Cavities Auto Recovery with BeamRF&Linac Section - ALBA Accelerators Division

Francis Perez

Angela Salom

ALBA Overview ALBA Accelerators RF plants of Booster and SR LLRF Conceptual Design

Automatic Recovery with beam Automatic startup without beam Automatic recovery with beam Trips postmortem analysis

Future Upgrade: Feedforward loops

Outline

Cavities autorecovery with beam – LLRF Workshop – October 2013 2/21

ALBA Overview

ALBA is a 3rd generation synchrotron light source, located at 20 km from Barcelona, Spain.

History2004. Start Acc + Building design2008 Linac Installation (turnkey system)2009 Booster Installation2010 Booster Commissioning and SR Installation2011 SR Commissioning2012 Operation with users

ALBA Overview

Cavities autorecovery with beam – LLRF Workshop – October 2013 4/21

LINAC100MeV

BoosterRamping at 3Hz

SR3GeV

ALBA Overview

Accelerators Main Parameters

Energy 3GeV

Circumference 268m

Beam Current 400mA

Emittance 4nm.rad

Lifetime ≈10h

RF Freq 500MHz

Beamlines up to 34

Cavities autorecovery with beam – LLRF Workshop – October 2013 5/21

RF Plants

Booster RF Plant

Ramping from 100MeV to 3GeV at 3Hz

Petra Cavity type (5 Cells)

Normal Conducting

500MHz

80kW CW - IOT

Injection Extraction

Cavity voltage 55kV 1000kV

Energy loss 0.001keV/turn 627keV/turn

Cavity power 0.1kW 33kW

Beam Power 0kW 2.5kW

Sync. Freq 13.7kHz 9.4kHz

Cavities autorecovery with beam – LLRF Workshop – October 2013 7/21

Storage Ring RF Plants

6 RF Plants of 160kW at 500 MHz

2 IOT Transmitters per RF cavity. Power combined in CaCo

Dampy CavityNormal ConductingSingle cell, HOM damped3.3 MΩ

Digital LLRF System based on IQ mod/demod

RF Parameters

U0 1.3MeV/turn

Vtotal 3.6 MV

q ≈ 2.5

fs ≈ 9kHz

PRF 960kW

Cavities autorecovery with beam – LLRF Workshop – October 2013 8/21

LLRF Conceptual Design

ALBA LLRF

Main Characteristics Based on digital technology using a

commercial cPCI board with FPGA

Signal processing based on IQ demodulation technique

Main loops: Amplitude, phase and tuning

RF diagnostics: Circular buffer for post-mortem analysis (0.5s)

Extra utilities Automatic conditioning for cavities Automatic soft start

Loops Resolution and bandwidth (adjustable parameters) Resolution Bandwidth Dynamic Range

Amplitude Loop < 0.1% rms [0.1, 50] kHz 30dB

Phase Loop < 0.1º rms [0.1, 50] kHz 360º

Tuning < ± 0.5º -- < ± 75º

Digital board: VHS-ADC from Lyrtech

Cavities autorecovery with beam – LLRF Workshop – October 2013 10/21

Analog Front End

Down Conversion

FPGA

Digital IQ Demodulation

Analog Front End

UpConversion

IOT2

RF Cavity Voltage (500MHz)

IF

RF Forward Cavity Voltage(500MHz)

Analog Timing System520MHz 500MHz

80MHz80MHz

CAVITY

Tuning Control Loop

Digital Timing SystemPCI Bus

DIGITAL LLRF - HARDWARE SCHEME

ICEPAPMotor Controller

13 RF Diagnostic Signals (500MHz)

IOT1CACO

DC

500MHz

8 ADCs80MHz

8 ADCs80MHz

FPGA

Digital IQ Demodulation and Control Loops

Control Inputs

Diagnostics Inputs

IQ Ctrl IOT1& IOT2

8DACs80MHz

Analog Front End

Down Conversion

FPGA

Digital IQ Demodulation

Analog Front End

UpConversion

IOT2

RF Cavity Voltage (500MHz)

IF

RF Forward Cavity Voltage(500MHz)

Analog Timing System520MHz 500MHz

80MHz80MHz

CAVITY

Tuning Control Loop

Digital Timing SystemPCI Bus

DIGITAL LLRF - HARDWARE SCHEME

ICEPAPMotor Controller

13 RF Diagnostic Signals (500MHz)

IOT1CACO

DC

500MHz

8 ADCs80MHz

8 ADCs80MHz

FPGA

Digital IQ Demodulation and Control Loops

Control Inputs

Diagnostics Inputs

IQ Ctrl IOT1& IOT2

8DACs80MHz

IQ Ctrl IOT1& IOT2

Analog Front End

Down Conversion

FPGA

Digital IQ Demodulation

Analog Front End

UpConversion

IOT2

RF Cavity Voltage (500MHz)

IF

RF Forward Cavity Voltage(500MHz)

Analog Timing System520MHz 500MHz

80MHz80MHz

CAVITY

Tuning Control Loop

Digital Timing SystemPCI Bus

DIGITAL LLRF - HARDWARE SCHEME

ICEPAPMotor Controller

13 RF Diagnostic Signals (500MHz)

IOT1CACO

DC

500MHz

8 ADCs80MHz

8 ADCs80MHz

FPGA

Digital IQ Demodulation and Control Loops

Control Inputs

Diagnostics Inputs

8DACs80MHz

Commercial Digital Board: cPCI

Commercial Digital Board: cPCI

LLRF Conceptual Design

Conceptual Design and Prototype

Analog Front Ends for Downconversion (RF to IF) and Upconversion (DC to RF)

Digital Commercial Board: cPCI with 16 ADCs, 8 DACs and Virtex-4 FPGA

Timing systems: 520MHz (500 + 20 MHz) for downconversion synchronized with digital 80MHz clock for digital acquisition

Analog Front End

Down Conversion

FPGA

Digital IQ Demodulation

Analog Front End

UpConversion

IOT2

RF Cavity Voltage (500MHz)

IF

RF Forward Cavity Voltage(500MHz)

Analog Timing System520MHz 500MHz

80MHz80MHz

CAVITY

Tuning Control Loop

Digital Timing SystemPCI Bus

DIGITAL LLRF - HARDWARE SCHEME

ICEPAPMotor Controller

13 RF Diagnostic Signals (500MHz)

IOT1CACO

DC

500MHz

8 ADCs80MHz

8 ADCs80MHz

FPGA

Digital IQ Demodulation and Control Loops

Control Inputs

Diagnostics Inputs

IQ Ctrl IOT1& IOT2

8DACs80MHz

Cavities autorecovery with beam – LLRF Workshop – October 2013 11/21

Automatic Recovery with Beam

One cavity -out of six- trips

Beam is not lost

One wants to recover the tripped cavity with heavy beam loading

RF Autorecovery with beam

Automatic Start up – Initial version:

- When RF trips LLRF Standby:- Low RF drive- Disable tuning - Open loops (I&Q)

- When RF ON LLRF smooth startup- Minimum RF Drive (low power to avoid high reflected power)- Tuning enabled- Amplitude and phase loops closed

- Smooth power increase

Main Inconvenients when applying this startup with beam:- After a trip, the cavity remains tuned it steals power from beam and

in some cases make it unstable

- Recovering the cavity with beam, when tuning the cavity, the beam induces more voltage in the cavity than the IOT Tuning loop becomes crazy

Cavities autorecovery with beam – LLRF Workshop – October 2013 13/21

RF Autorecovery with beam

New Automatic Start up – to take into account beam loading:

- When RF trip- Open loops (I&Q)- Disable tuning - Detune cavity (parking) by moving the plunger 30,000 steps up

- When RF ON:- IOT power high enough to induce more voltage in the cavity than the

beam loading after unparking- Amplitude and phase loops open because cavity is completely detuned

- Phase and amplitude of LLRF adjusted to have very similar conditions in open loop and close loop

- Plunger moved back 30,000 steps to tune cavity (unparking)- Tuning enabled

- Amplitude and phase loops closed- Smooth power increase

- Tested in all cavities at 130mA

Cavities autorecovery with beam – LLRF Workshop – October 2013 14/21

RF Autorecovery with beam

Cavity 10B trip:Other 5 cavities increase power

10B steals -20kW power to the

beam

After trip Parking Process starts

After 15s, 10B power = 0kW

1.552 1.554 1.556 1.558 1.56 1.562

x 104

-20

-15

-10

-5

0

5

10

15

20

25

30

t (s)

Pow

er to

the

beam

(kW

)

Cavity 10B trip

PBeam 06A

PBeam 06B

PBeam 10APBeam 10B

PBeam 14A

PBeam 14B

There are 6 cavities: 06A - 06B – 10A – 10B – 14A – 14B

1.5805 1.581 1.5815 1.582 1.5825 1.583 1.5835 1.584 1.5845 1.585

x 104

-15

-10

-5

0

5

10

15

20

25

t (s)

Pow

er to

the

beam

(kW

)

Cavity 10B autorecovery with beam

PBeam 06A

PBeam 06B

PBeam 10A

PBeam 10B

PBeam 14A

PBeam 14B

RF Autorecovery with beam

Cavity 10B autorecoveryRF ON in 10B: some power to the beamUnparking Process startsAfter unparking, 10B steals power to the beamTuning Loop Enable (10B power > 0kW)Amplitude and phase loop enable and power increased

10B RF ON Unparking Tuning

1485 1490 1495 1500 1505 1510 1515 1520-100

-50

0

50

t(s)

Tu

nin

gD

ep

ha

se (

º)

Cavity 10B Autorecovery with beam

TunDephase-10B

Tuning Dephase during autorecoveryRF ON : TuningDephase = -90ºUnparking Process starts → Tuning dephase approaching 0º and then overpasses this value

Cavities autorecovery with beam – LLRF Workshop – October 2013 16/21

1.581 1.582 1.583 1.584 1.585 1.586 1.587

x 104

140

160

180

200

220

240

t(s)

Bea

m P

hase

(º)

BeamPhase-06A

BeamPhase-06B

BeamPhase-10A

BeamPhase-10B

BeamPhase-14A

BeamPhase-14B

Cavity 10B autorecovery with beam

RF Autorecovery with beam

Cavity 10B autorecoveryRF ON: 10B Beam phase ~ 140ºUnparking Process finishes: 10B Beam phase ~ 230ºConclusion: the unparking process should move the plunger less steps than the parking process

10B RF ON Unparking Tuning

Beam Phase respect to RF ~ 150º

Cavities autorecovery with beam – LLRF Workshop – October 2013 17/21

0 0.5 1 1.5 2 2.5 3 3.5 4

x 10-3

100

110

120

130

140

150Beam Phase

t(s)

(º)

Beam Phase (º)

0 0.5 1 1.5 2 2.5 3 3.5 4

x 10-3

0

2

4

6

8

10

12Cav Dis - RvCav - Beam Power

t(s)

kW

Cav DisBeamPowerRvCav

Trips Post Mortem Analysis

Sometimes beam does not survive after RF trip

- The less # of cavities, the more likely to have beam dump due to RF trips- The higher the current, the higher reflected power in other cav. after a RF

trip

Will the beam survive at 400mA after a RF trip?

Transient of 06B after trip in 10A

Behaviour of 06B after a trip in 10B and no beam dump (61mA)Power to the beam increasesBeam phase gets reducedFrequency oscillations ~ 6kHz (synchrotron freq)Stabilization time ~ 3ms (longitudinal damping time)

Effect on beam trajectory (BPMs reading after RF ITCK – Data provided by A. Olmos)Cavities autorecovery with beam – LLRF Workshop – October 2013 18/21

Trips Post Mortem Analysis

Behavior of 06B after a trip in 10A and Beam Dump (100mA)Power to the beam starts to increaseBeam phase starts to get reducedBUT Reflected power reaches interlock level: 16kW

0 0.2 0.4 0.6 0.8 1 1.2 1.4

x 10-3

100

150

200

250Beam Phase

t(s)

(º)

Beam Phase (º)

0 0.2 0.4 0.6 0.8 1 1.2 1.4

x 10-3

-20

-15

-10

-5

0

5

10

15Cav Dis - RvCav - Beam Power

t(s)

kW

Cav DisBeamPowerRvCav

Provisional solution:Reflected power interlock level increased up to 23kWCavities detuned to avoid Robinson instability and to reduce the amplitude of reflected power in active cavities when one RF cavity trips

Maximum reflected power calculated when working at 400mA (data provided by Bea Bravo)Cavities β adjusted to have minimum reflected power at 400mAWorking with 6 Cavities, 600kV/cav, 400mA → RF trip causes:

Reflected Power transient of 73kW per cavity Circulator incapable to react against fast transients Reflected power will reach IOTsWill beam survive at 400mA after RF trip? Still don’t know, but not likely

Cavities autorecovery with beam – LLRF Workshop – October 2013 19/21

Future upgrade: Feedforward loops

Feedforward Loop

Feedforward loop to compensate transient when RF cavity tripsWhen cavity trips

- Cavity Voltage oscillates with frequency equal to synchrotron tune

- Transient time equal to damping time of machine

Compensation- Amplitude modulation triggered when

one cavity trips- Frequency, amplitude and phase of

modulation are adjustable parameters

Tests with beam still pending

Cavity Voltage & Beam Phase transient after RF trip

Cavity voltage, Power to Beam and Cavity Reverse Power

19 19.5 20 20.5 21 21.5 22 22.5 235000

5500

6000

6500

7000

7500

8000

8500

t(ms)

VC

av (

a.u.

)

Amplitude Modulation of Cavity after fake interlock

Cavities autorecovery with beam – LLRF Workshop – October 2013 21/21

Thanks for your attentionQuestions?