Recent observations of collective effects at KEKB

H. Fukuma, J. W. Flanagan, S. Hiramatsu, T. Ieiri, H. Ikeda, T. Kawamoto, T. Mitsuhashi, M. Tobiyama, S. S. Win, KEK

1. Effects of electron cloud in LER

2. Transverse coupled bunch instability in HER

30th Advanced ICFA Beam Dynamics Workshopon High Luminosity e+e- Collisions, October 13 - 16, 2003, Stanford, California

1. Effects of electron cloud in LER

The blowup is suppressed by solenoids installed around the ring at present operation condition.

At KEKB LER a vertical beam blowup caused by an electron cloud (e-cloud) has been observed since the beginning of the operation.

However, if the machine is operated with shorter bunch spacing, a threshold bunch current of the blowup decreases.

Two measurements were performed in the latest operation period to consider measures to suppress the e-cloud further,

A) Measurement of the blowup and the tune shift by changing the strength of solenoid field,

B) Measurement of the blowup and the tune shift by switching off the solenoid locally.

C) Detecting a head-tail motion by the e-cloud by a streak camera (preliminary).

Furthermore we are recently trying a rather academic measurement,

The blowup is still an issue of the near-future luminosity upgrade.

0.0

0.5

1.0

1.5

0 20 40 60 80 100 120

Th

eres

hol

d b

un

ch c

urr

ent

(mA

)

Solenoid Current [%] (100%<->4.5 A)

Oct. 2000

Mar. 20034 spacing

Jun. 20034 spacing

Jun. 20033 spacing

Jun. 20032 spacing 4 spacing

1) 3 and 4 bucket spacing : the threshold increases when the field strength increases.2 spacing : the threshold saturates.

A) Field strength of solenoid vs. blowup or tune shift

2) Assuming a present solenoid system, stronger field will be helpful in raising the threshold if bunch spacing is larger than/equal to 3 buckets.

a) Threshold current of the blowup

Central field or fringe field ?

3) Why stronger field is better ?

-150

-100

-50

0

50

100

150

0 10000 20000 30000 40000 50000

Field calculation of solenoid

Bz

at t

he

cen

ter

of s

olen

oid

(G

auss

)

Distance (mm)

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

0 20 40 60 80 100 120

L/L0(>20G)

L/L0(>10G)

L /

L(@

100%

exc

.)

Excitation(%)

Length covered by field larger than 10 or 20G

(1/8 arc in ring)

Normarized at 100% excitation

0.0

0.5

1.0

1.5

0 20 40 60 80 100 120

Th

eres

hol

d b

un

ch c

urr

ent

(mA

)

Solenoid Current [%] (100%<->4.5 A)

Oct. 2000

Mar. 20034 spacing

Jun. 20034 spacing

Jun. 20033 spacing

Jun. 20032 spacing 4 spacing

Stronger central field may be important.

The increase of the threshold current at stronger solenoid field is not explained by the increase of fringe field in arc sections.

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

0 20 40 60 80 100 120

L/L0(>20G)

L/L0(>10G)

L /

L(@

100%

exc

.)

Excitation(%)

Length covered by field larger than 10 or 20G

(1/8 arc in ring)

Normarized at 100% excitation

-0.002

0

0.002

0.004

0.006

0.008

0.01

0 100200300400500600700800

L/6/6/03_Solenoid

H_0%H_55%H_85%H_100%

Hor

izon

tal T

une

Shif

t

Bucket

-0.005

0

0.005

0.01

0.015

0.02

0 100200300400500600700800

L/6/6/03_Solenoid

V_0%V_55%V_85%V_100%

Ver

tical

Tun

e Sh

ift

Bucket

b) Effect of solenoid field on the tune shift

50% of full excitation, i.e. about 25 G is enough to saturate the tune shift.

4 trains, 200 bunches/train, 4 bucket spacing, bunch current 0.58mA

x,y re2

x,y ds K. Ohmi et al. (APAC01)

Tune shift by e-cloud

-0.005

0

0.005

0.01

0.015

0 100200300400500600700 800

L/6/6/03_sp V_4spV_3spV_2sp

Ver

tica

l Tun

e S

hift

Bucket

-0.001

0

0.001

0.002

0.003

0.004

0.005

0.006

0 100200300 400500600700800

L/6/6/03_sp H_4spH_3spH_2sp

Hor

izon

tal T

une

Shi

ft

Bucket

Bunch spacing

Large tune shift was observed in a fill pattern of 2 bucket spacing.

4 trains, 200 bunches/train, 2/3/4 bucket spacing, bunch current 0.5mA, with 100% solenoid

B) Location of solenoid vs. blowup or tune shift

The blowup and the tune shift were measured by turning off the solenoid locally.

Is there any difference in the effects of the solenoids in arc- and straight-sections ?

RF

Wiggler

Belle

Wiggler

Fuji

Tsukuba

OhoNikko

East

SouthWest

North

0

0.5

1

1.5

2

2.5

3

3.5

4

0 100 200 300 400 500 600 700

west arc off

all on

Ver

tica

l b

eam

siz

e @

I.P.

(mic

ron

)

Beam current (mA)

4/200/3 (regular)

0

0.5

1

1.5

2

2.5

3

3.5

4

0 100 200 300 400 500 600 700

Tsukuba off

all on

Ver

tica

l b

eam

siz

e @

I.P.

(mic

ron

)

Beam current (mA)

4/200/3 (regular)

0

0.5

1

1.5

2

2.5

3

3.5

4

0 100 200 300 400 500 600 700

Fuji off

all on

Ver

tica

l b

eam

siz

e @

I.P.

(mic

ron

)

Beam current (mA)

4/200/3 (regular)

West arc off Tsukuba straight section off

Fuji straight section off

all solenoid on

1. The solenoids in the straight sections are effective on the blowup, even in Fuji straight section where no wiggler magnets are installed.

2. Effect of solenoids on the threshold current of the blowup

1/4 arc > Fuji straight > Tsukuba straight

a) Blowup vs. location of e-cloud 4 trains, 200 bunches/train, 3 bucket spacing

b) Tune shift vs. location of e-cloud

Horizontal Vertical

Arc solenoids off 0.006 (1) 0.006 (1)

Straight solenoids off 0.0035 (0.58) 0.008 (1.33)

Tsukuba straight off 0.001 (0.17) 0.003 (0.5)

All solenoids on 0.0015 0.005

Sum 0.011 0.019

Horizontal (m2) Vertical (m2)

Arc sec. total 24600 (1) 24500 (1)

Straight sec. total 14800 (0.60) 20300 (0.83)

Tsukuba straight sec. 4000 (0.16) 11000 (0.45)

Total 39400 44800

x,y dsdrift

x,y re2

x,y ds -- (1)

(The vertical tune shift in Tsukuba straight section is consistent with (1). )

Tune shifts in arc or straight sections are consistent with (1) except for the vertical tune shift in the straight sections, assuming a fixed cloud density .

Calculation of

Measurement (4 trains, 200 bunches/train, 3 bucket spacing, bunch current 0.5 mA)

Large amount of e-cloud in high vertical beta sections in straight sections ???

1000 bunches, 4 bucket spacingSolenoid on

Train head

1000mA 938mA

983mA

Tail

899mA

C) An attempt to detect a head-tail motion by the e-cloud by a streak camera (preliminary)

Vertical

Longitudinal

1000 bunches, 4 bucket spacing

Train head

893mA 890mA

Solenoid off

Tail

900mA 897mA

Vertical beam size starts to increase at 3 or 4th bunch.

A tilt of a bunch is not clearly observed.

1. Addition of solenoids

a) Works in this summer

1) 215 solenoids at straight sections

2) 50 permanent magnets over BPM at Oho and Nikko straight sections

2. Changing the connection of the solenoid power cables to study the effect of polarity-changing-place.

Q

solenoid

Q

solenoid

D) Others

b) Near future plans

1. Increase of solenoid field

DC current : 4.5A 10 A

Temperature raise of solenoid coil: 100

C(Life time of enamel wire will be OK.)

2. Further solenoid winding in Fuji straight section (RF section)

3. Consideration of possibility to use electrodes to remove electrons inside magnets

No decision yet.

New power supplies are required.

Summary of observations of e-cloud effects

3. Suppression of e-cloud effects in 2 bucket spacing will be very difficult.

Large tune shift was observed in 2 bucket spacing operation.

Almost no effect was observed by increasing the solenoid field.

2. Substantial e-cloud is generated in straight sections according to the measurement of the blowup and the tune shift.

1. Increasing the solenoid field will improve the threshold of the blowup if bunch spacing is larger than/equal to 3 bucket spacing.

4. Clear vertical tilt along a bunch is not observed by the measurement of the streak camera.

2. Transverse coupled bunch instability in HER

In previous operation period, beam aborts accompanied by the horizontal oscillation sometime happened.

Tuning of the bunch-by-bunch feedback system looked to be OK.

Vacuum pressure especially around I.P. was also OK.

Abnormal temperature rise of vacuum components was not observed.

Beam oscillation was measured by a fast memory board after switching off the bunch-by-bunch feedback system.

Horizontal Vertical

bunch

turn

evolution of amplitude

4000 turn

Larger amplitude in tail-part Almost uniform amplitude

Horizontal growth rate < Vertical growth rate

1train, 1152 bunches/train, 4 bucket spacing, 600mA

r.m.s. amplitude along train abort gap

pilot bunch

VerticalHorizontal

Larger amplitude in tail-part Almost uniform amplitude

Saturation at large amplitude No saturation of amplitude

4trains, 200 bunches/train, 4 bucket spacing, 600mA

Growth rate

1 train, 1152 bunches, 4 bucket spacing

Growth rate : 0.002 (1/turn) or growth time : 5ms @700mA

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 200 400 600 800 1000

HER Horizontal

H (Jul01)

H max (Jul01)

H (Jun03)

Gro

wth

rat

e (1

/tu

rn)

Beam current (mA)

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 200 400 600 800 1000

V (Jul01)V max (Jul01)V (Jun03)

Gro

wth

rat

e (1

/tu

rn)

Beam current (mA)

HER Vertical

Oscillation mode

mode no.

lower sideband

upper sideband

magnified view of left part

1 train, 1152 bunches, 4 bucket spacing, 600mA

VerticalHorizontal

peak at about 10 peak at 0

400mA 500mA

600mA

1 train , 1152 bunches, 4 bucket spacing

Beam current dependence of horizontal mode spectrum

Peak did not move much.

Features

a. Oscillation amplitude along train

Horizontal: train head < train tail, Vertical:uniform

b. Growth rate

Horizontal :0.0006 turn-1(@600mA) < Vertical : 0.0029 turn-1(@600mA)

c. Mode

Horizontal :broad peak at about mode 10, Vertical :peak at mode 0

Different characteristics of the horizontal and vertical oscillations.

Sources of the instability may not be same in horizontal and vertical planes.

CO+ without magnetic field H + in a dipole field

Resistive wall Electron cloud

Bunch position along train

No amplitude growth along a train in resistive wall case.

Simulation of horizontal instability ( F. Zimmermann(PAC2003))1. CO+ without magnetic field 2. H + in a dipole field3. Resistive wall 4. Electron cloud

(1 train, 1200 bunches, 4 bucket spacing, 670mA)

CO+ without magnetic field H + in a dipole field

Resistive wall Electron cloud

Oscillation of a few bunches

Saturation of oscillation in ion cases.

Mode spectra

CO+ without magnetic field H + in a dipole field

Resistive wall Electron cloud

Peak appears near 10 in CO case.

Growth time

CO+ without magnetic field

H + in a dipole field

Resistive wall

Electron cloud

1ms

5ms

4000ms

2ms

@beam current 670mA (observation : 2ms max.)

Comparison with observation

Results of the simulation assuming CO+ ions is almost consistent with observations of the horizontal instability.

Peak of mode spectrum at about 10.

Saturation of oscillation amplitude.

Maximum growth time of order of 1 ms.

Can the simulation explain the vertical instability ?

Why the peak of the mode spectrum does not move by changing the bunch current ?

Expected future simulation work

Growing amplitude along the train.

Summary of observations of transverse coupled bunch instability in HER

1) Horizontal coupled bunch instability in HER sometime causes beam aborts which limit the beam current.

3) A question remains why the instability was not cured by the bunch-by-bunch feedback system.

2) Observations are consistent with the results of the simulation which assumes the instability is caused by CO+ ions.

4) Vertical coupled bunch instability is also observed in HER. Features of the instability are different from those of the horizontal instability, which suggests the vertical instability is caused by a different source than that of the horizontal one.