experience with feedback systems in modern synchrotron light … 2019-04-03 · experience with...
TRANSCRIPT
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Experience with feedback systems in modernsynchrotron light sources
D. Teytelman
Dimtel, Inc., San Jose, CA, USA
Accelerator Performance and Concepts WorkshopFrankfurt, December 10-12, 2018
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Outline
Introduction
Similarities Under Feedback Control
Feedback StoriesANKA Detective StoryNSLS-II Residual Motion Story
Future ChallengesMain IssuesTransverse Feedback and NoiseLongitudinal Instabilities and Harmonic Cavities
Extra Slides
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Learning From Many FacilitiesMachines
Ring C, m E, GeVMLS 48 0.1–0.6HLS 66 0.8LNLS UVX 93 0.5–1.37MAX IV 1.5 GeV 96 1.5DAΦNE 98 0.51Duke SR-FEL 108 0.2–1.2ANKA 110 0.5–2.5DELTA 115 1.5TLS 120 1.5ELSA 164 1.2–3.2Indus-2 173 0.55–2.5Photon Factory 187 2.5ALS 197 1.9Australian Synchrotron 216 3SPEAR3 234 3BEPC-II 238 1.89BESSY II 240 1.7TPS 518 3MAX IV 3 GeV 528 3CESR-TA 768 1.5–6NSLS-II 792 3SuperKEKB 3016 4/7
I Over the last 12 years I had apleasure of directly or indirectlyparticipating in commissioningbunch-by-bunch feedback in 22machines;
I A definite learning opportunity!I Helped me gain some
understanding of feedback limitingfactors;
I Becoming more important in futureaccelerators.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Learning From Many FacilitiesMachines
Ring C, m E, GeVMLS 48 0.1–0.6HLS 66 0.8LNLS UVX 93 0.5–1.37MAX IV 1.5 GeV 96 1.5DAΦNE 98 0.51Duke SR-FEL 108 0.2–1.2ANKA 110 0.5–2.5DELTA 115 1.5TLS 120 1.5ELSA 164 1.2–3.2Indus-2 173 0.55–2.5Photon Factory 187 2.5ALS 197 1.9Australian Synchrotron 216 3SPEAR3 234 3BEPC-II 238 1.89BESSY II 240 1.7TPS 518 3MAX IV 3 GeV 528 3CESR-TA 768 1.5–6NSLS-II 792 3SuperKEKB 3016 4/7
I Over the last 12 years I had apleasure of directly or indirectlyparticipating in commissioningbunch-by-bunch feedback in 22machines;
I A definite learning opportunity!I Helped me gain some
understanding of feedback limitingfactors;
I Becoming more important in futureaccelerators.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Learning From Many FacilitiesMachines
Ring C, m E, GeVMLS 48 0.1–0.6HLS 66 0.8LNLS UVX 93 0.5–1.37MAX IV 1.5 GeV 96 1.5DAΦNE 98 0.51Duke SR-FEL 108 0.2–1.2ANKA 110 0.5–2.5DELTA 115 1.5TLS 120 1.5ELSA 164 1.2–3.2Indus-2 173 0.55–2.5Photon Factory 187 2.5ALS 197 1.9Australian Synchrotron 216 3SPEAR3 234 3BEPC-II 238 1.89BESSY II 240 1.7TPS 518 3MAX IV 3 GeV 528 3CESR-TA 768 1.5–6NSLS-II 792 3SuperKEKB 3016 4/7
I Over the last 12 years I had apleasure of directly or indirectlyparticipating in commissioningbunch-by-bunch feedback in 22machines;
I A definite learning opportunity!I Helped me gain some
understanding of feedback limitingfactors;
I Becoming more important in futureaccelerators.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Learning From Many FacilitiesMachines
Ring C, m E, GeVMLS 48 0.1–0.6HLS 66 0.8LNLS UVX 93 0.5–1.37MAX IV 1.5 GeV 96 1.5DAΦNE 98 0.51Duke SR-FEL 108 0.2–1.2ANKA 110 0.5–2.5DELTA 115 1.5TLS 120 1.5ELSA 164 1.2–3.2Indus-2 173 0.55–2.5Photon Factory 187 2.5ALS 197 1.9Australian Synchrotron 216 3SPEAR3 234 3BEPC-II 238 1.89BESSY II 240 1.7TPS 518 3MAX IV 3 GeV 528 3CESR-TA 768 1.5–6NSLS-II 792 3SuperKEKB 3016 4/7
I Over the last 12 years I had apleasure of directly or indirectlyparticipating in commissioningbunch-by-bunch feedback in 22machines;
I A definite learning opportunity!I Helped me gain some
understanding of feedback limitingfactors;
I Becoming more important in futureaccelerators.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Steady State Spectra Under Feedback Control
Feedback Beam∑
Transverse/longitudinal position
Detection noise
∑
Disturbances
Error
I In transverse planes there arevery few steady-statedisturbances;
I Instabilities are damped to thenoise floor;
I Spectrum is determined by thedetection noise and thefeedback loop response;
I Open loop transfer function L(ω)peaks at beam resonance;
I Transfer gain from detectionnoise to the feedback input is
11+L(ω) — a notch.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Steady State Spectra Under Feedback Control
Feedback Beam∑
Transverse/longitudinal position
Detection noise
∑
Disturbances
Error
I In transverse planes there arevery few steady-statedisturbances;
I Instabilities are damped to thenoise floor;
I Spectrum is determined by thedetection noise and thefeedback loop response;
I Open loop transfer function L(ω)peaks at beam resonance;
I Transfer gain from detectionnoise to the feedback input is
11+L(ω) — a notch.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Steady State Spectra Under Feedback Control
Feedback Beam∑
Transverse/longitudinal position
Detection noise
∑
Disturbances
Error
I In transverse planes there arevery few steady-statedisturbances;
I Instabilities are damped to thenoise floor;
I Spectrum is determined by thedetection noise and thefeedback loop response;
I Open loop transfer function L(ω)peaks at beam resonance;
I Transfer gain from detectionnoise to the feedback input is
11+L(ω) — a notch.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Steady State Spectra Under Feedback Control
0.1 0.15 0.2 0.25 0.3−40
−30
−20
−10
0
10
20
Fractional frequency
Magnitude (
dB
)
I In transverse planes there arevery few steady-statedisturbances;
I Instabilities are damped to thenoise floor;
I Spectrum is determined by thedetection noise and thefeedback loop response;
I Open loop transfer function L(ω)peaks at beam resonance;
I Transfer gain from detectionnoise to the feedback input is
11+L(ω) — a notch.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Steady State Spectra Under Feedback Control
0.1 0.15 0.2 0.25 0.3−12
−10
−8
−6
−4
−2
0
2
Fractional frequency
Magnitude (
dB
)
I In transverse planes there arevery few steady-statedisturbances;
I Instabilities are damped to thenoise floor;
I Spectrum is determined by thedetection noise and thefeedback loop response;
I Open loop transfer function L(ω)peaks at beam resonance;
I Transfer gain from detectionnoise to the feedback input is
11+L(ω) — a notch.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
0 100 200 300 400 500 600
10−1
100
Freq (kHz)
Counts
jan0513/213833: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
310 315 320 325 330 335 340 345 350 35510
−1.8
10−1.7
10−1.6
Freq (kHz)
Counts
jan0513/213833: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
0 500 1000 1500 200010
−2
10−1
100
101
102
Freq (kHz)
Counts
oct2518/135004: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
1020 1040 1060 1080 1100 1120 1140 1160
10−1.9
10−1.8
10−1.7
Freq (kHz)
Counts
oct2518/135004: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
0 200 400 600 800 1000 1200
10−2
10−1
100
Freq (kHz)
Counts
apr0710/215244: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
400 450 500 550 600 650 700 750 800 85010
−3
10−2
10−1
Freq (kHz)
Counts
apr0710/215244: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
730 740 750 760 770 780 790 800 810 82010
−3
10−2
10−1
Freq (kHz)
Counts
apr0710/215244: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Steady State Spectra
390 400 410 420 430 440 450 460 470 48010
−3
10−2
10−1
Freq (kHz)
Counts
apr0710/215244: Signal spectrum averaged (quadratic) over all bunches
I Averaged spectra of allbunches under closed-loopfeedback control;
I BESSY II1 Y at 298 mA;I Zoomed in;I Aichi SR2 X at 300 mA;I Zoomed in;I TLS3 at 200 mA;I Two notches - dual plane (X
and Y) feedback;I Clean horizontal notch;I A line poking through the
vertical notch — evidence ofion-driven instabilities.
1h = 400, C = 240 m2h = 120,C = 72 m3h = 200,C = 120 m
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Outline
Introduction
Similarities Under Feedback Control
Feedback StoriesANKA Detective StoryNSLS-II Residual Motion Story
Future ChallengesMain IssuesTransverse Feedback and NoiseLongitudinal Instabilities and Harmonic Cavities
Extra Slides
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Initial Assessment
19:40 20:09 20:38 21:07 21:36 19:40 20:09 20:38 21:070
20
40
60
80
100
120
Time
Be
am
cu
rre
nt
(mA
)
ANKA, 2016−03−04
I A ramping machine, 500 MeVinjection, 2.5 GeV operation;
I Historically operated with onlytransverse feedback, withmoderate longitudinalinstabilities at injection energy;
I After a shutdown had problemsinjecting more than100–120 mA, limited bysudden partial beam loss.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Initial Assessment
19:40 20:09 20:38 21:07 21:36 19:40 20:09 20:38 21:070
20
40
60
80
100
120
Time
Be
am
cu
rre
nt
(mA
)
ANKA, 2016−03−04
I A ramping machine, 500 MeVinjection, 2.5 GeV operation;
I Historically operated with onlytransverse feedback, withmoderate longitudinalinstabilities at injection energy;
I After a shutdown had problemsinjecting more than100–120 mA, limited bysudden partial beam loss.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Initial Assessment
19:40 20:09 20:38 21:07 21:36 19:40 20:09 20:38 21:070
20
40
60
80
100
120
Time
Be
am
cu
rre
nt
(mA
)
ANKA, 2016−03−04
I A ramping machine, 500 MeVinjection, 2.5 GeV operation;
I Historically operated with onlytransverse feedback, withmoderate longitudinalinstabilities at injection energy;
I After a shutdown had problemsinjecting more than100–120 mA, limited bysudden partial beam loss.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Feedback in All Three Planes
I Turned on feedback in thelongitudinal plane;
I Still hitting a limit during injection,with partial beam losses;
I Feedback tuned near absolutelimit, growth time 2.3× Ts,damping time Ts;
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Feedback in All Three Planes
I Turned on feedback in thelongitudinal plane;
I Still hitting a limit during injection,with partial beam losses;
I Feedback tuned near absolutelimit, growth time 2.3× Ts,damping time Ts;
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Feedback in All Three Planes
0 0.2 0.4 0.6 0.8 1 1.2 1.40
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
Time (ms)
deg@
RF
mar0516/143809 Data, Fit and Error for Mode #45
Data
Fit
Error I Turned on feedback in thelongitudinal plane;
I Still hitting a limit during injection,with partial beam losses;
I Feedback tuned near absolutelimit, growth time 2.3× Ts,damping time Ts;
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Capturing the Beam Loss Event
0 5 10 15 20 25 30−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ounts
Longitudinal plane
0 5 10 15 20 25 30−100
0
100
200
300
400
500
600
Time (ms)
AD
C c
ounts
Vertical plane
I Noticed significant activity in thevertical plane during beam lossevents;
I Used baseband processor outputto trigger acquisition in all planes;
I Vertical correction signals arenormally small, only reachingfull-scale during beam loss;
I Longitudinal and vertical signalsfor bunch 140.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Capturing the Beam Loss Event
0 5 10 15 20 25 30−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ounts
Longitudinal plane
0 5 10 15 20 25 30−100
0
100
200
300
400
500
600
Time (ms)
AD
C c
ounts
Vertical plane
I Noticed significant activity in thevertical plane during beam lossevents;
I Used baseband processor outputto trigger acquisition in all planes;
I Vertical correction signals arenormally small, only reachingfull-scale during beam loss;
I Longitudinal and vertical signalsfor bunch 140.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Capturing the Beam Loss Event
0 5 10 15 20 25 30−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ounts
Longitudinal plane
0 5 10 15 20 25 30−100
0
100
200
300
400
500
600
Time (ms)
AD
C c
ounts
Vertical plane
I Noticed significant activity in thevertical plane during beam lossevents;
I Used baseband processor outputto trigger acquisition in all planes;
I Vertical correction signals arenormally small, only reachingfull-scale during beam loss;
I Longitudinal and vertical signalsfor bunch 140.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Capturing the Beam Loss Event
0 5 10 15 20 25 30−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ounts
Longitudinal plane
0 5 10 15 20 25 30−100
0
100
200
300
400
500
600
Time (ms)
AD
C c
ounts
Vertical plane
I Noticed significant activity in thevertical plane during beam lossevents;
I Used baseband processor outputto trigger acquisition in all planes;
I Vertical correction signals arenormally small, only reachingfull-scale during beam loss;
I Longitudinal and vertical signalsfor bunch 140.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event
0 5 10 15 20 25 30−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ounts
Longitudinal plane
0 5 10 15 20 25 30−100
0
100
200
300
400
500
600
Time (ms)
AD
C c
ounts
Vertical plane
I Which plane came first?I Zoom in, still too close;I Zoom more — looks like
longitudinal starts first, but couldbe trigger error;
I Longitudinal oscillation amplitudeexceeds 30◦;
I Modal analysis in Z shows mode46 rapidly running away under fullfeedback control.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event
8 8.5 9 9.5 10 10.5 11 11.5 12−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ounts
Longitudinal plane
8 8.5 9 9.5 10 10.5 11 11.5 12−100
0
100
200
300
400
500
600
Time (ms)
AD
C c
ounts
Vertical plane
I Which plane came first?I Zoom in, still too close;I Zoom more — looks like
longitudinal starts first, but couldbe trigger error;
I Longitudinal oscillation amplitudeexceeds 30◦;
I Modal analysis in Z shows mode46 rapidly running away under fullfeedback control.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event
9.6 9.7 9.8 9.9 10 10.1 10.2 10.3 10.4−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ou
nts
Longitudinal plane
9.6 9.7 9.8 9.9 10 10.1 10.2 10.3 10.40
100
200
300
400
500
600
Time (ms)
AD
C c
ou
nts
Vertical plane
I Which plane came first?I Zoom in, still too close;I Zoom more — looks like
longitudinal starts first, but couldbe trigger error;
I Longitudinal oscillation amplitudeexceeds 30◦;
I Modal analysis in Z shows mode46 rapidly running away under fullfeedback control.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event
9.6 9.7 9.8 9.9 10 10.1 10.2 10.3 10.4−150
−100
−50
0
50
100
150
200
250
Time (ms)
AD
C c
ou
nts
Longitudinal plane
9.6 9.7 9.8 9.9 10 10.1 10.2 10.3 10.40
100
200
300
400
500
600
Time (ms)
AD
C c
ou
nts
Vertical plane
I Which plane came first?I Zoom in, still too close;I Zoom more — looks like
longitudinal starts first, but couldbe trigger error;
I Longitudinal oscillation amplitudeexceeds 30◦;
I Modal analysis in Z shows mode46 rapidly running away under fullfeedback control.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event
I Which plane came first?I Zoom in, still too close;I Zoom more — looks like
longitudinal starts first, but couldbe trigger error;
I Longitudinal oscillation amplitudeexceeds 30◦;
I Modal analysis in Z shows mode46 rapidly running away under fullfeedback control.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event, Continued
I Spectrogram of vertical bunchsignal settles it!
I Longitudinal motion starts first,seen mostly as second harmonicin the amplitude detector channel;
I At beam phase excursionsexceeding 90◦ at detectionfrequency, vertical feedback gainflips;
I Positive feedback excites verticalmotion, causes beam loss.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event, Continued
I Spectrogram of vertical bunchsignal settles it!
I Longitudinal motion starts first,seen mostly as second harmonicin the amplitude detector channel;
I At beam phase excursionsexceeding 90◦ at detectionfrequency, vertical feedback gainflips;
I Positive feedback excites verticalmotion, causes beam loss.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event, Continued
I Spectrogram of vertical bunchsignal settles it!
I Longitudinal motion starts first,seen mostly as second harmonicin the amplitude detector channel;
I At beam phase excursionsexceeding 90◦ at detectionfrequency, vertical feedback gainflips;
I Positive feedback excites verticalmotion, causes beam loss.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Deciphering the Beam Loss Event, Continued
I Spectrogram of vertical bunchsignal settles it!
I Longitudinal motion starts first,seen mostly as second harmonicin the amplitude detector channel;
I At beam phase excursionsexceeding 90◦ at detectionfrequency, vertical feedback gainflips;
I Positive feedback excites verticalmotion, causes beam loss.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Fixing the Issue
I RF cavities in ANKA arecompression tuned;
I During injection, LLRF adjustscavity tuning to compensate forbeam loading;
I All cavity HOMs tune at the sametime;
I By adjusting cavity temperature weshifted the problematic HOMenough so that synchrotronsideband crossing happened atmuch lower beam current;
I Growth rates scale with current,feasible to control!
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Fixing the Issue
I RF cavities in ANKA arecompression tuned;
I During injection, LLRF adjustscavity tuning to compensate forbeam loading;
I All cavity HOMs tune at the sametime;
I By adjusting cavity temperature weshifted the problematic HOMenough so that synchrotronsideband crossing happened atmuch lower beam current;
I Growth rates scale with current,feasible to control!
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Fixing the Issue
I RF cavities in ANKA arecompression tuned;
I During injection, LLRF adjustscavity tuning to compensate forbeam loading;
I All cavity HOMs tune at the sametime;
I By adjusting cavity temperature weshifted the problematic HOMenough so that synchrotronsideband crossing happened atmuch lower beam current;
I Growth rates scale with current,feasible to control!
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Fixing the Issue
I RF cavities in ANKA arecompression tuned;
I During injection, LLRF adjustscavity tuning to compensate forbeam loading;
I All cavity HOMs tune at the sametime;
I By adjusting cavity temperature weshifted the problematic HOMenough so that synchrotronsideband crossing happened atmuch lower beam current;
I Growth rates scale with current,feasible to control!
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Fixing the Issue
I RF cavities in ANKA arecompression tuned;
I During injection, LLRF adjustscavity tuning to compensate forbeam loading;
I All cavity HOMs tune at the sametime;
I By adjusting cavity temperature weshifted the problematic HOMenough so that synchrotronsideband crossing happened atmuch lower beam current;
I Growth rates scale with current,feasible to control!
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
The Final Problem
18:43 19:12 19:40 20:09 20:38 18:43 19:12 19:40 20:090
20
40
60
80
100
120
Time
Be
am
cu
rre
nt
(mA
)
ANKA, 2016−03−05
I After adjusting cavity temperaturewe were able to keep the beamstable in X, Y, and Z;
I New problem — with all planesstable, injection saturated around110 mA due to poor Touscheklifetime;
I Streak camera, stabilized beam;I Applied quadrupole excitation
through the LFB;I Bunch lengthening leads to
Touschek lifetime improvement;I With both feedback and
modulation injected 160 mA andramped to 2.5 GeV.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
The Final Problem
18:43 19:12 19:40 20:09 20:38 18:43 19:12 19:40 20:090
20
40
60
80
100
120
Time
Be
am
cu
rre
nt
(mA
)
ANKA, 2016−03−05
I After adjusting cavity temperaturewe were able to keep the beamstable in X, Y, and Z;
I New problem — with all planesstable, injection saturated around110 mA due to poor Touscheklifetime;
I Streak camera, stabilized beam;I Applied quadrupole excitation
through the LFB;I Bunch lengthening leads to
Touschek lifetime improvement;I With both feedback and
modulation injected 160 mA andramped to 2.5 GeV.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
The Final ProblemI After adjusting cavity temperature
we were able to keep the beamstable in X, Y, and Z;
I New problem — with all planesstable, injection saturated around110 mA due to poor Touscheklifetime;
I Streak camera, stabilized beam;I Applied quadrupole excitation
through the LFB;I Bunch lengthening leads to
Touschek lifetime improvement;I With both feedback and
modulation injected 160 mA andramped to 2.5 GeV.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
The Final ProblemI After adjusting cavity temperature
we were able to keep the beamstable in X, Y, and Z;
I New problem — with all planesstable, injection saturated around110 mA due to poor Touscheklifetime;
I Streak camera, stabilized beam;I Applied quadrupole excitation
through the LFB;I Bunch lengthening leads to
Touschek lifetime improvement;I With both feedback and
modulation injected 160 mA andramped to 2.5 GeV.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
The Final Problem
00:28 00:57 01:26 01:55 02:24 00:28 00:57 01:26 01:550
50
100
150
Time
Be
am
cu
rre
nt
(mA
)
ANKA, 2016−03−06
00:28 00:57 01:26 01:55 02:24 00:28 00:57 01:26 01:550
0.5
1
1.5
2
2.5
Time
En
erg
y (
Me
V)
I After adjusting cavity temperaturewe were able to keep the beamstable in X, Y, and Z;
I New problem — with all planesstable, injection saturated around110 mA due to poor Touscheklifetime;
I Streak camera, stabilized beam;I Applied quadrupole excitation
through the LFB;I Bunch lengthening leads to
Touschek lifetime improvement;I With both feedback and
modulation injected 160 mA andramped to 2.5 GeV.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
ANKA SummaryI Historically operated with only transverse feedback, relied on
moderate longitudinal instabilities at injection energy to providesufficient lifetime to accumulate beam;
I Changes during a shutdown prevented injection above100–120 mA
I Investigation showed:I Longitudinal HOM in RF cavity crossed synchrotron sideband
around 110 mA;I Vertical beam loss mechanism — large longitudinal oscillation
moved vertical bunch-by-bunch feedback into positive range;I Longitudinal growth rates unfeasible to control with feedback;I Adjusted cavity temperature to move sideband crossing to lower
current, could run with full feedback control in all three planes;I With fully stabilized beam, lifetime dropped, injection saturated at
110–115 mA;I Used swept quadrupole excitation to control bunch length and
lifetime, injected to 160 mA, successfully ramped.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Outline
Introduction
Similarities Under Feedback Control
Feedback StoriesANKA Detective StoryNSLS-II Residual Motion Story
Future ChallengesMain IssuesTransverse Feedback and NoiseLongitudinal Instabilities and Harmonic Cavities
Extra Slides
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
80 85 90 95 100 105 110 11510
−1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel, loop gain 1
80 85 90 95 100 105 110 115
10−0.5
10−0.3
10−0.1
100.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel, loop gain 1
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
80 85 90 95 100 105 110 11510
−1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel, loop gain 1
80 85 90 95 100 105 110 115
10−0.5
10−0.3
10−0.1
100.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel, loop gain 1
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
80 85 90 95 100 105 110 11510
−1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel, loop gain 1
80 85 90 95 100 105 110 115
10−0.5
10−0.3
10−0.1
100.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel, loop gain 1
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
80 85 90 95 100 105 110 11510
−1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel, loop gain 2
80 85 90 95 100 105 110 115
10−0.5
10−0.3
10−0.1
100.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel, loop gain 2
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
80 85 90 95 100 105 110 11510
−1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel, loop gain 4
80 85 90 95 100 105 110 115
10−0.5
10−0.3
10−0.1
100.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel, loop gain 4
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
80 85 90 95 100 105 110 11510
−1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel, loop gain 8
80 85 90 95 100 105 110 115
10−0.5
10−0.3
10−0.1
100.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel, loop gain 8
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
80 85 90 95 100 105 110 11510
−1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel, loop gain 16
80 85 90 95 100 105 110 115
10−0.5
10−0.3
10−0.1
100.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel, loop gain 16
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Averaged Bunch Spectra vs. Feedback Gain 4
20 40 60 80 100 120 140 160 180
10−0.7
10−0.5
10−0.3
10−0.1
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Vertical feedback channel
20 40 60 80 100 120 140 160 180
10−0.4
10−0.3
10−0.2
10−0.1
100
Frequency (kHz)
Ma
gn
itu
de
(co
un
ts)
Monitor channel
Gain 1
Gain 16
I Two independent channelsmonitoring vertical motion, one inthe feedback loop, one out of theloop;
I Roughly similar sensitivities,250 mA in 1000 bunches;
I At low feedback gain a visibleresidual motion line due to ionexcitation;
I Double the feedback gain;I Again;I Again;I Once more;I A wider bandwidth comparison.
4Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Beam Size vs. Feedback Gain 5
0 2 4 6 8 10 12 14 1614
15
16
17
18
19
20
21
22
23
Feedback gain (arb. units)
Beam
siz
e (
µm
)
Vertical beam size measured by pinhole camera
I Vertical beam size measuredby a pinhole camera;
I A superposition of true beamsize and residual dipolemotion;
I Vertical emittance, calculatedfrom pinhole camera data;
I Beam lifetime is correlated withbeam size measurements,suggesting vertical sizeblow-up;
I Could get a better estimate oftrue beam size by subtractingknown dipole motion term.
5Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Beam Size vs. Feedback Gain 5
0 2 4 6 8 10 12 14 1614
15
16
17
18
19
20
21
22
23
Feedback gain (arb. units)
Beam
siz
e (
µm
)
Vertical beam size measured by pinhole camera
I Vertical beam size measuredby a pinhole camera;
I A superposition of true beamsize and residual dipolemotion;
I Vertical emittance, calculatedfrom pinhole camera data;
I Beam lifetime is correlated withbeam size measurements,suggesting vertical sizeblow-up;
I Could get a better estimate oftrue beam size by subtractingknown dipole motion term.
5Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Beam Size vs. Feedback Gain 5
0 2 4 6 8 10 12 14 168
10
12
14
16
18
20
Feedback gain (arb. units)
Em
itta
nce
(p
m.r
ad
)
NSLS−II vertical emittance
I Vertical beam size measuredby a pinhole camera;
I A superposition of true beamsize and residual dipolemotion;
I Vertical emittance, calculatedfrom pinhole camera data;
I Beam lifetime is correlated withbeam size measurements,suggesting vertical sizeblow-up;
I Could get a better estimate oftrue beam size by subtractingknown dipole motion term.
5Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Beam Size vs. Feedback Gain 5
0 2 4 6 8 10 12 14 168
10
12
14
16
18
20
Feedback gain (arb. units)
Em
itta
nce
(p
m.r
ad
)
NSLS−II vertical emittance
I Vertical beam size measuredby a pinhole camera;
I A superposition of true beamsize and residual dipolemotion;
I Vertical emittance, calculatedfrom pinhole camera data;
I Beam lifetime is correlated withbeam size measurements,suggesting vertical sizeblow-up;
I Could get a better estimate oftrue beam size by subtractingknown dipole motion term.
5Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Beam Size vs. Feedback Gain 5
0 2 4 6 8 10 12 14 168
10
12
14
16
18
20
Feedback gain (arb. units)
Em
itta
nce
(p
m.r
ad
)
NSLS−II vertical emittance
I Vertical beam size measuredby a pinhole camera;
I A superposition of true beamsize and residual dipolemotion;
I Vertical emittance, calculatedfrom pinhole camera data;
I Beam lifetime is correlated withbeam size measurements,suggesting vertical sizeblow-up;
I Could get a better estimate oftrue beam size by subtractingknown dipole motion term.
5Measurements courtesy of Weixing Cheng of NSLS-II.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Outline
Introduction
Similarities Under Feedback Control
Feedback StoriesANKA Detective StoryNSLS-II Residual Motion Story
Future ChallengesMain IssuesTransverse Feedback and NoiseLongitudinal Instabilities and Harmonic Cavities
Extra Slides
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
New Machines and Challenges
I Transversely, new light sources and colliders are going becomemore and more sensitive to residual dipole motion;
I Rule of thumb: residual dipole motion should be kept below 10% ofthe transverse beam size;
I Usually more critical in the vertical plane;I Low-noise techniques in RF front end and digitizer design are
required;I Most system designers do not care about spurs 100 dB below ADC
full scale;I Since bunch-by-bunch feedback settles to the front end/digitizer
noise floor, any spur can potentially ruin performance.I Longitudinally, harmonic cavities used for lifetime improvement
create major difficulties for bunch-by-bunch feedback:I HCs result in low synchrotron tune with large tune spread;I Conventional topology can handle tune spread of 2:1 at most.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
New Machines and Challenges
I Transversely, new light sources and colliders are going becomemore and more sensitive to residual dipole motion;
I Rule of thumb: residual dipole motion should be kept below 10% ofthe transverse beam size;
I Usually more critical in the vertical plane;I Low-noise techniques in RF front end and digitizer design are
required;I Most system designers do not care about spurs 100 dB below ADC
full scale;I Since bunch-by-bunch feedback settles to the front end/digitizer
noise floor, any spur can potentially ruin performance.I Longitudinally, harmonic cavities used for lifetime improvement
create major difficulties for bunch-by-bunch feedback:I HCs result in low synchrotron tune with large tune spread;I Conventional topology can handle tune spread of 2:1 at most.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Outline
Introduction
Similarities Under Feedback Control
Feedback StoriesANKA Detective StoryNSLS-II Residual Motion Story
Future ChallengesMain IssuesTransverse Feedback and NoiseLongitudinal Instabilities and Harmonic Cavities
Extra Slides
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Bunch-by-bunch Feedback
Controller
BeamKicker structure
Back−endFront−end
SensorBPM Actuator
I Sensor (pickup);I Analog front-end;I Controller;I Analog back-end;I Actuator (kicker).
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Bunch-by-bunch Feedback
Controller
BeamKicker structure
Back−endFront−end
SensorBPM Actuator
I Sensor (pickup);I Analog front-end;I Controller;I Analog back-end;I Actuator (kicker).
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
BPM Hybrid Network
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
B + C − A−D
B +D − A− C
A+B + C +D
A+B −D − C ∆Y
∆X
Σ
Q
A+ C
B +D
B −D
C −AA
C
B
D
I First stage of BPM signal processing — separating X/Y/Z signalsI Since we are digitizing in the end, why not digitize raw signals?I For X and Y we are dealing with small differences of large signals;I If we can reject the common-mode at 20–30 dB level, that is also
the gain of low-noise amplifier we can use to improve sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
BPM Hybrid Network
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
B + C − A−D
B +D − A− C
A+B + C +D
A+B −D − C ∆Y
∆X
Σ
Q
A+ C
B +D
B −D
C −AA
C
B
D
I First stage of BPM signal processing — separating X/Y/Z signalsI Since we are digitizing in the end, why not digitize raw signals?I For X and Y we are dealing with small differences of large signals;I If we can reject the common-mode at 20–30 dB level, that is also
the gain of low-noise amplifier we can use to improve sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
BPM Hybrid Network
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
B + C − A−D
B +D − A− C
A+B + C +D
A+B −D − C ∆Y
∆X
Σ
Q
A+ C
B +D
B −D
C −AA
C
B
D
I First stage of BPM signal processing — separating X/Y/Z signalsI Since we are digitizing in the end, why not digitize raw signals?I For X and Y we are dealing with small differences of large signals;I If we can reject the common-mode at 20–30 dB level, that is also
the gain of low-noise amplifier we can use to improve sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
BPM Hybrid Network
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
A (180◦)
B (0◦) C (Σ)
D (∆)
B + C − A−D
B +D − A− C
A+B + C +D
A+B −D − C ∆Y
∆X
Σ
Q
A+ C
B +D
B −D
C −AA
C
B
D
I First stage of BPM signal processing — separating X/Y/Z signalsI Since we are digitizing in the end, why not digitize raw signals?I For X and Y we are dealing with small differences of large signals;I If we can reject the common-mode at 20–30 dB level, that is also
the gain of low-noise amplifier we can use to improve sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Analog Front-end Design
D
C
B
A
BPM hybrid
Fro
mBPM
s×
Variableattenuator
Phase shifter
Frequency multiplierfrf
Bandpass filter
M × frf
Lowpass filter
To the ADC
MixerLNA
I Front-end requirements:I Low amplitude and phase noise;I Wideband to ensure high isolation between neighboring bunches.
I Input bandpass filter is an analog FIR filter that replicates BPMpulse with spacing, matched to detection frequency period;
I Detection frequency choice:I High frequencies for sensitivity;I Must stay below the propagation cut-off frequency of the vacuum
chamber.I Local oscillator adjusted for amplitude (transverse) or phase
(longitudinal) detection.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Analog Front-end Design
D
C
B
A
BPM hybrid
Fro
mBPM
s×
Variableattenuator
Phase shifter
Frequency multiplierfrf
Bandpass filter
M × frf
Lowpass filter
To the ADC
MixerLNA
I Front-end requirements:I Low amplitude and phase noise;I Wideband to ensure high isolation between neighboring bunches.
I Input bandpass filter is an analog FIR filter that replicates BPMpulse with spacing, matched to detection frequency period;
I Detection frequency choice:I High frequencies for sensitivity;I Must stay below the propagation cut-off frequency of the vacuum
chamber.I Local oscillator adjusted for amplitude (transverse) or phase
(longitudinal) detection.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Analog Front-end Design
D
C
B
A
BPM hybrid
Fro
mBPM
s×
Variableattenuator
Phase shifter
Frequency multiplierfrf
Bandpass filter
M × frf
Lowpass filter
To the ADC
MixerLNA
I Front-end requirements:I Low amplitude and phase noise;I Wideband to ensure high isolation between neighboring bunches.
I Input bandpass filter is an analog FIR filter that replicates BPMpulse with spacing, matched to detection frequency period;
I Detection frequency choice:I High frequencies for sensitivity;I Must stay below the propagation cut-off frequency of the vacuum
chamber.I Local oscillator adjusted for amplitude (transverse) or phase
(longitudinal) detection.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Examples of Front-End Sensitivities Achieved
Vertical Plane
Machine Atten. Calibration At nominal currentSPEAR3 0 dB 0.54 counts/mA/µm 0.96 counts/µmMAX IV 3 GeV 0 dB 0.98 counts/mA/µm 2.8 counts/µmASLS 2 dB 1.24 counts/mA/µm 0.83 counts/µmNSLS-II6 0 dB 1.5 counts/mA/µm 0.75 counts/µm
I LSB of the 12-bit ADC in Dimtel iGp12 is only 5 times larger thanthermal noise in the ADC bandwidth (wide for good isolation downto 2 ns bunch spacing);
I Not a lot of room for improved sensitivity, need to be smart withpickup selection, feedback algorithms.
6Older front-end design with lower sensitivity
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity and Noise
∑ Error
Transverse position (y)
Disturbances
Feedback Beam
Detection noise (vn)
I Complementary sensitivity function T (ω) = L(ω)/(1 + L(ω)) is thetransfer function between noise vn and beam motion y ;
I Assuming flat spectral density for vn can calculate amplification orattenuation of sensing noise;
I Qualitatively, faster damping corresponds to wider bandwidth→higher noise sensitivity;
I Rule of thumb: closed loop damping rate should be of the samemagnitude as open-loop growth rate.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity Functions Compared
0 500 1000 1500−80
−70
−60
−50
−40
−30
−20
−10
0
10
20
Frequency (Hz)
Ma
gn
itu
de
(d
B)
Complementary sensitivity function, noise gain −18.7 dB
I Growth and damping times inturns;
I τol = τcl = 300: −18.7 dBI τol = τcl = 30: −8.1 dBI τol = 30, τcl = 3.2: −6.0 dBI τol = 5.4, τcl = 5.4: 3.8 dBI Fast growth rates result in
higher noise sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity Functions Compared
0 500 1000 1500−80
−70
−60
−50
−40
−30
−20
−10
0
10
20
Frequency (Hz)
Ma
gn
itu
de
(d
B)
Complementary sensitivity function, noise gain −8.1 dB
I Growth and damping times inturns;
I τol = τcl = 300: −18.7 dBI τol = τcl = 30: −8.1 dBI τol = 30, τcl = 3.2: −6.0 dBI τol = 5.4, τcl = 5.4: 3.8 dBI Fast growth rates result in
higher noise sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity Functions Compared
0 500 1000 1500−80
−70
−60
−50
−40
−30
−20
−10
0
10
20
Frequency (Hz)
Ma
gn
itu
de
(d
B)
Complementary sensitivity function, noise gain −6.0 dB
I Growth and damping times inturns;
I τol = τcl = 300: −18.7 dBI τol = τcl = 30: −8.1 dBI τol = 30, τcl = 3.2: −6.0 dBI τol = 5.4, τcl = 5.4: 3.8 dBI Fast growth rates result in
higher noise sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity Functions Compared
0 500 1000 1500−80
−70
−60
−50
−40
−30
−20
−10
0
10
20
Frequency (Hz)
Ma
gn
itu
de
(d
B)
Complementary sensitivity function, noise gain 3.8 dB
I Growth and damping times inturns;
I τol = τcl = 300: −18.7 dBI τol = τcl = 30: −8.1 dBI τol = 30, τcl = 3.2: −6.0 dBI τol = 5.4, τcl = 5.4: 3.8 dBI Fast growth rates result in
higher noise sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity Functions Compared
0 500 1000 1500−80
−70
−60
−50
−40
−30
−20
−10
0
10
20
Frequency (Hz)
Ma
gn
itu
de
(d
B)
Complementary sensitivity function, noise gain 3.8 dB
I Growth and damping times inturns;
I τol = τcl = 300: −18.7 dBI τol = τcl = 30: −8.1 dBI τol = 30, τcl = 3.2: −6.0 dBI τol = 5.4, τcl = 5.4: 3.8 dBI Fast growth rates result in
higher noise sensitivity.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity vs. Feedback Gain
0 20 40 60 80 100−20
−18
−16
−14
−12
−10
−8
Gain (arb. units)
No
ise
ga
in (
dB
)
0 20 40 60 80 10010
0
101
102
103
104
Gain (arb. units)
Da
mp
ing
tim
e (
turn
s)
I 300 turns growth time, fractionaltune of 0.2, 5-turn feedback filter;
I No excitation, purely flat noisefloor;
I Minimum integrated sensitivity atτol = τcl;
I Highly peaked T (ω) at low gains,very wide at high gains.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity vs. Feedback Gain
2 3 4 5 6 7 8 9 10−19
−18.5
−18
−17.5
−17
−16.5
−16
Gain (arb. units)
No
ise
ga
in (
dB
)
2 3 4 5 6 7 8 9 1010
2
103
104
Gain (arb. units)
Da
mp
ing
tim
e (
turn
s)
I 300 turns growth time, fractionaltune of 0.2, 5-turn feedback filter;
I No excitation, purely flat noisefloor;
I Minimum integrated sensitivity atτol = τcl;
I Highly peaked T (ω) at low gains,very wide at high gains.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Sensitivity vs. Feedback Gain
0.15 0.2 0.25−40
−30
−20
−10
0
10
20
Fractional tune
Ma
gn
itu
de
(d
B)
Complementary sensitivity functions
1422 turns
846 turns
602 turns
467 turns
381 turns
322 turns
279 turns
245 turns
219 turns
78 turns
33 turns
14 turns
5 turns
I 300 turns growth time, fractionaltune of 0.2, 5-turn feedback filter;
I No excitation, purely flat noisefloor;
I Minimum integrated sensitivity atτol = τcl;
I Highly peaked T (ω) at low gains,very wide at high gains.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Feedback Filter Design
0.1 0.15 0.2 0.25 0.3 0.35−70
−60
−50
−40
−30
−20
−10
0
10
Fractional tune
dB
Magnitude response
5 taps
32 taps
64 taps
0.1 0.15 0.2 0.25 0.3 0.35−400
−300
−200
−100
0
100
200
300
400
De
gre
es
Fractional tune
Phase response
5 taps
32 taps
64 taps
I Transverse feedback FIR filters,tune of 0.2, adjusted for thesame closed-loop damping time(τcl = τol = 300 turns);
I Conventional wisdom — shorterfilter can generate fasterdamping, longer filter is quieterdue to narrower bandwidth;
I Let’s see what complementarysensitivity function tells us.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Feedback Filter Design
0.185 0.19 0.195 0.2 0.205 0.21 0.215−20
−15
−10
−5
0
5
Fractional tune
dB
Magnitude response
5 taps
32 taps
64 taps
0.185 0.19 0.195 0.2 0.205 0.21 0.2150
50
100
150
200
250
300
350
De
gre
es
Fractional tune
Phase response
5 taps
32 taps
64 taps
I Transverse feedback FIR filters,tune of 0.2, adjusted for thesame closed-loop damping time(τcl = τol = 300 turns);
I Conventional wisdom — shorterfilter can generate fasterdamping, longer filter is quieterdue to narrower bandwidth;
I Let’s see what complementarysensitivity function tells us.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Feedback Filter Design
0.185 0.19 0.195 0.2 0.205 0.21 0.215−20
−15
−10
−5
0
5
Fractional tune
dB
Magnitude response
5 taps
32 taps
64 taps
0.185 0.19 0.195 0.2 0.205 0.21 0.2150
50
100
150
200
250
300
350
De
gre
es
Fractional tune
Phase response
5 taps
32 taps
64 taps
I Transverse feedback FIR filters,tune of 0.2, adjusted for thesame closed-loop damping time(τcl = τol = 300 turns);
I Conventional wisdom — shorterfilter can generate fasterdamping, longer filter is quieterdue to narrower bandwidth;
I Let’s see what complementarysensitivity function tells us.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Residual Noise vs. Filter Length
0 10 20 30 40 50 60 70
−18.6
−18.5
−18.4
−18.3
−18.2
−18.1
−18
Number of taps
Nois
e g
ain
(dB
)
0 10 20 30 40 50 60 70−302
−301.5
−301
−300.5
−300
−299.5
−299
−298.5
−298
Number of taps
Dam
pin
g tim
e (
turn
s)
I Moderate increase of theintegrated noise gain with filterlength;
I Effect of the group delay;I T (ω) shapes very similar near
the tune — reflect identicalclosed-loop damping pole;
I As an added bonus, shorterfilters are less sensitive to tunevariation.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Residual Noise vs. Filter Length
0 10 20 30 40 50 60 70
−18.6
−18.5
−18.4
−18.3
−18.2
−18.1
−18
Number of taps
Nois
e g
ain
(dB
)
0 10 20 30 40 50 60 70−302
−301.5
−301
−300.5
−300
−299.5
−299
−298.5
−298
Number of taps
Dam
pin
g tim
e (
turn
s)
I Moderate increase of theintegrated noise gain with filterlength;
I Effect of the group delay;I T (ω) shapes very similar near
the tune — reflect identicalclosed-loop damping pole;
I As an added bonus, shorterfilters are less sensitive to tunevariation.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Residual Noise vs. Filter Length
0.195 0.2−15
−10
−5
0
5
10
Fractional tune
Ma
gn
itu
de
(d
B)
Complementary sensitivity functions
5 taps
6 taps
7 taps
8 taps
9 taps
10 taps
12 taps
15 taps
18 taps
21 taps
24 taps
32 taps
64 taps
I Moderate increase of theintegrated noise gain with filterlength;
I Effect of the group delay;I T (ω) shapes very similar near
the tune — reflect identicalclosed-loop damping pole;
I As an added bonus, shorterfilters are less sensitive to tunevariation.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Residual Noise vs. Filter Length
0.17 0.18 0.19 0.2 0.21 0.22 0.23−80
−70
−60
−50
−40
−30
−20
−10
0
10
Fractional tune
Ma
gn
itu
de
(d
B)
Complementary sensitivity functions
5 taps
6 taps
7 taps
8 taps
9 taps
10 taps
12 taps
15 taps
18 taps
21 taps
24 taps
32 taps
64 taps
I Moderate increase of theintegrated noise gain with filterlength;
I Effect of the group delay;I T (ω) shapes very similar near
the tune — reflect identicalclosed-loop damping pole;
I As an added bonus, shorterfilters are less sensitive to tunevariation.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Outline
Introduction
Similarities Under Feedback Control
Feedback StoriesANKA Detective StoryNSLS-II Residual Motion Story
Future ChallengesMain IssuesTransverse Feedback and NoiseLongitudinal Instabilities and Harmonic Cavities
Extra Slides
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Direct Energy Sensing
I Basic method: use horizontal signal from a pickup in a highdispersion location to sense energy oscillation directly;
I Eliminates the need to generate a 90◦ phase shift betweenlongitudinal position measurement and energy kick;
I Feedback filter is just a gain plus two constraints:I High-pass transition below the synchrotron frequency band to filter
out DC orbit offsets;I Low-pass transition above the synchrotron frequency band to
remove horizontal signals.I Can’t handle full lengthening (synchrotron frequency band extends
to DC);I Use two pickups with 0◦ or 180◦ relative phase advance to
eliminate horizontal signals?
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Direct Energy Sensing
I Basic method: use horizontal signal from a pickup in a highdispersion location to sense energy oscillation directly;
I Eliminates the need to generate a 90◦ phase shift betweenlongitudinal position measurement and energy kick;
I Feedback filter is just a gain plus two constraints:I High-pass transition below the synchrotron frequency band to filter
out DC orbit offsets;I Low-pass transition above the synchrotron frequency band to
remove horizontal signals.I Can’t handle full lengthening (synchrotron frequency band extends
to DC);I Use two pickups with 0◦ or 180◦ relative phase advance to
eliminate horizontal signals?
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Direct Energy Sensing
I Basic method: use horizontal signal from a pickup in a highdispersion location to sense energy oscillation directly;
I Eliminates the need to generate a 90◦ phase shift betweenlongitudinal position measurement and energy kick;
I Feedback filter is just a gain plus two constraints:I High-pass transition below the synchrotron frequency band to filter
out DC orbit offsets;I Low-pass transition above the synchrotron frequency band to
remove horizontal signals.I Can’t handle full lengthening (synchrotron frequency band extends
to DC);I Use two pickups with 0◦ or 180◦ relative phase advance to
eliminate horizontal signals?
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
APS-U Example
100
101
102
103
104
105
106
−60
−40
−20
0
20
Frequency (Hz)
Gain
(dB
)
100
101
102
103
104
105
106
−150
−100
−50
0
50
100
Frequency (Hz)
Phase (
degre
es)
I Possible filterresponse for APS-Uenergy sensing;
I fs = 620 Hz withoutharmonic cavities;
I Drops to100± 100 Hz underoptimal bunchlengthening;
I 2 dB gain and 12.5◦
phase ripple in20–800 Hz range.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
APS-U Example
100
101
102
103
104
105
106
−60
−40
−20
0
20
Frequency (Hz)
Gain
(dB
)
100
101
102
103
104
105
106
−150
−100
−50
0
50
100
Frequency (Hz)
Phase (
degre
es)
I Possible filterresponse for APS-Uenergy sensing;
I fs = 620 Hz withoutharmonic cavities;
I Drops to100± 100 Hz underoptimal bunchlengthening;
I 2 dB gain and 12.5◦
phase ripple in20–800 Hz range.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
APS-U Example
100
101
102
103
104
105
106
−60
−40
−20
0
20
Frequency (Hz)
Gain
(dB
)
100
101
102
103
104
105
106
−150
−100
−50
0
50
100
Frequency (Hz)
Phase (
degre
es)
I Possible filterresponse for APS-Uenergy sensing;
I fs = 620 Hz withoutharmonic cavities;
I Drops to100± 100 Hz underoptimal bunchlengthening;
I 2 dB gain and 12.5◦
phase ripple in20–800 Hz range.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
APS-U Example
101
102
103
−2
−1
0
1
2
Frequency (Hz)
Gain
(dB
)
101
102
103
−5
0
5
10
15
Frequency (Hz)
Phase (
degre
es)
I Possible filterresponse for APS-Uenergy sensing;
I fs = 620 Hz withoutharmonic cavities;
I Drops to100± 100 Hz underoptimal bunchlengthening;
I 2 dB gain and 12.5◦
phase ripple in20–800 Hz range.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Summary
I Bunch-by-bunch feedback can make different machines look alike,but it takes good system design and thorough understanding ofdriving terms;
I Never trust a “mild” instability — you never know when it will turninto a showstopper;
I Characterization of instabilities is a definite must for robustfeedback operation;
I New low-emittance machines demand noise-figure optimized RFfront-ends and digitizers;
I Simplified models are very helpful in understanding feedbackalgorithm trade-offs;
I Longitudinal feedback in presence of harmonic cavities ischallenging, new (so far untested) energy sensing technique mighthelp.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Summary
I Bunch-by-bunch feedback can make different machines look alike,but it takes good system design and thorough understanding ofdriving terms;
I Never trust a “mild” instability — you never know when it will turninto a showstopper;
I Characterization of instabilities is a definite must for robustfeedback operation;
I New low-emittance machines demand noise-figure optimized RFfront-ends and digitizers;
I Simplified models are very helpful in understanding feedbackalgorithm trade-offs;
I Longitudinal feedback in presence of harmonic cavities ischallenging, new (so far untested) energy sensing technique mighthelp.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Summary
I Bunch-by-bunch feedback can make different machines look alike,but it takes good system design and thorough understanding ofdriving terms;
I Never trust a “mild” instability — you never know when it will turninto a showstopper;
I Characterization of instabilities is a definite must for robustfeedback operation;
I New low-emittance machines demand noise-figure optimized RFfront-ends and digitizers;
I Simplified models are very helpful in understanding feedbackalgorithm trade-offs;
I Longitudinal feedback in presence of harmonic cavities ischallenging, new (so far untested) energy sensing technique mighthelp.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Summary
I Bunch-by-bunch feedback can make different machines look alike,but it takes good system design and thorough understanding ofdriving terms;
I Never trust a “mild” instability — you never know when it will turninto a showstopper;
I Characterization of instabilities is a definite must for robustfeedback operation;
I New low-emittance machines demand noise-figure optimized RFfront-ends and digitizers;
I Simplified models are very helpful in understanding feedbackalgorithm trade-offs;
I Longitudinal feedback in presence of harmonic cavities ischallenging, new (so far untested) energy sensing technique mighthelp.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Summary
I Bunch-by-bunch feedback can make different machines look alike,but it takes good system design and thorough understanding ofdriving terms;
I Never trust a “mild” instability — you never know when it will turninto a showstopper;
I Characterization of instabilities is a definite must for robustfeedback operation;
I New low-emittance machines demand noise-figure optimized RFfront-ends and digitizers;
I Simplified models are very helpful in understanding feedbackalgorithm trade-offs;
I Longitudinal feedback in presence of harmonic cavities ischallenging, new (so far untested) energy sensing technique mighthelp.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Summary
I Bunch-by-bunch feedback can make different machines look alike,but it takes good system design and thorough understanding ofdriving terms;
I Never trust a “mild” instability — you never know when it will turninto a showstopper;
I Characterization of instabilities is a definite must for robustfeedback operation;
I New low-emittance machines demand noise-figure optimized RFfront-ends and digitizers;
I Simplified models are very helpful in understanding feedbackalgorithm trade-offs;
I Longitudinal feedback in presence of harmonic cavities ischallenging, new (so far untested) energy sensing technique mighthelp.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Acknowledgments
I Thank you for your attention!I I would also like to thank physicists, engineers, and operators at
many machines around the world who directly or indirectlycontributed to measurements presented here.
I Special thanks to Weixing Cheng for fruitful discussions andNSLS-II measurements.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Root Locii in Complex Plane: Close Zoom
I Root locus on the complexplane:
I Starts at the open-loop pole(×), ends at the highest gainsetting (o);
I Real part corresponds togrowth (positive, right halfplane) or damping (negative,left half plane) rate;
I Imaginary part is thefrequency.
I Zoomed in around the dominantpole, all filters look the same.
Experience with feedbacksystems in modern
synchrotron light sources
Introduction
Similarities UnderFeedback Control
Feedback StoriesANKA Detective Story
NSLS-II Residual Motion Story
Future ChallengesMain Issues
Transverse Feedback and Noise
Longitudinal Instabilities and HarmonicCavities
Summary
Extra Slides
Root Locii in Complex Plane: Wider View
I Zooming out we see additionalpoles;
I These are due to the additionaldelay of the feedback controller;
I Added poles account forincreasing noise sensitivity.