Piezoelectric stack based system for Lorentz force ...
Transcript of Piezoelectric stack based system for Lorentz force ...
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Piezoelectric stack based system for Lorentz force compensation
Sękalski Przemyslaw1,2, Simrock Stefan2, Napieralski Andrzej11) Department of Microelectronics and Computer Science, Technical University of Lodz, Poland
2) Deutsche Elektronen-Synchrotron DESY-Hamburg, Germany
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Outline
• Electromechanical tuners variety
• Active elements
• Current problems and achievements
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Three main purposes of tuner system•Pre-tuning is necessary stage to reach proper frequency. Cavity with couplers are assembled at room temperature, during cooling down and pumping the cavity shape is deformed. Required system might work slow but must be able to change the cavity length in range of several millimeters (corresponds to a few MHz of detuning compensation). The pretuning phase will be performed i.e. once a week or even month.
•Lorentz Force compensation. During pulsed operation the cavity is reloaded with frequency of 1.3GHz. The current, which flows through cavity walls, interacts with electromagnetic field inside cavity and as a result changes its shape and its resonant frequency. Detuning depends on accelerating field gradient (∆fstatic~Eacc
2) and might be equal even 1000Hz. The Lorentz force is repetitive and periodic. The system needs to act quite fast (up to 2kHz). It will be operated each pulse (repetition rate up to 20Hz)
•Microphonics is a vibration of environment. It is fully stochastic. The detuning caused by microphonics is below 20Hz. A feedback loop is required. The system should work permanently.
Slow
tune
rFa
st tu
ner
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Electromechanical tuners varietyElectromechanical systems for Lorentz force & microphonics compensation
and for pre-tuning stage
CEA-Sacleycurrent design
• Double lever system• Stepping motor PHYTRON
with Harmonic Drive gear box• ∆Z=±5mm, ∆f=±2.6MHz• Theoretical resolution 1.5nm• Stiffness ~100kN/mm• Ready for piezoelectric and
magnetostrictive actuator
CEA-Sacleynew design
• Double lever with a screw-nut system • Stepping motor PHYTRON or SANYO
with Harmonic Drive gear box• ∆f=±2MHz @RT, ∆f=±460kHz @2K• Ready for piezoelectric
and magnetostrictive actuator• preload force for active elements is
almost decoupled with motor position
UMI Milan tunerCoaxial
• Three coaxial rings connected by blades• Stepping motor for pre-tuning stage• Piezos up to 72mm length• Shorter dead zone between cavities
350 283mm (total acceleratorlength reduction by 5%)
• Expensive (factor of 2-3)• Stiffer than others (easily
upgradeable)
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Control System setup
Distributed Object Oriented Control System
Function Generator memory 8x4kB
Low pass filter fstop=1kHz
PZD amplifierG=-40V/V
Piez
osta
ck
PZM amplifierG=0.5V/V
A/D converter C
avity
Probe antenna A/D converter
M A
T L
A B
Software Hardware (electrical part) mechanical part
LLRF system based on
FPGA or DSP
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Control Panel for Piezo in ACC1 Cav5, VUV-FEL
Forward power, reflected power and probe
Calculated detuning (single pulse method)
Function generator output voltage
Voltage supplied to piezo
Settings for pulse
shape
Info
Start/stop timer
Save/load settings
Pulse creation
Load settings to Function Generator
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Types of actuators (1/2)
Plunger & Belleville springs
Active magnetostrictive element with ferrite, s.c. coil and thermal
connectors
Niobium Cover
Magnetostrictive rod(made of Kelvin ALL®)
Ferrite necessary to close magnetic circuit
Superconductingcoil (Nb3Sn)
Thermal connectors
Dimensions: 10x10x36mmManufacturer: PI
Magnetostrictive actuatorsPiezoelectric actuatorsPZT
Dimensions: 10x10x30mmManufacturer: NOLIAC
Dimensions: 7.5x7.5x50mmManufacturer: PiezoMechanik
Dimensions: 7x7x30mmManufacturer: EPCOS
Dimensions: 6x6x20mmManufacturer: ETREMAMaterial: GalFeNOL
Dimensions: Φ10x20mmManufacturer: ENERGENMaterial: KELVIN ALL®
Look at the poster ThP60 for details
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Types of actuators (2/2)
sensor neededredundant actuator (normally sensor)Sensor
1 k€ + 0.6 k€0.8 k€ + 0.4 k€Price
Current driven (design effort higher)
Voltage driven,Electronic driving system
yes, without cavityyes, with and without cavityDemonstration
Low <0.1W, can be improved, probably active cooling needed
lowCryogenic heat loss
Small with niobium shielding, <25mG at 30 mm from actuator
Small,only from wires
Stray magnetic field
Low,magnetic field driven
High, electric break-through, safer when cooled down
Susceptibility to electric damage
Large range, 0.2-10kNSmall range, 0.7-1.5kNNOTE: few kN per 1mm (416kHz)
Preload
Large>108 cycles (no damage)
Large, if preload correct>1010 cycles (small damage)
Lifetime
Large, 0.14 -0.2% (need to be checked)Small, ~0.05% Max strain at 2K
Magnetostrictive tunerPiezoelectric tuner
sensor neededredundant actuator (normally sensor)Sensor
1 k€ + 0.6 k€0.8 k€ + 0.4 k€Price
Current driven (design effort higher)
Voltage driven,Electronic driving system
yes, without cavityyes, with and without cavityDemonstration
Low <0.1W, can be improved, probably active cooling needed
lowCryogenic heat loss
Small with niobium shielding, <25mG at 30 mm from actuator
Small,only from wires
Stray magnetic field
Low,magnetic field driven
High, electric break-through, safer when cooled down
Susceptibility to electric damage
Large range, 0.2-10kNSmall range, 0.7-1.5kNNOTE: few kN per 1mm (416kHz)
Preload
Large>108 cycles (no damage)
Large, if preload correct>1010 cycles (small damage)
Lifetime
Large, 0.14 -0.2% (need to be checked)Small, ~0.05% Max strain at 2K
Magnetostrictive tunerPiezoelectric tuner
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Problem with piezoelectric devices preload force
The lifetime of the piezo element depends on preload force
1,E+00
1,E+03
1,E+06
1,E+09
1,E+12
1,E+15
0 500 1000 1500 2000 2500preload force [N]
lifet
ime
[cyc
les]
The preload force applied to piezostack implemented in cavity tuner, was roughly
calculated and/or assumed but never measured
Four methods of the static force measurement at 1.8÷4 Kelvin are proposed:
1. Resonance position on the impedance curve2. Capacitance change3. Strain gauge sensor (metal)4. Piezoresistive sensor (semiconductor crystal)
No proofat LHe temperature
1 0 1 0 00
2 0
4 0
6 0
8 0
1 0 0
1 2 0
1 4 0
Cap
acita
nce
chan
ge [n
F]
A p p lie d fo rce [N ]
e xp e rim e n ta l d a ta lin e a r fit y= 7 1 *x - 4 7
60
65
70
75
80
85
0 50 100 150 200 250kg
kHz
T = 4.2 K
T = 15 K
Applied Force [kg]Re
sona
nce
Freq
uenc
y
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Resonance position shift versus applied force
Again the same behavior for both piezos,
0 N 2 kN
Again the same behavior for both piezos,
0 N 2 kN
Electrical resonance around 20kHz @RT
(two NOLIAC piezostacks)
Electrical resonance around 35kHz @RT
(two NOLIAC piezostacks)
The same behavior for both piezos,
0 N 2 kN
The same behavior for both piezos,
0 N 2 kN
Electrical impedance curve
Electrical impedance curve
18kHz 25kHz
31kHz 38kHz
phas
e [d
eg]
phas
e [d
eg]
Am
plitu
de [a
u]
Am
plitu
de [a
u]
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
System identification
Eacc Eacc
Vpiezo Vpiezo
∆ω ∆ω
Eacc Eacc
Vpiezo Vpiezo
∆ω ∆ω
Let’s assume that system is linearDetuning caused by the piezoelement
might be calculated as a difference between detuning caused by RF field with
piezo action and RF field only.
Detuning is measured using forward power and probe signal. Forward power last only
1,3 ms, therefore there is need to shift piezo pulse versus RF field by T and
perform next measurement.
T
To eliminate microphonics and other noises there is need to average data from
several measurements
X1(s) Y(s)L(s)
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
System identificationE a c cE a c c
V p i e z o V p i e z o
∆ ω∆ ω
E a c c
V p i e z o
∆ ω
E a c c
V p i e z o
∆ ω
E a c c
∆ ω
E a c c
p i e z o
∆ ω
E a c cE a c c
V p i e z o V p i e z o
∆ ω∆ ω
E a c c
V p i e z o
∆ ω
E a c c
V p i e z o
∆ ω
E a c c
∆ ω
E a c c
p i e z o
∆ ω
Piezostack and Lorentz forceact on detuning
Only Lorentz forceact on detuning
Subtraction of previous results is Piezo
influence on detuning.(Time shift is needed for
time sweep)
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
System identification
Spectrum of pulse response(pulse is half sinewave 800Hz) Spectrum of step response
Proper approximation will be finished soon.
Piezoelectric element control signal will be tested during next available access time (probably September 2005)
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Recent results
Remaining detuning is less than 20Hz, for field gradient almost 20MV/m (VUV-FEL, module ACC1, cavity 5)
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
ConclusionsThere are several options for cavity tuners (UMI – coaxial tuner, CEA tuners – old and new one). All tuner are simultaneously developing by CEA Sacley, France and
University of Milan, Italy. A lot of problems are solved so far i.e. neutral point, force measurement at 2K (using i.e. resonance shift monitoring), but there are still
plenty difficulties, which need to be worked out.
There are two types of actuators: magnetostrictive and piezoelectric one. The second one has been tested with cavity with success. The detailed study needs to be
performed to compare both solutions and choose the best one. Both of types were tested successfully at LHe temperature.
First generation of CEA tuner with EPCOS piezostack is already mounted in ACC1 cavity 5 (TTF II). It is possible to reduce detuning caused by LF from 170 to 20Hz
during flat-top (almost 90%).
The control signal is set manually nowadays. However, LF is very repetitive from pulse to pulse (according performed test the same settings are adequate for months).
The feed-forward and feedback algorithm are under developing.
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Acknowledgement
We acknowledge the support of the European Community-Research Infrastructure Activity under the
FP6 “Structuring the European Research Area” program (CARE, contract number RII3-CT-2003-506395),
and Polish National Science Council Grant ”138/E-370/SPB/6.PR UE/DIE 354/2004-2007”.
P. Sekalski, S. Simrock, A. NapieralskiPiezoelectric stack based system for Lorentz force compensation
SRF 2005 Workshop, July 10-15
Thank you for your attention