Novel Instrumentation for future TeV e+e- colliders Detectors for bunch length measurement and Beam loss monitoring Mayda Velasco, Anne Dabrowski, Alex

Novel Instrumentation for future TeV e+e- colliders

Detectors for bunch length measurement

and Beam loss monitoringMayda Velasco, Anne Dabrowski,

Alex CarterNorthwestern University

DOE site visitSeptember 2006

A. Dabrowski, September 05 20061/33

Overview CTF3 Facility

A. Dabrowski, September 05 2006

Drive Beam Injector

Drive Beam Accelerator

PETS Line

30 GHz source Stretcher

Delay Loop TL1 (2006)

CR (2006)

TL2

(2007)

CLEX

(2007)

30 GHz tests

RF photo-injector test

(2006-2007)

Electron beam facility• 1.5microsecond pulse• 150MeV electron LINAC• 3.5 Amp current • Newly commissioned delay loop• Very rich environment for beam instrumentation

development

2/33

Overview Northwestern CTF3 Activities


Drive Beam Injector

Drive Beam Accelerator

PETS Line

30 GHz source Stretcher

Delay Loop TL1 (2006)

CR (2006)

TL2

(2007)

CLEX

(2007)

30 GHz tests

RF photo-injector test

(2006-2007)

‘Gallery’

Timing & 100MHz ADC’s

0-2kVPower supply

Cables Patch Panel

Beam position monitor

Acceleratingstructure

Quadrupoles

e-

‘Tunnel’

Beam Loss Monitoring

Pickup for bunch length measurement

3/33

Outline

• RF pickup for bunch length measurement

– Principle of the measurement – Advantages of this type of bunch length measurement?– Status of the hardware designed, installed and tested in

2005/2006– Hardware and design improvements still to be installed

before the Fall data taking.– Work planned for the Fall:

• Final setup installed, tested• Analysis performed in normal beam operation

• Beam loss Monitoring

– Reminder devices installed and fully instrumented since 2003

– Ongoing work in optimization of setup



Principle of the measurement I

The frequency spectrum of the bunch contains all the information about the bunch shape and length.

The field induced in the waveguide has a temporal evolution directly related to that of the electron bunch

Hence it is possible to derive the spatial charge distribution from the frequency spectrum of the field excited in the waveguide

In this measurement, one takes advantage of the power spectrum in the frequency domain. For a given beam current, the larger the power spectrum amplitude, the shorter the bunch length.

The RF-pickup detector measures the spectrum of the electromagnetic field of the bunch collected by a rectangular RF-pickup connected to a waveguide.

5/33

Principle of the measurement II

22

)( e

The smaller the bunch length the larger the signal at a given frequency

The smaller the bunch length the larger measured power

To reach the < ps sensitivity, need to have reach of high frequencies

Solid: σt = 1 ps

Dash: σt = 2 ps

Dash-dot: σt = 3 ps

Pow

er

Sp

ectr

um

[a

.u.]

Pow

er

Sp

ectr

um

[a

.u.]

• Consider an electron bunch of N particles with longitudinal charge distribution Λ(z)• The frequency spectrum is found by Fourier transformation the fields, e.g.



• Advantages– Non-intercepting / Non destructive– Easy to implement in the beam line– Relatively low cost (compared to streak camera and RF

deflector)– Relatively good time resolution (ns) follow bunch length

within the pulse duration– Measure a single bunch or a train of bunches– Relative calibration within measurements

• Short comings in the calibration– Beam position sensitive– Sensitive to changes in beam current

• If used with the RF deflector and/or a streak camera (method of measuring bunch length at CTF3)– Can provide an excellent cross calibration of device

Advantages of the RF-Pickup

7/33


• An RF pickup was installed in CTF2– Rectangular waveguide coupled to a rectangular

hole made on the beam pipe surface– Using the mixing technique it measured bunches as

short as 0.7ps. It was limited by a maximum mixing frequency of 90 GHz.

• This device was dismantled in 2002, and is no longer being used at CTF3.

• Goal is to re-install the device with an improved design– Increase maximum frequency reach to max mixing

at 170 GHz, to reach bunch length measurements of 0.3ps

– Design a ceramic/diamond RF window for good vacuum and transmission at high frequency

– Spectral analysis by single shot FFT analysis from a large bandwidth waveform digitizer

RF-pickup device installed in CTF2

C. Martinez et al, CLIC note 2000-020

8/33


mixing technique in CTF3

Install one WR-28 pickup, with WR-28 waveguide to transport the signal to the gallery

Use a series of filters, and waveguide pieces on the same mounting board in the gallery to separate the signal from the beam in the following frequency ranges:

WR-28 Waveguide components, and mixer at 26.5 GHz

WR-12 Waveguide components and mixer at 56.5 GHz

and at 74.5 GHz

(27 – 39) GHz

(46 - 54) GHz

(75 – 90) GHz

(60 – 72) GHz

WR-6 Waveguide components with mixer at 157 GHz

(143 – 157) GHz

(157 – 170) GHz

Recuperated from CTF2

NEW

NWU9/33

Investment in new hardware


Components purchased by Northwestern– Local oscillator (Down

converter) • LO 157 GHz • RF 142-177 GHz

– D-band waveguide components (waveguide WR-6 1.65 mm x 0.83 mm, cutoff 110 GHz)

• D-band Horn (gain 20dB)• D-band fixed attenuator (10

dB)• D-band waveguide 5cm Status:

Installed and tested - OK

10/33

Investment in new hardware – Filters


• Fabrication of Brass filter at 157 GHz– Conical edge at angle of 20 degrees,

thickness 3 x diameter of hole• Used as a high/low pass filter to study

frequency band:– 157 GHz-14 GHz – 157 GHz+14 GHz

• Other filters to be used from CTF2 setup: 26.5 GHz, 40 GHz, 56.5 GHz, 74.5 GHz

Scale (mm)

Status:Installed and tested - OK

11/33

Investment in new hardware – Digitizer


Acqiris DC282High-Speed 10-bit PXI/CompactPCI® Digitizer•

4 channel

3 GHz bandwidth with 50 Ω high-frequency front end

2 GHz bandwidth with 50 Ω standard front end

2-8 GS/s sampling rate

Auto-install software with application code examples for C/C++, MATLAB, National Instruments LabVIEW etc

Provide a single shot FTT in real timeStatus:Currently being installed

12/33

Installation of Pickup and waveguidesInstallation of Pickup and waveguides

• October 2005 Shutdown

• Pickup with Al203 window (3.35 mm thickness) installed in the CT line

• Signal from BPR0523 split, and ½ taken to the gallery along ~ 15m of WR-28 waveguide


CT

.BH

D 0

510

CT

.BP

M 0

515

CT

. WC

M 0

525

CT

.VV

S 0

512

CT

.QD

D 0

520

CT

.QF

D 0

530

CT

. PH

M 0

560

CT

.MT

V 0

550

CT

S.M

TV

060

5

CT

.BH

E 0

540

Dum

p

Dump

CT

.DH

D05

05

CD

.DH

L 01

01

CD

.BP

M 0

105

CD

.VP

I 010

4

CT

.BP

R 0

532

13/33

Installed analysis station in GalleryInstalled analysis station in Gallery


1st mixing @ 26.5 GHz

2nd mixing @ 56.5 GHz

3rd mixing @ 74.5 GHz

Analysis board for

(28 – 90) GHz

14/33


Examples of Signals from May 2006Examples of Signals from May 2006

Mixer: 56.5 GHzLO: 9.6 GHz

Peaks: 0.1 GHz Beam signal 66 GHz

Equipment

recuperated from

CTF2 installed

correctly &

functions well

Mixer: 26.5 GHzLO: 9.6 GHz

Peaks: 0.1 GHzBeam signal 36 GHz

15/33

Plots Alex Carter


Signal for analysis to the Oscilloscope

Input signal from the Local Oscillator for the second stage of mixing

Beam signal from WR-28 waveguide

Output of 1st mixing stage at 157 GHz

157 GHz brass filter.

Analysis of

(143 – 170) GHz

(setup in transmission or reflection)

Installed analysis station in GalleryInstalled analysis station in Gallery

16/33


Analysis of Signals from May 2006Analysis of Signals from May 2006

Mixer: 157 GHzLO: 11.4 GHz

Peaks: 0.4 GHzBeam signal 168 GHz !

New 157 GHz mixer works well!

Showed proof of principle that power measurements at high frequency can be done at > 150 GHz!

Reach sub-pico second bunch length measurement

17/33

Optimization of Window

rf

c

Material Thickness Epsilon

Al203 window 3.35 ± 0.07 mm 9.8

CVD diamond window 0.500 ± 0.005 mm (~6 at 30 GHz measured @ CERN


@ 90 GHz through Al203 λ is effectively ~ 1 mm

Al203 window not optimized for good transmission designed a thin (0.5mm) diamond window with lower εr. (~6 at 30 GHz measured @ CERN)

•Design prototype completed.

•Machining and brazing test happening now.

plan install window in machine before Fall data taking.

0.5mm

rf

c

18/33

Work needed before Fall data taking

• Complete machining and brazing of new diamond Window.

• Redesign analysis board in gallery:– All 4 mixing stations read out at once with newly

installed large bandwidth waveform digitizer.– Careful attention must be paid to alignment

• Write software:– For new Acquiris data acquisition.– To remotely control the local oscillator

• Write analysis code to use the new system for data taking during the run.


Calibration of device - Fall data taking


Chicane optics & bunch length measurements - 2004

Magnetic chicane (4 dipoles)

RF Deflector Screen

Betatron phase advance(cavity-profile monitor)

Beta function at cavity and profile monitor

Beam energy

RF deflector phase

RF deflectorwavelength

Deflecting Voltage

Bunch length

y0

y

Deflecting mode TM11

RF deflector off RF deflector on

Slide from Thibaut Lefevre (CERN)

20/33



Bunch Length Measurement with the Streak Camera

OTR@ MTV0550SR@ MTV0361

Two different bunch compression settings in the Frascati Chicane

= 8.9ps = 4.5ps

Maximum Sweep speed @ 10ps/mm

21/33



OTR@ MTV0550SR@ MTV0361

= 8.9ps = 4.5ps

Betatron phase advance(cavity-profile monitor)

Beta function at cavity and profile monitor

Beam energy

RF deflector phase

RF deflectorwavelength

Deflecting Voltage

Bunch length

y0

y

Deflecting mode TM11

Calibration Strategy:

For various settings on the chicane, take bunch length measurements using both the RF deflector, and the RF-pickup. Calibrate the response function of the pickup.

Once calibrated, the pickup can be installed anywhere else in the machine where a bunch length measurement is needed.

The RF-pickup is a much less expensive device than the RF deflector & Streak camera.

Should have better resolution than the Streak camera.22/33

Summary of work on Bunch Length detector


• An RF-pickup has been installed in the CT line in CTF3• The setup should provide a bunch length measurement

up to 0.3ps• The RF-pickup is installed close to the RF deflector for a

good cross calibration of bunch length.• Proof of principle during May 2006:

– Measuring high frequencies with a WR-28 waveguide pickup was possible!

– The Mixer at 157 GHz was tested and works well.– The filter machined at 157 GHz works well.

• Improved RF diamond window for high frequencies is being machined and brazed, and on schedule to be installed in machine before Fall data taking.

• The new acquiris data digitizing scope is ready for installation, and programming of the DAQ is schedules for the fall.

• The RF-pickup device is flexible, and can be installed in various locations once calibrated

23/33

BLM s installed along the LinacBLM s installed along the Linac



Quadrupoles

e-

Yz

x

Typical Linac section

Design beam optics

Goal : To install the BLM at a position where the beam could be lost

Install 3 chambers and 1 Faraday Cup per Girder which have an accelerating structure

A. Dabrowski, November 24 2004

Chamber installed

24/33

Beam Loss monitoring


‘Gallery’

Timing & 100MHz ADC’s

0-2kVPower supply

Cables Patch Panel



Quadrupoles

e-

‘Tunnel’

•Full setup since 2003

•3 detectors installed per girder.

•Cross calibration with Faraday cup

•Electronics with 2 gain ranges (x10 difference)

25/33



Goal is to measure secondary shower, and from there determine if possible where the original source of the shower was:

From Geant3.21 simulation [1]

Secondary e+/e- flux distribution at 100cm from the point of the beam loss in the X/Y plane transverse to the beam line

Secondary e+/e- flux versus the longitudinal position. Each curve corresponds to the output signal amplitude of each of the 4 detectors located at different z positions along a LINAC module.

[1] M. Wood, CTF3 note 064, (2004)

26/33



Energy distribution of photons and electrons produced by 0.1% loss of 25MeV electrons in the LINAC coming out at an angle of 30mrad

Signals in the detector will have dominant contributions from both photons and electrons.

Cross sections, at low photon energy < 1 MeV photoelectric effect; > 10 MeV pair production.

Signals has contributions from Photoelectric effect, secondary emission and ionisation27/33

Calibration PlateauCalibration Plateau for chambers filled with He or Ar gas taken in the CTF3 machine, normalized to the Faraday cup

Linear response between Chamber and Faraday cup

Calibration Factor ~ 6 depending on beam loss shower shape


0 100 200 300 400 5000

2

4

6

8

10

12

He Ar

Am

plitu

de (

a.u.

)

Bias Voltage (V)

4000 5000 6000 7000 8000 9000 10000-24

-16

-8

0

V= 450V V= 400 V V= 300 V V= 200 V V= 100 V V= 0 V

Out

put

Vol

atg

e (m

V)

Time (ns)

Length of Plateau greater for Ar as expected. He breaks down at > 450 V

28/35

Girder 6

Girder 5

Beam loss with nominal beam opticsChamber signals for nominal beam optics have good signal and time response with only low x20 gain low x20 gain amplifieramplifier (and 50 Ω resistor)

Linear response between Chamber and Faraday cup

Calibration Factor ~ 6 depending on beam loss shower shape

System Provides a Mapping of beam loss in Z along girder !


Girder 6

High gain (x200) for the electronics

Girder 5

Very Low lossesVery Low losses on Girder 5 /6

Losses at the limit of the BPM resolution … signals on Girder 6 visible with high gain electronics, but on Girder 5 we are not sensitive to losses BUT perhaps there were no losses?

High gain (x200) for the electronics


Very Low Losses on Girder 6

gain (x200(x201K)) for the electronics

Data taken with very low loses in the beam …

Modified the electronics to increase the sensitivity… 50Ω input impedance to the amplifier replaced by 1kΩ (x20 more)

• still with the amplifier gain of x20 or x200

• Drawback : we loose the time resolution (x20 slower)

Since the Calibration Factor between SEM and Faraday cup X5, and input impedance factor is x20, here we are in the less than ‰ region of beam loss A. Dabrowski, November 24 2004

In this condition… SEM sees the beam loss, but the Faraday Cup in the noise.

31/35

Alternative detector for region of low beam loss

1000 1500 2000 2500 3000

-300

-200

-100

0

100

PMT as a Cherenkov monitorA

mpl

itude

(m

V)

t (ns)

Signal from PMT

Relative calibration with Faraday Cup x200 at 300 V


In the 1st cavity

Girders 11/12 …

Energy ~ 60-100 MeV•From Simulation, beam loss is more collimated at higher energy …

•A PMT as Cherenkov detector was installed in the 1st Cavity, together with a Faraday Cup.

•Change PMT voltage, depending of sensitivity required

•PMT have relatively low dynamic range (103) within a given voltage gain•At 300V bias voltage, the PMT is x200 more sensitive than the Faraday cup

•Here we estimated the loss to be lower than 10-4 of the beam current

32/33

Summary … Beam Loss on Linac


Chambers are sensitive to beam loss along the girder …. Additional monitoring complementary to BPM has been used since 2003 In regions of very low loss, Chambers sensitive if one sacrifices resolution, and integrates over large impedance.

Alternate electronics are being considered for the Fall, for a test in Low loss environment. Amplifier installed inside chamber box (discussions with Jim Crisp at Fermilab Beams division lot of experience building electronics for beam current monitors). Benefit from reducing the RC (capacitance dominated by cable increase input impedance and maintain time resolution)

33/33


Backup Slides


Why is this measurement needed?

• Optical radiation• Streak camera -------------------- xxxxxxx xxxxxxx > 200fs• Non linear mixing ----------------- xxxxxxx xxxxxxx Laser to RF jitter : 500fs• Shot noise frequency spectrum -- xxxxxxx xxxxxxx Single bunch detector

• Coherent radiation• Interferometry ------------------- xxxxxxx xxxxxxx• Polychromator --------------------- xxxxxxx xxxxxxx

• RF Pick-Up -------------------------------- xxxxxxx xxxxxxx xxxxxxx > 500fs

• RF Deflector ----------------------------- xxxxxxx xxxxxxx xxxxxxx

• RF accelerating phase scan -------------- xxxxxxx xxxxxxx xxxxxxx High charge beam

• Electro Optic Method• Short laser pulse ------------------ xxxxxxx xxxxxxx xxxxxxx Laser to RF jitter : 500fs• Chirped pulse ---------------------- xxxxxxx xxxxxxx xxxxxxx > 70fs

• Laser Wire Scanner ---------------------- xxxxxxx xxxxxxx xxxxxxx Laser to RF jitter : 500fs

1 n! Limitations

Performances of Bunch Length detectors Performances of Bunch Length detectors (table thanks to Thibaut Lefevre, CERN)(table thanks to Thibaut Lefevre, CERN)

Documents

Novel Instrumentation for future TeV e+e- colliders Detectors for bunch length measurement and Beam loss monitoring Mayda Velasco, Anne Dabrowski, Alex