measurement and control of charged particle beams in the relativistic heavy ion collider

Measurement and Control of Charged Particle Beams in the

Relativistic Heavy Ion Collider

ESS/AD seminar - April 16th , 2014

Michiko MintyInstrumentation Systems Group Leader

Collider-Accelerator DepartmentBrookhaven National Laboratory

OUTLINE

The Relativistic Heavy Ion Collider (RHIC)

Maximizing the scientific output of RHIC

Accelerator physics challenges

Feedback-based beam control orbits tunes and coupling

Impact on RHIC performance

Summary

High Energy Colliders in the US

Stanford Linear Collider, e+ e- (1989 – 1998)

2 miles

TeVatron, p+p (1987 – 2011)

1.2 miles

RHIC (>2001)

0.75 miles

RHIC: versatile collider in terms of species (p, d, Cu, Au, U,…) and beam energies (maximum of 100 GeV/n for ions, 250 GeV for protons); the only high energy polarized proton collider

RHIC

LINAC

Booster

AGS

Tandems

STAR

PHENIX

EBIS

Relativistic Heavy Ion Collider (RHIC)

RHIC beams: 110 bunches, each bunch contains ~1E9 ions or 1E11 protonsRHIC bunches are guided and focused using ~ 1750 superconducting magnetsRHIC bunches are very small (~100 m at interaction points)RHIC bunches circulate ~ 80,000 times per second

INJECTIONACCELERATIONCOLLISIONS

RHIC consists of 2 separate superconducting accelerators, 2.4 miles (3.8 km) long






Summary


N1 N2

RHIC performance (ions or protons) is characterized by the rate at which particles collide, the

Ncol

x y

Luminosity ~

f

N

xy

NNcol

is the number of colliding bunches

fis the collisionfrequency


RHIC performance with protons is also characterized by the beam’s polarization

Uhlenbeck and Goudsmit (1926): protons possess a spin angular momentum

the spin of a proton responds like a magnetic dipole; it precesses in magnetic fields

at RHIC we preserve the average orientation of all the proton’s spins, the polarization

spin






Summary

Challenges - orbits

offsets between the bunchesdegrade luminosity

beam’s orbits (positions and angles) must be controlled

bunches should collide head-on to maximize collision probability

x

40% loss ifx = 100 m(y = 0)

zoom

L / L0 = e-(x/4x)2

e-(y/4y)2

correction factorfor position errors

Challenges

The stability of beams in a circular accelerator depends on the so-called “tune” of the accelerator

the tune, Qequals the number of oscillations made by a bunch in one revolution around the accelerator

in this sketch, the vertical (y) tune is Qy = 13.5 and Q = 0.5

1 23

45

67

8910

1112

13

0

13.5

oscillations about the ideal trajectory

we monitor and controlthe fractional tune

Q

ChallengesResonances! These characterize the tendency of a system to oscillate at a greater amplitude at certain frequencies of excitation

improperly timed pushes …

properly timed periodic forces …

if the forces are too large …

… will not rock the chair

… will rock the chair

In accelerators, resonances must be avoided

……………………

Challenges

first order

second order

order 0 - driven by dipole magnets

Qx

Qy

resonance diagram

Qx

Qy

resonance diagramorder 0 - driven by dipole magnets

Qx

Qy

resonance diagram

Qx

Qy

resonance diagram

Qy

order 0 - driven by dipole magnets

resonance diagramresonance diagram

Qx

third order

the (fractional) tunes should be irrational

In an accelerator, resonances can occur if perturbations act on a bunch in synchronism with its oscillatory motion.

resonance condition:

m Qx + n Qy = p (m, n, and p are integers)

The errors arise from imperfections (or misalignments) of the magnets

Challenges

bounded by strong 3rd and 4th order resonances

beam 2

beam 1

zoom

“working point” in RHIC (protons, at full energy)

for polarized proton operation, the resonance at 0.70 is critical during acceleration

the operating point is therefore moved during acceleration






Summary

17:40 22:55

WHY AT RHIC

correct

(unavoidable) persistent currents and hysteresis effects

thermal effects

magnetic field errors - including power supply variations, bit limitations, response time and magnet alignment errors

yrms during acceleration, run-9timeof day

23:50 10:20 14:30

time (s)

startacceleration

end






Summary

v

Precision beam position measurements

vacuumchamber

“stripline” beam position monitor (BPM)

23 cm length

precision of average orbit measurements improved by > factor 10

added digital equivalent of a single-pole, low pass filter (IIR filter) to effectively average out predominantly ~ 10 Hz variations in the closed orbit

yold ynew

XoldXnew

precision of measurement now ~ 5 m

4 km full scale

zoom

400 m full scale

+/- 60 microns full scale

at RHIC we use 600 BPMs (150 /plane) to measure the orbits along accelerator

(smaller than the diameter of 1 red blood cell)

before Run-10 acquisition rate: nominally 0.5 Hz nondeterministic

After Run-10 acquisition rate: 1 Hz deterministic

BPM data delivery

measurement based on existing beam position monitors using new and improved algorithm for measuring average orbit using original survey (e.g. offset) data deterministic data delivery

feedback design orbit correction algorithm (“singular valued decomposition”) extended to application at 1 Hz rate during energy ramp reference orbits specified in terms of BPM data (not corrector strengths)

Orbit Feedback

x x = M

x = vector of ~ 320 BPM measurementsM = matrix of transfer functions

= vector of angular deflection of ~ 230 correctors

Mij

= M-1 xMij is the transfer matrix between the

ith BPM and jth corrector dipole

orbits well controlled, reproducible, and well below the 200 m tolerance

acce

lera

tion

sta

rt

250

GeV

collisio

ns

~400 m

~20 m

no feedback

with feedback

proof-of-principle for orbit feedback using existing infrastructure (2010) energy feedback principle improved (2011) constrain average horizontal corrector strengths use all arc BPMs for energy offset determination implementation of orbit and energy feedback on all ramps (2011)






Summary

Precision tune measurementsapply (using a “kicker”) a broadband excitation near the beam’s natural frequency

Q = fres / frev

“kicker” BPM

signal processin

g

frequency

generator

fres

the beam responds at it’s natural resonant frequency, fres

the fractional tune, Q, isfrev = (known) revolution frequency

measurement precision: 2E-5

Precision coupling measurementsresonance-free region has Qx ~ Qy … precisely where coupling effects are strongest

with Qx ~ Qy and nonzero coupling (C = 0) beam control in one plane affects the other and produces unexpected results

C

/

> factor 10 improvement in measurement resolution

run-08

run-11

measurement

based on direct-diode detection (BBQ = base-band tune) for precision measurements - M. Gasior , R. Jones (2005)

feedback design

uses methodology of coupling angle measurement – Y. Luo (2004) distinguishes between eigenmodes - R. Jones, P. Cameron, Y. Luo (2005)

history

demonstrated at RHIC in 2006 - P. Cameron et al (2006) successfully applied for all ramp developments in 2009 used regularly by operations for ramp development in 2010 used together with orbit and energy feedback for all ramps in 2011

Tune and Coupling Feedback

before:

8 periods………………………………... (repeat) ……………………………..…….………

1 period usedfor BBQ/BTF

1 period usedfor BBTF

1 (possibly corrupted) period used BBQ/BTF

1 in 16 periods of data (AFE output, I/Q demodulator input) used for BBQ/BTFintermittent corruption of this data due to CPU-limits and data overwrites with BBTF (ADOs removed)

after:

average of 8 periods usedfor BBQ/BTF

…………………………………….………………..... (repeat) ……………………………..…………………….………

8 periods

tunes and coupling well controlled, reproducibility is excellent

Cmultiple super-imposed ramps






Summary

Impact on RHIC performance(1) Accelerator availability

~ $ 100k eliminated need for dedicated re-optimization efforts

~ $ 100k per operational mode change - particle species, energies or optics (3-4 per fiscal year)

~ $ 100k savings for initial beam setup

at least 1 extra week for physics operation with electrical costs at 25 MW at $60/MW-hr

Time required to successfully accelerate beams to full energy reduced from > 3 days to 2 hours

(2) Operation under extreme conditions: near-resonance

acceleration

With routine orbit, energy, tune, and coupling feedback on every acceleration of protons to high energies, the vertical tune could be lowered towards dangerous 2/3 orbital resonance (and away from spin resonance at 7/10).

Qy= 0.0062/3 resonance

end ofacceleration

25 % increase in relative polarization of each beam

equivalent to 14 additional weeks of RHIC operations ($3.5 M) for same level of statistical uncertainty for physics program

sinc

e ru

n-9

(3) acceleration/decelerationA dedicated study was performed to confirm the degree of residual polarization loss during acceleration

executed with complete suite of feedbacks demonstrating fully automated beam control, enabled an otherwise impossibleexperiment

SummaryThe resolution of all measurements (beam position, energy deviation, tune, coupling, and chromaticity) has been improved by more than a factor of 10 and is nearing the limitations of the instrumentation

Control of the parameters affecting beam properties during acceleration in RHIC has transitioned from being pre program-med to based on measurements of the beam’s properties

Feedback-based beam control is now the norm: all beams in RHIC are now established using orbit, tune, coupling and energy feedback. Precision control of these parameters has expanded the parameter space accessible during acceleration. This allows for more extreme operating conditions and is now essential for polarized proton operation.

AcknowledgementsBPM support P. Cerniglia, A. Marusic, K. Mernick,

, T. Satogata, P. ThiebergerOrbit feedback T. D’Ottavio,

Tune/coupling feedback A. DellaPenna, M. Gasior (CERN), L. Hoff, R. Jones (CERN), , C. Schultheiss, C.Y. Tan (FNAL), S. TepikianA. Curcio, C. Dawson, C. Degen, Y. Luo, G. Marr, B. Martin, P. Oddo, T. Russo, V. Schoefer,

Chromaticity feedback S. Tepikian

10 Hz feedbackP. Cerniglia, A. Curcio, L. DeSanto, C. Folz, C. Ho, L. Hoff, , C. Liu, Y. Luo, W.W. MacKay, G. Mahler, W. Meng, , C. Montag, R.H. Olsen, P. Popken, V. Ptitsyn, G. Robert-Demolaize, and many others

Energy feedback A. Marusic,

Operations G. Marr, V. Schoefer

Management W. Fischer, T. Roser

R. Hulsart, R. Michnoff

A. Marusic, V. Ptitsyn, G. Robert-Demolaize

P. Cameron, Y. Luo, A. Marusic

A. Marusic, K. Mernick, M. Wilinski

A. Marusic,

K. Smith

, R. Smith, J. Ziegler

Run coordinatorsM. Bai, K. Brown, H. Huang, G. Marr, C. Montag, V. Schoefer

R. HulsartK. Mernick, R. Michnoff

P. Thieberger

Accelerator Reproducibilityparameter no feedback

FeedbackWHY to automate well-defined processes to

reduce sensitivities to external influences

HOW compare measurement with desired valueapply correction

cruise control

desired speed

measured speed

difference apply required

change in gas

cruise

GOAL maintain steady conditions

tune/coupling feedback at RHIC

kicker BPM

signal processin

g

frequency

generator

QF

Qx

conversion to currents

QD

model

phase lockloop Qy

QxQy

desired desired

Qx

10 Hz feedback at RHICreduce orbit changes due to triplet magnet vibrations

x (m

)

collision point

triplet triplet

time time

< run-9 feedback on relative beam positionshistory

10 Hz feedback at RHIC

run-10 new 10 Hz feedback, proof-of-principle with new correctors high speed daughter cards for BPMs dedicated networking digital signal processingrun-11 routine application

higher integrated luminosity

essential for (possible future) RHIC operation with near-integer tunes

Chromaticity Feedback

The chromaticities, x and y represent the coupling of transverse and longitudinal motion and may be defined as

x,y = Qx,y/ (p/p)

where Q = the spread in betatron tunes within the bunch of momentum spread p/p

Equivalently the chromaticity may be expressed as

x,y = - ( - 1/2) Qx,y/ (frf/frf) or x,y ~ Qx,y/ (frf/frf)

where Q is the change in betatron tune with change in accelerating frequency frf . We use this form for measurement of the chromaticity: we measure the change intune with applied change in accelerating frequency. Corrections are applied to thesextupoles (no skew sextupoles to date).

Chromaticity Feedback: WHY

1) initially thought to be a requirement for operational tune feedback (tune peak too broad and flat if too large)2) beam stability requires < 0 below transition energy and > 0 above 3) dynamic aperture issues if too large

Chromaticity Feedback: HOW

tune/coupling feedback chromaticity feedback

Measurement:

Vary rf frequency(specifically, adda frequency modulation ofamplitude frf

with periodicityf) and measuretunes

With feedback (and T/Cfeedback too), the corrections sent to the quadrupoles (the “filtered tunes”) are usedas input to themeasurementalgorithm

measurement and control of charged particle beams in the relativistic heavy ion collider

Documents