MD400: LHCemittancegrowthin presence ofanexternal sourceofnoise during collision

Transcription

1 CERN ACC NOTE javier.barranco@cern.ch MD00: LHCemittancegrowthin presence ofanexternal sourceofnoise during collision J.Barranco,X.Buffat,T.Pieloni,C.Tambasco,J.Qiang,G.Trad,D.Valuch,M.Betz,M. Wendt,M.Pojer,M.Solfaroli,B.Salvachua,K.Fuchsberger,andM.Albert École polytechnique fédérale delausanne, Particle Accelerator Physics Laboratory, CH 0 Lausanne, Switzerland CERN,CH Geneva, Switzerland Lawrence Berkeley National Laboratory, Cyclotron Road, Berkeley CA970 Abstract Theinterplay between head on beam beam interaction andexternal sources of noise canbe asignificant sourceofemittancegrowth,especiallywhenconsideringlargebeam beamtuneshiftsasforthehl LHC upgrade project. In this experiment the emittance growth of colliding bunches with different brightness and therefore different beam beam parameters in the presence of an external white noise source with different amplitudes is measured for different gains of the transverse feedback. Keywords: CERNLHC,Beam Beam Effects. This is an internal CERN publication and does not necessarily reflect the views of the CERN management.

2 CERN ACC NOTE Contents Motivation Experimentalconditions MDresults Summary Acknowledgement References i

3 CERN ACC NOTE Motivation Themaximum achievable luminosity inacollider asthelhcanditsupgradesisdictatedbyidentifying the limits due to beam beam interactions. One of these limitscouldcomefromanemittanceincrease duetotheinterplay ofhead on collisions andexternalnoise. Themaximumachievable head oncollision and its dependency on external noise, as for example the one introduced by crab cavities, is therefore an essential study to identify possible limitation and define tolerances on hardware (i.e. crab cavities) that could deteriorate the beam emittances and consequently the luminosity reaches of a collider. The present MD goal is to understand the interplay between beam beam head on and an external noise source. As a direct consequence one will be able to estimate acceptable noise tolerances to be introduced byfuture newelements asforexample crabcavities foreseeninthehl LHCproject. Aswell the results presented here will be used to validate existing analytical and numerical estimations. This to gain confidence in our predicting tools for the up grade andfutureacceleratordesigns.accordingto [, ], under certain assumptions, the expected emittance growth in colliding bunches under the influence of an external noise is proportional to the frequency spread introduced and therefore to the beam beam parameter. The emittance growth is given by, dǫ, x ǫ 0 dn = s 0 ( ) ) g S(, () π ξ with s 0 aconstant, is the noise amplitude in [σ], g is the transverse feedback gain defined as g =/τ withτ the number of turns todamp, ξ is the beam beam parameter and S(x)=/(+x). Thereexistaswellanextensivecampaignofself consistentsimulationsdonewiththecombi[] code that show very good agreement against Eq. () (Fig. ) forlargeenoughdampergains.butinter esting deviations for smaller ones []. Numerical limitations are under investigation also to identify the limits of numerical simulations of such complex systems as colliding beams. Damper Gain g=0. (0 turns) 0 / 0 [per hour] COMBI LHC Noise=0 COMBI LHC Noise= 0 COMBI LHC Noise=0 Alexahin Noise=0 Alexahin Noise= 0 Alexahin Noise= total Fig. : Simulations by COMBI where the emittance growth of a single bunch colliding in a single IP is evaluated for different noise amplitudes and beam beam parameters. The damper gain is set to 0 turns inthis case. The agreement is fairly good witheq.(). Experimental conditions Threebunches perbeamwereinjected: witnessnominal, nominalwithincreased intensity andahigh brightness one (see Table ). Damper gain was initially set to0 turns. The octupole current was kept low at 6. A as well as chromaticity at units which are known tohaveaclearimpactonthemeasured emittance growth. The two colliding bunches will meet the partners in both IP and IP. After collision

4 CERN ACC NOTE Table : Target beam parameters and expected beam beam parameterforthedifferentbunchestobeused during the MD. Bunch I[0 ppb] ǫ n,x,y [0 6 µrad] ξ bb /IP HighBrightness Fig. : The MD in a nutshell. Various TIMBER signals corresponding to the three fills done during the noise MD. Three bunches, two colliding and one witness, are injected per fill (three steps in intensity signals b green andb yellow). Afterwards theyarebrought intocollision (luminosity signals ATLAS redandcms blue). Oncethecollisions areoptimized theexternal noiseisintroduced inbothbeamsand both planes as shown by the BBQsignals (b cyan and b purple). Table : Feedback voltage and amplitude of the white noise produced in [σ]. Voltage [V] NoiseAmp.[σ] optimization white noise with different amplitudes was introducing using the feedback. The emittances are monitored and recorded during 0 minutes per noise step. The same procedure was used in the three fills only changing the noise amplitude and the damper gain. For the first fill two noise amplitude stepsweredone60and0v(tableandfig.)withthefeedbackgainsetto0turns.sincewerealised that these noise amplitudes were too aggressive we backed off a bit for the next fills.for the second three noise amplitude steps were done, and 8 V (Table and Fig.)withthefeedbackgainsetto0 turns. For the third and last three noise amplitude steps were done, and 8 V(Table and Fig.) withthe feedback gainset to 00 turns. The tune shift due to the beam beam interaction (Fig. ) will be varying during the fill due to the large variation of emittances and intensities (Fig. ). Verygood agreement between expected coherent

5 CERN ACC NOTE Fig. : BBQ signal for the third fill for Beam and horizontal plane. The high intensity beams expe riences a total Q 0.08 which is almost twice larger value for the HL LHC. This is consistent with the expected coherent tuneshift from the high brightness colliding bunch in beam (I=.7 0 ppb and ǫ =.µm) colliding attwoipsand assuming ayokoya factor of Beam I [0 ppb] Time [min from ::00.] Beam I [0 ppb] Time [min from ::00.] Fig. : Measured intensities for the fills.

6 CERN ACC NOTE tune shift and the measured one is found. Fig. shows the coherent tuneshift measured in beam horizontal during the rd fillof ξ bb,high int 0.08 which agrees with the beam parameters of I=.7 0 ppb and ǫ =.µm(figs. and Fig.7)colliding at IPs with a Yokoya factor of. [6]. This tune shift is almost twice the expected ξ bb,hl LHC. MD results DuringthewholeMDtheemittances weremeasured bythetransverse synchrotron light monitors(bsrt) andwerelaterpost processed tobepresented in[%/per hour]. Duringthefirstfill(Fig.)onlytwosteps in amplitude noise were done so the last stretch there was actually no noise applied. In the case of wit ness beams since there is no beam beam interaction all the emittance growth measured is in principle purely duetotheexternal noise andother sources related tosingle beam(intra beam scattering, different working point and the external noise). A first look at the emittance growth rates between colliding and non colliding shows a clear dependency on the beam beam parameter. Bunches colliding show a much fastergrowthrate,howeverthewitnessornon colliding bunchesshowanon negligible emittancegrowth that can be explained only. The growth rate increases for increasing amplitude of the noise. Moreover the thehigh intensity bunches, andconsequently larger beam beam parameter, always present larger growth rates. Inthesecondfillthreesteps ofnoiseamplitude weredone(,and8v)sincewerealisedthat the range used in the first fill was too aggressive leading quickly to very large emittances and degrading the beam beam parameter. Figure6showstheemittancegrowthin[%/perhour]forBeamand. Thedampergainiskeptto 0 turns. Comparing Fig. (lower amplitude noise) to Figure 6(larger amplitude noise) we can observe that the emittance growth rates are reduced for the case with reduced noise amplitude. The growth rate scales with the amplitude of the noise as expected. A more detailed analysis and comparison with the analytical estimates can be found in [7]. Finally, in the third fill we wanted to explore the impact of thetransversefeedbackgain.again three steps of noise amplitude were done (, and 8 V) similar to the second fill. The damper gain was set to 00 turns for this case. Figure 7 shows the emittancegrowthin[%/perhour]forbeamand. Asintheprevious fillsacleardependence betweenthenoiseamplitudeandthebeam beamparameter appears. However from expectations of [] it would reasonable to think that a weaker damper would allow larger emittance growth rates for the same noise level,which iswhat happens inbeam,but not in Beam. As mention in the motivation there is currently an extensive campaign of self consistent simulations that will help us to better understand the interplay between the transverse feedback and the beam beam and noise driven emittance blow up. A detailed comparison to expectations can be found in[7].. Summary Apreliminary study of the effect of external noise on colliding bunches was studied at injection energy in the LHCusing fills amounting for a total of hours. Different noise amplitudes were introduced using the transverse feedback as a noise generator. Theamplitudes wereinthe order of0 σ. Theexperimentaimedtoquantifytheemittancegrowthduetotheinterplaybetweenbeam beam head on tune spread and an external noise. The main goal was tobenchmarkexistinganalytical andnumerical estimations asdone in[7,8] and todefine tolerances fornewerdevices, i.e. the crab cavities for the HLLHC. Inadditiontothenoiseamplitude,thetransversedampergain (0 and 00 turns) and different beam beam parameters (ξ bb =[0.0,0.0]) were scanned to get the dependency on these machine and beam parameters influence since from the analytical estimates they play an important role. Chromaticity waskept at units since the time waslimited and this is another parameter which has very strong impact on the emittance growths in the presence of noise. Simulations show an important dependence onthisparameter that aswellisnot taken into account in Eq.()and seems

7 CERN ACC NOTE st Fill Beam Horizontal. st Fill Beam Vertical = % = 6 % = 9 %.. = 8 % = 690 % = % = 6 % = % = %. = 69 % = 6 % = 8 % Time [min] from 0:: = 7 % = % = 8 % = 8 % = % = 8 % Time [min] from 0::00.97 (a)beam. st Fill Beam Horizontal st Fill Beam Vertical. = 89 % = 69 % s = 0 %. = 669 % = 0 % = 7 %.. = 9 % = % = % = % = 9 % = %. = 6 % = 9 % = 9 %. = 6 % = 9 % = 7 % Time [min] from 0:: Time [min] from 0::00.97 (b)beam. Fig.:Emittance growth in [%/hour]during the first fill.transverse feedback gain set to 0 turns. First stepnoise of 60V,second 0 Vand 0 Vin the last one for comparison (corresponding amplitude inσ in Table ). also very non linear. For this reason we will like to repeat this MD in the future with different values of chromaticity to validate the expectations. Non collidingbuncheswereusedduringthemdforcomparisonandeveninthecaseofnoexternal noise applied a non negligible emittance growth rate of 0%/hour depending on which beam and plane. Beamseemstobemoresensitivetothenoiseintroduced. All studies have been performed at injection energy to save time and have the possibility to explore and scan the parameters since the problem has too many uncertainties to be addressed. A test at top energy could clarify the impact of energy dependent processes. Adetailedanalysisandacomparisonbetweenanalyticalmodel and self consistent simulations and the LHC data collected during this MD has been presented at[7].aclearmismatchbetweenmea surements and expectations have been stressed. It has been showed that the emittance growth measured

8 CERN ACC NOTE nd Fill Beam Horizontal. nd Fill Beam Vertical = 89 % = % = 6 %.. = 9 % = 0 %. = 8 % = 9 %. = %. = 0 % = 8 % = 9 % = 6 %. = 6 % = 6 % = 9 % = 96 % = 6 % = % Time [min] from 06::00.8 (a)beam Time [min] from 06:: nd Fill Beam Horizontal 6 nd Fill Beam Vertical = % = 9 % = 70 % = 8 % = % = 60 % = 8 % = 8 % = % = 0 % = 66 % = 6 % Time [min] from 06::00.8 (b)beam. = 07 % = 7 % = 9 % = 66 % = % = 7 % Time [min] from 06::00.8 Fig. 6: Emittance growth in [%/hour] during the second fill. Transverse feedback gain set to 0 turns. Firststep noise of V,second Vand 8 Vin the last one (corresponding amplitude in σ in Table ). inthe LHCduring the MDis by factors larger than what expectedfromsimulationsandanalyticalmod els. The emittance growth measured in the LHC shows the expected beam beam dependency as in [] but the amplitude is consistent with a factor larger noiseamplitude.thispointstoeitheranoise sourse at injection energy that already acts on the beams and that needs to be understood and reduced. Orcould be due to some interplay with mechanism present at injection. A detailed analysis of the data is still on going to understand how to quantify the effects onemittancegrowthonthesinglebeamthat is normally subtracted from the colliding beam effects. The finding in this MDare worrying since they show a much stronger sensitivity to noise of colliding beams.the conditions during the MDwere also different respect to the normal operation since we were testing the noise effects at injection energy. Acknowledgement Many thanks go to the OP crews, in particular the SPS shifts fortheexcellentbeamsandallthelhc experiments for providing luminosity measurements at injection energy during this experiment. A fun 6

9 CERN ACC NOTE rd Fill Beam Horizontal. rd Fill Beam Vertical = 8 % = 9 % = 8 %... = % = % = 9 %. = 8 % = 06 % = 6 % = 9 % = % = 7 % = 8 % = % = 8 % = 0 % = 7 %.. = 9 % Time [min] from 08:8: Time [min] from 08:8:.7 (a)beam. 6 rd Fill Beam Horizontal 6 rd Fill Beam Vertical = 60 % = % = 7 % = 778 % = 78 % = 7 % = 98 % = 09 % = 99 % = 7 % = 9 % = % = 77 % = 7 % = % = 9 % = 6 % = % Time [min] from 08:8: Time [min] from 08:8:.7 (b)beam. Fig. 7: Emittance growth in [%/hour] during the third fill. Transverse feedback gain set to 00 turns. Firststep noise of V,second Vand 8 Vin the last one (corresponding amplitude in σ in Table ). damental verification will be to check the effect of noise at top energy to reduce the parameter space to be studied. 7

10 CERN ACC NOTE References [] W. Herr et al. Head on beam beam tune shifts with high brightness beams inthelhc",cern ATS Note 0 09 MD. [] Y.I.Alexahin, Onthe Landau damping anddecoherence of transverse dipole oscillations incollid ing beams", Particle Accelerators, Vol. 9, pp. 7. [] V. Lebedev et al., Emittancegrowthdue tonoise anditssuppression withthefeedback system in Large Hadron Colliders", SSCL Preprint 88. [] X. Buffat, Transverse beams stability studies at the Large Hadron Collider", EPFL PhD Thesis, 0. [] J. Barranco et al., NoiseStudies",Beam beamandluminositystudies,june0. [6] K. Yokoya, H. Koiso, Tune shift of coherent beam beam oscillations". Particle Accelerators, 990. Vol. 7, pp [7] J.Barrancoetal., Effectofnoiseoncollidingbeams.LHCexperienceandimplicationtoHL LHC" HL LHCworkshop, June0, CERN,Switzerland. session//contribution//attachments/777/706/hilumi_noise_barranco. pdf [8] JiQiang, UpdatesonSimulationofNoiseEffectsUsingBeamBeamD",Beam beamandluminos ity working group, September 0. 8