101 GeVAutomatic controlJx = function (URF)
For highest energy reach: Reduce Jx.
10/25/2001 Cornell October 2001 18
Scaling empirically fitted by Keil, Talman, Peggs, …
Several points in a given machine, similar configuration for LEP.
Independent cross-check of previous results, however:
• Beam-beam limit reached at 45.6 GeV• Beam-beam limit not reached
Can we infer the beam-beam limit at high energy?
Look at functional dependence of beam-beamparameter on bunch current…
What Is the Energy Dependence of the Beam-beam Limit?
10/25/2001 Cornell October 2001 19
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 200 400 600 800 1000
Ver
tical
bea
m-b
eam
par
amet
erBunch current [µA]
98 GeVSimple model used to fit unperturbedemittance and beam-beam limit:
Two fit parameters A and B:
( )21
y bbA B
ii
ξ = ⋅+ ⋅
20
*0
*
2y
xx
e y
eA fr
β επ γ εβ
= ⋅ ⋅ ⋅
( )1
y b
Biξ
=→ ∞
No BB blow-up
ξy (asymp) = 0.115εy (no BB) = 0.1 nm 0
50
100
150
200
250
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Lum
inos
ity [1
030 c
m-2
s-1
]Vertical emittance [nm]
With BB limit
No BB limit
Limited gainin luminositywith εy:
Vertical Beam-beam Blow-up
10/25/2001 Cornell October 2001 20
Dependence of vertical beam-beam tune param.on bunch current I (in the regime of strongsynchrotron radiation, K. Cornelis): ( )2
1y i
A B iξ = ⋅
+ ⋅
Two fit parameters A and B:
Knowing all other parameters, A is just givenby the unperturbed vertical emittance. Withouta beam-beam limit:
20
*0
*
2y
xx
e y
eA fr
β επ γ εβ
= ⋅ ⋅ ⋅
1y i
Aξ = ⋅
1( )y
Biξ
=→ ∞
B gives the asymptotic beam-beam limit of the vertical beam-beam parameter:
• Beta beat due to beam-beam not included• Tune dependent resonances are not included• Beam-beam tune shift might see other limits
Model of Beam-beam Parameter Versus Bunch Current:
For a given BB limit, the increase of luminositywith current is proportional to the energy γ(el.-magn. field of beam scales as 1/γ)
( )
2
* 22b
y
b
be
i
ir BeL
A
n γβ
= ⋅
+ ⋅
*2b
ye y
bnr
L ie
γ ξβ
∞ = ⋅ ⋅
Use model to predict luminosity:
10/25/2001 Cornell October 2001 27
Compare BB fit to luminosity data: 98 GeV
• Very well described• Simple “squared scaling” not adequate
0
50
100
150
200
250
300
350
0 200 400 600 800 1000
Lum
inos
ity [1
030 c
m-2
s-1
]
Bunch current [µA]
10/25/2001 Cornell October 2001 28
What happens for emittance (unperturbed) improvement:
0
50
100
150
200
250
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Lum
inos
ity [1
030 c
m-2
s-1
]
Vertical emittance [nm]
With BB limit
No BB limit
(unperturbed)
99 98
DFSlimit
(6 mA current)
LEP luminosity limit due to beam-beam: 2.0 1032 cm-2 s-1
(expected at maximum possible current)
(unperturbed)
(8 mA current)
0
50
100
150
200
250
300
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Lum
inos
ity [1
030 c
m-2
s-1
]
Vertical emittance [nm]
With BB limit
No BB limit
10/25/2001 Cornell October 2001 29
Optimisation
• Horizontal beam size given by synchrotron radiation and optics• Working point – beam-beam • Vertical emittance
– Coupling, global & local– Residual dispersion - golden orbits and dispersion free steering
• Vertical beam size at interaction point:– β*
X and β *y
– Dispersion at IP
Thereafter: Reproducibility
10/25/2001 Cornell October 2001 30
Vertical emittance:1999/2000: βy
* = 5 cm ( )2rmsy xyC D KEε ε∝ ⋅ ⋅ + ⋅ +K
E∝ (solenoids)
• Initial tuning of coupling, chromaticity, orbit, dispersion, …• Vertical orbit to get smallest RMS dispersion• Coupling to get smallest global coupling• Local dispersion, coupling, β-function at IP
“Golden orbit” strategy for optimization: Trial and error! Complement with:
Dispersion-free steering (DFS): 1) Measure orbit and dispersion2) Calculate correctors to minimize both
Peak luminosity}}
Note: Global correction generally also improves local dispersion/coupling!
Luminosity balance
10/25/2001 Cornell October 2001 31
-4
-2
0
2
4
0 100 200 300 400 500
y [m
m]
BPM number
-40
-20
0
20
40
0 50 100 150 200 250 300
θ y [µ
rad]
Corrector number
ORBIT DISPERSION CORR. KICKS
DFS: Simultaneously optimize orbit, disp., corr.
-4
-2
0
2
4
0 100 200 300 400 500
y [m
m]
BPM number
-15
-10
-5
0
5
10
15
0 100 200 300 400 500
Dy
[cm
]
BPM number
-15
-10
-5
0
5
10
15
0 100 200 300 400 500
Dy
[cm
]
BPM number
-40
-20
0
20
40
0 50 100 150 200 250 300
θ y [µ
rad]
Corrector number
Measured single beam performance of DFS in LEP:
(same algorithm as implemented for the SLC linac)
10/25/2001 Cornell October 2001 32
00.10.20.30.40.50.60.70.80.9
1
0 1 2 3 4 5 6
ε y [n
m]
Dy (rms) [cm]
1999 1998
98 GeV
0
0.2
0.4
0.6
0.8
1
4500 5000 5500 6000 6500 7000
Ver
tical
em
ittan
ce [n
m]
Fill number
1998
1999
(Data 500-550 µA)
Reduction of RMS dispersion
E [GeV]
94.5
96
98
100
101
Reduction ofvertical emittance
Emittance ratio: 0.5%
(simulated)
(DFS + change of separation optics)
19992000
1998
Vertical optimization
10/25/2001 Cornell October 2001 33
Damping partition number Jx used to reduce horizontal beam size σx:
Good for luminosity and backgrounds in experiments…
In 2000: Keep RF frequency shift small (~ -50 to +20 Hz).
(ii) Choice of RF frequency:
/ rmsxx x x x xJ D Eβ εσ β∝ ⋅= ⋅ Increase with
beam energy.
10/25/2001 Cornell October 2001 34
(E.g. H. Burckhardt, R.Kleiss. Beam Lifetimes in LEP. EPAC94)
05
101520253035404550
18:0018:00 19:0019:00 20:0020:00 21:0021:00Li
fetim
e [h
]
Time
ElectronsPositrons
Lifetime withoutcollision
Putting intocollision
Lifetime during collision(increase with current decrease)
Different regimes:
1) Without collision:Lifetime τ0 due to particles lost in Compton scatteringon thermal photons, beam-gas scattering.We assume 32 hours.
2) In collision:Lifetime due to particles lost in radiative Bhabha scatt.or beam-beam bremsstrahlung.
LEP lifetime without surprises:
10/25/2001 Cornell October 2001 35
Formulas in convenient units for LEP2 parameters (94.5 GeV):
−⋅⋅=−−
][1
][1][2.671]10[
0
1230
hhmAiscmL bunch ττ
BCT][3
1hy τ
ξ⋅
≈
30
35
40
45
50
55
19:00 19:05 19:10 19:15 19:20 19:25 19:30
Lum
. [10
30 c
m-2
s-1
]
BCT
40
45
50
55
60
19:00 19:05 19:10 19:15 19:20 19:25 19:30
Lum
. [10
30 c
m-2
s-1
]
Time
ALEPH DELPHI OPAL
Luminosity decay
Load new orbit + cancel
Correctorbitdrift
Optimizevert. tune
No effect from tails, resonances, …
Lifetime at High Energy Used As Fastest Luminosity Signal:
10/25/2001 Cornell October 2001 36
Operations
• Standard techniques:– Measure & correct beta*– Beta beating, coupling…– Essential, of course, good diagnostics, established measurement
techniques: Q-loop, Fast displays of lifetimes, beam sizes, Orbit feedback, Bunch current equalisation
• First years:– Lack of basic high-level control facilities– Poor data management– Interfaces to crucial beam instrumentation missing in control room– Poor and unreliable, incoherent data acquisition systems
8 additional Cu RF units + 0.14 GeVHigher RF gradient + 0.96 GeVLess RF margin + 1.50 GeVReduced RF frequency + 0.70 GeVBending length + 0.20 GeVTotal + 3.50 GeV
Maximum energy: 101.0 GeV ⇒ 104.4 GeV
Improvements:
10/25/2001 Cornell October 2001 42
LEP operated in “discovery mode”:
Beam energy increased by 3.4 GeV• Increase of RF voltage (3650 MV), excellent stability• Change of operational strategy (ramp during physics fill, …)• Reduced shift of RF frequency• Increase of average bending radius
Push beam energy on cost of luminosity• Reduce beam current (5 mA instead of 6.2 mA)• Run with small Jx, large horizontal beam size• Mini-ramp to quantum lifetime limit
(zero margin in RF voltage)• Lose all fills with RF trips
Luminosity production rate lower than 1999 but still excellent (as in 1998)Luminosity improvement in 1999 with better tuning: + 20 %Price to pay for energy increase in 2000: - 20 %
2000 was the second productive LEP year
2000: conclusions
10/25/2001 Cornell October 2001 43
Unique at LEP:
Large range of energies 22 GeV to 104.5 GeVPolarization studied from 41 GeV to 98.5 GeV
Explore spin dynamics in unique regime
Bench marking of theoretical predictions
Sharp drop-off!
E [GeV]
P [%
]
LEP1
TRISTAN
HERAPETRA
VEPP4, DORIS II, CESR
SPEAR
VEPP-2M, ACO
LEP2
0
25
50
75
100
0 10 20 30 40 50 60 70 80 90
With...Without...
Harmonic Spin Matching
Transverse spin-polarization in LEP
10/25/2001 Cornell October 2001 44
Precise determination of the LEP beam energy (10-5 relative accuracy, ~ 1 MeV)Precise measurement of the Z mass and width
∆ E
[M
eV]
11. November 1992
∆ E
[M
eV]
29. August 1993
Daytime
11. October 1993
-5
0
5
23:00 3:00 7:00 11:00 15:00 19:00 23:00 3:00
-5
0
5
11:0
0
13:0
0
15:0
0
17:0
0
19:0
0
21:0
0
23:0
0
18:0
0
20:0
0
22:0
0
24:0
0
2:00
4:00
6:00
8:00
Axis of earth rotationCERN, Geneva
Moon
Ecliptic
Small changes of energy accurately measured(energy change from 1mm circumference change)
Use of Polarization at LEP:
10/25/2001 Cornell October 2001 45
MeV6486.440E
=ν 519109.31 ντ
λ ⋅⋅== −
p
6 26.67 10E
Eνσσ ν ν−= ⋅ ≈ × ⋅
( )
2 22
22 2,
( / )1118
k mp
k md
w T
k m
γ
γ γ
ντν
τ ν ν ν
∆=
− − −
∑
−⋅
= 2
2
2
22
2exp
2 γ
ν
γ
ν
νσ
νσ
mm IT
13
2
<<=γνλναCondition for correlated
spin resonance passings:true
Spintune
Polarizationbuildup rate
Spin tunespread
Theory by Derbenev, Kontratenko, Skrinsky (With LEP Parameters):
Synchrotron tune γν
Resonance strength 2102 1094.1 ν⋅×≈ −kw
P [%
]
E [MeV]
440 MeV
ν
N
σE = 31 MeV (LEP I)
σE = 51 MeV (DWIG)
σE = 125 MeV (LEP II)
0
10
20
30
40
50
60
70
80
44400 44500 44600 44700 44800 44900 45000
0
0.2
0.4
0.6
0.8
1
100.75 101 101.25 101.5 101.75 102 102.25
SITFSODOMsimulations
10/25/2001 Cornell October 2001 46
0
10
20
30
40
50
60
35 40 45 50 55 60 65 70 75 80
Pol
ariz
atio
n [%
]
Energy [GeV]
SITF simulations (50 seeds)
τp/τd = (ν/88)4
τp/τd = (ν/95)2
First order theory: Includes spin resonances with kx, ky, ks=1
Nkkkkkkk syxssyyxxdepol ∈⋅±⋅±⋅±= ,,,νννν
Machine tunes
Synchrotron sidebandsdetermine polarizationdegree in LEP
Simulation confirms 1/E4
dependence of polarization!
Energy Dependence of Polarization:
10/25/2001 Cornell October 2001 47
0
10
20
30
40
50
60
70
80
40 50 60 70 80 90 100
Pol
ariz
atio
n [%
]
Energy [GeV]
Linear
Higher order
Measurements• With 90/60, 60/60 and 102/45 optics.
• Goal for energy calibration: > 5%
• Polarization not always fullyoptimized.
1998: Polarization and energy calibrationhas been extended to 60.6 GeV (P = 7% measured)!
Drop in polarization degree consistent with higher-order theory…
Polarization Measurements in LEP:
10/25/2001 Cornell October 2001 48
44.7 45.0 47.0 51.3 57.4
Equivalent E [GeV]
Higher-order
Linear
Bl [Tm]
P ∞ [
%]
0
10
20
30
40
50
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Wigglers increase spin tune spread and thus allow “simulating” energy increase...
Higher-Order Theory also Confirmed with Wigglers:
10/25/2001 Cornell October 2001 49
With: 26
GeV44065.01076.6
⋅⋅= − E
νσ• LEP enters uncorrelated regime with high
energy and small Qs!
• If spin resonance passing is uncorrelatedit is completely uncorrelated for LEP!
• We can stay in the correlated regime byincreasing the value of Qs!
Evaluate Correlation Criteria for LEP:
10/25/2001 Cornell October 2001 50
MeV6486.440E
=ν 519109.31 ντ
λ ⋅⋅== −
p
6 26.67 10E
Eνσσ ν ν−= ⋅ ≈ × ⋅
( )
2 22
22 2,
( / )1118
k mp
k md
w T
k m
γ
γ γ
ντν
τ ν ν ν
∆=
− − −
∑
−⋅
= 2
2
2
22
2exp
2 γ
ν
γ
ν
νσ
νσ
mm IT
13
2
<<=γνλνα γν νσ >>
[ ]
24 22
3 2
108 exp( 2 )11 154 11
p
d
w νν
τ σπ ντ π π ν λ
− −= ⋅ ⋅ ⋅ +
2p k
d
wτ πτ λ
=
1<<νσ
Condition for correlatedspin resonance passings:
true
true
false
false
Condition for completeuncorrelation
true
Spintune
Polarizationbuildup rate
Spin tunespread
Synchrotron tune γν
Resonance strength 2102 1094.1 ν⋅×≈ −kw
Theory by Derbenev, Kontratenko, Skrinsky (with LEP
parameters):
10/25/2001 Cornell October 2001 51
Expected polarization: Very low, but possible increase at high energies?
New polarization optics (101.5/45 degrees) for measurements at low AND high energy
Sharp drop after LEP1 in agreementwith theory/simulations.
Transverse spin polarization crucialfor precision measurements of theW and Z properties (energy calibration)
First measurement in regime of uncorrelated spin resonance crossing.No sign of transverse polarization.
New varieties of Harmonic SpinMatching gave up to 57% polarization.
We can trust the polarization theories in LEP regime!
Precise predictions for future projects…
Achievements at LEP:
10/25/2001 Cornell October 2001 53
MeV6486.440E
=ν 21 51 8.7 10p
λ ντ
−= = ⋅ ⋅
6 22.4 10E
Eνσσ ν ν−= ⋅ ≈ × ⋅
( )
2 22
22 2,
( / )1118
k mp
k md
w T
k m
γ
γ γ
ντν
τ ν ν ν
∆=
− − −
∑
−⋅
= 2
2
2
22
2exp
2 γ
ν
γ
ν
νσ
νσ
mm IT
13
2
<<=γνλναCondition for correlated
spin resonance passings:true
Spintune
Polarizationbuildup rate
Spin tunespread
Theory by Derbenev, Kontratenko, Skrinsky (With VLLC33 Parameters):
Synchrotron tune γν
Resonance strength 2102 1094.1 ν⋅×≈ −kw
Build-up time τp: 1.9 h
Spin tune ν: 417.5
Spin tune spread σν: 0.42
Synchrotron tune: 1/7
10/25/2001 Cornell October 2001 54
What does this mean for VLLC?
0
25
50
75
40 60 80 100 120 140 160 180 200
Pol
ariz
atio
n [%
]
Energy [GeV]
LinearLEP HO
VLLC33 HOMeasurements
Large spin tune spread Enhancement of depolarization(as in LEP at high energy)
10/25/2001 Cornell October 2001 55
P [%
]
E [MeV]
440 MeV
ν
N
σE = 31 MeV (LEP I)
σE = 51 MeV (DWIG)
σE = 125 MeV (LEP II)
0
10
20
30
40
50
60
70
80
44400 44500 44600 44700 44800 44900 45000
0
0.2
0.4
0.6
0.8
1
100.75 101 101.25 101.5 101.75 102 102.25
10/25/2001 Cornell October 2001 56
Linear / higher-order theory for different Qs…
Qs = 0.2: Expect sufficient polarization up to 80-85 GeV!
Raising Qs improvespolarization for high energies!
Why?Imagine Qs = 1
Qs satellites overlayinteger resonances
(ν = k + i · Qs)
High Qs for LEP
10/25/2001 Cornell October 2001 57
0
10
20
30
40
50
60
40 60 80 100 120 140 160 180 200
Pol
ariz
atio
n [%
]
Energy [GeV]
Ultra-highenergy
Qs = 1/5Higher-order theory
5 %
Uncorrelated passingsof spin resonances
(small Qs)
Spin tune spreadσν >> 1
(probably not true at 100 GeV)
Theory predicts: Polarization comes back at ultra-high energies!
Why? Fast increase of polarization build-up, increase in depolarization slows down!
Very uncertain regime (who knows what really happens)…
Polarization increase at Ultra-high energies:
10/25/2001 Cornell October 2001 58
Strong transverse damping: Very nice beam dynamics regime (performance)
- Less tails- Less effects from resonances (we can jump them)- Ramp colliding beams at high energy- Higher beam-beam limitTwo thirds of all LEP luminosity collected in the last 3 years (out of 10.5y)
LEP data would indicate a beam-beam limit of 0.17 for VLLC33.
Optimization of vertical orbit to the limit (dispersion/coupling correction for LEP)
Need operational overhead in RF voltage (>= 6 % in LEP) - optimize # klystrons
Do not expect significant radiative spin-polarization (even linear level is very low)
Some preliminary thoughts:
10/25/2001 Cornell October 2001 59
Sociology
• Good support from equipment groups, good motivation, close interaction with machine in-house expertise.
• Common control room – operations as focus for machine physicists, equipment groups and experiments.Regular informal contact at all levels.
• Comprehensive annual workshops - Chamonix.• Cross-fertilisation from other labs.• Stimulated by close contact with experimental physicists.• Makeup of operations. Ph.Ds on shift