Download - CALIFES linac

1C. Simon CTF3 Meeting

CavityBPM for CALIFES

C. Simon, M. Luong, D. Bogard, W. Farabolini,

J. Novo

CTF3 Meeting – 27-29 January 2009

227/01/09 CTF3 Meeting

CALIFES linac

CA

.BH

B 0

40

0

CA

.SN

H 0

11

0

CA

.SN

I 0

12

0

CA

.VV

S 0

10

0 CA

.DH

G/D

VG

01

30

CA

.VV

S 0

20

0C

A.M

TV

02

15

CA

.BP

M 0

22

0C

A.D

HG

/DV

G 0

22

5

CA

.SN

G 0

23

0

CA

.SN

G 0

25

0

CA

.BP

M 0

24

0

CA

.BP

M 0

26

0

CA

.BP

M 0

31

0

CA

.BP

M 0

38

0

CA

.DH

G/D

VG

02

45

CA

.DH

G/D

VG

02

65

CA

.DH

G/D

VG

03

20

CA

.DH

G/D

VG

03

85

CA

.DH

B/D

VB

02

70

CA

.IC

T 0

21

0

CA

.VV

S 0

30

0

CA

.QF

D 0

35

0

CA

.QF

D 0

36

0

CA

.QD

D 0

35

5

CA

.BP

R 0

37

0

CA

.BP

M 0

41

0CA

.MT

V 0

39

0

CA

.MT

V 0

42

0C

A.V

VS

05

00

CA

.SD

H 0

34

0

CA

.DH

B/D

VB

02

30

CA

.DH

B/D

VB

02

50

ACS 0230 ACS 0250 ACS 0270

DUM

0450

TBTS

ITL

CA

.MT

V 0

12

5(v

irtu

al c

ath

od

e)

CA

.FC

U 0

43

0

6 BPMs are installed on the CALIFES linac


Reentrant Part

Reentrant Cavity BPM for CALIFES

Bent coaxial cylinder designed to have:

- a large frequency separation between monopole and dipole modes

- a low loop exposure to the electric fields

• It is operated in single and multi-bunches modes

• The cavity is fabricated with titanium and is as compact as possible :

~125 mm length and 18 mm aperture 4 mm gap


RF characteristics

Eigen modes

F (MHz) Ql (R/Q) (Ω) (R/Q) (Ω)

Calculatedwith HFSS

in eigen mode

Measured in the CLEX

Calculatedwith HFSS

in eigen mode

Measuredin the CLEX

CalculatedOffset 5

mm

CalculatedOffset 10

mm

Monopole mode

3991 3988 24 26.76 22.3 22.2

Dipole mode

5985 5983 43 50.21 1.1 7

• RF characteristics of the cavity: frequency, coupling and R/Q

Due to tolerances in machining, welding and mounting, some small distortions of the cavity symmetry are generated.

This asymmetry is called cross talk and the isolation is evaluated > 26 dB.

BPM

• Similar results from one BPM to another, and in-situ results

are comparable to Saclay in-lab measurements

Monopole and dipole transmission measured by the network analyzer

Monopole and dipole reflection measured by the network analyzer


Signal Processing Hybrids installed close to BPMs in the CLEX

Multiport switches used to have one signal processing electronics to control six BPMs.

Analog electronics with several steps to reject the monopole mode

BPMs in the probe beam linac tunnel

beam positionmonitor 1

x1

x2

y2

y1

180°hybrid

180°

hybrid

phaseshifter

3 dBcombiner


x1

x2

y2

y1

180°hybrid

180°

hybrid

3 dBcombiner

.

.

.

6 to 1multiportswitch

phaseshifter

x

y

6 dB

6 dB

18 dBm leak.

18 dBm leak.

Filter 5997 MHz600 MHz BW

0 to 70 dB

0 to 70 dB

18 dBm leak.

Lowpass Filter120 MHz

0 to 70 dB

3 dBsplitter

reference2.891 GHz

14 dBm

RF electronics in the hall

8channel

videoamp

Gv = 10

4channel10 bits2 Gs/sADC


DAQ in the hall

control unit

VME Crate

signals for switchesand variable

attenuators control

Ix

Qx

Iy

Qy


RF

LO

II-Q

demodulator

Q

RF

LO

II-Q

demodulator

Q



AcqirisDigitizers

Hybrid

couplers

RF electronics used synchronous detection with an I/Q demodulator.


Control Command of BPMs1/Sampling with acqiris boards and readout signals on OASIS

3/Results sent on graphic windows under JAVA and in a file to be used with Matlab

6 BPMs implemented

sincos QI

2/Signal processing: Digital Down Conversion

- raw waveform multiplied by a local oscillator of the same frequency to yield a zero intermediate frequency - real and imaginary parts of each IF are then multiplied by a 60 coefficient, symmetric, finite impulse response (FIR), low pass filter with 40MHz 3dB bandwidth

P

A

Qsin

22 QIA

Beam position

A

Icos


Signal in 32 bunches mode behind RF electronics

Signal in single bunch mode behind RF electronics

Signal voltage determined by the beam’s energy loss to the dipole mode.

Dipole mode signal depends on frequency fi and external coupling Qi of this mode

Simulations (1)

Φ(t) = heaviside function, q = bunch charge, R0 = 50 Ω, (R/Q)i = coupling to the beam and ζi = 2 (dipole mode)

iaiQ

tiatiiQ

tiVtiS

ii..2

).sin(...cos)

.2

.exp()(

iQ

RqQRiV

i

oii

.

..)/.( 22

24

11

iQiia

single bunch

Sd(n)

Dipole mode signal in 32 bunches mode

Dipole mode signal in single bunch mode

~ 30 ns~10 ns

Rdn

0

32

I

Phi n Te( )

Vod e

2 Fd1n Te

2 Qd1

cos ad1 n Te( ) 2 Fd1sin ad1 n Te( )

2 Qd1 ad1

32 bunches


Re-entrant Cavity at CALIFES

Environment Noise RMS: 66 µV (measured at FLASH)

Δ signal (gain adjusted to get an RF signal ~ 0 dBm, simulated in single bunch) with 5 mm offset : 590 mV (simulated), Noise: 0.5 mV (calculated) Resolution : 3.2 µm (simulated)

with 0.1 mm offset : 555 mV (simulated), Noise: 0.1 mV (calculated) Resolution : 100 nm (simulated)

Noise is determined by :

Thermal Noise :

kb = Boltzmann constant, BW (Hz) = Bandwidth, T (K) = Room Temperature.

Noise from signal processing channel :

with

Simulations (2)

.** thn PGNFP

BWTb

kPth **

...*

11

21

3

1

21

GG

F

G

FFNF

Pth = Thermal noise, NF=Total noise figure of the signal processing, Fi and Gi respectively the noise factor and the gain of component i.


BPM calibration (1)

All BPMs have same electronics and same losses coefficients should be the same

Charge calibrated with charge calculated by ICT

CA

.BP

M 0

22

0

CA

.IC

T 0

21

0


Beam studies (1)

channels

Amplitude of signals corresponds to 6 nC with 150 bunches determined thanks to ICT

• τ too long on channel (green) problem with diode

• Impedance problem? -> Move lowpass filter

BPMs in the probe beam linac tunnel


x1

x2

y2

y1

180°hybrid

180°

hybrid

phaseshifter

3 dBcombiner


x1

x2

y2

y1

180°hybrid

180°

hybrid

3 dBcombiner

.

.

.

6 to 1multiportswitch

phaseshifter

x

y

6 dB

6 dB

18 dBm leak.

18 dBm leak.


0 to 70 dB

0 to 70 dB

18 dBm leak.


0 to 70 dB

3 dBsplitter

reference2.891 GHz

14 dBm

RF electronics in the hall

8channel

videoamp

Gv = 10



DAQ in the hall

control unit

VME Crate

signals for switchesand variable

attenuators control

Ix

Qx

Iy

Qy


RF

LO

II-Q

demodulator

Q

RF

LO

II-Q

demodulator

Q



AcqirisDigitizers

ICT


Beam studies (2)

Charge ~ 6 nC read by the first and second BPM. ~ 100 % transmission behind 1st section

channels

y I and Q channels

Signals minimized

beam offset minimized


Beam studies (3)

Channel

y I and Q channels

• Signal frequency: 100 MHz

• Shape of y shows in the pulse train :

- charge not constant

or

-beam position not constant

Simulation of I and Q signals


BPM calibration (2)

Magnets switched off between steerers and studied BPM to reduce errors and simplify calculation.

Move beam with one steerer in horizontal and vertical frame.

Average of 500 points for each steerer setting.

R = transfer matrix from steerer to BPM

Calculate for each steerer setting, the relative beam position in using a transfer matrix between steerer and BPM :

Dx = R12*Dx’(angle at steerer)

CA

.BP

M 0

380

CA

.DH

G/D

VG

032

0


Summary

Reentrant cavity BPM features for CALIFES:

• Operated in single and multi-bunches

• Single bunch resolution potential < 1 µm

• Charge of beam measured

Software under development.

First beam seen by BPMs Dec. 2008

New beam tests end of March 2009


Acknowledgements

Thanks to CERN and CEA teams

Thanks to Pascal ContrepoisAline CurtoniStéphane DeghayeMick DraperPatrick GirardotFrancis HarrraultPierre-Alain LeroyFabienne OrsiniFranck PeaugerAnastasiya Radeva Lars Soby

Thank you for your attention