SPEAR 3 Beam Stability and Stabilization R. Hettel NSLS-II Stability Workshop April 18-20, 2007.

R. Hettel SPEAR 3 Orbit Stability and Stabilization NSLS-II Stability Workshop April 18-20, 2007

SPEAR 3 Beam Stability and StabilizationR. Hettel

NSLS-II Stability Workshop

April 18-20, 2007


AcknowledgmentsPrimary developers:

T. Straumann: real-time processing, system architecture and communication

A. Terebilo: accelerator physics, system operation development

J. Sebek: turn-turn BPM processing

D. Martin: BPM systems

F. Rafael, G. Leyh: corrector power supply development

Main contributors:S. Allison

J. Corbett

R. Hettel

E. Medvedko

G. Portmann

T. Rabedeau

J. Safranek

C. Wermelskirchen

E. Daly, N. Kurita, J Langton, A. Ringwall, J. Tanabe (SPEAR 3 mech des)

EDM electrical support group


Stability Requirements

experiment parameters

E/E(coher) < 10-4t < 0.1% ttiming, bunch length

E/E(coher) < 5 x 10-5

E/E(rms) < 10-4

(und n = 7)

x < ~5 rad

y < ~1 rad(undulator)

< 10-4 photon energy resolution

E/E(coher) < 10-4

E/E(rms) < 10-4

x,y < 0.1% x,y

x,y < 0.1% x,y

x,y < 5% x,y

x,y < 5% x,y

< 0.1% intensity

steering to small samples

beam energy/

energy spreadbeam sizebeam orbit

experiment parameters

E/E(coher) < 10-4t < 0.1% ttiming, bunch length

E/E(coher) < 5 x 10-5

E/E(rms) < 10-4

(und n = 7)

x < ~5 rad

y < ~1 rad(undulator)

< 10-4 photon energy resolution

E/E(coher) < 10-4

E/E(rms) < 10-4

x,y < 0.1% x,y

x,y < 0.1% x,y

x,y < 5% x,y

x,y < 5% x,y

< 0.1% intensity

steering to small samples

beam energy/

energy spreadbeam sizebeam orbit

• Stability requirements for small beams may be relaxed if beam size at experiment is limited by beam line optics (e.g. mirror slope error, point-spread function, etc.)

• Stability requirements depend on time interval


• Disturbance time scale << experiment integration time: Orbit disturbances blow up effective beam and , reduce intensity at

experiment, but do not add noise

For / = cm/o < ~10%: ycm(rms) < ~0.3 y ycm(rms) < ~0.3 y'

Note: can have frequency aliasing if don't obey Nyquist….

• Disturbance periods experiment integration time:Orbit disturbances add noise to experiment

For / = ~2 cm/o <~10%: ycm(rms) < 0.05 y ycm(rms) < 0.05 y'

• Disturbance periods >> experiment time (day(s) or more):

Realigning experiment apparatus is a possibility

• Sudden beam jumps or spikes can be bad even if rms remains low

Peak amplitudes can be > x5 rms level

• Most demanding stability requirements:

Orbit disturbance frequencies approximately bounded at high end by data sampling rate and a low end by data integration and scan times

noise not filtered out

Stability Time Scales


• Short stiff girders and magnet supports (>20 Hz)

• Chamber constrained vertically and horizontally at BPMs

• Invar supports for key BPMs (~3 m/oC)

• 18”-24” concrete floor

• Tunnel temp stable to ± ~1oC/day

SPEAR 3 Mechanical Design


SPEAR 3 Electrical Design – Power SuppliesDipole:

stability: 50ppm (or better); 3 ppm/OC diurnalripple: 0.2% pk-pk of full output voltage ripple (DC-1 MHz)chopper freq: 20 kHz

Quadrupole and Sextupole:stability: 100 ppm; 6 ppm/OC diurnal ripple: 0.2% pk-pk of full output voltage ripple (DC-1 MHz)

chopper freq: 40 kHz

Correctors:stability: 500ppm; 30ppm/OC diurnalnoise: 17 ENOB, 0.001 Hz – 4 kHz chopper freq: 40-60 kHzDAC resolution/update rate: 24-bit (>18 bit for ± 1 mrad corrector) / 4 kHzbandwidth: ~1 kHz

RF HVPS (90 kV)stability: < 0.1% FSripple: < 1% pk-pk (<0.2% rms) above 60 kV

Power supply stability requirements depend on ring design


SPEAR 3 BPMs and ProcessorsBergoz MX-BPM (modified)• mux'd button processing (16 kHz)

• ADCs sample baseband button signals (before internal analog position calc circuit)

• SPEAR 3 version has:o 5 dB more input attenuation than standard

module for 500 mAo wider IF filter to sample turn-turn orbit (2.2

MHz vs. 0.4 MHz)o ~2 mm res for injected beam (0.03 mA)

Echotek Digital Receivers • parallel I/Q processing of down-converted

button signals (8 chan/module = 2 BPMs)

IF = 16.65 MHz (13 frev)

sample freq = 64.02 MHz (50 frev)

• provision for simultaneous processing of test tone calibration signal

• ~0.3-mm res for injected beam (0.03 mA)

• 2-µm turn-turn resolution at >~10 mA

• nanometer resolution in 100 Hz BW

84 mm44

.2

mm 34

m

m

24 mm13 mm18.8 mm

12-mm diam buttons

BPM processor temperature regulated to < ± 0.4oC pk


SPEAR 3 BPM Processing and Fast Orbit Feedback

Corrector Power Supplies - Bldg. 118 rev. 4/12/07

Central BPM/Orbit Feedback

StationBldg. 117

EPICSnetwork

Injecttrig

6 Hcorrs

6 Vcorrs

spare/misc.

IOC

/ctr

lIO

C/c

trl

IOC

/ctr

l

IOC

(PP

C)

MC

OR

Com

m(1

00 M

b E

’net

no T

CP

/IP)

BP

M/F

dbk

CP

U/IO

C(P

PC

)

T-S

tam

p/S

ync

out

BP

M C

omm

(2ea

100

Mb

E'n

etno

TC

P/IP

)

West Pit (quadrants 1 & 4) East Pit (quadrants 2 & 3)

8 Hcorrs

8 Vcorrs

8 Hcorrs

IOC

/ctr

lIO

C/c

trl

IOC

/ctr

l

8 Hcorrs

8 Vcorrs

8 Hcorrs

IOC

/ctr

lIO

C/c

trl

IOC

/ctr

l

8 Vcorrs

8 Vcorrs

8 Vcorrs

IOC

/ctr

lIO

C/c

trl

IOC

/ctr

l

8 Vcorrs

8 Vcorrs

8 Vcorrs

IOC

/ctr

lIO

C/c

trl

IOC

/ctr

l

frev =1.28 MHz

sync/4 kHz T-stamp

BPM74

BPM 46 BPM 73 BPM92

BPM27

BPM 1 BPM 28 BPM45

4 kHz sync

LO LO

IF C

lk

4 kH

z sy

nc

LO LO

IF C

lk

4 kHz / T-stamp

MCOR comm

EPICS network

fRF = 476.300 MHz

frevout

RF/ClockSig Gen

LOout

IF Clkout

syncsync

IOC

/A

DC

s

error

sumphotonBPMs

29-BPMMUX'd Button Processing

baseband buttondetect ADCs

18-BPMParallel Button

Processing(in progress)

digital IF detect(1st turn, turn-turn,

closed orbit)

T-Stmp/Sync

IOC(PPC)

ctrl

Ene

t

BP

M E

net

sync

out

frev

syn

c/T

-stm

p

27-BPMMUX'd Button Processing

baseband buttondetect ADCs

18-BPMParallel Button

Processing(in progress)

digital IF detect(1st turn, turn-turn,

closed orbit)

T-Stmp/Sync

IOC(PPC)

ctrl

Ene

t

BP

M E

net

sync

out

frev

syn

c/T

-stm

p

Note:

SPEAR 3 was commissioned and operated until recently using slow orbit feedback running on MATLAB

(Corbett, Portmann)


SPEAR 3 Fast Orbit Feedback

Rx TUSVR

x T1)( UVS

Static orbit correction Dynamic orbit correction

nxK̂n PI T1)( USV

naKmnaKnaK̂ P0m

IPI

· 4 kHz update · latency (pipeline delay + deadtime) = 0.7-1 ms· RTEMS realtime OS · EPICS control and monitoring

x

xref

S-1UTx KPI V

2 ea remote IOCs (+ phBPM IOC)

Central CPU(1 GHz powerPC+altivec)

18 ea remote IOCs (8 correctors/IOC)

100 Mb/s E’net broadcastno TCP/IP

2 ea 100 Mb/s E’netno TCP/IP

+

-


Distributed weak perturbations:

• Uncorrelated small (~1μm) vibrations of individual magnets and supports cause orbit motion is concentrated in the modes with large singular values and frequency range 1-200Hz.

SPEAR 3 Orbit Motion

Localized strong perturbations:• Gap or phase changes in undulators occur

on a ~1s time scale. Local feed-forward correction was implemented using ID trim coils, adjacent quads (tune) and skew quads (coupling). Cause global orbit distortions of a few m rms without FOFB.

• Vehicle traffic on the overpass bridge causes slow (~1s) motion of the floor and microns of orbit instability.

• RF power supply ripple inducing synchrotron oscillations

08:00 08:30 09:00 09:30 10:000

0.5

1

1.5

2

2.5

3x 10

-3

time

r.m.s

. orb

it 0.

5s a

vera

ge [

m]

Horizontal

Vertical

SOFB FOFB

FOFB correction of ID gap changes and bridge traffic effects. Based on 2 hours of averaged (0.5s)


SPEAR 3 Phase Oscillations and RF HVPS Ripple

• Phase oscillations measured with turn-turn BPM:

3.6 mrad rms = ~1.2 ps rms

bunch length = 17 ps rms

• Working to implement mode-0 feedback

• RF HVPS ripple induces 0-mode longitudinal phase oscillations

• Problem with RF HVPS causes extra oscillation amplitude @ 60 Hz nominal ripple: 0.4% rms of 70 kV

100 ms

4 ms


FOFB effect on distributed weak wide bandwidth perturbations. Based on 1s of 4kHz BPM data

Fast orbit feedback in operation since June, 2006. Integrator loop gains set conservatively for start of operations. Studies ongoing to find optimal tuning.

SPEAR 3 Fast Orbit Feedback – Bandwidth

100

101

102

103

10-4

10-2

100

102

Frequency [Hz]

Inte

gra

ted

R.M

.S.

Po

we

r S

pe

ctr

um

[ m

2 ]

FOFB Off

FOFB OnFOFB On, more gain

100

101

102

103

10-4

10-3

10-2

10-1

100

Frequency [Hz]

R.M

.S.

Po

we

r S

pe

ctr

al

De

ns

ity

[ m

2 /Hz

]

FOFB Off

FOFB OnFOFB On, more gain

-10

-8

-6

-4

-2

0

2From: u1 To: y1

Mag

nitu

de (

dB)

102

103

-135

-90

-45

0

Pha

se (

deg)

Bode Diagram

Frequency (rad/sec)

Data

3zero-3pole model

Limiting factors:

• Corrector field penetration in vac chamber (copper with CuNi inlays for bandwidth to ~200 Hz)

• Time delay - 3 clock cycles


0 20 40 60 80 100 12010

-5

10-4

10-3

10-2

10-1

100

101

102

Singular Values

SPEAR 3 Fast Orbit Feedback - Eigenmodes

1. Uncorrected orbit error from ‘real’ sources

2. ‘Spilling’ from other modes accumulating in corrector magnets

050

100150

200

0

20

40

60-4

-2

0

2

4

x 10-3

Time [min]

Orbit Error in BPM space

BPM #

y [

mm

]

050

100150

200

0

20

40

60-2

0

2

4

6

8

10

x 10-3

Time [min]

Orbit Error in eigenmode space

Eigenmode #

y U

[m

m]

Eigenmode spectrum

Ignoring even a single eigenmode results in gradual buildup of error:


0 50 100 150 200 250 300 350 400 450 5000

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

PS

D [

m2 ]

/Hz

BL11 pBPM PSD

No FOFB

Currrent Ki only

Reduced Gain

0 50 100 150 200 250 300 350 400 450 5000

5

10

15

20

25

30

35

40

Frequency [Hz]

IPS

[m

2 ]

BL11 pBPM Integrated PS

No FOFB

Currrent Ki only

Reduced Gain

3. When modal Ki and Kp gains are tuned to reduce motion seen by in-loop electron BPMs, out-of-loop photon BPMs suffer

10 20 30 40 5010

-1

100

101

102

Vertical eigenmode #

Inte

gra

tor

frquency [

Hz]

Ki values for different eigenmodes

integrator bandwidths for different eigenmodes

SPEAR 3 Fast Orbit Feedback – Eigenmodes – cont.

4. Feedback gain/BW is reduced for higher eigenmodes to reduce orbit noise but to still allow modal “mop-up”


• Vertical motion at photon BPMs (~15-20 m from source) not included in feedback can be 10s of microns even though stability shown by electron BPMs is <1 m

• “Beam Line Dynamic Steering” (BLDS) has been introduced:

• Response of photon BPMs to a local angle bump in 2 electron BPMs is measured offline

• Photon BPM data averaged for 1 min for each beamline

• Once a minute apply calculated correction to the electron BPM FOFB target.

• BLDS is not perfect: 1 degree of freedom does not exactly correct source motion; combination of position and angle could be tuned to maximize performance

• FOFB architecture allows to bring in pBPM data at 4 kHz rate and response matrix can be extended to include pBPMs

• Practical issue for including pBPMs in response matrix: need to reconfigure matrix (add/remove rows) on the fly when beam lines open and close

-0.45 -0.4 -0.35 -0.3 -0.25 -0.2-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

eBPM [mm]

pB

PM

[m

m]

pBPM/eBPM measured off-line

Measured

linear

SPEAR 3 Photon Monitor Feedback


Beam Line Mirror FeedbackT. Rabedeau, SSRL

• error signal obtained from position sensitive detector near beam focus

• error signal used to control piezo high voltage

• piezo provides mirror fine pitch control with typical full range of motion +/- 30 rad or +/- 0.6mm or more focus motion.

focus 1.4 m rms

source 17.3 m rms


SPEAR 3 Fast Orbit Feedback – Operator Interface


2oC

RF frequency (green) changes by 1 kHz (C/C = ~0.5 mm/234 m) for a 2oC tunnel temperature variation (red) over 1 month period

RF Frequency Feedback

RF frequency (green) changes ~30 Hz twice daily from lunar tide

(9oC pk-pk outside diurnal temperature over 4 days shown in violet)


SPEAR Floor Motion

-0.00040

-0.00030

-0.00020

-0.00010

0.00000

0.00010

0.00020

0.00030

0.00040

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

June 2003 - July 2003 Nov 2003 - July 2003 Sep 2004 - July 2003

400 µm

-0.00040

-0.00030

-0.00020

-0.00010

0.00000

0.00010

0.00020

0.00030

0.00040

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

June 2003 - July 2003 Nov 2003 - July 2003 Sep 2004 - July 2003

400 µm

Floor monument changes in first year of operation

HLS (Georg Gassner)

• Data correlation analysis over 1 year suggests external temperature is the main factor for short term floor movement, not the internal temperature of the tunnel.

• More HLS sensors to be added


SPEAR 3 Orbit Stability and Feedback – Future Development

• Plan to characterize diurnal instability of floor, ring and beam line components using high resolution sensors (HLS, etc). This information might be included in feedback/feedforward

• Studying to potential improvement gained by adding a roof over SPEAR (and possibly subsequent air conditioning)

• Beam line dynamic steering to be integrated into FOFB

• Better photon monitors are being developed

• More parallel BPM processors will be added (will pay attention to new SLS/DESY design)

• Plan to continue developing feedback to incorporate ring and beam line sensors and actuators

SPEAR 3 Beam Stability and Stabilization R. Hettel NSLS-II Stability Workshop April 18-20, 2007.

Documents

orbit stability

stability workshop

beam stability

hettel spear

hettel nsls

stabilization nsls

tanabe spear

bpm processing