LLRF

ILC GDE MeetingFeb.6,2007

Shin Michizono

LLRF

- Stability requirements and proposed llrf system- Typical rf perturbations- Achieved stability at FLASH- R&D items- Schedule

- Other comments

2

LLRFStability Requirements for Main Linac

phase tolerance limiting

luminosity loss (deg)

phase tol. limiting incr. in energy spread

(deg)

amplitude tolerance limiting

luminosity loss (%)

amplitude tolerance limiting increase in energy spread (%)

Related fluctuations

correlated BC phase errors .24 .35 HV

uncorrelated BC phase errors .48 .59 Microphonics

correlated BC amplitude errors 0.5 1.8 HV, Ibeam

uncorrelated BC amplitude errors 1.6 2.8 Microphonics

correlated linac phase errors large .36 HV

uncorrelated linac phase errors large 5.6 Microphonics

correlated linac amplitude errors large .07 HV, Ibeam

Uncorr. linac amplitude errors large 1.05 Microphonics

Summary of tolerances for phase and amplitude control. These tolerances limit the average luminosity loss to <2% and limit the increase in RMS center-of-mass energy spread to <10% of the nominal energy spread.

Ref. Mike Church

3

LLRFLLRF system configuration at ILC

8 x

MMM

CREV

FWD

KLY

XFMR

ModulatorDriver

LLRF crate 1:- CPU, PS- 3x LLRF- BPM- Timing Distribution- (HOM coupler)

Modulator Control & Interlock (Rack)

Klystron Control, Interlock & Water

Cryogenic Controls (Rack)

REV FWD

2x e--field

Temp (RTD)

Arc (air)

PMT (vac)

Arc

RF Phase Reference Line

WG pressure

Vacuum Controls (Rack)

M M M M M M M MBeamPickup

24x Cavity Probe (plus evt. 48x HOM coupler)

M M M M M M M M

BPM & Magnet

M M M M M M M M

8 x

MMM

CREV

FWD

8 x

MMM

CREV

FWD

8x Coupler

Cryomodule

Accelerator Tunnel

Service Tunnel

Timing & Synchronization

Inter-System Feedback Network/Bus

MPS Network/Bus

RT Control Network

Local Oscillator

LLRF crate 3:- CPU, PS- Motor Controller- Piezo Controller

Temp Temp Temp

Temp

LLRF crate 2:- CPU, PS- Klystron Protection & Interlock Systems

LLRF & Instrumenation Rack

TempPMT

WG pressKlyArc

7xFWD/REV

HOM (1..48x)CavREV (24x)CavFWD (24x)CavProbe (24x)Reference (3x)

BeamPickup (3x)BPM (3x)

48x signal

72x coax

24x fiber

96x Motor

48x Piezo

Temp

WG pressureWG pressure

REV FWD REV FWD

Arc

Po

wer

& H

igh

Leve

l Sig

nal P

ene

trat

ion

RF

& L

ow L

eve

l Sig

nal P

enet

ratio

n

All electronics are located at service tunnelVector-sum control of 26 cavitiesTotal 26x3+6=84ch RF monitors

4

LLRFLLRF Rack Detail

5

LLRFFB algorithm

26x

26xIn the case of proportional control Output = Gain*Error+FF-> Sufficient dynamic overhead is necessary at high gain operation (>50)

LLRF

Klystron operation point and overhead

@31.5 MV/m operation, llrf overhead is <16.5% in power (8.25% in amplitude)@33 MV/m operation, overhead becomes 11% in power-> It is used for fluctuation compensation, detuning compensation and so on.

0

20

40

60

80

100

0 10 20 30 40 50Pin [W]

Pow

er [

%]

76.5% for cavity input

7% rf distribution loss

16.5% for llrf control*

*neglect all other factors such as HV ripple

Extra rf drive (Gx(error)) is necessary at FB.Proportional FB gain is limited around 80 (when we can pick-up 0.1% error).And the error can be suppressed 1/G=1/80

LLRF

ILC GDE MeetingFeb.6,2007

Shin Michizono

LLRF

- Stability requirements and proposed llrf system- Typical rf perturbations- Achieved stability at FLASH- R&D items- Schedule

- Other comments

8

LLRFSources of Perturbations

o Beam loading o Cavity dynamics

- Beam current fluctuations - cavity filling

- Pulsed beam transients - settling time of field

- Multipacting and field emission

- Excitation of HOMs o Cavity resonance frequency change

- Excitation of other passband modes

- thermal effects (power dependent)

- Wake fields - Microphonics- Lorentz force detuning

o Cavity drive signal- HV- Pulse flatness o Other- HV PS ripple - Response of feedback system- Phase noise from master oscillator - Interlock trips- Timing signal jitter - Thermal drifts (electronics,

power- Mismatch in power distribution amplifiers, cables, power

transmission system)

9

LLRFTypical Parameters in a Pulsed RF System

10

LLRF

Lorentz Force detuning compensation

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

0 10 20 30 40 50 60detuning [Hz]

addi

tona

l am

plitu

de

Detuning of 30 Hz require additional 2% rf power.

11

LLRF

Microphonics

From Thesis of Thomas Schilcher

12

LLRFSuppression of fluctuations by FB

*When the system can detect 0.1% error.**only FF

Larger dynamic overhead is desired for the larger FB gain.

Dynamic overhead[%] in power* 0 6 20

FB gain 0 30 100 req.(BC) req.(ML)microphonics &Lorentz force

detuninguncorrerated 30Hz 9.2 0.3 0.09 0.48 5.6 [deg.]

Kly HV stability both 1% 12 0.4 0.12 0.24 0.36 [deg.]

Dynamic overhead[%] in power* 0 6 20

FB gain 0** 30 100 req.(BC) req.(ML)microphonics &Lorentz force

detuninguncorrerated 30Hz 1.28 0.043 0.013 1.6 1.05 [%]

average beamcurrent correated 1% 1 0.033 0.010 0.5 0.07 [%]

Kly HV stability both 1% 1.25 0.042 0.013 0.5 0.07 [%]

Amplitude

Phase

LLRF

ILC GDE MeetingFeb.6,2007

Shin Michizono

LLRF

- Stability requirements and proposed llrf system- Typical rf perturbations- Achieved stability at FLASH- R&D items- Schedule

- Other comments

14

LLRFField Regulation at FLASH

By T. Schilcher

LLRF

ILC GDE MeetingFeb.6,2007

Shin Michizono

LLRF

- Stability requirements and proposed llrf system- Typical rf perturbations- Achieved stability at FLASH- R&D items- Schedule

- Other comments

16

LLRFR&D items

FPGA board development having >26 ADCs RF field stabilities <0.5% in amplitude and <0.24 deg. in phase Crate evaluation (VXI, ATCA, ….)

– Redundancy, board size Software development

– Feedback algorithm

– Klystron linearization

– Exception detection and handling

– Warnings and alarms High IF study

17

LLRFDESY SIMCON3.1 Controller

18

LLRF

FPGA & DSP boards @KEK

10 16bit-ADCs10 16bit-ADCs

FPGAFPGA

2DACs2DACs

Quench etc.Quench etc.

Output maxOutput max

RF offRF off(by diagnostics in DSP)(by diagnostics in DSP)

Real time intelligent diagnostics by DSP boardReal time intelligent diagnostics by DSP board

Custom FPGA board: Mezzanine card of the commercial DSP board10 16bit-ADCs and 2DACs + 2Rocket IO40 MHz clock

Commercial DSP board (Barcelona) (same to J-PARC system):4x TI C6701 DSPsCan access to FPGA like an external memory of DSP

19

LLRFNow, the number of ADCs in a FPGA board is limited due to the substrate. (maybe ~15 with 16 layers in substrate)The idea is based on the ‘digital radio’ and obtaining cavity signals with a ADC.

Mixture of two signals decrease the resolution of analog signals but averaging increases the resolution.

IF1(8 MHz)

IF2(12 MHz)

Triger(48 MHz)

Mixed signal(IF1+IF2)

averaging(IF2 1)-> IF1 signal

averaging(IF1 1)-> IF2 signal

R&D: Proposal of IF mixture

Over-sampling: IF 8 MHz & 12 MHz with 48 MHz sampling Over-sampling: IF 8 MHz & 12 MHz with 48 MHz sampling -> include averaging effect ->increase resolution-> include averaging effect ->increase resolution

Cavity signals do not change during averaging (due to high Q values)→ Enough IF separation

LLRF

ILC GDE MeetingFeb.6,2007

Shin Michizono

LLRF

- Stability requirements and proposed llrf system- Typical rf perturbations- Achieved stability at FLASH- R&D items- Schedule

- Other comments

21

LLRFSchedule

Support for (test) facilities (XFEL,SMTF,STF) Crate evaluation FPGA board development having >26 ADCs. Software development High IF study

I II III IV I II III IV I II III IV

DESY* XFEL design &R&DFPGA board for XFELATCA board developmentConversion of LLRF to ATCAImplement and evaluate ATCA LLRFFinalize LLRF for the XFEL

Software development with SIMCON- DSP

FNAL33 ch FPGA board for ILCTA(NML)Operational Simcon systemATCA development Depending on success with BPM projectFB algorithmhigh IF and sampling

KEK STF- 0.5 STF- 18 cavities vector sum control32ch ADCs FPGA board (ATCA)**ATCA I/ O interlock**IF mixture**24 caivities vector sum

* http:/ / xfel.desy.de/ content/ e154/ upload/ upload_file/ TDR/ XFEL- TDR- Ch- 10.pdf (XFEL TDR)** J FY budget is still unknown.

2007 2008 2009

LLRF

ILC GDE MeetingFeb.6,2007

Shin Michizono

LLRF

- Stability requirements and proposed llrf system- Typical rf perturbations- Achieved stability at FLASH- R&D items- Schedule

- Other comments

23

LLRF

24

LLRF

LLRF

Rf distribution error v.s. max. cavity gradient in case of the 2 cavities

Only rf distribution variation

100

101

102

103

104

105

106

107

108

109

0 5 10 15distribution error [%]

cavi

ty fi

eld

[%]

10% error in rf distribution induces 8.5% higher cavity field

100

100.5

101

101.5

102

102.5

103

103.5

104

104.5

0 2 4 6 8 10 12

Ql error [%]

cavi

ty fi

eld

[%]

10% error in loaded Q induces 4% higher cavity field

Examples: 5%pk-pk Ql variation + 0.2dB (2.3%)pk-pk distribution variation-> 2%+2%=4% cavity field overshoot 31.5*1.04=32.8 MV/m10%pk-pk (1.7%rms)+0.07dBrms (4.6%pk-pk) -> 8% overshoot 34 MV/m

Although Ql and RF distribution ratio control can helpful for flattening each cavity field, This does not work without beam condition.And some residual errors exist due to the imperfect setting.

26

LLRFLimitation of coupling adjustment method

By Julien Branlard

Coupling adjustment method does not work at no-beam condition.In order to satisfy both beam/no-beam condition, complex technique (including detuning control) will be necessary.

Vector sum

Lower cavity