RF Control for the DESY UV-FELcasa.jlab.org/seminars/2004/slides/simrock_040206.pdf• Set and maintain accelerating fields during TTF II operation as • UV FEL user facility •

CASA Seminar 2/6/04 Stefan Simrock DESY

RF Control for the DESY UV-FEL

Stefan Simrock

DESY


• DESY UV-FEL

• Configuration of the RF Systems

• Requirements for RF Control

• Design of the LLRF

• Issues for the European X-FEL

• Outlook

21 Jan. 2004 J. Rossbach: TTF2 Status Report 3

Status of beamline installations

250 m

RF gun

FEL experimenta

l area

bypass

undulators

collimatorbunch compressorLaser

M1 M2 M3 M4 M5 M6 M7

bunch compressor

1000 MeV4 MeV 150 MeV 450 MeV


General LLRF Requirements

• Set and maintain accelerating fields during TTF II operation as

• UV FEL user facility • Tesla Test Facility

• Cavities to be controlled: • RF Gun (nc) • 3rd harmonic cavity (3.9 GHz) • Vector-sum of cryomodule 1 • Vector-sum of cryomodules 2+3 • Vector-sum of cryomodules 4,5,6,(7) • S-Band cavity (nc) at 2.856 GHz • Provide stable phase reference for Laser, and diagnostics


Electron gun for minimum emittance: PITZ

PITZ gun installed into TTF Jan 2004

TTF2 RF GUNAT PITZ

Jean-Paul Carneiro Behavior of the TTF2 RF with long pulses and high repetition rates.


TTF 2 Laser Upgrade

diode-pumpedNd:YLF preamplifier

Pulse shaper(T = 5%)

Diode-pumpedNd:YLF oscillator

Emicro = 16µJP = 16 W

AOM

fround trip= 27 MHz

EOMAOM

Faraday

pulsepicker

pulsepicker

2-stage flashlamp-pumpedNd:YLF booster amplifier

fastcurrentcontrol

shot-to-shotoptimizer

2-stage diode-pumpedNd:YLF amplifier

fastcurrentcontrol

fourthharm.

Emicro = 200µJP = 200 W

20 ps flat-top4 ps edges

Emicro = 30µJEburst = 24 mJUV (262 nm)

pum

p di

ode

pum

p di

ode

pum

p di

ode

pum

p di

ode

pum

p di

ode

pum

p di

ode

pum

p di

ode

pum

p di

ode

Pulse Stacker

Note:

longitudinal flat profile by pulse

stacker only

• Together with Max-Born-Institute, Berlin (I. Will et al.)

• Upgrade has been tested at PITZ


Transverse Emittance Measurement @ PITZ-

1.71.5

Frank Stephan TESLA meeting, Zeuthen, Jan. 2004 21

transverse profile ( D=1.2 mm )

FWHM � 21 ps; rise/fall time � 7 psmm 0.020.61σmm 0.020.55σ

y

x

��

��

Fine Tuning of Laser Parameters

( measured in UV )


General LLRF Requirements (Cont’d) • The LLRF System for the FEL user facility must be • Reliable • Operable • Reproducible • Maintainable • Well Understood • Meet (technical) performance goals

• The LLRF System for the TESLA Test facility must • demonstrate high gradient operation at 35 MV/m

- requires piezo tuners for Lorentz force compensation - requires operation close to klystron saturation - exception handling

• Demonstrate of operation with electronics installed in tun-nel


LLRF Requirements: Support by:

ELHEP GroupWarsaw University of Technology

Institute Electronic System

Team:About 20 people scientists, engineers and students (Ph.D. and M.Sc.)working at:Warsaw, CERN, DESY


Amplitude and Phase Stability

• Typically requirements are

• σA/A < 10-3 amplitude • σφ < 0.3 deg. for phase (fast fluctuations) • Must distinguish correlated and uncorrelated errors, intra-pulse, inter-pulse,

and long-term (thermal > minutes). Long term stability of better than 1 deg. is difficult to achieve.

• Other requirements • ACC1: cav. 1-4 at 12.5 MV/m, cav. 5-8 at 20 MV/m phase

of accelerating field -10.8 deg. • ACC39 at 14 MV/m at 183 deg. • S-Band cavity at 2856 MHz phase stability < 1 deg. • RF Gun operation without field probe. Rep. rate, pulse

length and power must be variable.

TTF2 linac

2nd bunch compressor (internally called BC3)

bypass

undulators

collimator

M1 M2 M3 M4 M5 M6 M7

Laser

RF-Gun

1st bunch compressor (internally called BC2)

bunch compressor (internally

called BC2)

250 m


Magnetic chicane forlongitudinal compression


Temporarybeamline forseeding

Beam dump


Longitudinal bunch shape measurements at TTF2• Streak camera

–FESCA 200 (Hamamatsu)• Transverse deflecting cavity

–S-band travelling wave cavity from SLAC• Coherent radiation

–Interferometer from RWTH Aachen• Electro-optical sampling


LOLA Shipment from SLAC to DESY• Vertically deflecting cavity for bunch length

measurements at TTF2

• Built in the late 60ies by G. Loew et al.

• SLAC contribution to TTF2

“Intra-Beam Streak Camera”

ee−−

σσzz

3.66 m3.66 mVV((tt)) σσyy

RFRF‘streak’‘streak’

SS--bandband

transverse RF deflectortransverse RF deflectorββcc

ββpp∆∆ψψyy


LOLA shipment

LOLA installed


Electro-optical sampling at TTF1

• Ti:Saphire laser (15 fs) and ZnTe crystal• Electron bunch can be scanned by varying the delay of the laser• Polarization of the laser changes depending on the amplitude of the

“beam fields“

Signal


Task Name

vacuum installationACC1 string assemblyACC1 tunnel installationcoupler conditioning (warm)cool downcoupler conditioning (cold)cavity conditioninggun commissioning (nighttime)gun commissioning with beaminjector commissioning incl. BC2tunnel closedbeam through ACC1final installationcommssioning entire machinebeam through bypass into dumpbeam through undulator into dumfirst lasing (30 nm)FEL saturation (30 nm)

4/21

9/11

9/27

12/10

2

1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21 1 11 21Jan '04 Feb '04 Mar '04 Apr '04 May '04 Jun '04 Jul '04 Aug '04 Sep '04 Oct '04 Nov '04 Dec '04 Jan '05 Feb '05

Major schedule for TTF2 installation and commissioningMore detailed schedule (400 items) to be presented in WG3 by M. Körfer

JLAB, Nov. 2001 Stefan Simrock DESY

Typical Parameters in a Pulsed Linac

time

flat-topfill

cavity phase

accelerating voltage

incident power

cavity detuning

beam pulse

ampl.

...

Beam Loading

Beam pulse pattern(micro and midi pulse structure)


Sources of Perturbations

o Beam loading - Beam current fluctuations - Pulsed beam transients - Multipacting and field emission - Excitation of HOMs - Excitation of other passband modes - Wake fields

o Cavity drive signal - HV- Pulse flatness - HV PS ripple - Phase noise from master oscillator - Timing signal jitter - Mismatch in power distribution

o Cavity dynamics - cavity filling - settling time of field

o Cavity resonance frequency change - thermal effects (power dependent) - Microphonics - Lorentz force detuning

o Other - Response of feedback system - Interlock trips - Thermal drifts (electronics, power amplifiers, cables, power transmission system)


Pulsed Operation at High Gradients

0 0.5 1 1.5 20

5

10

15

20

25Cavity Gradients (First Cryomodule)

time [ms]

Eacc [M

V/m

]

0 0.5 1 1.5 2

−150

−100

−50

0

50

100

150

Cavity Phases (First Cryomodule)

time [ms]

ph

ase

[d

eg

.]

+140deg./ms0 µs flattop -100deg./ms

900 µs flattop

-22.5 MV/m

Gradient (8 cavities)

short pulse

Phase (8 cavities)


Measurement of Cavity Q L and Detuning

0 200 400 600 800 1000 1200 1400 1600 1800 200010

−2

10−1

100

101

102

time [us]

log(

E_a

cc )

field decay: ~ e-t/τ

τQL

π f⋅-----------=

slope:1τ---–

rel. error < 1%0 200 400 600 800 1000 1200 1400 1600 1800 2000

50

60

70

80

90

100

110

120

130

140

150

time [us]

phas

e [d

eg]

phase with respect to LO⇒ ∆ω = dϕ

dt------ 2π ∆f⋅=

slope:dϕdt------

rel. error < 2 Hz

Loaded Q Detuning


Lorentz Force Detuning

0 500 1000 1500 2000−300

−200

−100

0

100

200

300

time [µs]

detu

ning

[Hz]

Lorentz Force Detuning of D39 in Chechia

fill: 500 µs

flat: 800 µs

15 MV/m20 MV/m25 MV/m30 MV/m

fillflat-top

decay


Digital Control at the TTF

DAC

DAC

ReIm

Cavity 32......8x

Cavity 25

klystronvector

modulator masteroscillator

1.3 GHz Cavity 8......8x

Cavity 1

cryomodule 4

...cryomodule 1

. . . .

LO 1.3 GHz + 250 kHz

250 kHz

ADC

f = 1 MHzs

. . . . ...

vector-sumΣ( )ab

a -b

1 8( )ab

a -b

25( )ab

a -b

32( )ab

a -b

DSPsystemsetpoint

tablegaintable

feed

tableforward

++digital

low passfilter

ImRe ImRe ImRe

clock

LO

ADC

LO

ADC

LO

ADC

ImRe

power transmission line

1.3GHzfield probe


240

270

300

330

0

Module 2

1

2

34 5

6

7

8

Beam Transient based Phase and Gradient Calibration

time

Eacc

Pinc

Ibeam

(open loop)

φcav

beam induced transient

cavity filling

field decay

beam induced transient

Re(Eacc)φbeam∆Vind I ∆t

rQ----

π f⋅ ⋅ ⋅⋅=

for ∆t << τcav :


Single Bunch transient Detection

field probe (1300 MHz)

delay line

powerdivider

phaseshifter

combiner

60 dB

LO

I

Q

Magnitudeof transient

TransientVector Detector

Klystron

few ns

cavity

Principle of high resolution transient detector

−40 −20 0 20 40 600.24

0.245

0.25

0.255

0.26

0.265

0.27

0.275

0.28

0.285

0.29

time [nsec]

am

plit

ud

e [a

rb. u

nits

]

Amplitude

Single Bunch Transient ( ∆A/A=0.1%)


Single Bunch transient Detection

field probe (1300 MHz)

delay line

powerdivider

phaseshifter

combiner

60 dB

LO

I

Q

Magnitudeof transient

TransientVector Detector

Klystron

few ns

cavity

Principle of high resolution transient detector

−40 −20 0 20 40 600.24

0.245

0.25

0.255

0.26

0.265

0.27

0.275

0.28

0.285

0.29

time [nsec]

am

plit

ud

e [a

rb. u

nits

]

Amplitude

Single Bunch Transient ( ∆A/A=0.1%)


LLRF for TTF II

• Digital Feedback • C67 based DSP board • 8 channel ADC boards • 8 channel DAC board • Gigalink interface between boards

• Downconverter based on AD 8343 • Master Oscillator and Frequency Distribution • New Frequencies 13.5 MHz and 2856 MHz

Zeuthen, Jan. 2004 Stefan Simrock DESY

C67 DSP board


C67 DSP board


Rack Layout and Cabling for TTF I


Installation Status of LLRF for ACC 2-6

FB and FF Scheme

Daniel Kotthaus

No probe in the cavity

Vacc=Vfor+Vref

Feedback on Vfor (later

with Vacc ?) with variable

Gp and GI

Feedforward on Vfor

adaption of FF with Vref

Detectors for Gun Control● The same detectors for forward and reflected power

● Field control with IQ detectors (AD 8347)

● Measurement of Loaded Q and detuning with logarithmic amplitude and phase detector (AD 8302)

● Modern phase detector (HMC 439) for phase monitoring

● Detector diode for amplitude monitoring

Daniel Kotthaus

Performance of the IQ Detector

Daniel Kotthaus


FPGA based RF Gun ControllerFPGA


Performance at TTF (1)

200 400 600 800 1000 1200 1400 1600 18000

5

10

15

400 600 800 1000 120011

11.5

12

12.5

time [µs]

Zoomed Region

acce

lera

tin

g g

rad

ien

t [M

V/m

]

only feedback (gain = 70)

with feedback and feedforwardcompensation

beam

0 500 1300900 1700

300 500 1100

200 400 600 800 1000 1200 1400 1600 1800−15

−10

−5

0

5

Zoomed Region

with feedback and feedforwardcompensation

only feedback (gain = 70)

400 600 800 1000 1200−1

−0.5

0

0.5

1

beam

time [µs]0 500 1300900 1700

300 700 1100

acce

lera

tin

g p

ha

se[d

eg]

Amplitude Phase


Adaptive Feedforward

time [µs] time [µs]

Step Step Response

0 200 400 600 800 1000 12000

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

measured

Closed LoopSystem

Vec

tor

Sum

Rea

l

0 200 400 600 800 1000 12000

200

400

600

800

1000

1200

calculated for alinear time-invariantSystem

∆E τ1( )

∆E τ2( )

…∆E τn( )

T11 T12 … T1n

T21 T22 … T2n

… … … …Tn1 Tn2 … Tnn

∆ ff 1

∆ ff n

…∆ ff n

=

Par

t (∆

E(t

) )

[ΜV

/m]

∆ff t( ) ∆ ff jΘ t t j–( ).j

∑=

Fee

d F

orw

ard

Rea

l Par

t (∆ ff(

t) )

[bits

]


Reproducibility of Subsequent Pulses of Vector-Sum

800 850 90090.35

90.4

90.45

90.5

90.55

90.6

time [us]

acce

l. vo

ltage

[MV

]

10 consecutive rf pulses

peak-to-peak: ± 0.03 MV∆Vacc (rms) < 0.02 MV

800 850 900 950

24.3

24.35

24.4

24.45

24.5

time [us]ph

ase

[deg

]

10 consecutive rf pulses

peak-to-peak: ± 0.03 o

∆Φ(rms) < 0.02 o

Gradient Phase


System Identification (1)

smoothed

Cavity-Field

Forward-Power

Beam-Current

smoothed&

derived

smoothed

Detuning during Pulse

Beam-Phase

Bandwidth

V˙ ω1 2⁄– ∆ω

∆ω ω1 2⁄–V Rω1 2⁄ I G I B+( )+=

V˙ ω1 2⁄– ∆ω

∆ω ω1 2⁄–V Rω1 2⁄ I G I B+( )+=

Differential Equation


0 5 10 15 20 250

20

40

60

80

100

120

140

160

180

No of Measurement

beam

phas

e/° 40˚ 50˚ 50˚ 60˚ 70˚

System Identification (2)• force Pf to be zero at

end of pulse by sub-straction of df*Pr(complex)

• the whole shapechanges

PrPf

400 600 800 1000 1200 1400 1600 1800 2000−150

−100

−50

0

50

100

150

time/µs

∆ f/H

z, ∆

f/H

z2 , V2 /2

MV

2

1st Order ID on Cavity 7

∆ω(t)

Single Pulse DetuningMeasurement

Vacc

Beam phase of 4 cavities fordifferent phase of Vacc

Correct for directivity of couplers


Piezo-Actuator:

Umax=150Vl = 39 mm

∆l≈ 4 to 5µm at 2K∆fmax, static≈ 500Hz

Active Compensation of Lorentz Force Detuning (1)

He-tank+ cavity

tuning mechanism

piezo


Active Compensation of Lorentz Force Detuning (2)

−200

−100

0

100

200

300

400

detu

ning

[Hz]

0 500 1000 1500 2000time [µs]

fill time 900 µs constant gradient

without compensation

with compensation

“beam on”- time9-cell cavityoperated at23.5 MV/m

Lorentz forcecompensatedwith fastpiezoelectrictuner

26/01/2004Lutz Lilje DESY

RF signals at 35 MV/m

Blue: With piezo

Red: Without piezo

26/01/2004Lutz Lilje DESY

NEW: Frequency stabilization at 35 MV/m

Blue: With piezo

Red: Without piezo

Frequency detuning of ~1000 Hzcompensated with resonant excitation of a mechanical cavity resonance at 230 Hz.

NOTE: This is rather an demonstration of the capability of active tuning. Application in a real machine is probably difficult/impossible.

11th SRF Workshop, 2003 Stefan Simrock DESY

Integration of Piezo Tuner for TTF


Measurement of Mechanical Preload

Force sensor

Lifetime of piezo depends stronglyon mechanical preload. Optimumaround 1 kN/ cm^2.

Piezo 1Piezo 2

lifetime

preload

1 kN/cm^2

2

(cycles)

1015

1010

3

Characterization of the load cellCharacterization of the load cellA new insert was designed to host different load cells and the load generating device. Our goal is the characterization of the sensor at 4 K up to 2kN.

A load cell under test – from Burster

• The button on the cell is pushed by stainless steel rod, 20 mm diameter.

• The loading force is generated by a screwing device provided with washer springs at the top of the insert.

• The loading force is measured by a calibrated load cell placed in the cross junction, working at room temperature.

Some resultsSome resultsUp to now some 2kN load cells from Burster have been tested at 4K.

Now some points are clear:

•High offset and low reproducibility are the main critical problems

•Linear range reduces and reproducibility fails at cold => we’ll test cells with specific cryogenic features and higher RT range(10-15 kN)

The measured TF of two 8415 model and their beaviour vs time

T=300 K

y = 2.55E-05x - 3.89E-04R2 = 1.00E+00

y = 2.38E-05x + 3.59E-05R2 = 1.00E+00

-0.001

0.000

0.001

0.003

0.004

0.005

0 50 100 150 200Kg

V8415-6002 s/n:1850218415-6002 s/n:184992

T=4 K

y = 1.70E-05x + 1.32E-02R2 = 9.62E-01

y = 1.84E-05x + 1.22E-02R2 = 9.82E-01

0.0120

0.0133

0.0145

0.0158

0.0170

0 50 100 150 200Kg

V 8415-6002 s/n:1850218415-6002 s/n:184992


New facility transfer functionsNew facility transfer functions

-70

-60

-50

-40

-30

-20

-10

0 500 1000 1500 2000Hz

dB

Amplitude

The single copper cell system has been caractherized measuring the transfer function between the piezoelectric actuator voltageand the phase detuning of the cavity.

Closed loop tests seem possible up to 2 kHz.

-600

-500

-400

-300

-200

-100

0

0 500 1000 1500 2000Hz

Deg

Phase


Microphonics Control

~~~klystron

cavity

piezoactuator

LO

RF

Microphonics

Controller

C(s)

IF

P(s)

uc(s)yc(s)

yp(s)

up(s)

piezosensor


Controller Design

P(s)

C(s)

frequency [kHz]

|D(s)|

1 10

Choose open loopresponse D(s)

C s( ) D s( )P s( )-----------= ⇒

D(s) : stability criteria fulfilled high gain at low freq. fast roll-off at high freq.


Feedback Successfully Applied to QWR

• C6701 processor from TI on PCI board (M67) with 4 ADCs and DACs (200kHz sampling rate)

• Programmed state space equation for 20th order sys-tem:

• Latency only 20 µs for 20x20 matrix multiplication (C++) • Applied only notchfilter (672 Hz) and low pass (1kHz) to

control microphonics in QWR

xk 1+ Axk Buk+=yk 1+ Cxk 1+ Duk 1++=

MIT Optics & Quantum Electronics Group

Balanced Cross-CorrelatorOutput

(650-1450nm)

Ti:sa

Cr:fo

3mm Fused Silica

SFG

SFG

Rep.-RateControl

(1/496nm = 1/833nm+1/1225nm).

Δt

0V


Measuring the residual timing jitter

Output(650-1450nm)

JitterAnalysis

SFG

Ti:sa

Cr:fo

3mm Fused Silica

SFG

SFG

Rep.-RateControl

(1/496nm = 1/833nm+1/1225nm).

GD

-GD/2


Measuring the residual timing jitter

Output(650-1450nm)

JitterAnalysis

SFG

Ti:sa

Cr:fo

3mm Fused Silica

SFG

SFG

Rep.-RateControl

(1/496nm = 1/833nm+1/1225nm).

GD

-GD/2


Experimental result: Residual timing-jitter

The residual out-of-loop timing-jitter measured from 10mHz to 2.3 MHz is 0.3 fs (a tenth of an optical cycle)

1.00.80.60.40.20.0C

ross

-Cor

rela

tion

Am

plitu

de

-100 0 100

Time [fs]

100806040200Time [s]

Timing jitter 0.30 fs (2.3MHz BW)

Long Term Drift Free


Timing Stabilized Fiber Links (<1km)

Assuming no fiber length fluctuations faster than 2L/c.

Timing Stabilized Fiber Links (<1km)

CrossCorrelatorFiber Link

PZT

Fixed Length L

532 nm Pump laser

10 fs Ti:sapphire laser

Two wavelength filter

Dichroic Beam Splitter

All Fiber Implementationsat 1.5 µm with high repetition rates


Self-balanced sub-10 fs RF-Synchronization

λopt/4

(2n+1) λRF/2


Balanced Cross-Correlator



LLRF Subsystems/Components

o RF phase reference - from main driveline - LO for downconvertero Timing Systemo Vector modulatoro Downconvertero Digital Control (Fdbck + FF) - ADC, DSP, DAC - includes exception handlingo Redundant simple feedforwardo Redundant monitoring systemo Transient detectiono Interfaces to other subsystems - includes interlocks

o Waveguide tuner and controlso Cavity resonance control - slow (motor) tuner - fast (piezo) tunero CPU in VME crateo Network to local controlso Cabels and connectorso Power supply for electronicso Airconditioning in rackso Software - DSP (FPGA) code - server programs - client programs - LLRF Parameters - Finite State Machine

Page 1

19 January 2004 8:46 pm

/home/simrock/doc/frame/ttf_meeting/ttf_meet_jan_04/ref/llrf_team.fm

LLRF Team

Name Field of Expertise

Ayvazyan, Valeri Software, FSM, DOOCS, Controls, Applications,Linac Operation

Bienkowski, Andrej RF Hardware, analog and digital hardwareBrandt, Alexander Finite state machineBruns, Thomas Computer (Unix) administrationCichalewski, WojciechKoseda, Boguslaw

FSM and applications

Czarski, Tomasz RF Modelling, FPGA development, optimal controlCzuba, Krzysztof M.O. and Distribution, Fiber optic linkEints, Frank HiwiFelber, Matthias HiwiFroelich, Thomas Installation, Documentation, MaintenanceHensler, Olaf DOOCS control system (deputy of K. Rehlich)Grecki, Mariusz TUL-DMCS group leaderIgnachin, NikolaiSytov, Sergei

Analog, digital, and rf electronics

Jezynski, Tomasz FPGA control for RF Gun/XFELKierzkowski, FPGA hardware and programmingKotthaus, Daniel RF Gun ControlLilje, Lutz Piezo tuner, high gradient cavitiesLorbeer, Bastian Master Oscillator and DistributionMakoswki, Dariusz Radiation issues for electronicsMatsumoto, Toshiyushi RF System Modelling, LLRF DevelopmentMoeller, Guenter RF Hardware, Downconverter, vector-mod, rf-gatePawlik, Pawel Single bunch transientPetrosyan, Gevorg DSP programming, DSP code and serverPetrosyan, Lyudvig Timing expert, ADC serverPosniak, Krzysztof FPGA hardware and softwarePucyk, Piotr DOOCS control of FPGARehlich, Kay DOOCS control system (group leader)Romaniuk, Ryzard WUT-ISE group leaderRutkowksy, Peter DOOCS control of FPGARybka, Dominik Radiation damage to electronicsSchrader, Matthias RF ControlSekalski, Przemyslaw Piezotuner and controlSimrock, Stefan LLRF (group leader)Vetrov, Piotr DSP hardware (DSP board, Gigalink, ADC, DAC)Wagner, Richard Hera Protonen HF, NT AdministrationWeddig, Henning RF Hardware, M.O. and Distribution, Analog and

digital electronics, RF MeasurementsZabolotny, Wojciech FPGA hardware and programming


Summary

• Commissioning of LLRF for TTF II is well underway • Feedforward for ACC 4,5,6 (old IQ drivers) available • New C67 based DSP System for RF Gun and ACC1 under

commissioning. In operation with cavity simulator • New “field” detectors for RF Gun • Prototype of FPGA based controller and cavity simulator

• Master oscillator and frequency distribution are presently being installed

• New frequencies (2856 MHz, 13.5 MHz) • Temperature stabilized coaxial distribution • Highly stable fiber optic monitoring system

• Automation of LLRF operation under development

Documents

RF Control for the DESY UV-FELcasa.jlab.org/seminars/2004/slides/simrock_040206.pdf• Set and maintain accelerating fields during TTF II operation as • UV FEL user facility •