Jan Uythoven with input from RAssmann MBarnes FCaspers SClaudet HDay LGentini AGrudiev EMétral BSalvant CZannini Heating due to Beam Impedance ID: 779865
Download The PPT/PDF document "Beam Induced heating and bunch length de..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
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
Slide2Heating 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
Slide3LHC 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
Slide415/07/2011 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
Slide5Check 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
Slide6Bunch 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
Slide7Different 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
Slide9MKI 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
Slide10Compare 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
Slide11Cryogenic 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
Slide12Injection
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
Slide13Singularities 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
Slide14Conclusions 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
Slide15Temperature 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
Slide16Collimator Heating by EM fields
Bunch Length
Collimator
temperature
25 deg.
0.5 ns
1h
Slide17Measured 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
Slide18Calculated 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
?
Slide19Collimator 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
Slide20Bunch 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
Slide21Conclusions
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
Slide22Conclusions
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