Transcript RF Systems for ThomX

15th European Synchrotron Light Source
Radio-Frequency Meeting
5 - 6 October, 2011
ESRF, Grenoble
RF systems for ThomX
P. Marchand - Synchrotron SOLEIL
The ThomX Project
Compact source of hard X-rays (40 – 90 keV)
Flux of up 1013 photons / sec, generated by Compton Back Scattering
(CBS : collisions between e- bunches and laser pulses  ω dif ~ 4 γ2 ω laser )
Applications
- Medical sciences (imaging + therapy)
- Cultural heritage sciences (Louvre Museum, for instance)
Compactness for accommodation in hospitals and museums
Funding of 12 M€ for Phase 1 : building of a prototype  feasibility proof
 Phase 2 : industrialization
Work supported by the EQUIPEX program from the Research Ministry,
Région Ile de France, CNRS-IN2P3 and University of Paris-Sud
Contributions from LAL-Orsay CNRS-IN2P3, SOLEIL, CELIA Bordeaux,
ESRF, C2RMF-CNRS, UDIL CNRS, INSERM Grenoble, Thales TED,
Institute Neel Grenoble
LAL-Orsay & SOLEIL in charge of the accelerator complex, housed inside the
former DCI building on the university site in Orsay (~ 5 km from SOLEIL)
ThomX accelerator complex
7m
SR optics :
4-fold
symmetry
Double Bend
Achromat
Interaction
Region
FP
optical
cavity
10 m
Injection of a single e- bunch (20 mA), which collides at each turn with
laser pulses at the IP, inside the FP optical resonator  X rays from CBS
CBS  fast degradation of the e- beam quality  storage for ~ 20 ms
 Injection rate of 50 Hz (after 20 ms, extraction to BD & new injection)
The LINAC Injector
Photocathode RF gun :
• Replica of the CERN-CTF3 gun, built by LAL
• Ec = 100 MV/m with 10 MW
• Mg cathode (Q up to 1 nC)
• Laser : l = 266 nm, E ~ 100 µJ, st ~ 5 ps
Accelerating structure :
• LIL type (4.5 m long) AS, spare from SOLEIL
• P = 10 (20) MW  E = 50 (70) MeV
2.5 cell 3 GHz gun
RF Power source :
• 35 MW TH 2100 klystron from Thales
• Solid state modulator (3 µs, 50 Hz)
• Power splitting : 10 MW  Gun
20 MW  AS
Expected beam performance (PARMELA)
• E ~ 50 MeV (max 70)
• sE/E < 0.4 %
• en ~ 5 p mm.mrad
HV modulator
Klystron
RF system of the
Storage Ring
SR RF parameters
At 50 MeV, Urad ~ 2 eV / turn  Pbeam ( Ib = 20 mA ) ~ 0
• No power to be delivered to the beam (s = 0)
• RF system only generates VRF for suitable longit. acceptance
Selected RF frequency  500 MHz
Good compromise
- VRF = 500 kV  1 single cell cavity  PRF (dis) ~ 35 kW
- Availability of power sources & other RF components
- Reasonable equipment size
- Zhom  will dictate the choice of cavity design
HOM impedances and
instability thresholds
U rad ~ 0  t damping ( ~ 1 s ) >> t storage ( ~ 20 ms )
To preserve the beam quality  Instability growth time, ti > 20 ms
Longitudinal  HOM in resonance, til = 2 Qs E/e / (a Io Rs fm)
Rs . fm (HOM) natural : 0.1 - 1 M . GHz  til ~ 10 µs !!
Transverse  HOM in resonance, tit = 2 E/e / (bT frev Io RT)
RT (HOM) natural : 1 - 10 M / m  tit ~ 10 µs !!
In both, longitudinal and transverse cases,
damping of Zhom by a few 103 is required !!
(more critical than in 3rd generation LS  x 10)
Cures to HOM impedances
1) De-Qing of the HOM (HOM couplers)  a few 102 - 103
 Not enough & cumbersome equipment around the cavity
« DAMPY » cavity - ALBA
PEP-2 cavity (LBNL)
Cures to HOM impedances
2) HOM tuning  Prevent resonant excitation by the beam
ELETTRA cavity with its 3 tuning means
- Temperature control of fHOM
- Lcav (mech. deformation)  fo
- Movable plunger on the equator
• 1 single cavity
• ~ No beam loading
• Small circumference
Power coupler
Tw ± 0.1°C
Lcav
Well suited to
HOM tuning
 Beam spectrum lines : df = 18 MHz
 HOM resonance BW : a few 10 kHz
 fHOM (tuning) : a few MHz
 Rs (f = 0) / Rs (f ) = 1 + (2 Q f / f )2
 A few 103 to 104  Ok for ThomX
Plunger
HOM Spectrum
Ok
Ok
Ok
Ok
Ok
2Q s E o / e
I b .a .t l
tl = 20 ms
ELETTRA cavity L-HOM spectrum (9 modes) over the 18 MHz base band
RF power source
VRF = 500 kV, using 1 ELETTRA cavity  PRF (dis) = 35 kW
At 500 MHz  Klystrons, IOTs, Solid State Amplifiers (SSA)
SOLEIL technology
- Well proven (6 years op.)
- No HV
- Modularity  redundancy
- …
35 kW SSA of the SOLEIL Booster
147 modules of 330 W @ 352 MHz
~ 35 000 runing hours over 6 years
Operational availability of 100 %
Minor pbs on 5 modules only
without impact on the operation
H = 2.50 m ,  = 2 m
 For ThomX, make it at 500 MHz
SOLEIL - LNLS collaboration
Two amplifiers of 50 kW @ 476 MHz for the LNLS storage ring
with components designed by SOLEIL (RF modules of 400 W)
April 2010 : the SOLEIL - LNLS team in Campinas-Brazil,
after successful tests of the amplifiers
LNLS 50 kW RF plants
The two 50 kW SSA have run satisfactorily on the LNLS SR for ~ 1 year
SOLEIL R&D’s with SSA
@ 352 MHz
6th generation transistors (Vdc = 50 V) + SOLEIL expertise  fast progress
 At 352 MHz, Pmod ~ 700 W, G > 20 dB,  > 70%
[Current LR301mod. (Vdc = 28 V) : P = 315W, G = 13 dB,  = 62 % @ 352 MHz ]
 Huge improvement : Pmod x 2.2 , better performance (G ,  , linearity)
& thermal stress strongly reduced (T : - 60 °C)  longer lifetime
 Beg. 2009, transfer of technology agreement concluded with ELTA-AREVA
 ESRF contract for 7 SOLEIL type amplifiers of 150 kW (14 x 75 kW towers)
 June 2010 : A 10 kW unit (16 modules) successfully tested at SOLEIL
 June 2011 : First 75 kW tower passed the acceptance tests ( ESRF)
SOLEIL SSA : Evaluate 6th generation transistors of lower power (~ 330 W)
from NXP & Freescale  replace LR301 with min. modification
In view of storing 500 mA using a single cryomodule :
• Combination of two 180 kW SSA for powering one cavity
• Input power coupler (P > 300 kW) developt  CERN/ESRF/SOLEIL collab.
R & D’s with SSA
@ frequencies other than 352 MHz
Prototypes of 500 MHz module : P = 650 W, G = 18 dB, η = 67 %
Components design is completed
 1 x 50 kW for ThomX
 First tower : by the end of 2012
 4 x 150 kW for SESAME
Extend the technology to frequencies from FM to L band
 VALVO/SOLEIL  set of circulators covering the whole freq. range
 Prototype of 88 MHz module : P = 900 W, G = 25 dB, η ~ 80 %
 BBEF : 20 kW CW – 1.3 GHz SSA for the Beijing University
 Collab. Agreement under finalization with CERN for a prototype of
20 kW @ 200 MHz in anticipation of 2 x 1.6 MW
New features :
 Modular high efficiency 230 V_ac / 50 V_dc power converters
 Option for housing the complete SSA inside a cabinet
 Waveguide-to-coaxial combiner (WaCCo)  adjustable coupling
 Possibility of matching variable number of modules
Waveguide-to-Coaxial
Combiner (WaCCo)
2 coaxial inputs
dl
WG
output
 Two 6 inches coaxial input ports (2 x 80 kW)  1 WG output
 Replace a coaxial combiner + a coaxial-to-WG transition
 Design optimization with HFSS and Microwave Studio
 A 500 MHz prototype is being fabricated by BBEF
 Movable SC  can ensure a good matching for different configurations
wit diff nb of dissipaters per tower or diff nb of modules per dissipater
ThomX LLRF system – slow loops
Compensation of slow perturbations >> tfcav = 40 µs
 Conventional LLRF (frequency, phase, amplitude loops)
 Replica of the actual analogue SOLEIL design, adapted for 500 MHz
Phase loop
Frequency tuning loop
Amplitude loop
RF
SWITCH
3dB

40 kW
AMPLIFIER
Coupler
Drive
CAVITY
500 MHz
PID
RF ON / OFF
PID

Phase
control
o
dV
d
cav

in
df
+
Voltage
control
Tuning
control
cav
-
Vcav
PID
Tuner
Fast phase / energy oscillations




Injection errors, dEi , di
Mismatch between injected bunch and RF bucket
HOM excitations
Transient beam loading
- d = 8° (  divided by Gfbk )
(Ib : 0 - 20 mA instantly)
- Only first injections (stationary after ~ 1 s)
dE/E
Oscillations in phase & energy
@ fs, the synchrotron frequency
with damping time, td  1/Urad
di
d(t)
t
dEi
di
d
Either Phase or Energy errors  Phase & Energy oscillations (quadrature)
Fast phase / energy oscillations
e- bunch length : 20 – 30 ps rms
Laser pulse duration : 5 ps rms
Synchro e- / laser

t < 5 ps   < 1°
(E / E)inj =  0.5 % (LINAC)  inj =  8° (AS)
Without oscillation damping :
 Emittance growth
 Bad bunch / laser overlap
Still amplified by
mismatch & HOM
Loss of efficiency in the
e-/laser interactions @ IP
 td ~ 1 s >> tst = 20 ms  ~ no natural damping during tst
 fs = 500 kHz >> BWcav = 25 kHz  damping through the cav. impossible
 3 means for generating some damping :
1) Longitudinal FB using an additional broad band cavity
2) Harmonic cavity  Landau damping
3) Direct RF FB on the main cavity  increase its effective BW (> 500 kHz)
No need for additional cavity
Direct RF Feedback principle
G
Z(ω)
-
Pin
+
C
Ig
Rg
Rs
Ib
Vc
L
With FB : Z' ω  
Z ω 
1  G ω  Z ω 
At resonance (  r) , Z' 
Z
1  G0
; G ω  
G0
Rs
e
 jω Δ T
Loop delay
 BW' c av  BW
Gain limitation ( stability criterion )  1  G 0 
c av
x
1  G 0 
π QL
2 ω r ΔT
Ampli-cav
distance
ThomX : ampli - cavity distance ~ 10 m  T ~ 150 ns
 Glimit ~ 60  BW ~ 1.5 MHz >> fs
Cavity transfer function with RF FB
0
Amplitude [dB]
-10
T = 150 ns
-20
Gain 0
-30
Gain 44
Gain 66
-40
fr - fs
-50
-60
495.00
496.00
497.00
498.00
499.00
fr + fs
500.00
501.00
502.00
503.00
504.00
505.00
Frequency (MHz)
200
Phase [deg.]
150
100
T = 150 ns
50
Gain 0
0
Gain 44
-50
Gain 66
-100
-150
-200
495.00
496.00
497.00
498.00
499.00
500.00
501.00
Frequency (MHz)
502.00
503.00
504.00
505.00
RF FB + fast beam phase loop
BPM
d
G
Phase
comparator
Ib
90°
MO
500 MHz
RFSwitch

Driver
AMPLI 50 kW
3 dB
CAVITY
Phase
Shifter
Interlocks
RF FB  BWcav x (1 + Go) > fs
 Modulate Vcav at f > fs
Phase loop
(BW > fs)
PUcav
RF feedback
Go
Att

- Phase comparison between Vc (PU cav) & Ib (BPM)
- The error signal, d (+ 90°) controls a phase shifter
Alternative : Modulate the MO with d
BW ?
Phase / energy oscillations
with RF FB + fast phase loop
inj = 10°, Go = 50, G = 5 , T = 150 ns
Damped after 20 µs
 T_damping = 3 µs for G = 30 (stability limit)
Complete LLRF
MO
500 MHz
50 kW
AMPLI
RF
SWITCH
3dB
Coupler
3dB

PA
Att
PID
A
Tuning
control
PID

CAVITY
Tuner

Direct RF Feedback
Phase
control
Go
+
Att

-
PID
Beam
PU
Voltage
control
Beam
phase
G
Conventional system with 3 « slow » loops around the cavity
Frequency
Amplitude
Phase
Oscillations @ fs
(500 kHz)
 inj , HOM, …
- Direct RF Feedback (G ~ 50)
- Fast beam phase loop (BW > 500 kHz)
Summary & Conclusion
RF system of ThomX SR
1) One 500 MHz ELETTRA type cavity (HOM tuning)
2) 500 kV with 35 kW, supplied by a SOLEIL type SSA
3) LLRF : conventional system with 3 slow loops (fr , V, v)
+ high gain RF feedback & fast phase loop (b)
Rem : ThomX is a small machine, but quite complex and challenging,
in particular as regards to the electron beam dynamics
Planning : RF equipment available for installation in ThomX by mid-2013
- Amplifier & LLRF  designed & supplied « turn key » by SOLEIL
- Cavity  one of the ELETTRA cavities, dedicated to SESAME,
made available for ThomX until mid-2016 (validation on the
machine and then fabrication of another one, modified or not)