ff

Quadrupole Pick-Ups at CERN & Fermilab

A. Jansson FNAL

Andreas Jansson, US-LARP meeting, May 9, 2003 2

ff

Talk outline

The quad pick-ups in the CERN PS

The quad pick-up in the Fermilab AA

Possibilities for LHC (and Tevatron)

Andreas Jansson, US-LARP meeting, May 9, 2003 3

ff

What is a quadrupole pick-up?

A pick-up sensitive to the r.m.s. beam size.

Uses the small non-linear terms in electrode response to particle position to measure quadrupole moment.

Quadrupole moment is a measure of ellipticity.

Beam

Pick-up seen along beam path

A

BD

C

Vacuum chamber

position horizontalDCBA

DB

x

moment quadrupoleDCBAC)(A-D)(B 2222

x yxy

Andreas Jansson, US-LARP meeting, May 9, 2003 4

ff

PS pick-ups

Induction loop

Pick-up seen along beam path

A

BC

D

Flux lineBeam

Magnetic coupling. Insensitive to radiation Signal on 50

Intensity signal is suppressed by pick-up geometry (coupling to the radial field component).

Bandwidth ~25 MHz (covers full bunch spectrum at injection)

Two pick-ups installed in machine.

Andreas Jansson, US-LARP meeting, May 9, 2003 5

ff0

20

40

60

80

100

120

140

Number of machine revolutions

Qua

drup

ole

mom

ent (

mm

^2)

QPU4

QPU3

Beam: EASTBDate: March 28, 2001

Ouput signals (Time Domain)

Oscillations give information about mismatch

DC value give information about emittance

(Data corrected for beam position)

Andreas Jansson, US-LARP meeting, May 9, 2003 6

ff

Output signals (Frequency Domain)

Peaks are wide due to fast decoherence (caused by space charge tune spread).

Working point in PS machine often make signals overlap in frequency domain.

Need two pick-ups in optically different locations to separate H/V quad signal components! 0

0.05

0.11

0.16

0.22

0.27

0.33

0.38

0.44

0.49

0.55 0.6

0.66

0.71

0.77

0.82

0.88

0.93

0.99

Frequency (tune)

Sign

al p

ower Qh Qv

2Qh 2Qv

Andreas Jansson, US-LARP meeting, May 9, 2003 7

ff

Position contribution to quad moment

Data with only beam position oscillations (taken after filamentation).

Quadrupole moment versus its expected position contribution x2-y2 should be straight line with unit slope, as obesrved.

The position contribution can be subtracted with good accuracy!

y = 0.9699x - 3.8003

-20

-15

-10

-5

0

5

10

15

20

-20 -10 0 10 20

Position contribution (mm^2)

Qua

drup

ole

mom

ent (

mm

^2)

Andreas Jansson, US-LARP meeting, May 9, 2003 8

ff

Comparison with Wire Scanners

Comparison Quad PU vs. Wire-scanner on stable beam.

Several different beam types.

Systematic error bar from: Beta function ~10% Dispersion ~10% Mom. spread ~3%

QPU3

0

20

40

60

80

100

120

140

160

180

0 20 40 60 80 100 120 140 160 180WS + tom oscope [m m2]

QP

[mm

2]

SFTPROEASTBEASTCADLHCEASTAlab meas

QPU4

-20

0

20

40

60

80

-20 0 20 40 60 80WS +tom oscope [m m 2]

QP

[mm

2]

Andreas Jansson, US-LARP meeting, May 9, 2003 9

ff

)(

)()(

2

2

2

222

22

22

yxx

x

yx

q

y

q

x

q

xxp

q

yyy

q

xxx

yx

DDDDD

0

10

20

30

40

50

60

-1 1 3 5 7 9

Turn

Qua

drup

ole

mom

ent [

mm

2]

10 free parameters, 20 data points

Measurement of matching

Simultaneous fit to the two pick-up signals gives: Injected emittances. Betatron mismatches. Horizontal dispersion

mismatch.

“Best fit” tunes gives information on space charge.

Fixed tune give wrong fit results for matching parameters.

Andreas Jansson, US-LARP meeting, May 9, 2003 10

ff

Injection matching measurement Betatron mismatch

Dispersion mismatch

k

'DD

D

kD

Emittance

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Horizontal

Vert

ical

ReferenceQNO40 -10A

QNO40 -20A

Horizontal Matching

-0.5

0

0.5

-0.5 0 0.5

db/b

a db

/b -d

a

Vertical Matching

-0.5

0

0.5

-0.5 0 0.5

db/b

a db

/b -d

a

Dispersion Matching

-0.5

0

0.5

-0.5 0 0.5

dD/b^1/2

a dD

/sqr

t(b)

+ d

D' s

qrt(

b)

Andreas Jansson, US-LARP meeting, May 9, 2003 11

ff

Measurement of filamented emittance For a stable beam, the

emittance can be calculated from only two pick-up readings.

22yx

0

0.5

1

1.5

2

2.5

QPU 3/4 WS H54 WS H64

Emitt

ance

(um

)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

QPU 3/4 WS V65

Emitt

ance

(um

)

Horizontal

Vertical

Wire-scanners

Quad PUs

Statistical error

22

2222

21

2111

xpyyxx

xpyyxx

D

D

Different horizontal/vertical beta function ratios at the two pick-ups are required.

Signal noise can be reduced by averaging over many turns.

Andreas Jansson, US-LARP meeting, May 9, 2003 12

ff

Measurements within the bunch

Normalized for intensity in each point separately.

Variation in quad moment along bunch mainly due to dispersion and momentum spread.

QPU4

-30

-20

-10

0

10

20

30

-150 -50 50 150

time [ns]

quad

mom

ent [

mm

2]

QPU3

50

60

70

80

90

100

110

-150 -50 50 150

time [ns]

quad

mom

ent [

mm

2]

raw data

corrected for dispersionBunch shape

-150 -50 50 150

time [ns]

curr

ent

Momentum spread

00.20.40.60.8

11.21.41.61.8

-150 -50 50 150

time [ns]

mom

. spr

ead

1e-3

Andreas Jansson, US-LARP meeting, May 9, 2003 13

ff

Fermi Pbar Accumulator Q-pickup

Un-terminated strip-line. Motors to center pickup on beam. LP preamps (only 1st harmonic) in

tunnel. Can inject calibration signal to

balance preamps. No hybrid used. Plate signals

sampled directly on 14 bit ADCs at up to 10 MHz (16x per revolution).

Injected beam is a train of typically 7-35 bunches at 53 MHz (1/12-5/12 of the circumference).

Ref: Vladimir Nagaslaev

Andreas Jansson, US-LARP meeting, May 9, 2003 14

ff

Lessons from beam measurements Initial transient signal due to

intensity step and limited bandwidth.

Beam loss on electrodes at injection.

Fast decoherence due to large chromaticity at extraction orbit.

This makes it hard to interpret data from parasitic measurements

With lowered chromaticity (dedicated measurements) performance is adequate.

Ref: Vladimir Nagaslaev

Andreas Jansson, US-LARP meeting, May 9, 2003 15

ff

Measurements with low chromaticity

Ref: Vladimir Nagaslaev

Time domain data fits the model

Frequency componenets behave as expected when changing steering and matching

Andreas Jansson, US-LARP meeting, May 9, 2003 16

ff

Quad pick-ups in the LHC? Are they wanted/needed? What is the time resolution required?

Single turn (BW=frev) is the minimum, and may be adequate e.g. for dedicated measurements (pilot bunch?). Relatively straight-forward to build.

Single bunch (BW=frep) is desirable e.g. for parasitic measurements during normal operation. Requires some R&D.

Intra-bunch resolution (BW=1/tbunch) helpful to diagnose e.g. head-tail motion, but may not be needed if beam is properly set up. May be difficult to achieve for LHC (~GHz).

The better the time resolution, the easier it is to understand what’s going on (less dependent on assumptions).

Andreas Jansson, US-LARP meeting, May 9, 2003 17

ff

How build a wideband quad pick-up?

The PS pick-up’s common mode rejection is limited at high frequencies by resonances in the magnetic induction loop.

Reducing the loop dimensions could perhaps gain a factor 3-4 in frequency, but also reduces coupling.

Removing the resonances require breaking the loop and adding matched terminations strip-lines.

With a good WB hybrid, can get >60dB CMRR over ~200 MHz bandwidth (see CERN PS/BD Note 99-09).

Andreas Jansson, US-LARP meeting, May 9, 2003 18

ff

Wide-band hybrid coupler

Four-decade hybrid,J.M. Belleman,CERNPS/BD/Note 99-09

60dB

200 MHz

Andreas Jansson, US-LARP meeting, May 9, 2003 19

ff

Summary In the PS:

Betatron mismatch of a few percent could be detected (both amplitude and phase).

Emittance was measured, in rather good agreement with wire-scanners.

In the Fermi pbar accumulator:Clean signals (mismatch) obtained in dedicated studies.Difficult to interpret parasitic measurements.

For the LHC:Single turn (narrow band) should be straight-forward.Single bunch (wide band) needs some work, but seems

possible using e.g. stripline couplers and WB hybrids.Tevatron could act as test-bench and benefit from LHC

design work.