Document Number
Date
Revision
State
Classification
Page
June 12, 2015
2.4
Released
1 (31)
Procedures for ESS Superconducting Cavities Turn on
Name
Author
ESS: Rihua Zeng, Anders Sunesson, Morten Jensen, Stephen Molloy, Christine Darve,
David McGinnis
CEA Saclay: Franck Peauger, Abdallah Hamdi,
IPN Orsay:
Reviewer
Mohammad Eshraqi, Guillaume Olry, Pierre Bosland
Chess Controlled Procedure for Superconducting Cavities Ed: 2.4
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
REVISION HISTORY
Version
Reason for version
Date
1.0
New Document
20-Sep-2013
2.0
Corrections from S. Molley.
16-Oct-2013
Add procedures for with beam.
2.1
Corrections from M. Eshraqi.
14-Mar-2014
Add “drifting beam method” procedures.
Add appendix from Julien Branlard(DESY), on test proposal for ESS
cavities,.
Add Appendix of identifying beam modes limitation under beam
loading with different beam peak current and different beam pulse
length.
Add references.
2.2
Add procedures recommend by Mircea Stirbet (JLAB), on power
coupler and cavity conditioning
15-Mar-2015
Add content in motivation and background
Modify title order for sub-sections.
2.3
Comments from C. Darve.
4-Jun-2015
Change to formal format used at ESS.
2.4
Modify title and arrange order for sub-sections
Add description and background for high precision RF measurement of
basic cavity parameters, and some corresponding preliminary
procedures.
Add appendix 3: effects of inaccurate parameter measurement on
cavity field under feedforward-only control
Minor modification on appendix 1.
2 (31)
12-Jun-2015
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
TABLE OF CONTENT
PAGE
1. MOTIVATION AND BACKGROUND ...........................................................5
2. CAVITY TURN ON PROCEDURES ...............................................................8
2.1. Procedure for without Beam[9,10,11] ..................................................8
1. RF Cables Calibration ................................................................................. 8
2. Technical Interlock / Sensors. .................................................................... 8
3. RF source / Waveguides / LLRF. ................................................................. 8
4. Cooldown to 2K ......................................................................................... 8
5. Cavity Spectra measurements.................................................................... 8
6. Cavity Tuners Test ...................................................................................... 8
7. Cavities On Resonance. .............................................................................. 8
8. Input RF Couplers and Cavities Conditioning. ............................................. 8
9. Cavity Gradient Ramp up. ........................................................................ 11
10. Cavity Quench Threshold Identification. ................................................ 11
2.2. Procedures for with Beam ..................................................................12
Scheme A: Phase Scan ----- ΔT method[15,17] ............................................... 12
Scheme B: Phase Scan -----signature matching [16]........................................ 13
Scheme C: Beam Loading ------Drift beam [12] ............................................... 14
Scheme D: Beam Loading ------Zero Crossing [13] .......................................... 14
Scheme E: Single bunch transient [14] ........................................................... 15
3. PROCEDURES TO IDENTIFY RF&CAVITY DYNAMICS ...............................17
3.1.
11.
12.
13.
14.
15.
16.
High Precision Measurement of Cavity Basic Parameters ...................17
Single point Ql and Detuning measurement in decay curve ................... 18
Dynamic Ql and Detuning measurement in open loop ........................... 18
Dynamic Ql and Detuning measurement in closed loop......................... 18
Dynamic Ql and Detuning measurement in noisy environment ............. 18
R/Q Calibration ...................................................................................... 18
RF based phase & amplitude calibration ................................................ 19
3.2. Cavity & RF Dynamics .........................................................................19
17. Cavity pass band modes ......................................................................... 19
18. Klystron input-output characteristics (power) at different
modulator voltage ............................................................................ 19
19. Klystron input-output characteristics (phase) at different
modulator voltage ............................................................................ 20
3 (31)
Document Number
Date
Revision
State
Classification
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
June 12, 2015
2.4
Released
Klystron ripple frequency and amplitude ............................................... 20
Lorentz force detuning at different cavity field levels ............................ 20
Lorentz force to cavity tuning transfer function ..................................... 20
Piezo tuner to cavity tuning transfer function (frequency
domain) ............................................................................................ 20
Piezo tuner to cavity tuning transfer function (time domain) ................ 20
Moto tuner to cavity tuning transfer function ....................................... 20
Microphonics spectrum ......................................................................... 20
System open loop matrix ....................................................................... 20
System closed loop matrix ..................................................................... 20
Phase and amplitude setting .................................................................. 21
Cavity field behaviour close to and at quench ........................................ 21
Fast fault recovery ................................................................................. 21
4. OTHER PROCEDURES .............................................................................21
5. REFERENCE: ...........................................................................................21
APPENDIX 1: BEAM MODES LIMITATION IDENTIFICATION STUDIES
IN PHASE & AMPLITUDE SETTING...................................................23
APPENDIX 2: PROPOSAL OF CAVITY TESTS FOR ESS CAVITIES (JULIEN
BRANLARD, DESY) ..........................................................................25
APPENDIX 3: EFFECTS OF INACCURATE PARAMETER MEASUREMENT
ON CAVITY FIELD UNDER FEEDFORWARD-ONLY CONTROL .............27
4 (31)
Document Number
Date
Revision
State
Classification
1.
June 12, 2015
2.4
Released
MOTIVATION AND BACKGROUND
The start point of this document is trying to solve the problem “When the cavity is
installed in tunnel at ESS, how to turn on RF power to reach required gradient, and
what to be done when the beam comes?”
For superconducting cavities, a complex set of procedures are required in different
situation to check, test and calibrate every components, determine configuration
parameters, and identify operational limitations. When it is better to automate most of
procedures in software, which ensures procedures reuse and consistent result
interpretation, it is still acceptable to get sophisticate procedures describe in text in
documentation when automation is not applicable.
When preparing these procedures, it is important to keep in mind ESS challenges and
tough facts expected to face during the phases of commissioning, normally operating and
maintaining for all the normal conducting cavities and superconducting cavities at ESS.
During the stage of procedures investigation and formalization, it is equally important, if
not more, to identify and find suitable solutions to address these challenges, and to
understand better cavity system and get to know its limitations, thereby testing,
controlling and operating the cavity system efficiently and effectively.
The challenges we have to address at ESS are:
1. Cavity Control Challenges. How to deal with the new challenges: longer pulse,
higher beam intensity, higher beam power, higher gradient, spoke cavities and
high demands for energy efficiency and availability.
2. Cavity Operation Challenges. How to deal with wide spread of cavity parameters:
maximum operable gradient, Q load value, operating gradient and synchronous
phase, Lorentz force detuning coefficients, major mechanical modes, power
overhead variations, etc. The same type of cavities (especially superconducting
cavities) behaviour not identically as in theory, due to the fact of complexities and
uncertainties in cavity system manufacturing, surface processing, and assembling.
3. Cavity Test Challenges. How to learn as much as possible from a variety of RF
tests carried out at different test stands and final accelerator tunnel, in order to
better understand the cavity system and know its limitations, thereby operating
the cavity system efficiently and effectively.
When facing up to new challenges, the traditional control, operation and test method
and algorithms fall short in many areas. For example, for the long pulse (~3.5 ms RF pulse,
almost 3 times longer than SNS and DESY), Lorentz force detuning compensation with
driving the piezo tuner by a simple half-cycle sinusoid pulse might no longer suit for ESS
superconducting cavities; For the higher beam intensity, the beam loading in cavities
become heavier. The resulting field stability issue in normal conducting cavities and
serious power overshoot in superconducting cavities can not be well addressed in
feedback plus simple adaptive-feedforward system, especially during beam commission
5 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
phase; For the higher beam power, the same situation of RF setting errors at SNS (up to
2° in phase and 2% in amplitude) might not be acceptable at ESS due to probably higher
beam loss at high power linac of 5MW; For the spoke cavities, which will be the first time
to be operated in the linac, uncertainties require a flexible scheme to quickly change
operation points to avoid any unexpected consequences; For the high energy efficient
requirement, much effort should be paid to minimize the power overhead required for
cavity field control, aiming to reduce 30% power overhead to 10%; For the high
availability requirement, it is essential to establish the schemes of fast recovery from
quench and fast recovery from single/multiple LLRF, klystron, modulator, cavity,
cryomodule failures.
An extensive investigation has been carried out regarding to these new challenges at ESS,
and the design of cavity control strategies has been made with keeping these new
challenges firmly in mind, highlighting the advantages and disadvantages of different
methods and novel ideas newly-proposed in other accelerator facilities, and exploring
possible applications at ESS for these methods and novel techniques through discussion
with experts and detailed review in the literature. The overall control strategies to deal
with new challenges are listed but not limited as follows:















Adaptive compensation for Lorentz force detuning at long pulse via piezo tuner
Beam loading compensation at commissioning phase with different beam peak
current and different beam pulse length.
Beam loading compensation at normal operating phase
Klystron ripple and droop compensation
Residual cavity detuning compensation from pulse to pulse
Klystron linearization
Phase drift calibration under big temperature variations
Strategies and method to deal with slow variations of system environment and
operation conditions
Iterative learning (adaptive feed forward) and optimal control to reduce tracking
error in an optimal way, while keeping the deviation of each step from power
amplifier output small.
Improve cavity phase and amplitude setting accuracy
High precision measurement of basic cavity parameters: Q load value, R/Q, and
dynamic detuning.
Energy efficiency increasing in cavity control and cavity operation
Ability to work at nonlinearities
Ability to work close to limitation
Ability to change operation point quickly and correctly
The methods and novel ideas listed above become possible as consequences of advances
in modern technologies (flexible FPGA, faster CPU, bigger memory, and faster
communication speed), novel control concept, as well as innovative measuring
techniques. High quality data with high resolution, high accuracy and completeness are
6 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
the essential basics to succeed in using these methods and novel ideas. For example,
adaptive compensation for Lorentz force detuning requires accurate measurement of
transfer function from piezo tuner to cavity detuning, and accurate measurement of
cavity detuning; Feed forward compensation for beam loading requires information of
beam current and beam pulse length, as well as high resolution data of beam arrive time
at cavities; Feedforward compensation for klystron ripple and droop requires accurate
measurement of modulator voltage variations; Feed forward compensation for residual
requires accurate and real-time measurement of cavity detuning and QL; Klystron
linearization requires good knowledge of klystron input-output power characteristic
curves at different modulator voltages; Other advanced control and demanding operation
methods require even more strict data quality and special data sets, which will be seen in
the later sections.
Further more, wide spread of cavity parameters make it impossible to employ an single
“uniform’ control algorithm and control configuration data even for the same type of the
cavities. Instead, adjustments and modifications are needed for each cavity and
customized approaches are preferred to reflect cavity’s individual performance and
characteristics. Under this context, a large number of data are expected and required for
all the cavity systems in ESS linac. A controlled, centralized and searchable data storage is
therefore necessary to save the data effectively and ensure to obtain the data
immediately when using them.
When preparing to collect data, it is important to keep in mind which kinds of data are
required, and to what extent the data quality should be. Although it is the best way to
verify the data and refine the requirements in the normal operating cavities, it is always
not practical and too late to generate requirements at final stage. Instead, the data and
general data quality required could be estimated earlier from results of test stands or
similar cavity tests in other accelerator facilities, as well as the prediction of theory
models. As an example, the time-domain transfer function data of piezo tuner to cavity
detuning would be required to compensate the longer pulse Lorentz force detuning,
according to the experiments carried out at Fermilab. Another example is that, the data
resolution of better than 100ns can be concluded from theory model prediction for
feedforward table to compensate beam-loading effects in normal conducting cavities.
In a word, data is the central core to understand the cavity system and to develop
advanced methods and algorithms to address the challenges. Moreover, as it will be seen
in the following sections, the combination and interaction of data, model and
tests/experiments will make great contribution to better system development.
7 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
2.
CAVITY TURN ON PROCEDURES
2.1.
Procedure for without Beam[9,10,11]
This procedure includes the steps for cavity test and normal cavity operation. In practical
operation, turning on the cavity after a short-term trip can start from step 10, when all
preparations such as RF cables calibration, interlock check, system check, cavity
conditioning and cavity on resonance have already done.
1. RF Cables Calibration
 Time Domain Reflectometer (TDR) cables check
 Directional Couplers / Circulators: get calibration data.
 Calibrate RF power measurement cables with
352.21MHz/704.42MHz
 Make RF calibration summary table
attenuators
at
2. Technical Interlock / Sensors.
 Check the sensors (e-, arc detector, water flow, temperature, etc)
 Set the hardware interlock thresholds
 Check the interlock
3. RF source / Waveguides / LLRF.
 Klystron / LLRF check on the load
 Waveguides visual check (how to check those in tunnel?)
 System check / RF leak check at low power
4. Cooldown to 2K
5. Cavity Spectra measurements
(how to measure it for cavity in cryomodule?)



Measure the fundamental mode spectra
Measure the cavities HOMs spectra and Qload
Calibrate the cold RF cables at 2K
6. Cavity Tuners Test
 Test the cavities step-motor frequency tuners and record the motor tuner to
cavity detuning transfer function.
 Tune the cavities to the 352.21MHz/704.42MHz using the Network Analyzer
7. Cavities On Resonance.
 Cavities fine-tuning to the 352.21MHz/704.42MHz by using piezo tuner
 Qload, Kt calibration ( Eacc=Kt(Ptrans)1/2 )
8. Input RF Couplers and Cavities Conditioning.
 Find the cavities/couplers limits at low repetition rate
8 (31)
Document Number
Date
Revision
State
Classification

June 12, 2015
2.4
Released
Run the standard cavity/coupler conditioning program (the following is an
example at Desy):
20, 50, 100, 200 s pulse lengths up to 1MW (minimum 700kW, different for
different sections of ESS linac),
300, 400 s up to 330 kW,
500 s + 100, 200, 400, 800 s flat top pulse up to 250 kW,
Cavities high peak power (HPP) test is part of the conditioning (automatic),
if the cavities limits are lower, run up to the minimum cavity limit.
Prof. Mircea Stirbet recommend the following procedures for input couplers and
cavities conditioning at ESS:
8.1
a)
b)
c)
d)
Check coupler instrumentation and interlocks:
If vacuum gauges are operational
If electrometer for electron activity are operational
If radiation monitors are functional
If cooling fluid (water on the antenna, He gas on the outer conductor) have
the right specifications (flow, temperature). By the way, cooling FPC antennae
with water is a risky choice in SRF applications.
e) If the temperature sensors are reading supposed temperature ( RT, 50 or 70K
intercept, 2K)
f) If arc detector(s) on couplers are operational
g) If ice ball prevention heater is operational (this is a power supply, heating
tape and a temperature sensor which should keep the air side of the coupler
at 25 – 29 C. If this heater is not operational, moisture could condense,
eventually will froze and overnight a nice ice ball could be generated).
h) If the coupler has capacitor for DC bias, this should be OFF (short circuited)
during RF conditioning.
8.2
a)
b)
c)
Proceed with RF conditioning:
Start with low pulse duration
Start with low RF amplitude
Find the cavity and lock on the Pi mode resonance. Presume active cavity
tuning.
d) Set vacuum feedback loop threshold at a convenient initial value (1e-8 – 5e-8
mbar).
9 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
e) Increase RF amplitude (computer controlled operation). If there is vacuum
and electron activity perform RF processing using vacuum feedback loop.
f) Once vacuum and electron activity is less intense, continue increasing RF
pulse amplitude up to the cavity specification.
g) Eventually change vacuum feedback loop threshold to a higher value (1 10-7
mbar). Do not go higher than 2.5 e-7 mbar.
h) Once maximum RF power amplitude is reached, perform a “phase
“conditioning, changing the RF phase in steps of 2-5 degrees.
i) Go back to the initial phase and decrease the RF pulse amplitude.
j) Change pulse duration and repeat steps 8.2 (steps d - g).
k) Continue until the coupler is operational at maximum pulse duration and RF
amplitude is achieved. OBS. The cavity could interfere with FPC conditioning.
Sometimes, helium boiling, temperature runaway, field emission onset or
cavity quench events could manifest before reaching the maximum RF power
amplitude. On cavity, probably the administrative limit on Eacc max should be
used.
l) If field emission manifest during cavity processing (higher than background
ionizing radiation signal) perform eventually cavity RF processing. This step
could be time consuming.
m) Once RF coupler processing was demonstrated with maximum pulse duration,
decrease RF amplitude, turn the RF OFF and set DC bias ON.
n) Turn RF On and ramp the using the maximum pulse duration, up to the
administrative Eacc. Observe vacuum and electron activity. There should be
none.
8.3 If possible, perform operations specified at 8.1 and 8.2 on all couplers, at the
same time (in parallel) on a cryo-module. If conditioning is done sequential,
one coupler after the other, once all couplers and cavities are conditioned, a
test should be done with all 2 (or 4) couplers RF powered at the same time.
Perform a long term run (at least one hour) at constant cavity gradients. Keep
an eye on Helium level, helium temperature, ionizing radiation levels, and
coupler’s temperatures.
8.4 The cryo-module should be ready for beam operation. And this is another
story.
10 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
9. Cavity Gradient Ramp up.
 Load feed forward tables for low gradient operation. Simulated beam with
synchronous phase is included in the feed-forward signals, and preliminary
injection time is calculated according to measured QLoad.
 Modulate the phase of RF power during filling time to track the cavity
resonance frequency, in order to minimize RF power required compensating
detuning effect during filling time.
 Adjust motor tuner and adjust injection time to obtain a constant cavity field
at low gradient operation where Lorentz force detuning is small. The adjusted
injection time is recorded and employed in feedforward tables and feedback
setpoints.
 Set a proper pre-detuning via motor tuner to compensate synchronous phase
operation and to minimize the Lorentz force detuning effect at nominal
gradient. The Lorentz force detuning is measured in advance or is predicted by
models.
 Scale up the feed forward tables to reach higher gradient.
 Raise gradient slowly.
 When rising close to nominal gradient, update feedforward table by
measuring dynamic cavity detuning in previous pulses from forward, reflected,
and transmitted powers. In such a way, Lorentz force detuning and predetuning are expected to be compensated.
 Optimize the pre-detuning to reduce the RF power.
 Apply feedback once the open loop response is close to the desired closed
loop response.
 Increase gain to nominal.
 Optimize feedback parameters for minimum residual amplitude and phase
fluctuation during flat-top.
10. Cavity Quench Threshold Identification.
 Cavity maximum gradient measurement at low repetition rate
 Cavity maximum accelerating gradient measurement at nominal repetition
rate with cryo losses (Qo) and radiation measurements
 Radiation / Dark Current measurements, if needed.
 Quench detection can be made by detecting Qload , since there is sharp drop in
Qload when quench occurring. Fast quench handling can be therefore made in
next RF pulse by reducing or shut off the RF power.
 Detailed and high-resolution quench level identification is essential to work
close to limitation. Firstly, ramp up the cavity gradient with 1 MV/m until a
quench is detected and RF safely turned off. After a brief cryogenic recovery
time, a second ramp-up is performed using smaller gradient increments (0.1
MV/m). The two measurements are correlated and archived, along with the
forward, reflected and probe waveforms leading to the two quenches. Nearquench and safe-operation thresholds are calculated for every cavity as a
function of their measured quench threshold.
11 (31)
Document Number
Date
Revision
State
Classification
2.2.
June 12, 2015
2.4
Released
Procedures for with Beam
1. RF off. Drift low intensity beam (5mA) through Linac.
2. Ramp up the cavity field as mentioned in step 10 in procedure without beam.
3. Set cavity phase and amplitude.
Several methods are applicable here to set the cavity phase and amplitude. At least 2~3
methods expect to be implemented, so as to cross check. These methods are expected to
investigate in details to find proper implementation in order to improve the setting
accuracy. The accuracies obtained at SNS for different methods are listed in Table 1. The
description and general procedure for these methods are given in the following sections.
Scheme A: Phase Scan ----- ΔT method[15,17]
Phase scan methods are referring here to the way of calibrating setting point for RF
cavities by scanning RF phase and amplitude, measuring beam arrival times at
downstream locations, comparing measured phase to model predicted data, and
identifying the best-matched data for calibration. ΔT method is a classical phase scan
method and used widely in normal conducting linac such as in LAMPF, Fermilab, JPARC
and SNS.
Linear system response is assumed in ΔT method and it is only valid in the vicinity of
design phase and amplitude. ΔT method is a cavity-by cavity operation, assuming that the
cavities upstream to the one being adjusted are “on”, and the cavities downstream are
“off”. Beam phases (or beam arrival time) are provided by two downstream BPMs. The
two BPMs can be neighbouring each other, or separated by several cryo-modules, which
depends on the specific location of cavity (the sensitivity of beam velocity to energy gain
becomes low as beam energy goes high) being adjusted. The cavities between two BPMs
should be detuned by 20 (?) cavity bandwidth.
The general procedures of ΔT method are listed below:
1. Find approximate phase and amplitude set point, by observing BPM signals and
beam loading effect, and doing RF based calibration.
2. Cavity being adjusted is off. Record two downstream BPMs phases ϕbpm1-0 and
ϕbpm2-0.
3. Ramp the cavity being adjusted to nominal field calibrated by RF power based
measurement (amplitude accuracy in RF based calibration is around 10%).
4. Turn on beam with repetition rate 14Hz, beam intensity <10mA(see attached
table for DTL. For superconducting cavity, it could be relaxed due to higher cavity
bandwidth), and beam pulse length < 20μs.
5. Record two downstream BPMs phases ϕbpm1 and ϕbpm2.
12 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
6. Calculate relative changes of BPMs phases between cavity “on” and “off” Δϕbpm1 =
ϕbpm1-0 -ϕbpm1 and Δϕbpm2 = ϕbpm2-0 - ϕbpm2. Plot Δϕbpm1 and Δϕbpm2.
7. Scan the cavity synchronous phase with step 0.5°(?) over the range ±5°(?) of
design phase, and repeat 1~5 at each step, to generate a constant-amplitude,
variable-phase curve in (Δϕbpm1, Δϕbpm2) plane.
8. Calculate the slope of the curve, which depends on cavity amplitude, and compare
it with the slope values of model predict curves at different amplitude. These
predicted curves have a common point of intersection.
9. Use some fitting algorithm to determine best-fit amplitude.
10. Having determined proper amplitude, it is now possible in model to calculate the
transfer function relating Δϕbpm1 and Δϕbpm2 to phase deviation Δϕ and energy
deviation ΔW at the entrance of cavity with respect to nominal value. Δϕ and ΔW
can then be determined.
11. Correct the phase set point, and if necessary, correct as well the input energy at
cavity entrance according to the result in step 10.
Scheme B: Phase Scan -----signature matching [16]
Unlike ΔT method having a linear system response and small input energy displacement
restriction, signature matching method can work at large displacement of initial
conditions. In high energy part, signature matching methods can easily scan the phase
over 360° at different amplitude, and make good match with model predict curve.
However, at low energy linac, cavity phase scan can only be several ten degrees where
beam stay sufficiently bunched to produce good signals at downstream BPMs, and the
accuracy indicated at SNS for low energy part is not good enough. ΔT method is probably
necessary to get a good setting accuracy for ESS spoke cavities.
The general procedures for signature matching are listed below:
1. Ramp the cavity being adjusted to nominal field calibrated by RF power based
measurement (amplitude accuracy in RF based calibration is around 10%).
2. Detuning the downstream cavities by 20 cavity bandwidth to bypass the beam,
which locate between two downstream BPMs.
3. Turn on beam with repetition rate 14Hz, beam intensity <10mA(see attached
table for DTL. For superconducting cavity, it could be relaxed due to higher cavity
bandwidth),, and beam pulse length <20μs.
4. Record two downstream BPMs phases ϕbpm1 and ϕbpm2.
5. Scan the cavity synchronous phase with step 0.5°(?) over the full range 360°, and
repeat 4~5 at each step, to generate a constant-amplitude, variable-phase curve
in (Δϕbpm1, Δϕbpm2) plane.
6. Predict the values in model for BPM phases (ϕbpm1_calc and ϕbpm2_calc) as a function
of synchronous phase.
7. Spline fit the measured phase difference (ϕbpm1 - ϕbpm2).
13 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
8. Match the model predict values with measured ones, by minimizing the difference
between (ϕbpm1 - ϕbpm2) and (ϕbpm2_calc - ϕbpm1_calc) over the range of scanned
phase. Phase deviation Δϕ, input beam energy deviation at entrance of cavity ΔW,
and cavity amplitude deviation ΔV are adjusted in this matching procedure.
9. Correct the phase and amplitude set points according to the result in step 8.
Scheme C: Beam Loading ------Drift beam [12]
The drifting beam technique is based on very strong beam-cavity interactions in the SC
cavity for high current beams. It was proposed several years ago and recently realized at
SNS. It uses measured beam currents and pulse shapes with a beam current monitor
(BCM), and beam induced signals in the SC cavity with the cavity control circuit. Using the
measured beam current in a beam-cavity model that simulates the beam-loading in the
cavity, by comparing model simulation results with the actual measurement of the cavity,
cavity phase and the field amplitude are determined precisely.
The general procedures are listed below:
1. Measure accurately beam current and beam pulse shape by BCMs.
2. Tune the cavity to resonance frequency as close as possible (<50Hz?)
3. Turn off RF. Turn on beam with repetition rate 1Hz(?), beam intensity 10mA(?),
beam pulse length > 50 μs.
4. Measure the phase and amplitude of beam-induced signal. Measure the phase
and amplitude of noise signal before next beam pulse coming. Subtract noise
signal from beam-induced signal.
5. Repeat the measurement in step (4) for ~10 beam pulses and average the results.
6. Predict the beam-induced signal in model by measured beam current and beam
pulse shape.
7. Determine the phase offset and amplitude calibration coefficient by comparing
measured result with model calculations.
8. Set amplitude and phase.
Scheme D: Beam Loading ------Zero Crossing [13]
The goal is to adjust the cavity phase such that the beam passes the cavity at the zero
crossing of the accelerating field. This is accomplished by variation of the cavity phase set
point until the average (microphonics !) transients are nulled. The cavity phase is now at 90 deg. or +90 deg. The sign can be determined from a change in transient when
increasing the phase setpoint. The phase is calibrated by adding a phase offset such that
the measured phase reading is correct at zero crossing.
14 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
This method might have higher accuracy since no transient from beam loading cause
perturbations to cavity field.
Scheme E: Single bunch transient [14]
The concept of the transient detector for the X-FEL is based on nulling method, where the
cavity probe signal is split into two branches, one delayed by a up to 100 ns and phase
shifted by 180 degrees before adding the two signals. The nulled signal is amplified by 6080 dB with an rf amplifier and the transient induced by a single bunch is detected by a
schottky diode based rf vector detector to achieve the required low noise performance.
15 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
Table 1: Overview of cavity phase and amplitude setting methods used at SNS
Warm Linac
Superconducting Linac
Beam Based Calibration
Energy
degrader
(only for
DTL)
Delt-T
Beam Based Calibration
Phase-scan
signature
RF based
Calibration
Phase-scan
signature
Drift Beam
±5%[1]
Amplitude
1.7% (at
most) differ
from Delt-T
results[7]
Varied 0~5%
with Delt-T
results
±2.4%[1]
±4%(better
results after
precise
calibration for
BCM)[1]
Phase
1.5 °(at
most) differ
from Delt-T
results[7]
Varied 0~11°
with Delt-T
results
±1°
±1°
Beam pulse
<50us[5]
(Calibration
uncertainties
could be
checked in
test stand)
~20us[2]
<50us
<50us[8]
>50us[4]
N/A
<20mA[4]
N/A
1Hz
1Hz
5~20us[3]
<20mA[2]
Beam
Current
<20mA[5]
Repetition
rates
1Hz
Error Source
>5mA
1Hz
Beam
parame
-ters
5mA is
threshold*[6]
1Hz
1Hz
BPM
Calibration,
Beam
parameters,
Beam
loading in
unpowered
cavities
BPM
Calibration,
Beam
parameters,
Beam loading
in unpowered
cavities
16 (31)
BCM
Calibration,
Beam
Parameters,
LLRF noise,
Cavity detuning,
Passband
modes
Measurement
errors, Cavity
parameters
errors
Document Number
Date
Revision
State
Classification
3.
June 12, 2015
2.4
Released
PROCEDURES TO IDENTIFY RF&CAVITY DYNAMICS
The deeper understanding of RF & cavity we want to have, the more details on RF& cavity
dynamics would require to figure out. Significant benefit will be then followed:




Significant Energy Efficiency improvement in LLRF control becomes possible. Big
effort has been made to investigate how operate LLRF at 90% of klystron
saturation.
High quality measurement data on basic cavity parameters: Ql, R/Q, dynamic
detuning, phase, amplitude
Give fundamental insight into RF & Cavity dynamics
Trace all the way from cavity test, cavity commissioning, and cavity operation,
which in turn provide great benefit for cavity manufacturing.
This chapter will discuss tests for individual modules separately, and focus on the
following issues:




3.1.
Why do we need these tests
Theory and method behind test and data processing to obtain required data
Standard test procedure, which not only include how to carry out test, measure
signals, but equally important are the data analysis, data organization and data
storage
Automation
High Precision Measurement of Cavity Basic Parameters
More details can be seen in attached thesis from Philip Jönsson “High Precision RF
Measurement for ESS Cavity Parameters”.

The ESS cavity control and operation methods/algorithms are challenging due to
the use of long pulse, higher beam intensity, high beam power, high gradient,
uncertainties in spoke cavities and high demands for energy efficiency and
availability. Suitable and effective solutions is been identifying at ESS to address
these challenges, taking fully use of advantage gained at ESS, with great benefit
from advantages in digital technology revolution, progress in high performance in
hardware, achievement in system-level optimization, and remarkable
achievement in accelerator community on high precision RF measurement and
innovative automatic control for superconducting cavities. All these potential
solutions rely on high precision measurement of basic cavity parameters and
consequent high quality data with high resolution, high precision and
completeness. The following procedures are given in this sections on high
precision measurement of Ql, R/Q, dynamic detuning, phase, and amplitude.
17 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
11. Single point Ql and Detuning measurement in decay curve
DESY example: early than 1991(Ql) and early than 1998(detuning)
12. Dynamic Ql and Detuning measurement in open loop
Ql and detuning can be derived from power measurement and careful calibration of
forward power. Detuning can also be measured from the phase difference between
incident wave and probe wave. The accuracies of each method have to be compared.
13. Dynamic Ql and Detuning measurement in closed loop
Ql and detuning are derived from the power measurement (forward, reflected, probe), it
is also found in other labs that better result can be obtained in second derivation:
(proposed earlier than 1998)
dP
 1/ 2  i P  21/ 2 F
dt
1/ 2  
  dP  
Re  P * 

  dt  
Re P * P  2 F 
  dP

Im P * 
 21/ 2 F  
 dt

  
*
PP
14. Dynamic Ql and Detuning measurement in noisy environment
15. R/Q Calibration
A direct way to determine R/Q is the “bead pulling” field profile measuring method, by
monitoring π-mode frequency offset when perturbing cavity field using a small metal
bead. This method is suitable for offline measurement but not for operation.
18 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
An alternative method to calibrate R/Q could make use of the voltage induced by a short
beam or RF pulse. For the very short pulses TB (TB≪1/ω1⁄2, TB<<1/∆ω, ω1⁄2 is the cavity
half bandwidth), the maximum value that pulse induced voltage reaches can be
approximately written as:
Vmax »
w0
4
( R Q) TB × I
With carefully beam-based calibration data of cavity phase and amplitude as mentioned
in cavity turn on procedures with beam, by measuring voltages induced by short beam/RF
pulses with different currents and using linear regression, R/Q is then determined. The
calibration procedure is as follow.
16. RF based phase & amplitude calibration
3.2.
Cavity & RF Dynamics
17. Cavity pass band modes
18. Klystron input-output characteristics (power) at different modulator voltage
By choose appropriate RF input waveform in cavity filling time, an online klystron inputoutput characteristics can be obtained with direct measurement on klystron input power
and output power. The procedure is as follow (see attached thesis from Staffan Rydén,
“Beam Loading Studies for Cavity Phase and Amplitude Setting”).
19 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
19. Klystron input-output characteristics (phase) at different modulator voltage
20. Klystron ripple frequency and amplitude
21. Lorentz force detuning at different cavity field levels
22. Lorentz force to cavity tuning transfer function
23. Piezo tuner to cavity tuning transfer function (frequency domain)
24. Piezo tuner to cavity tuning transfer function (time domain)
earlier than 2011, test in Fermilab, FLASH, JPARC…
Test Steps:
1. Excite the piezo with a series of impulses and sweep the impulse-to-RF delay
(Equivalent to measuring the piezo-to-detuning transfer function using CW)
2. The forward, probe and reflected RF waveform recorded at each delay and used
to calculate detuning.
3. The response 2-D matrix is then obtained (detuning vs. time at different peizo
pulse dealys)
4. The response matrix can be inverted to give waveform required to produce a
given detuning profile.
25. Moto tuner to cavity tuning transfer function
26. Microphonics spectrum
27. System open loop matrix
28. System closed loop matrix
Start earlier than 1998 (mature system used in Flash and will be in XFEL)
20 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
Test step:
1. find an appropriate excitation signal to excite all system dynamics to be
modelled.(time domain, frequency domain?)
2. remove the trend and means from observed data, by averaging, filtering, shift
data sequences or remove outliers.
3. assign model structure(the level of complexity)
4. choose identification algorithm
5. validate the model(compare with the measured system data not used in
identification, or with other physical known parameters )
29. Phase and amplitude setting
30. Cavity field behaviour close to and at quench
31. Fast fault recovery
4.
OTHER PROCEDURES
5.
REFERENCE:
[1] Y. Zhang, et al., Comparision of SNS Superconducting Cavity Calibration Methods.
[2] D. Jeon, et al., Beam Loading Effects on Phase Scan for Superconducting Cavities.
[3] S. Henderson et al., The Commissioning Plan for the Spallation Neutron Source Ring
and Transport Lines.
[4] Y. Zhang et al., Determination of field amplitude and synchronous phase using the
beam-induced signal in an unpowered superconducting cavity
[5] J. Galambos, SNS Commissioning Strategies and Tuning Algorithms.
[6] J. Galambos, Pasta- an RF phase scan and tuning application
[7] Dong-o Jeon, Comparision of phase scan vs acceptance scan for the SNS DTL
[8] A.V. Aleksandrov, SNS Warm Linac Commissioning Results
9. Superconducting Accelerating Cryo-Module Tests at DESY.
10. OPERATIONAL ASPECTS OF THE RF CONTROL SYSTEM FOR THE TESLA TEST FACILITY.
11. LLRF TESTING OF SUPERCONDUCTING CRYOMODULES FOR THE EUROPEAN XFEL.
12. Determination of field amplitude and synchronous phase using the beam-induced
signal in an unpowered superconducting cavity.
21 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
13. TRANSIENT BEAM LOADING BASED CALIBRATION OF THE VECTOR-SUM FOR THE
TESLA TEST FACILITY.
14. SINGLE BUNCH TRANSIENT DETECTION FOR THE BEAM PHASE MEASUREMENT IN
SUPERCONDUCTING ACCELERATORS
15. DEVELOPMENT AND IMPLEMENTATION OF ∆T PROCEDURE FOR THE SNS* LINAC
16. PASTA an RF phase scan and tuning application
17. The Delta-T Tuneup Procedure for the LAMPF 805 MHz Linac,
22 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
APPENDIX 1: BEAM MODES LIMITATION IDENTIFICATION STUDIES IN
PHASE & AMPLITUDE SETTING
The table below shows whether the perturbations, caused by different beam pulse length
and peak currents, are within cavity field stability ±1%, 1°. (For DTL tank 2. First table is
result with feedback control. Second table is result without feedback control). For
superconducting cavity, perturbation expects to be less due to large cavity voltage time
constant). Tables are from Staffan Rydén’s thesis, “Beam Loading Studies for Cavity
Phase and Amplitude Setting”).
, 084- 89: 46; 21.6< 64140/4/1; =1=60.>1?; .4: 301: 2>1@5: /01=; 81! 1A 1: 2>1! 1>038001088; 81.6< 64/7
B0C71*1D1, : //0/19; 45140/4/E1! 1D1F: 6./19; 45140/4/E1" 1D1F: 6./1=60.>140/4E1#1D1F: 6./1@5: /0140/4
G0: < 1H- 880241621 , - ./01.023451621- /7
<I7
! " # $ % & ' ( ) !* !! !" !# !$ !%!& !' !( !)
" +% * * * * * * * * * * * * * * * * * * *
$+% * * * * * * * * * * * * * * * * * * *
&+% * * * * * * * * * * * * * * * * * * *
( +% * * * * * * * * * * * * * * * * * * *
! *+% * * " " " " " " " " " " " " " " " " "
! " +% * " " " " " " " " " " " " " " " " " "
! $+% * " " " " " " " " " " " " " " " " " "
! &+% * " " " " " " " " " " " " " " " " " "
! ( +% * " " " " " " " " " " " " " " " " " "
" *+% * " " " " " " " " " " " " " " " " " "
" " +% " " " " " " " " " " " " " " " " " " "
" $+% " " " " " " " " " " " " " " " " " " "
" &+% " " " " " " " " " " " " " " " " " " "
" ( +% " " " " " " " " " " " " " " " " " " "
#*+% " " " " " " " " " " " " " " " " " " "
#" +% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#$+% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#&+% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#( +% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$*+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$" +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$$+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$&+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$( +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%*+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%" +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%$+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%&+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%( +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
&*+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
&" +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
23 (31)
"*
*
*
*
*
"
"
"
"
"
"
"
"
"
"
"
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
, 084- 89: 46; 21.6< 64140/4/1; =1=60.>1?; .4: 301: 2>1@5: /01=; 81! 1A 1: 2>1! 1>038001088; 81.6< 64/7
B0C71*1D1, : //0/19; 45140/4/E1! 1D1F: 6./19; 45140/4/E1" 1D1F: 6./1=60.>140/4E1#1D1F: 6./1@5: /0140/4
G0: < 1H- 880241621 , - ./01.023451621- /7
<I7
! " # $ % & ' ( ) !* !! !" !# !$ !%!& !' !( !)
" +% * * * * * * * * * * * * * * " " " " "
$+% * * * * * " " " " " " " " " " " " " "
&+% * * * " " " " " " " " " " " " " " ! !
( +% * * " " " " " " " " " ! ! ! ! ! ! ! !
! *+% * * " " " " " " ! ! ! ! ! ! ! ! ! ! !
! " +% * " " " " " " ! ! ! ! ! ! ! ! ! ! ! !
! $+% * " " " " ! ! ! ! ! ! ! ! ! ! ! ! ! !
! &+% * " " " " ! ! ! ! ! ! ! ! ! ! ! ! ! !
! ( +% * " " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
" *+% * " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
" " +% " " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
" $+% " " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
" &+% " " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
" ( +% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#*+% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#" +% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#$+% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#&+% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
#( +% " " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$*+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$" +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$$+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$&+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
$( +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%*+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%" +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%$+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%&+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
%( +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
&*+% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
&" +% " ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
24 (31)
"*
"
"
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
APPENDIX 2: PROPOSAL OF CAVITY TESTS FOR ESS CAVITIES (JULIEN
BRANLARD, DESY)
2013-12-06 (From Julien Branlard, DESY)
1. Cavity fundamental mode measurements
a. 5 minute / cavity
b. Measure spectrum with network analyzer
c. Check the center frequency is according to specs
d. Can also measure Ql from width of resonance peak
e. Measure other sub-pi modes at the same time
2. Calibration of cold cables
a. Quick (5 mins)
b. Required in order to get accurate gradient measurement
c. Network analyzer measurement, s11 off resonance

3. Coupler conditioning
a. Crucial step
b. Necessary (especially for 1.5 MW per coupler)
c. Has to be done warm, then cold
d. Time consuming (several days)
e. Requires automation (exception handling is what makes it hard)
f. NOTE: for cold conditioning, cavity gets gradient  more complicated
(Xrays, field emission, field breakdown, quenches, etc…)

4. Cavity gradient calibration
a. Need to measure loaded Q and compare against specifications
b. From loaded Q, you derive kT parameter
c. Vcav proportional to kT SQRT(Pfwd) (I think Q0 is also needed)
d. Just a computation , not an actual test but it’s required
e. Check gradient against specifications
f. Could perform gradient check and verify quench limit
g. Makes sense to do this before cold coupler conditioning

5. Heat load
a. Time consuming because cryo time constant is big
b. Let cavity sit at one gradient, measure heat load
c. Increase gradient, repeat etc…
25 (31)
Document Number
Date
Revision
State
Classification
d.
6. Tuner
a.
b.
c.
d.

7. Piezo
a.
b.
c.
June 12, 2015
2.4
Released
Check dynamic heat load against specifications
Check functionality
Check tuning range, tuning constant
Performed with network analyzer,
observe peak frequency change as a function of tuner steps
Check functionality
Apply DC bias and verify frequency shift
Verify maximum tuning range  has an implication on where the slow
tuner center frequency should be

8. Lorentz force detuning
a. Relative fast test
b. Place cavity on resonance
c. Ramp up gradient
d. Measure detuning (phase slope) as a function of gradient
e. Derive LFD constant
f. Could be done in parallel with cold coupler conditioning


26 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
APPENDIX 3: EFFECTS OF INACCURATE PARAMETER MEASUREMENT ON
CAVITY FIELD UNDER FEEDFORWARD-ONLY CONTROL
Pictures are cut from Philip Jönsson’s thesis “High Precision RF Measurement for ESS
Cavity Parameters”.
Inaccurate amplitude
27 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
Inaccurate phase
28 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
Inaccurate R/Q
29 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
Inaccurate Ql
30 (31)
Document Number
Date
Revision
State
Classification
June 12, 2015
2.4
Released
Inaccurate dynamic detuning
31 (31)