Direct measurement of the transverse long-range wake-field of

Walter Wuensch SAREC review meeting 30-31 January 2012

Direct measurement of the transverse long-range wake-field ofthe CLIC main linac accelerating structure

G. De Michele1,2,3, E. Adli1,4, A. Grudiev1, A. Latina1, D. Schulte1, W. Wuensch1

1CERN, 2PSI, 3EPFL, 4The University of Oslo

W. WuenschSAREC review meeting

30-31 January 2012


MotivationThe main performance issues for the CLIC main linac accelerating structures are:

• Accelerating gradient• Wakefields and higher-order-mode suppression• rf-to-beam efficiency

(The three issues are marvelously interconnected)

The performance of the accelerating structures drives the efficiency and cost of CLIC.

We dedicate significant resources to demonstrate high-gradient performance.

Walter Wuensch SAREC review meeting 30-31 January 2012Break down rate scaled to 100 MV/m, CLIC pulse

At KEK and at SLAC

High-gradient testing

Status end 1011


We now propose an experimental program to investigate long-range wakefields. Why? Do we expect surprises? Well no, but…

Wakefields , and specifically higher-order-mode suppression, is probably being computed very accurately. But we’re currently close to beam dynamics limits and as we push the design towards lower cost we will depend on the accuracy of the simulations more and more.

Also accelerating structures are complex objects, for example with lots of silicon carbide loads. The only experimental verification of a waveguide damped wakefields was in ASSET in late 1999 – we feel it is essential to check our latest design, material choice etc.

Finally the wakefield measurement is a concrete exercise for our beam dynamics team and their tools. It can only be healthy for those dealing with wakefields in simulation to deal with them in a real machine.

Motivation 2


There is a long tradition of measuring accelerating structure

wakefields in ASSET

XB-10, ICFA Mini-Workshop, R.M. Jones, Cockcroft Inst., 30th Nov-3rd Dec 2010 66

DS

Qcu

RDDS1

ASSET Data

Conspectus of GLC/NLC Wake Function Predictionand Exp. Measurement (ASSET dots)

DDS3 (inc 10MHz rms errors)DDS1

RDDS1 H60VG4SL17A/B-2 structure interleaved

Refs: 1. R.M. Jones,et al, New J.Phys.11:033013,2009. 2. R.M. Jones et al., Phys.Rev.ST Accel. Beams 9:102001, 2006.3. R.M. Jones, Phys.Rev.ST Accel. Beams, Oct.,2009.

1. GLC/NLC Exp vs Cct Model Wake

An Asset Test of the CLIC Accelerating Structure, PAC2000

Higher-order mode damping demonstration in

ASSET 1999

150 cells/structure, 15 GHz

24 cells/structure, 12 GHz(loads not implemented yet)

Then

NowDouble band circuit model

C L I CC L I C

30 January 2004Alexej Grudiev, GdfidL for TDS wakefield calculations

Full length TDS results comparison


The CLIC baseline accelerating structure and its wakefield


The CLIC baseline structure

Average loaded accelerating gradient 100 MV/mFrequency 12 GHzRF phase advance per cell 2π/3 rad.Average iris radius to wavelength ratio 0.11Input, Output iris radii 3.15, 2.35 mmInput, Output iris thickness 1.67, 1.00 mmInput, Output group velocity 1.65, 0.83 % of cFirst and last cell Q-factor (Cu) 5536, 5738First and last cell shunt impedance 81, 103 MΩ/mNumber of regular cells 26Structure length including couplers 230 mm (active)Bunch spacing 0.5 nsBunch population 3.72×109

Number of bunches in the train 312Filling time, rise time 67 ns, 21 nsTotal pulse length 243.7 nsPeak input power 61.3 MWRF-to-beam efficiency 28.5 %Maximum surface electric field 230 MV/mMaximum pulsed surface heating temperature rise

45 K

Hs/EaEs/Ea Sc/Ea2

Wakefield simulations• Different EM codes have been used to simulate the wake-

fields in the CLIC RF prototype with dispersive materials: GdfidL, CST Particle Studio®, ACE3P

• Different materials (i.e. different EM properties) have been tested in order to meet the beam dynamics requirements

11

CST: superPC, 128 GB of RAM and 24 CPUs

GdfidL: distributed and shared memory

Transversewake-field

The predicted wake-field

12

Transversereal impedance

Transversewake-field

Three main dipolar bands in theImpedance spectrum

Absolute valueof the envelope ofthe transverse wake

2nd bunch

What we need from measurements

• Roll-off speed to be confirmed up to 0.25m• Wake-field threshold• Identification of frequencies from wake-field

measurements: high sampling of the wake-fields with the witness bunch

13

Roll-off

threshold

14

Critical issue for simulation in time domain: dispersive materials• GdfidL

– The permittivity can be expressed with an N-th order Lorentz medium with resonant angular frequencies ωn and damping angular frequencies γn:

– This feature has been debugged in late 2011!

• CST 2012 version– the transient solver has undergone many improvements, in particular with respect to,

e.g., broadband constant loss tangent. The broadband sensitivity analysis with respect to geometry and material property variation has also been enhanced.

4th order Lorentz medium


Wakefield measurements of the CLIC baseline accelerating structure in ASSET/FACET

• Possibility of having driving and witness bunches with positrons and electrons.

• Adjustable bunch spacing for a timing span behind the driving bunch.

• Bunch length flexibility: ideally shorter than 1mm in order to resolve the 3rd dipole band which shows up a peak around 40GHz.

16

Why FACET?

Layout of the ASSET facility

POSSIBLE LOCATION

Where in FACET

17

ASSETCLASSE

LI02 sector


Young people from the CLIC study (not me) are becoming familiar with the linac and the wakefield

measurement

Experimental procedure

19

dw

ccdw

y ytWE

eeLQqwd

22

22

tW

dy

wq

dQ

LwE

transverse wakefield

drive bunch offsetwitness charge

drive bunch charge

active length of the structureenergy of the witness bunch

The transverse wake-field can be extrapolated from the measurements of the deflecting angle of the witness bunch

y

d

wdrive bunch lengthwitness bunch length

Transversewakefield

y

Experiment simulations with tracking code PLACET

• BPMs upstream and downstream of thestructure are needed in order to determine the bunch offset in the structure, as well asdipole corrector magnets to generate this offset.

• only BPMs downstream of correctors are sensitive to the correction itself. That means that the lattice response is triangular.

20

Response matrix from thesouth extraction kicker in SRD to LI02

A. LatinaG. De Michele

21

Positrons: offset scan to drive wakefield

The BPMs just before and after the RF prototype are LI02, 201 and LI02, 211. The correctors used are only two upstream these two BPMs. With only two correctors seems possible to shift the positron beam in the RF prototype. To be verified the maximum voltage that can be applied to the correctors.

Electrons: Deflected by wakefield• We can compute the witness bunch deflections from betatron oscillation

fits to data from the BPMs downstream of the test structure, and corrected for incoming orbit jitter using the results from similar fits to data from BPMs in the NRTL line [C. Adolphsen et al., PRL 74 13, 1995]

• For the fit of trajectory parameters, one needs the knowledge of BPM locations and lattice parameters, i.e. quantities which have no stochastic errors. For the fit of lattice parameters, however, one has to correlate BPM data with trajectory parameters, i.e. quantities which will in general have significant stochastic noise components. In this case, the fit must take into account the noise on the measured trajectory parameters due to finite BPM resolution [SLAC-CN—371].

• Once we have simulated all the experiment with the tracking code PLACET, our routines will be Interfaced with the SLAC Control Program (SCP). Hopefully in March 2012 first tests.

22

23

BPMs resolutionFrom the experiment proposal:

An accurate data analysis is required

• Good knowledge of the optics is essential• Orbit fit to determine the kick

y

Lattice simulations and verification in 2012 run

• The experiment can be reproduced numerically with all imperfections. Data acquisition as well as data analysis can be simulated.

• From the physical point of view the experiment can be completely simulated.

• We have only deal with technical problems and software integration on SCP.

24


The test structure – a simplified assembly since the dipole mode

Q’s are low

main strongback in the ASSET area

FACET prototype scheme

26

• six TD26 accelerating structures• simple vacuum tank• no power will be sent to the test structure and the double feed couplers will be terminated by the loads

• clamped aluminum cells• available length: 76.80 inches

vacuum tank

1.5m

76.80 inches

beam pipe transition

bellow

vacuum

pump

vacuum

pump

Support and alignment system

vacuumflange

ASAS

ASAS

ASAS

CLIC accelerating structure for ASSET

6 sections, 24 cells each, compact coupler designClamped aluminum disksTo be tested at SLAC

Ø180mm

TD26 with Compact Couplers (x6), clamped

Cell shape accuracy 20µm

Aluminium diskDamping

material (SiC)

Threaded rod for AS clamping

Clamp for SiC fixation

Aluminium disks

Features for Tank alignment on the girder

Features for AS alignment in the Tank

Reference spheres BEAMA. Solodko

• Uses existing vacuum tank • Estimated total load of the CERN installation is about 400 kg• Two vacuum ion pumps (noble diode pumps, 53-55 l/s)• The RF structure is aligned mechanically inside the vacuum tank• The vacuum tank will be aligned to the beam axis by means of adjustable feet

(drawings needed)• Two tooling balls on both ends are present on top of the vacuum tank

28

Mechanical design: features

CLIC meeting Walter Wuensch 6 May 2011

What full prototype cells look likeMachining: OFHC copper diamond milled and turned disks with micron

precision.

+- 5 micron tolerance lines

G. Riddone, S. Atieh

• Available longitudinal space 76.80inches• RF structure disks diameter: 180mm• RF structure longitudinal dimensions: 6 X

250mm• Beam pipe: 1inch outer diameter• Vacuum flanges: 2 ¾ inches• Distance between beam-pipe center to main

strongback: 8.795inches• Accelerating cell inner radius: 2.35mm

30

Mechanical design: dimensions


Schedules

32

Coordination with FACET runs(may already be obsolete)

• 2012 - CLIC personnel participate in linac measurements in preparation for wakefield measurements.

• 2013 - After commissioning of positrons, we have two options:– Re-measure the installed DDS then install the CLIC RF

structure– Install and measure only the CLIC RF structure

• But in any case the CLIC RF structure will be ready for September 2012.

Overall Schedule2012 2013

Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4Changing machine parameters and BPMs data acquisitionShipping of the CERN prototype to SLACFinal assembly at SLAC

Re-measurements of the DDS structures which are still in place in LI02InstallationDirect measurements of wake-fields for CLIC RF structure

33

Direct measurement of the transverse long-range wake-field of

Documents

Transcript of Direct measurement of the transverse long-range wake-field of