LHC beams from the injectors - PowerPoint Presentation

Slide1

LHC beams from the injectors – Chamonix outcome

H. Bartosik summarizing presentations from Chamonix 2017 workshop:

MSWG 17.02.2017Slide2

LHC Beam Brightness in the PSB

LHC performance workshop 20171

LHC 25ns

BCMS

25ns

Change of WP from

(4.28, 4.45) -> (4.42, 4.45)

r

ecovering 2015 brightness

LHC

100ns

LHC 25ns

Brightness per PSB Ring (

normalised

to LHC bunch intensity)

Time Slide3

25 ns Standard Beam in 2016Operated close to optimum brightness (in 2016 one batch only

…)Intensity limited mainly by beam loading and available RF power during SPS rampLHC performance workshop 2017

2

LHC fills in 2016Slide4

Transv. Emittance Growth for 25 ns Standard Beam in 2016

PSB emittances on brightness curveImportant blowup of ~40% (H) / 15% (V) between PSB and PS; H ~10% expected from dispersion mismatch and V ~5-10% from dispersion+betatron mismatch

Should apply better dispersion correction of measured profile (H plane)Optics of transfer line will be improved for LIU (in particular additional quadrupole

)LHC performance workshop 2017

3

dispersion mismatch?

Injector

data from wire scanners

Absolute calibration still an issue

SPS data at FT not reliable

LHC data from BSRT at injection

A

vg. 2016 value at LHC start of collision

~3.5

μmSlide5

25 ns BCMS beam in 2016Not yet at optimum brightness due to

emittance blow-up  transport of low-emittance beams more challengingMargin to increase intensity to 1.3e11 p/b at SPS extractionLHC performance workshop 2017

4

LHC fills in 2016Slide6

Transv. Emittance Growth for 25 ns BCMS beam in 2016Important blowup of

~50% (H) / 10% (V) between PSB and PS; H ~20% expected from dispersion mismatch and V ~5-10% from

dispersion+betatron mismatchVertical growth mainly along PS cycle

(see next slides); ~25%LHC performance workshop 2017

5

dispersion mismatch?

Avg. 2016 emittance at LHC start

o

f collisions:

~2

μmSlide7

MD on vertical blow-up in PS (I)

LHC performance workshop 20176

B

low

-up on inj.

Plateau

S

maller

average

emittance after 2

nd

injection (new bunches not yet

blown up)

B

low-up during first ramp (enhanced space charge due to bunch shortening)Slide8

MD on vertical blow-up in PS (II)

LHC performance workshop 20177

Optimized working point

Constant bucket area during first ramp (to keep bunches longer)

Still to optimize voltage program to minimize lossesSlide9

Remarks Emittance MeasurementsWire scanner measurement comparison between machines still problematic

Dispersive component of measured profile (particularly important for small beams and beams with large momentum spread  BCMS and future LIU beams)Should rather use measured than simulated optics input parametersSPS: Resolution of wire scanners and wire damage threshold (SPS FT data to be taken with care…)

Calibration of wire scannersCorrect settings of wire scannersLHC

supertable displays since 2016 emittances/brightness from injectors referenced to LHC bunch values and beam type for filteringSPS measurements should be split up in FB and FT (

emittance growth during cycle)

Standardised

way of logging – use always same wire scanner with pre-configured settings

Include automatic BSRT (or wire scanner) measurement data at LHC injection in

supertable

LHC performance workshop 2017

8Slide10

Special beams (I)80 bunches

7 PSB bunches injected into PS  triple splitting  one of 21 bunches eliminated  splitting into 80

Tested in LHC MD (1.15e11 p/b in 2.6 μm)Increased losses at LHC injection (scraper settings?

) - optimization neededUsed PS extraction kicker for eliminating bunch (dirty…);

f

urther

MDs needed for using transverse damper

Potential for larger number of bunches in LHC

(320b per

injection

after LS2

)

,

or for mitigating potential total current limits in SPS for same LHC performance (240b per LHC injection)8b4e7 PSB bunches injected into PS  double instead of triple splitting  flat top splitting into 56Tested in LHC MD (1.15e11 p/b in 1.8

um)Before LIU potential for higher intensity per bunch

Interesting to overcome e--cloud issues

(mainly for LHC)

LHC performance workshop 2017

9

LHCB1

LHCB1Slide11

Special beams (II)Doublet

beam‘Scrubbing beam’ for LHCBased on LHC 25 ns standard beamWith final bunch rotation, using 40 MHz, but no 80 MHz cavitiesSPS injection on unstable phase to split bunches into doublets

Not an easy beam for the SPS (beam stability and losses are critical)Could be made available again for LHC tests by mid 2017

LHC performance workshop 2017

10

25 ns

Non adiabatic splitting at SPS injection

20 ns

5 nsSlide12

Main 2016 ImprovementsPS: Introduction of second 40 MHz cavity for LHC beam transfer to SPS

Losses in SPS (even on flat bottom!) mostly due toun-captured beam as result from PS bunch rotationLongitudinal tails and therefore losses reduced using

additional 40 MHz cavity in PS

Loss reduction in SPS by ~40%Longitudinal losses at LHC inj. reduced by ~factor 10SPS: Increase of 800 MHz voltage at flat top  more margin to cure longitudinal instabilities

LHC performance workshop 2017

11

Simulated distribution at SPS injectionSlide13

Limitations for 2017 from InjectorsLimitation

of number of bunches into LHC from transfer line collimators (TCDIs) due to insufficient attenuationStandard beam OK with 4x72, BCMS limited to 3x48 (144) bunches

Max. intensity:PS

: Limited to ~1.6e11 p/b due to longitudinal instabilities (or ~2e11 p/b with new Finemet-based longitudinal feedback if prepared for operational use)

SPS

:

Limited to

~1.3e11 p/b

mainly because of beam loading and lack of RF

power resulting in

losses

SPS Q22 optics with intermediate transition energy to be tested in 2017; hoping to increase intensity

out of SPS due to less required RF power

 increased margin for beam loading compensation

LHC performance workshop 201712Slide14

Expected 2017 ImprovementsPS

:Improve transverse emittance preservation along the cycle (optimisation of voltage program for constant bucket area, working point,…) – can it be made operational in second half of the run?SPS:N

ew TIDVG internal dumpNew beam dump

limit OK for 2017 beams (1h continuous dumping once per supercycle)Will require some conditioning with beam for >36 LHC bunches, at each intensity increase and for FT beam; crucial for trains of 288 bunches for LHC scrubbing run

Hoping to increase intensity

out of SPS due to less required RF power with

Q22 optics

(to

be tested in

2017)

Minimum

batch spacing in 2017:

200

ns MKP (SPS injection kicker) gap  higher number of bunches in LHCSuccessfully tested in LHC p MDs and operationally used for 2016 p-

Pb runIn particular interesting for BCMS LHC filling schemesRelies on optimal synchronization of MKP switches (requires regular checks)

LHC performance workshop 2017

13Slide15

TIDVG4 – From Paper to RealityAll parts at CERN; being assembled and welded

Bake-out start: 1st MarchLHC performance workshop 2017

14Slide16

LHC physics beams in 2017LHC performance workshop 2017

15

Intensity

[1e11 p/b]

Emittance [um]

pattern

25 ns standard (like 2016)

1.15

2.5

(2.4)

1-4 x 72

 288

25 ns standard (max. intensity)

1.30

2.8

(2.7)

1-4 x 72

 288

25 ns BCMS (like 2016)

1.15

1.7

(1.4)

1-3 x 48

 144

25 ns BCMS

(

max. intensity)

1.30

1.9

(1.6)

1-3 x 48

 144

25 ns 80 bunches (like 2016)

1.15

2.6

(2.4)

1-3(4) x 80

 240

25 ns 80 bunches

(

max.

intensity

)

1.30

2.8

(2.7)

1-3(4) x 80

 240

8b4e (like 2016)

1.20

1.8

(1.6)

1-3 x 56

 168

8b4e

(

max. intensity

)

1.60

2.4

(2.1)

1-3 x 56

 168

E

mittances

in parentheses should be achievable,

to

be demonstrated operationally Slide17

Scrubbing Requirements for 2017

Sector 12

has been

warmed-up

(vented) to replace the dipole with the inter-turn short:

Based on the LS1 experience we have to assume that the

SEY will be reset

(might be better thanks to the larger accumulated dose but no direct experience is available)



We

might

need to

start gently (short trains

)

Nevertheless scrubbing in

S12 will be more efficient than in 2015

since:

It will

be easier to preserve the beam quality

: only 1/8 of the arcs with high SEY

Better management of

heat load transients from the cryogenics

7 days

allocated in the present schedule

Intensity

increase will also be limited by

conditioning time

of injection kicker (MKI2D)

which was exchanged

after the end of the p-p run

We plan to use the longest available bunch trains up to

4x72b per

injection

:

Important occasion to assess the

scrubbing potential of the nominal bunch scheme

which could not be used up to now in Run 2 (issues with TDI in 2015, SPS dump in 2016)Slide18

2017 physics production: what changes?

SPS beam dump will be replaced

more intensity

per injection:Standard scheme

(low brightness): 288b per injection (4x72b)

BCMS scheme

(high brightness

):

144b per injection (3x48b)

Improved

vacuum in inj.

k

icker (MKI)

regions:

Possible to

increase the bunch intensity

(up to ~1.3e11 p/bunch)

Improved

rise-time in both LHC and SPS injection kickers

200 ns

spacing between PS batches (225 used in 2016)800 ns

spacing between SPS injections (900 used in 2016

)

In the following we will also assume that the Abort Gap Keeper Length is adjusted to the actual train length (as done in 2016)

Slide19

BCMS

2017: 2556b, 144

b/injection (3x48)

Standard 2017

:

2760b,

288

b/injection

(4x72, ~40% lower brightness)

BCMS

2016

:

2220b

,

96 b/injection (2x48)

15% more bunches

w.r.t

. BCMS 2016

7% more bunches

w.r.t. BCMS 2017

Thanks to C.

Schwick and J. Boyd

2017 physics production: what changes?

Filling schemesSlide20

The

impact of the filling scheme can be estimated knowing the e-cloud rise-time from the RF stable phase measurements

Bunches at the head of the train generate significantly less heat load

w.r.t. bunches in the tailsFilling

schemes having

short trains and lots of

gaps

are more favorable

The effect of the

bunch intensity

is modeled based on data acquired in MD (2016)

We

assume that the

scrubbing status is the same as at the end of 2016

Might

not be true right at the beginning of the year due to S12 recovery…

Heat load estimates: BCMS vs standard

e-cloud saturation

~bunch 30

e-cloud saturation

~bunch 20

1us

gap

250 ns

gap

RF stable phase measurement

J. Esteban MullerSlide21

Heat load estimates: BCMS vs standard

BCMS 2016

2x48b per injection

2220

2220

2220

Case 1: “BCMS 2016”, 2220b., 2x48b per inj

ection

(in case SPS dump is not replaced)

Some margin

w.r.t

.

cryo

cooling capacity for 1.1 x

10

11

p/bunch

(as in 2016)

Bunch intensity could be increased up 1.3 x 10

11

p/bunch

without limitations on the number of bunches

Cryo

limit

Estimate for the sector with the highest heat load (S81)

2

016Slide22

BCMS 2016

2x48b per injection

BCMS

2017

3

x48b per injection

2220

2556

2220

2556

2220

2556

2260

Dashed bars:

max. allowed by filling scheme,

Full bars:

max. allowed by heat load limit

Heat load estimates: BCMS

vs standard

Cryo

limit

Estimate for the sector with the highest heat load (S81)

2

016

Case 2: “BCMS 2017”, 2556b., 3x48b per injection

Still

within the

cryo

capacity

limit for bunch intensities up to

1.2x10

11

p/bunch

Limit is exceeded by 10% if the bunch intensity is increased to 1.3

x10

11

p/bunchSlide23

2220

2556

2760

2610

2280

2220

2556

2760

2220

2556

2760

2030

2260

BCMS 2016

2x48b per injection

BCMS

2017

3

x48b per injection

Standard 2017

4x72b per injection

Cryo

limit

Dashed bars:

max. allowed by filling scheme,

Full bars:

max. allowed by heat load limit

Heat load estimates: BCMS

vs standard

Estimate for the sector with the highest heat load (S81)

2

016

Case 3: “Standard 2017”, 2760b., 4x72b per injection

C

ryo

capacity

limit is already reached

for a bunch intensity of

1.1x10

11

p/bunch

For larger bunch intensity the standard scheme is limited to a

number of bunches

that is even

lower than BCMSSlide24

2220

2556

2760

2610

2280

2220

2556

2760

2220

2556

2760

2030

2260

BCMS 2016

2x48b per injection

BCMS

2017

3

x48b per injection

Standard 2017

4x72b per injection

Dashed bars:

max. allowed by filling scheme,

Full bars:

max. allowed by heat load limit

Heat load estimates: BCMS

vs standard

Case 3: “Standard 2017”, 2760b., 4x72b per injection

C

ryo

capacity

limit is already reached

for a bunch intensity of

1.1x10

11

p/bunch

For larger bunch intensity the standard scheme is limited to a

number of bunches

that is even

lower than BCMS

S

tandard scheme

(lower brightness)

does not really allow for a larger number of bunches!

BCMS

seems to be the

natural choice for 2017-18

. Moreover:

Intensity

ramp-up

will most likely be

fast

(

2016-like)

It

will be easier to deal with

S12 recovery

if

needed

But probably

we

will not see more conditioning than in

2016

Not

much impact on Run 2

performance but concerns for Run 3 and HL-LHC

A period with long bunch trains (4x72b) could be envisaged if this scheme shows to be promising during the scrubbing run (possible if we keep

b

*=40) Slide25

Possible tests with “doublet” scrubbing beams

20 ns

5 ns

In 2015, due to strong transverse instabilities, it was possible to

accumulate only trains of 24 doublets

(up to ~250 doublets in total)

This schemes becomes interesting only if it is possible to store significantly

more bunches

(>1000 doublets) and in

longer trains

(48-72 doublets/train)

Doublets

could not be used in 2016

due to limitations from the SPS

beam

dump

We plan to

restart these studies in MD

with

m

ain goals:

Identify optimal settings to

stabilize the beam

(Q, Q’,

octupoles

, using the knowledge from 2015-16)

Assess

the achievable beam intensity

In case of positive outcome, we could think of

longer test period

to probe the

scrubbing efficiency

(in 2017 or later)

Losses observed in 2015 on trains of 24 doubletsSlide26

SummaryTransverse emittance preservation along the chain

of great importance, also in view of LIU and HL-LHCNeeds constant monitoring (maybe special control room application to react timely to degradation), adequate beam instrumentation and data quality (better dispersion correction), refined simulations and beam studies,

optimised matching between machines, longitudinal mitigation of space charge effects (controlled emittance blow-up)

Intensity:Improvement of losses from use of additional 40 MHz PS cavity (reduction of PS-to-SPS transfer losses and LHC injection losses)To become operational, suggest dedicated application to switch between 1/2 40 MHz cavities (no spare 40 MHz cavity) and

review spare policy

(additional anode power converter and other spare parts)

SPS remains intensity bottleneck

T

ransfer of

BCMS bunches

SPS

 LHC

limited to 144 due to insufficient protection with high brightnessSuccessful installation and operation of new TIDVG key to lift 2016 intensity limitations for LHC and FT physics

LHC performance workshop 2017

25Slide27

26

LHC scrubbing:

4 x 72 bunches

conditioning of new TIDVG

+ setting up of LHC beams

LHC physics: 3 x 48 bunches BCMS up to 1.3e11 p/b (200 ns batch spacing!)

LHC scrubbing MD later in the run: doublet beams with 0.6-1.6e11 p/doublet