1

CC

& M

P -

CC

10

- C

ER

N1

6.1

2.2

01

0

Crab Protection @ LHCJ. Wenninger

CERN Beams Department

for the LHC Machine Protection Panel

Beam excursions1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

2

The principle of the CC is to kick head and tail of the bunches to enhance the luminosity. The more effective the CC is, the larger the kicks, the larger the excursions of head and tail.

Local crabs : SLHC-V3

][2.1][ˆ sx sx xx 5.2|ˆ| max Assuming a full length of 2 s

s

xIn case of ‘failures’ (cavity trips, control errors etc) the resulting particle excursions should not lead to component damage, and if possible not even to quenches (efficiency).

Performance gain and criticality of machine protection issues are strongly correlated !

Stored Energies

16

.12

.20

10

CC

& M

P -

CC

10

- C

ER

N

3

The LHC : a new regime for Machine Protection.Even the beam halo can be dangerous !

LHC 2010

LHC 2011

Already today MP largely drives progress, MDs etc(unfortunately)

Tails @ LHC 20101

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

4

y = 0.27 mm

~ 5.7 ~ 4.7 ~ 3.7 3.5% of beam within 1.5

Tail populations of 1% to 3% were measured with primary collimator scans (end of fills) in the ‘last’ sigma:

For a nominal beam this would represent around 3-10 MJ !

R. Assmann

MPS & Collimation1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

5Courtesy C. Bracco

Although the primary design goal of the collimators is beam cleaning, they also play an essential role for MP.

Collimators define the machine aperture. But phase space coverage is not complete !

The large majority of failures leads to a primary particle impact at one of the collimators.o Collimators are robust to survive limited beam impact.o Collimators must cope with beam impact from CC failures.o Beware of (massive) quenches downstream of collimation sections !

LHC beam dump delays1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

6

Achievable response time ranges between 100 s and 270 s.

>> Triggering a dump is not the end of the story,

must be able to survive up to another 3 turns without damage.

UserSystemprocess

a failure has been detected…

beam dump

request

Beam Dumping System waiting for beam gap

max 89μs

Signalssend

to LBDS

t2 t3

Beam Interlocksystemprocess

max 100 μs

t1

> 10μs

Kicker fired

t4

all bunches have been extracted

max 89μs

KEKB CC trip1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

7

1. D11-F klystron out

2. D11-F cavity voltage

3. D11-F cavity tuner phase

4. HER DCCT

RF off

Beam Abort

Y. Morita

Complete cavity trip in less than one LHC turn !

Machine Protection

Ultra-fast failures1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

8

~ 1 turn failures are too fast for the MPS to react => protection requirements for ultra-fast failing objects:

‘Harmless’ => no damage

and/or

‘Never occur’ => really rare

“consequences” x “rate” must compatible with SIL3.

(less than 1 ‘catastrophic’ event in 1000 years)

SIL3 systems at the LHC : BLMs, beam interlock system, beam dump system.

Protection of the LHC against CC failures

16

.12

.20

10

CC

& M

P -

CC

10

- C

ER

N

9

Options to limit the machine protection footprint of CCs :

Make failures ‘slow’ and

provide fast interlock signals.

Reduce the effect of a trip.

Reduce the density of tails.

…

Reduce movement (loss) to a fraction of a beam sigma / turn

Detailed info on failure modes is required !

Failure mitigation: divide and conquer !

16

.12

.20

10

CC

& M

P -

CC

10

- C

ER

N

10

Possible scenario for CC would be to split them into a sufficient number of INDEPENDENT sub-systems (power+control+cavity).

o Failure of one cavity / unit becomes acceptable.

o Impedance?

o Watch out for common-cause failures that break the independence !

- What about mains failures??

Failure mitigation: get rid of the tails!1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

11

Assuming primary collimation set to ~6 sigma, one could try to deplete the amplitudes between 3-4 sigma and primary collimation:

o Those particles do not contribute significantly to luminosity.o If they are ‘no’ particles out there, failures may be acceptable.

Use CCs in combination with a hollow electron lens acting as tail scrapper?o CC operation would be coupled to e-lens.o How to verify that tail population is acceptable?

Safe(r) CCs1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

12

Limit losses/movements per turn to a fraction of beam sigma equivalent (~per mill of total beam).

This will probably require some optimization:o failure modes (time constants),o no. cavities (effect of trip),o tail control (?),o collimation (upgrades),o …

Do not forget:o Commissioning time (also MP aspects !),o Availability (trip rate !!).

16

.12

.20

10

CC

& M

P -

CC

10

- C

ER

N

13

Times scales1

6.1

2.2

01

0C

C &

MP

- C

C1

0 -

CE

RN

14

Best failure detection time = 40 us = half turn

NC magnetcircuit

QuenchKickersHuman beings

Failures

Interlocks

UFOs