Beam Induced heating and bunch length dependence - PowerPoint Presentation

Slide1

Beam Induced heating and bunch length dependence

Jan Uythoven – with input from

R.Assmann

,

M.Barnes

,

F.Caspers

,

S.Claudet

,

H.Day

,

L.Gentini

,

A.Grudiev

,

E.Métral

,

B.Salvant

,

C.Zannini

Heating due to Beam Impedance

Significant temperature rise due to heating by the beam has been measured for

LHC injection kickers MKI

Cryogenic beam screensCollimator(s)For these three systems a strong dependence on the bunch length was measuredPresent operation with slightly longer bunches seems to solve the issue... for the momentGeneral beam induced heating – any suspicion of specific Higher Order Modes?Consider these three systems in more detail

15/07/2011 Mini Cham

RF Heating, Jan Uythoven

LHC Injection Kickers MKI

Reached 54 degrees

Long time constants, order of 10 hours

Strong dependence on bunch lengthLarge variation between measured temperatures Expected to be a measurement artifact15/07/2011 Mini Cham

Measured magnet temperatures in May and June

55

C

20

C

RF Heating, Jan Uythoven

15/07/2011 Mini Cham

Ferrite Heating: Measurements in Clean Room

PT100 on ground plate of magnet

PT100 on ferrite of “tube”

Ceramic tube

PT100 setup for installed

MKI’s

Both PT100’s installed gave same reading to within ± 4

˚

C.

Conclusion: for measurements in clean room, Curie point of ferrite corresponded to

106

˚

C measured

on ground plates for dissipation of

172 W/m

in “heater element”.

SIS injection interlock presently 60

˚C,

because of uncertainty in temperature reading

LIBD: May 17, 2011

4

Magnet in vacuum tank; tank at atmospheric pressure; no bake-out jacket.

Magnet Inductance

M.Barnes

RF Heating, Jan Uythoven

Check of MKI rise time during soft start B2

15/07/2011 Mini Cham

Above

Tcurie

T = 5 C will give:

30 % reduction of magnet inductance

30 % reduction of kick strength –

can be checked with pilot injections

20 ns reduction in kicker rise time –

can be checked without beam

No correlated reduction in pulse rise time found so far

RF Heating, Jan Uythoven

Bunch length: Injection kicker magnets MKI

6

B2 bunch length 1.2 ns

MKI B2 temperature

plateau at 49 degrees

B2 bunch length 1.13 ns

MKI B2 temperature

@ 54 degrees and still increasing

Pulsed MKI at 54 degrees, without beam. Rise time ok: increased temp IL to 60 degrees

15/07/2011 Mini Cham

RF Heating, Jan Uythoven

Different Measurements of MKI Impedance

15/07/2011 Mini Cham

H.Day

RF Heating, Jan Uythoven

Measurements in March 2010 are a MKI with 15 screen conductors before bake-out

Measurements in Jan 2011 are a MKI with 15 screen conductors after operation in the machine

Some variation over the different measurements

Real

Increasing impedance for higher

frequencies = shorter bunches

Impedance peaks from 1 GHz onwards

Measured and Calculated Power Spectra

RF Heating, Jan Uythoven

15/07/2011 Mini Cham

Bunch length of ~ 1.2ns (t is half the bunch length) gives good agreement

with measurement below

1.2GHz.

At

higher frequencies it underestimates

The beam impedance indicates that we should investigate the bunch spectra above 1GHz given the larger impedance in this regime

Most

theoretical

spectra have negligible higher frequencies (>1GHz)

Should we worry about this?

H.Day

MKI Integrated Power Loss

RF Heating, Jan Uythoven

15/07/2011 Mini Cham

For 2808 bunches (20 % reduction rel. to figure):P_loss

≈ 87 W/m (measured spectrum)P_loss

≈ 70 W/m (cos^2 spectrum)

Earlier estimates: P calc ≈ 90 W/m

f

0

= 40MHz



3554 bunches

N

b

= 1.15x10

11

Compare to temperature measurement

To be compared to 172 W/m to reach

Tcurie

during lab testsStrong dependence on bunch length not completely understoodPresently with 1400 bunches, half of nominal. Expect about 45 W/m power deposition. Lab. Test 172 W/m heater power. Assume 120 W/m transferred to ferrites, which gave dT

= 80 C.

For 45 W/m expect 45/120 *

80

C = 30 C, which is what we measured with 1400 bunches

Nominal operation (90 W/m) is expected to give T = 85

C, expected to be acceptable

More critical will be if the

bunch current will be significantly increased !

RF Heating, Jan Uythoven15/07/2011 Mini Cham

Cryogenic Beam Screens

LHC 8:30 meeting

7/15/2011

Injection

Ramp

Longer bunch length on purpose => lower heat loads confirmed

Appreciated as well by collimators and kickers !!!

18K

Temperature increase (Peak - before injection) for 1092

bunches

:

Maxi: +4.5K (1836,

short bunch: 1.12ns-1.14ns

)

Avg

: +3.6 K

Mini: + 2.6K (1859,

longer bunch:1.22ns-1.27ns)

About 2.0K for injection, rest for ramp effect

∆T

S. Claudet

Bunch length

Injection

Ramp

Too low bunch length => direct impact on image current heat loads!

ARCs

34

&

56

mostly penalised

Injection

Ramp

Longer bunch length => lower heat loads

Very homogeneous behaviour of all arcs

Bunch L B1

1.1 ns

1.2 ns

13 K

22 K

13 K

22 K

Singularities in the Arcs

RF Heating, Jan Uythoven

15/07/2011 Mini Cham

Majority of

beam screen loops behave similarlySome strange effects in sector 23 that could be linked to orbit

changes. To be confirmed.

Few zones per sector showing higher heat loads

than average,

could be diagnosed with

tomograph

to see if some PIM’s could be the origin?

No sharp HOM expected in present LHC beam screen impedance model, to be confirmed for the PIMs

B.Salvant

Conclusions from Serge

The installed capacity for beam screen cooling and local limitations are as expected and will not change in the coming months.

The present operation scenario make use only of some 5 K temperature rise (13K to 17K with peaks at injection below 25K) while some 15K where considered (5K to 20K)

This is part of our present strategy to try to learn how to use it at best without perturbing the beamsWe have much more to gain in extending the present operating range rather than talking about real limitations...We are doing our best with the time we have to be away of the critical path (hopefully succeeded) with appropriate cooling, bake-out, operational settings and updated logic

15/07/2011 Mini Cham

RF Heating, Jan Uythoven

My conclusion: The higher temperature excursions do not form a hard limitation, but

cryo

operation is a lot easier with longer bunches

Temperature rise seen on many

collimators

RF Heating, Jan Uythoven

15/07/2011 Mini Cham

Strongest T rise and hitting the interlock limit of 50

C

only on TCTVB_4R2

7 Days

12 minutes

25

50

Collimator Heating by EM fields

Bunch Length

Collimator

temperature

25 deg.

0.5 ns

1h

Measured collimator impedance

15/07/2011 Mini Cham

No evidence for a resonance up to 3

Ghz

with this kind of measurement

Real part longitudinal impedance (relative)

F.Caspers

Measurement without transition beam pipes

RF Heating, Jan Uythoven

Calculated collimator impedance

15/07/2011 Mini Cham

B. Salvant,

L.Gentini

Work Ongoing

Most heating expected near transitions, but not at the collimator

blocks

RF Heating, Jan Uythoven

Peak shifted

upwards

by ferrites

behind

the transitions

?

Collimator heating - TCTVB beam 2

Also seen on other collimators, but less important

Unlikely collimator block heats up that rapidly

Higher Order Mode could be possibleNo peak in measured impedanceSmall peak around 1.4 GHz in actual impedance modelCan be the heating of the temperature sensor (support)Temperature interlock limit can be increased if requiredPresent interlock limit set as a warning level, vacuum effects (not seen). Block can stand extremely high temperatures.

RF Heating, Jan Uythoven

15/07/2011 Mini Cham

Bunch Length

Important parameter for

Cryogenics stability

Collimator heatingInjection kicker heating...Work ongoing to improve blow-up control during the ramp by the RF-teamBetter reproducible results -> test operation with longer bunchesDisadvantage of longer bunches is possibly more debunched

beam when a cavity trips, but not an issue at the moment

RF Heating, Jan Uythoven

15/07/2011 Mini Cham

Ramp

Conclusions

Injection Kickers, Collimators and Beam Screens have reached temperatures higher than desired or expected

Presently ‘solved’ by using increased bunch length of about 1.2 ns

During the energy ramp and at 3.5 TeVOnly disadvantage is possible increase of longitudinal losses, but not a concern for the moment (abort gap population)For the MKI and collimators we see some trapped, Higher Order Modes. Still to be confirmed if they

can explain the strong dependence on bunch lengthNo reason at the moment not to use longer bunches (1.2 ns)

Impedance team working on these three issues with high priority

15/07/2011 Mini Cham

RF Heating, Jan Uythoven

Conclusions

MKI

Measured temperature rise is in agreement with present models

Seems enough margin to operate with nominal parametersHowever, if bunch current significantly above nominal might become an issueCount on further increasing the actual temp. interlock limit (60 C to 100 C)

Cautious because of large damage potentialIncrease temperature interlock in small steps!

Development of thermal models and effect of different beam screens

Can gain a factor two in screening by increasing the number of screen conductors from 15 to 24 (24 = original design)

Cryogenics

No hard operational limit, but ‘easier’ when temperatures more stable

Suspicion on some RF fingers – non conformity

Collimators

Suspicion on temperature measurementPossibly Higher Order Modes due to reduced damping by ferrites

Can increase temperature interlock limit when requiredRF Heating, Jan Uythoven

15/07/2011 Mini Cham