Transcript LHC Emittance Preservation V. Kain, G. Arduini, B. Goddard, B. J. Holzer, J.

LHC
Emittance Preservation
V. Kain, G. Arduini, B. Goddard, B. J. Holzer, J. M. Jowett,
M. Meddahi, T. Mertens, F. Roncarolo, M. Schaumann,
R. Versteegen, J. Wenninger
1
Introduction
LHC
Performance of collider is measured in luminosity
kN f 
L

F
*
4  
2
Small  to maximize
luminostiy
Produce small  in
injectors → need to
keep  small in the LHC
Will present analysis of emittance preservation from SPS extraction to
start of LHC collisions (mainly 50 ns p+)
2
Injector Performance
LHC
o
Impressive performance of injectors
−
From design report (25 ns):
PSB 2.5 mm – PS 3 mm – SPS 3.5 mm
0.5 mm
0.5 mm
Points to be addressed:
o
PSB: poorer
performance ring 1,
ring 2
o
PS: more blow-up with
higher intensity
o
SPS: larger
measurement spread
at extraction
50 ns 2011: blow-up PSB to SPS 0.3 – 0.4 mm
3
LHC Performance
LHC
Analysed ~ 60 fills between mid July to mid August (50 ns, 1.2 ×1011, *=1.5 m)
Comparison of convoluted emittance from LHC luminosity with SPS wire scan for
144 bunches:
On average ~ 20 - 30 % growth between SPS flattop and
collisions
4
Measurement Limitations 2011 (1)
LHC
Results of analysis based on SPS/LHC wire scanners, LHC synchrotron light
monitor and luminosity of ATLAS and CMS
o
Wire scanners in the SPS:
−
Measured in the SPS routinely when setting up (intermediate 12 bunches + nominal 144
bunches)
−
No synchronous measurement SPS-LHC
−
Large spread in short time scales
Preparation of fill 2240
> 50 % spread
o
Wire scanners in the LHC:
−
Measured routinely intermediate batch (12 bunches)
−
Sometimes measured first nominal batch (144 bunches) – intensity limit at 2.5 × 1013 p+
−
No measurements at 3.5 TeV for physics beam
5
Measurement Limitations 2011 (2)
LHC
o
o
Continuous beam size measurement with Synchrotron Light Monitor (BSRT):
−
Bunch-by-bunch: 3 s per bunch → 69 min (!) per ring;
−
Good for relative measurement under same conditions

Absolute calibration not obvious (wire scanners are used for cross-calibration)

Cannot compare data at different energies – cannot see effect of ramp

Could not compare data for different beams/planes
Emittance from luminosity
−
Single emittance value for different beams and planes
−
Not always fully optimized /not publishing the correct value
−
Assumes Gaussian beams
6
Improvements for 2012
LHC
o
Long list of measurement improvements planned for 2012
o
Instrumentation:
o
o
−
Faster and better calibrated BSRTs
−
Commission Beam-gas Ionization Profile Monitor (BGI)
−
Pre-prepared wire scanner hardware settings for circulating intensity
−
Bunch-by-bunch wire scans in the SPS
Methods:
−
Automatic wire scans during ramp
−
Synchronous measurements across the accelerator complex
Analysis and GUIs:
−
Make measured betas available
−
More reliable fits
−
Write fit results into logging database
7
LHC
CURRENT UNDERSTANDING
OF LHC EMITTANCE BLOW-UP
8
Mismatch at injection?
LHC
Used:
o
Wire scan data from SPS 12 bunches for fills 1960 - 2025
o
Wire scan data from LHC beam 1 12 bunches for fills 1960 - 2025
LHC/SPS = 1.07 ± 0.11
LHC/SPS = 0.99 ± 0.12
Emittances conserved at injection within measurement
accuracy
9
Beam 1 versus Beam 2
LHC
o
Comparing beam 1 with beam 2 wire scans of 12 bunches
LHC1/LHC2 = 0.96 ± 0.08
LHC1/LHC2 = 1.06 ± 0.08
Using the measured beta functions at wire scanners:
Beam 1 and beam 2 emittances are consistent
10
Growth @ 450 GeV
LHC
o
Emittances are growing at injection – reasonably consistent with IBS although
slighty faster
Fill 1897, beam 1
BSRT gated over 12 bunches
Fill 2028, beam 1
BSRT gated over single bunch
Simulations of IBS, uncoupled T. Mertens
Horizontal emittance: ~ 10 % in 20 minutes
Filling about 30 minutes → D/ 0 – 10% in H
More studies to come in 2012 to understand
faster growth of bunch length and transverse
emittance
11
→ Effect on bunch-by-bunch luminosity
LHC
A. Ryd, CMS
Specific luminosity 10 %
lower for first bunches
consistent with growth
injection
12
The Ramp
LHC
o
Cannot use BSRT data, BGI was not commissioned
o
Dedicated fills are necessary: low number of bunches to do wire scans through
the ramp
First indication: Abort gap cleaning test fill (12+12+12 bunches, 50 ns)
H0 = 1.6 mm
V0 = 1.3 mm
H0 = 1.6 mm
V0 = 1.4 mm
Used measured  at injection and flattop and linear interpolation between
Blow-up during the ramp: measurement indicates > 20 % all planes
13
The Ramp - continued
LHC
Last BI MD in 2011:
Ramp of 4 bunches per ring – different emittances
Unfortunately no useful wire scan data for beam 1
emittance D ~ 1 mm
emittance D ~ 0.7 mm
> 20 % growth different,
Relative
but…different emittances grow by the same amount
Although larger than for 12 bunch trains
14
Possible sources for growth @ ramp
LHC
o
Ramp still needs to be further optimized
−
o
Chromaticity,…
Example proton
Fill 2254 ramp?
(1380 bunches per beam)
Effect of reduced damper
gain during
damper gain hor beam 1
(linear scale) HIGH @450 GeV
before prepare for ramp
drop of damper
gain
increase of damper
(electronic) gain in ramp
To maintain approx.
same damping rate
Reduce damper gain
before ramp for tune
feedback
→Increase of BBQ
amplitude
ramp (energy)
increased BBQ amplitude
= more residual beam oscillations
=> potentially leading to blow-up;
but signal needed for tune feedback
which is switched on here
BBQ hor Beam 1
amplitude
W. Hofle
Injection plateau
prepare for
ramp
ramp
15
Effect of damper gain change
LHC
Emittance evolution during fill with single bunch (loss map fill).
Evolution during injection and switch of damper gain during ramp preparation:
@ 450 GeV
@ 450 GeV
3.5 min
Effect of damper gain change not clear (data inadequate)!
Required test 2012: change gain and wait at injection
16
Emittance blow-up at 3.5 TeV
LHC
o
Emittance evolution from BSRT
o
Looked at the Abort Gap Cleaning Test Fill (12 + 12 + 12 bunches per ring)
−
o
Integration: 3 s per bunch; averaged over 12 bunches → good resolution
Blow-up during squeeze for beam 1 H
−
Took measured beta at 3.5 m and 1 m from optics team into account!!
−
Was there anything different in this fill?
17
Blow-up during the squeeze
LHC
o
Looked at more fills since Evian workshop
−
if emittances of all bunches similar
evolution in time
→
bunch-by-bunch scan of BSRT gives emittance
Bunch emittances similar
Looked at several fills for end of 2011
→ BSRT data consistently shows blow-up for beam 1 H
between * 5 m and 1.5 m. No obvious source.
Needs to be followed up during squeeze commissioning 2012
18
Dependence on bunch intensity?
LHC
Approximately constant absolute growth between SPS
extraction and LHC collisions for different bunch intensities
19
Ions
LHC
o
Ions also experience blow-up during the ramp
o
Wire scans during the ramp not possible with physics beam
Data taken during quench test on 7th of December, single bunch analyzed
Wire Scanner data - B1 - horizontal plane
normalized Eps. X (μm.rad)
2.5
2
Blow-up of ~ 20 %
1.5
Similar to proton trains
1
Ramp
0.5
0
14:15
14:20
14:25
14:30
14:35
14:40
14:45
14:50
Time
20
Summary
LHC
o
Excellent performance of injectors: 1.9 mm for 1.5 × 1011 p+ per bunch
o
Emittances grow 20 % - 30 % from SPS extraction to LHC collisions
o
LHC injection: emittances preserved within measurement accuracy
o
LHC injection plateau: emittance growth apparent (~ 10 % in 20 minutes for
horizontal plane)
−
Continue to minimise time at injection
−
Dedicated filling cycle? Reduce time in “prepare ramp”…
o
LHC ramp: blow-up in H and V > 20 % (50 ns), more for single bunches
o
LHC squeeze: blow-up for beam 1 H, > 20 %
o
2012: understand and (hopefully) correct blow-up
−
Expect improvements from measurements, tools and analysis
−
Need to sort out during commissioning: INSTRUMENTS, blow-up @squeeze, ramp
optimization, minimize time at injection
−
MDs: effect of damper gain and working point,…, IBS
21
LHC
EXTRA SLIDES
22
Challenge for 2012
LHC
Understand and correct LHC
blow-up
…Make (good) outliers
routine
G. Arduini
23
Beta errors for emittance
LHC
o
Error emittance only taking error on beta function into account.
o
Plot 3 sigma error
24
Dependence on bunch intensity?
LHC
beam 2 similar
High pile-up MD, fill 2201 (2.4e+11):
Delta similar to BI MD,
but more than with 12 bunch trains
Delta of ~ 1 mm in H
Delta of ~ 0.6 mm in V
H0 = 2.4 mm
V0 = 2.2 mm
Ib = 2.4 x 1011
25