BeamCurrentMonitors JeanClaudeDenard (SynchrotronSOLEIL) - - PowerPoint PPT Presentation

BeamCurrentMonitors

JeanClaudeDenard

(SynchrotronSOLEIL)

DITANETSchoolonBeamDiagnosticTechniques 30March– 3April2009 RoyalHollowayUniversityofLondon(UK)

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 2

☼ Electromagnetic field associated tocharged particle beams

Summary

☼ Destructivemonitors:faradaycup …. ☼ Nondestructivemonitors;electromagnetic interaction

Wall current monitors Current transformers Cavity monitors,SQUID

☼References

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 3

LongitudinalEFieldDistributionof aPointChargeina Conducting Tube

• Chargeproduces Efield
• Efield induces image

chargesof oppositesign on the wall Note:there is no Efield

• utside the tube

Ultrarelativistic charge

σw ≈0

v≈ c

• The Wall Current distributionis

the imageof the beam distribution butof oppositesign and without DCcomponent

Static charge

a

v=0

rms length σw =a/√2

• Moving Efield creates Hfield

inside the tube

• The imagechargesmove along

with the inner charge.

• Wall current +beam current =0.

Then,Ampère’slaw, ∫ (Hdl)=iindicates that H=0

• utside the tube(except forDC

field).

Moving charge

σw =a/γ√2 a

v=ßc

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 4

Example of Wall Current LongitudinalDistributionfora Point!like Moving Charge

55GHz 7.5GHz 1.8GHz Wall current BWlimitation(*) 2.9ps 21ps 90ps ∞ rms length (ps)=a/(β γc√2) 0.9mm 6mm 15mm 18mm σw =a/(γ√2) 0.999 0.94 0.55 β =(11/γ2)½ 20.6 3 1.2 1 Lorentzfactor γ 10MeV 18GeV 1MeV 1.8GeV 100keV 184MeV kinetic energy Eforelectrons forprotons

Numerical Examples with aTubeDiam.2a=50mm

(*) The actual distributionis not gaussian,butforthe sake of simplicity,its Bandwidth has beenapproximated tothat of agaussian distributionof same rms length

Ifσl >>a/ γ√2,wall current distribution=beam distributionσl

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 5

Fieldsassociatedtoachargedparticlebeamfora beamlengthσl >>a/ γ√2

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 6

TransverseFieldDistribution

i dl H =

∫

) . (

r

E

Static charge v=0

ρ

Moving charge v<c

no magnetic field inside magnetic field appears

r

ρ E H

r

ρ E H

Ultrarelativistic charge v≈ c

• utside the pipe:no Efield

no magnetic field (except DC)

ty permeabili

• magnetic
• vacuum
• ty

permittivi

• electric
• vacuum
• impedance
• vacuum
• 377
• /
• With

: wave TEM 2 / ) (

• r
• radius

at . : Law s Ampère'

• =

= = Ω = = = = → =

∫

µ ε ε µ η η π

• H

E

• r

i r H i dl H

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 7

TEMWave inVacuumChamber is Like inanAirFilled CoaxialTransmissionLine

☼ Similarities:TEMwave carries the same EMenergy (Pointing vector P=E×

H):

☼ amonitorcan be realistically tested inacoaxialline structure. ☼ Some differences:

At High frequencies (cutoff frequencies aredifferent inthe two cases)

i(t)

E H

P

E H

P

i(t)

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 8

FaradayCup

ammeter i Current Source

• Destructive
• Absolute measurement of DCcomponentwith anammeter
• AnoscilloscopeorSample &Hold measures the peak current in

caseof pulsed beam.

• Canbe used forthe calibrationof nondestructivemonitorsthat

provide relativemeasurements.Forexample FCcalibrates anRF cavity current monitoronCEBAFinjector (CWsuperconducting Linac).

gun ammeter e Faradaycup

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 9

FaradayCup;DesignIssues

☼ Absolute accuracy is usually around 1%,it is difficult toreach 0.1%. ☼ Needs toabsorb all the beam:blockwith largeentrancesize and thickness >>

radiationlength.AFCbuilt at DESYand presently used onalow current 6 GeV beam at JLAb uses1m3 of lead (12tons).

☼ Backscattered particles (mostly e):narrow entrancechannel,bias voltageor

magnetic field redirect the backscattered e onthe FC.Accuracy evaluation requires MonteCarlosimulations(EGGSfrom SLAC;GEANTfrom CERN).

☼ Power (W)=EinMeV × Iin]A.

Example:5MeV FCinCEBAFinjectorwith200]ACWbeam→ 1000W. Acoolingcircuittakesthepowerout.Theisolationisdonewithdeionized waterandinsulatingrubbertubes.

☼ Safetyissues:FCneedstobealwaysterminatedbyaDCcircuittoavoid

arcingandapotentiallydangeroushighvoltagethatwoulddevelopatcable end.ApairofhighimpedancediodescanbeconnectedinparallelontheFC

• utput.

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 10

SLSWide Bandwidth CoaxialFaradayCup (0!4GHz)

M.Dachetal.(SLS); BIW2000 beam 50ohm Coaxial structure

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 11

Calorimeter

☼ Calorimetry refers toadirectmeasurement of the totalenergy delivered toa

massiveblockof metal (silver ortungsten)over aperiod of time.

☼ Totalenergy is determined bymeasuring the temperature rise of the object if:

The average beam energy is precisely known Any energy losses can be accounted forbyreliable calculation ordirect measurement.

☼ Acalorimeter hasbeendeveloped forCEBAFCWbeam (A.Freyberger,to

be published)

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 12

Wall Current Monitor:

Beam and Wall Current Spectra forUltraRelativistic Beams Timedomain ↓ Frequencydomain ←ibeam→ ←iwall→

t

t T τ

T τ

1/T 1/2πτ πτ πτ πτ

f f DC NoDC

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 13

Wall Current Monitor:Concept

r

iwall ibeam vout vout =iwall*r r iwall vout

t

vout

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 14

WCM:FromConcepttoActualImplementation

☼ Forvacuumquality:ceramic gap ☼ ris made of several resistors in

parallel,distributed around the gap. Special chipresistors still behave asresistors inthe GHzfrequency range.

☼ Electrical shield:avoids parasitic

external currents flowing through r, and prevents the beam EMfield from radiating outside the monitor.

☼ High ]material fills the space

between gapand shielding forlow frequency response.

r

Cgap

iwall

]

iwall

L

logamplitude

L r

gap

rC 1

f π ω 2 =

) (log scale ω

]

) / ( 2 a b Ln h ; ; L

r

• π

=

WCMresponse

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 15

Implementation Example:6kHzto6GHzWCM:(R. WeberBIW93)

Ceramic gapand surfacemounted resistors ferrites ferrites Resistors arenot pureresistors at high frequencies r

Resistor with 5mmwire connections ≡

r

≈ 1pF ≈ 6nH

Inthe GHzrange,standardresistors arereplaced bySurfaceMounted Resistors that havesmaller inductanceand capacitance.

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 16

6kHzto6GHzWCM:(R.Weber)

☼ r=1.4ohm(80resistors inparallel). ☼ rCgap circuitat high frequencies

Ceramic gapconsidered asalump capacitor: dm=90mm;t =3.2mm;and w =4.5mm =>Cgap =33pF Ceramic gapbehaves asaradial transmissionline matched toits 1.4a characteristic impedance:fh >6GHz (measured)

m r gap

d w S t S C π ε ε 2 * ≅ =

t(thickness)

5 . 9

• :

alumina ≈

r

ε

w

dm

GHz 4 . 3 2 1

• and

= =

gap h

rC f π

]H 40 with kHz

• 6

. 5 2 = = = L L r flow π

Shield =1turn

h

a

b

) / ( 2 a b Ln h ; ; L

r

• π

=

☼rL circuitat low frequencies

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 17

WCM:OutputConnection;Beam PositionDependence

☼

Offcenter beam yields higher wall currents near the beam.

☼

There is adifference signal (iwall top – iwall bottom)that propagates around the gap.The wave velocity is reduced bya factor √εr =3with alumina.

☼

With our example,propagationtime is 2.8ns. Positiondependence starts around f>>300 MHz.We want no positiondependence up to several GHz!

☼

Apractical solutionis tocombinesignals from fourquadrants.

E

combiner Vout

t

☼

A20GHzWCMs is indevelopment at CERN. 3Delectromagnetic simulationcodes(MAFIA, GDFIDL,HFSS)arenecessary fordamping the high frequency componentsof the wake fields without loading significantly the useful signal (L.Soby etal.,EUROTeVReport2006104,2006).

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 18

Beam Current Transformerinlow and mid frequency range

☼ Beam is a1turnprimary winding

The magnetic field is : H≈ beam current / toroid circumference ib

Primary

☼

Beam current ib is « transformed »into ib/nonthe secondary winding.n=10or20are common values.

☼

Vout =ib/n * [R/(1+R/jn2L1ω)];with L1=oneturn selfinductance

☼

Inmid rangefrequencies,L1ω >>R,and Vout ≈ Rib /n

☼

Allbeam current transformers need aninsulating gapinorder toleave the magnetic field reach the toroid.

R/n2 L1

Secondary

ib/n

R n2L1

vout

beam

R

vout

Η

nturns

2 4 6 8 10

2 4 6 8 10

time

beamcurrentib responsevout

timecte R/n2L1

ib

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 19

WCM=Current Transformerwith n=1

r

ib ]

≡ ≡

ib r

Primary and Secondary equivalent circuits

) / ( 2 a b Ln h

• L1

r

• π

π π π = = = =

]

The electrical shield is equivalent toa1turnwinding Butthe high frequency analysis is better done using the WCMconcept

Electricalshield crosssection

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 20

Integrating Current Transformer(ICT)=Beam Charge Monitor

ib

3beampulsesofsamecharge(areaQb =∫ ib.dt )

0,1 1 0,01 0,1 1 10

Pulsespectra

time

Ib (t) Ib (f)

f LowPass Filter R

ib/n

n2L1

vout

t (Transformer&filter)impulseresponse

vout

0,1 1 0,01 0,1 1 10

Filterbandwidth<<beampulsebandwidth

f

Filterbandwidth Amplitudepulsechargewhen responsetime>>beampulseduration

forshortpulseslongrep.rate: warmLinacs &transferlines beam

Η

nturns

LowPass Filter vout

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 21

FastCurrentTransformer(FCT)

beam

R

n=10to20turns 50Ω coaxialcable

Highfrequencyequivalentcircuit

• Thegapcapacityusuallysetsahighfrequencylimit.
• RC representstransformercorelosses,itdependson

frequencyandfieldamplitude

• Rs istheseriesresistanceofthesecondarywinding
• L2istheleakageinductance
• CistheadditionofstraycapacitiesandCgap/n2;a gap

madeofAl2O3,afewmmthickisusual(Cgap≈30pF).

• L3istheinductanceofthecoaxialcableconnection
• Coaxialcablehasincreasinglosseswithfrequency
• Veryhighfrequencylossesoccurintotheshield

cavitywherethebeamexcitesmanymodes.

• Thehighfrequencyequivalentcircuitdoesnottake

intoaccountthecavitymodelosses.Wakefieldloss evaluationpossiblewith3Dsimulationcodes(GDFIDL…)

is=ip/n

n2L1

RC Rs

L2

Cgap/n R= 50Ω

L3 vout Cgap ib

0,1 1 0,01 0,1 1 10

F(GHz)

ib vout spectra

Example:Soleil BoosterandTranfer Line

• Transformerbandwidthisnarrowerthanbeamspectra.
• Partofthesignalpowerislostathighfrequencies

intothemagneticcore.

• Thereislittlepowertakenfromthebeam;thenbeam

stabilityormagneticcoreheatingarenoissues.

Last2turns inSoleilBoosterand first turn inStorage Ringseen onFCTs

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 22

FCTIssuesonhighcurrent StorageRings

☼

Magneticcoreheatedover110°Cwitha 300mA beam(¾ Ringfilling)

☼

Theheatingproblemhasbeensolvedbyinstalling additionalcapacitorsinparallelonthegapand improvingtheaircoolingofthetoroid.

☼

Butthereducedbandwidthaffectsthe information:thesignaldoesnotreturnto zerobetweenbunchesat2.8nsintervals

☼

Resistorsaroundthegap(likeinaWCM)can reducesthepowerenteringthecavity.

☼

Anothersolutionistofillthecavityspacewith microwaveabsorbingmaterial.

☼

Athoroughstudyofthesesolutionswouldneed GDFIDLandHFSSsimulationsandplentyof time.

☼

Agoodandcheapsolutionseemstousea photodiodeilluminatedbythesynchrotron radiationeasilyavailableinthediagnostic hut.

☼

Then,thegapcouldberemovedwhichwillreduce thehighordermodelossesonthebeam.

Example:Soleil StorageRingFCT

time Ringrevolution period

FCTresponse photodioderesponse

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 23

DCcurrentMonitors:DCCTs,alsocalledPCTs

☼

DCCTs areimportantmonitorsforStorageRings. The1µA resolutionisusuallylessthan1E5of thebeamcurrent.WithaCWsuperconducting Linac beamof200µA (CEBAF),the1µA resolutionisonly0.5%.

☼

ADCCTisazeromagneticfluxsensor.Itfeeds backaDCcurrentinthebeamcurrentopposite directioninordertocancelthemagneticfluxina setoftoroids.Aprecisionresistorinthefeedback currentpathyieldstheoutputvoltage.

☼

TheDCmagneticfluxinducedbythebeaminto themagneticcircuitsdoesnotdependonthebeam position(Ampère’s law).

☼

DCCTs arecommerciallyavailable.

☼

Thezerofluxsensingisaffectedbyexternal magneticfields:magneticshieldings arenecessary

☼

Thezerocurrentdriftswithtemperature:power losses,especiallythroughthegapmustbeprevented.

☼

Isolationgapcanbeverynarrowandgapcapacitor high(afewnF).

☼

GapExample:Elettra andSoleil havecustommade isolationgap.Itisacapton foilsandwichedbetween twoaluminumhalfgaskets.

IssuestoaddressbeforeinstallingaDCCT

N.Rouvière vacuumgasket design

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 24

OtherNonDestructiveDCMonitors

☼

CryogenicCurrentComparator (A.Peters,GSI).Usedforlowcurrentionbeams.ACCCuses aSQUIDasnulldetectorforthemagneticfield(SQUID=SuperconductingQuantum InterferenceDevice).TheSQUIDdetectsextremelysmallchangesofthemagneticfield.A fractionofnA resolutionfor100nA beamshasbeenobtained.Itperformsanabsolute measurement.Like withalltransformers, Ampere’slawmakesitindependentofbeam position.Butit averydelicateinstrumenttoimplement.

☼

Cavitycurrentmonitor forCWbeammeasurement.Itisapassivecavityinfundamentalmode likeanacceleratingcavity.Theoutputpickupvoltageisproportionaltothebeamcurrent. Likeinalinac wherethebeamenergydoesnotdependonbeamposition,theoutputpowerof apassivecavitydoesnotdependonbeampositioninaratherlargecentralarea.Itissensitive tonA beamsbutneedsanexternalcalibration. Example:CEBAF,stainlesssteellowQcavities(Q≤ 8000)forlowcurrents,calibrated againstaDCCTat≈ 100µA.

DITANETSchoolon BeamDiagnostics BeamCurrentMonitors JeanClaudeDenard 25

References

☼ Muchofthecontentofthislecturehasbeenextractedfromthefollowing:

WebberR.C.,“TutorialonBeamCurrentMonitoring”BIW2000,pp 83101. Also“ChargedParticleBeamCurrentMonitoringTutorial”BIW1994,pp323 and“LongitudinalEmittance:AnIntroductiontotheConceptandSurveyof MeasurementTechniques,IncludingDesignofaWallCurrentMonitor”BIW 1993. R.Shaffer,BIW1993. Unser K.,noteISRCO/696,March1969;EPAC1990;BIW1991. HofmannA.,“FrontierofParticleBeams;Observation,Diagnosis and Correction”.Proceedings,Anacapri 1988. Talman R.,BIW1993. Littauer R.,USacceleratorschool,SLAC1985. PetersA.etal.BIW2008. Denard JC.etal.PAC2001. Cassinari L.,privatecommunications.