Jorge GINER NAVARRO CLIC Workshop 2015 26012015 26012015 J Giner Navarro CLIC WS2015 1 Overview Introduction Conditioning data from test stands Magnitudes to describe conditioning status ID: 1044652
Download Presentation The PPT/PDF document "Statistical analysis of RF conditioning ..." 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.
1. Statistical analysis of RF conditioning and breakdownsJorge GINER NAVARROCLIC Workshop 201526/01/201526/01/2015J. Giner Navarro - CLIC WS20151
2. OverviewIntroductionConditioning data from test standsMagnitudes to describe conditioning statusComparison of different structure conditioningsConclusions26/01/2015J. Giner Navarro - CLIC WS20152
3. IntroductionPerformance of CLIC accelerating structures has been tested in klystron-powered test stands at KEK, SLAC and CERN.New prototypes of accelerating structures need to be conditioned to achieve the CLIC main requirements of gradient 100 MV/m at a pulse length of around 200 ns and a low breakdown rate (BDR) of 10-7 bpp/m.RF conditioning process needs to be understood in order to minimize time and costs.26/01/2015J. Giner Navarro - CLIC WS20153
4. Xbox-1: TD26CC#1 conditioning historyAutomatic operation by a conditioning algorithm [see J.Tagg presentation]26/01/2015J. Giner Navarro - CLIC WS2015411168 BDsB. Woolley - CLIC WS2014
5. Rescaled gradient26/01/2015J. Giner Navarro - CLIC WS20155BDR= 7e-5 bpp2e-5 bpp2e-6 bppCONDITIONINGBDR measureTD26CC#1 raw dataEquivalent gradient curve with constant pulse length of 250 ns since the beginning Equivalent gradient curve with constant pulse length of 250 ns since the beginning and constant BDR of 2e-5 bpp
6. Describing conditioning statusAnalogously, we can define a “scaled BDR” :which gives the BDR evolution at a fixed gradient and pulse width. 26/01/2015J. Giner Navarro - CLIC WS20156We can define a “scaled Gradient” to describe the conditioning level of the structure:which gives the Gradient curve keeping constant pulse width and BDR. Using the scaling law: [ A.Grudiev et al, Phys. Rev. ST AB 12, 102001,1 (2009) ] These magnitudes gives us the conditioning status of the structure according to our requirements.
7. NEXTEF (KEK): TD24R05#4 test historyConditioning and fixed-gradient tests carried out in NEXTEF test stand for the TD24R05 structure provides a source of comparison with our data. Data courtesy of T. Higo.26/01/2015J. Giner Navarro - CLIC WS20157Normalized gradient hereScaled gradient TD24R05_#4
8. Comparison of conditioning evolution26/01/2015J. Giner Navarro - CLIC WS20158Scaled gradient vs cumulative number of PULSESScaled gradient vs cumulative number of BREAKDOWNSConditioning to high-gradient is given by the pulses not the breakdowns! #Pulses#BDs
9. Further studiesAccording to these results, pulsing at constant gradient would slowly decrease the breakdown rate in the structure, which means that the surface is well influenced by the RF high-powered pulses.Study of long term trends, whether there is an asymptotic BDR or not, and the time needed to complete the conditioning is hard to determine. High-repetition rate systems are more efficient in this study.[see A. Korsback presentation]26/01/2015J. Giner Navarro - CLIC WS20159
10. HRR Fixed-Gap system data analysis26/01/2015J. Giner Navarro - CLIC WS201510In the Fixed-gap system, at DC Spark lab (CERN), high electric fields are reproduced between two Cu electrodes, pulsing at a repetition rate up to 1 kHz. Analogous studies to RF accelerating structure tests can be driven in less time.Here we compare its conditioning evolution in terms of surface electric field.Data courtesy of N.Shipman
11. RF Breakdown statistics26/01/2015J. Giner Navarro - CLIC WS201511118.8e-6 bpp6.5e-6 bpp2.0e-6 bpp2.1e-3 bpp2.3e-3 bpp3.1e-4 bpp1.1e-4 bppAnalysis in Breakdown statistics shows different regimes of the BDR.[See Anders Korsback presentation for full analysis in a DC system]
12. ConclusionsWorking on data analysis from test stands provides a better understanding about the conditioning process, the goal of which is the feasibility and the proper performance of the accelerating structure in the linear collider.Scaling laws are used to compare different structure conditionings and the same trends are found with the number of pulses, but not with the number of breakdowns.Different models (dislocations, local tips…) are being studied to describe the effect that the wall’s surface resists more power with increasing pulses.High repetition rate systems would provide valuable information in this study. The Fixed-Gap system in the DC Spark lab is running to acquire new fresh data.Results from this study lead to more strategies to carry out during the conditioning of the RF structure. Optimization in time (and cost) will be essential when producing new structures.Thank you for your attention!26/01/2015J. Giner Navarro - CLIC WS201512Acknowledgement to A. Degiovanni and W. Wuensch for their contribution in this work!
13. EXTRA SLIDES26/01/2015J. Giner Navarro - CLIC WS201513
14. Pulse and BD dependence26/01/2015J. Giner Navarro - CLIC WS201514A. Degiovanni
15. Normalized BDR in LOG-LOG scale 15
16. Normalized BDR in LOG-LOG scale – Linear fitA’ = -8.35 +/- 0.02A’ = -7.74 +/- 0.02 16
17. Pivot modelBDR = 1 (limit for operation by definition)Emax (assumed limit for gradient)The exponent X increases with the number of pulses (n)log(BDR)=X(n)*log(E0)Emeas = α Emax17 70 80 90 100 110 120 130 140 155A.D 02.04.2014
18. Pivot model fit results (TD26R05CC)27/05/201418Pivot model: X0=8.0q=6.2e-8 pulse-1E*max=Emax tp1/6 =396 MV/m ns1/6 50ns100ns150ns200ns250nsX0=8.0X=14.7X=18.6X=20.7X=24.0X=26.6