1. What we may gain with the sorting at MEB Presented by L. Bottura for the MEB Session 4 - Magnetic Requirements for Commissioning Divonnix, January 2006

2. Outline Our mission statement Sorting in practice: the MB’s Macro-sorting Skimming/sifting the FQ Geometry classes Examples SSS’s Examples The other magnets (DS/MS/IR) Examples Issues Conclusions and perspective

3. Mission statement If all magnets performed to (beam) specifications, we could install any magnet anywhere In reality, we are faced with magnets performing worse, as, or better than (production) specified, available as produced and requested as from installation schedule Although a global sorting is out of the question, delays in the installation have provided an appropriate stock of magnets (e.g. in excess of 400 MB’s) Mission: Find suitable slots for the available magnets that perform better than specified, as specified or out-of-tolerance Preserve and (if possible) optimize the machine performance Include provisions to face day-to-day requirements (faults during processing the magnets) Follow the planned installation schedule with a suitable flow of allocated magnets

4. Slot allocation for MB’s 580 MB’s have been assigned to a slot in the tunnel (nearly 1/2 of the LHC) A stock of ≈ 250 magnets is available for macro- and local- sorting Working mode was drastically modified at the end of 2004 to semi-automatic assignment by batches (see later) to match the demands of transport and installation teams Semi-automatic assignment started as proposed by S. Fartoukh

5. Macro-sorting for MB’s Pre-select batches of 1 sector (154 MB’s + some 10…20 spares) among the available stock (1) that have: The appropriate diode type (R/L) A 50/50 split between corrector packages (A/B) The same inner cable (01B and 01E show slight differences in b 1 at injection and initial ramp) Minimum b 1 and b 3 random (see next slides) An appropriate split among golden/silver/mid-cell geometry (see next slides) Random mixing of Manufacturer (Alstom/Ansaldo/BNN) Outer cable type (02B/02C/02D/02E/02G/02K) NOTE: (1) CM and CR magnets. Magnets with delivery/completion date within few weeks of allocation are also considered for pre-selection Based on proposals from S. Fartoukh and E. Todesco

6. Xs1 Xs2 Xs3 Negative trend of b3 for Ansaldo magnets and Alstom magnets with non-nominal shims Initial production with non-nominal shims Change of cross section Skimming/sifting the FQ - 1/2 Production  b 3  1.9 units (vs. 1.4 units target) Courtesy of E. Todesco

7. Skimming/sifting the FQ - 2/2 Optimized choice can be used to select batches with  b 3  1.0 … 1.6 units Initial production with non- nominal shims and change of cross section Mixing of cross-section 2 and 3 Inner cable 01E

8. mid-cell silver mid-cell silver Geometry classes - 1/3 In one sector: not more than 46 (23+23) MC at least 10 (5+5) G Preferable (to allow sorting): At least 20 (10+10) G Not more than 20 (10+10) MC Beam size Based on a proposal from S. Fartoukh and J.B. Jeanneret

9. Geometry classes - 2/3 Geometry of as-built MB’s More non-silver magnets than allowed A bit less golden magnets than desired Distribution of classes in allocated sectors

10. Silver-right Geometry classes - 3/3 Silver-leftGolden-right Golden-left We can take advantage of the change of beam waist in the cell ! OK ! Classes devised and defined by S. Fartoukh, J.B. Jeanneret and the WGA

11. Example: MB geometry The case of MB1148: Assigned to DS slot (geometry-critical) LBBLQ.8L1 based on anticipated geometry Unique type of interconnect (slot swapping not feasible) Central foot blocked at cryostating (WP02), producing mid-cell geometry Flanges out of tolerance (interconnect issue) Foot at the limit of the adjustment range Solution: installation shift  x = 0.7 mm r-parameter of MB1148 as built r-parameter of MB1148 with installation shift V1V2 V1V2 Courtesy of J.B. Jeanneret

12. Magnetic sorting Local sorting on TF, b3, a2 to: Insure that the CO can be corrected with < 30 % of the corrector strength Minimize the driving terms of 3 rd order resonance Control the driving terms of of coupling resonance and vertical dispersion Method: No more than 3 MB’s with |b 1 | > 10 units in a raw Form self-compensating sequences of MB’s to absorb |b 1 | > 15 units Flip-flop pairing magnets with b 3 above/below the Pairing at  magnets with large or small b3 Flip-flop pairing at 2  magnets with a 2 above/below the Pairing at  /2 magnets with a 2 above/below the Algorithm devised by S. Fartoukh, discussed at FQWG

13. Example: b1 local-sorting b1 distribution in sector 7-8 V1, MCBH strength at 7 TeV and residual CO error Courtesy of S. Fartoukh Gain: MCBH budget necessary for b1 correction limited to +/- 15 % of the available strength XS2 and XS3 magnets XS1 BNN magnets

14. Example: b3 local-sorting XS1 BNN magnets XS3 magnets  -paired flip-flop paired b3 distribution in sector 7-8 Effect of flip-flop pairing Effect of  -pairing Courtesy of S. Fartoukh 3rd order resonance driving terms Gain: effective random b3 and driving terms reduced by a factor 3

15. SSS allocation SSS come in many different types, with reduced sorting possibility Batch selection and qualification is performed in advance to cold test Pairing at  /2 magnets with b 2 above/below the (or pairing at 3  /2, or flip-flop at , 2  issue with D-beating) 110/362 SSS (nearly 1/3 of the main ring) allocated to date Courtesy of M. Modena

16. SSS geometry Specifications devised and defined by J.B. Jeanneret and WGA The available aperture is tighter in the MQ’s, with no difference among cells Specification based on D(H) quadrupole g r SSS58 The present situation requires care (see next slide) to avoid aperture loss at the level of 0.5 to 1 mm

17. Example: SSS geometry The case of SSS95: Assigned to slot Q25R8 BPM support out-of- tolerance by 0.25 mm (cannot be corrected) Field angle 1.6 mrad (i.e. a 2 = 32 units) Solution Installation shift and roll  z=-0.1 mm,  =-0.9 mrad Negligible aperture loss (of the order of 50  m, not critical because the quadrupole is F) V1V2 V1V2 r-parameter of SSS95 as built r-parameter of SSS95 with installation shift and roll Courtesy of E. Wildner, Y. Papaphilippou

18. Example: SSS b2 sorting Courtesy of Y. Papaphilippou b2 distribution in sector 7-8 as from warm measurements Collars with permeability out of specification have large apparent deviation from the production average Gain: total beta-beating kept well within (factor 2 to 3) the allocated budget Total  -beating (2 planes, 2 apertures)

19. The other magnets About 200 magnets: DS/MS (Q4…Q11) and IR (Q1…Q3, D1…D4) Correction dipoles for IP8,IP2 Discussed one-by-one, based on the specific requirements of the proposed slot (e.g. SSS607 in Q5L8) Allocated 6/114 DS/MS quadrupoles (< 10 %) 4/4 cold D1 6/8 D2 (the remaining 2 are preallocated) 5/24 IR quadrupoles (Q1/Q2 of IR8 R+L and Q3 or IR8 L) 3/6 warm compensation dipoles (IP8 spectrometer) In addition MQW pre-sorted based on b2 and geometry Maximum operating Current: 3453 A SSS607 training curve

20. Example: D1 geometry The case of D1 at right of IP8 (D1L105) Pre-assignment based on field quality and geometry inferred from measurements taken on the cold mass skin Large deviations from straightness found in the cold bore  x=1.7 mm,  z=2.7 mm Critical n1 = 5.7  at collision vs. 7  target (with  *=1 m) Solution: Installation shift  x=-0.6 mm Marginal n1 = 6.3  at collision (with  *=1 m, but this is an extreme case not used) D1L105 horizontal geometry D1L105 vertical geometry Courtesy of M. Giovannozzi

21. Issues Replacement of magnets at installation Risk: we may lose the advantages of sorting MB batch selection, cold test planning and fiducialisation Risk: reduced flexibility as the production ends and the “sorting buffer” is depleted SSS installation vs. production Work: meeting the installation needs requires swift action (days) Quads in the DS and MS Work: documentation, automation, organization, as for MB’s and SSS’s IR magnets, most critical elements in the machine at collision Work: qualify cold D3/D4, Q1/Q2/Q3 Warm magnets Work: document, qualify, sort and assign to slot Assist the coordination work through anticipation Cold test planning Pre-assignment of MQ’s in the SSS Quench level in MQTL correctors

22. Results and perspective - 1/2 So far we met our goals, and, when possible, we did better… Maintaining the magnetic properties under control (using sorting and compensation on field quality) Preserving the mechanical aperture (using sorting on geometric classes and installation shifts/rolls) Negligible aperture loss in MQ’s, 0.1 mm (D) to 0.2 mm (F) MB’s in the shadow of MQ’s Optimizing the installation sequence to gain margin (limiting the corrector strength, resonance driving terms)

23. Results and perspective - 2/2 … but we are only half way (at most) IR, MS and DS are in front of us (and there is work to be done to specify aperture targets and qualify magnets) Changes of transport/installation scenarios and needs result in pressure on magnet delivery, we are in the middle of this process. The situation will escalate during 2006

24. Acknowledgements MembersAlternate Members Luca Bottura (Chairman) Massimo Giovannozzi (Scientific secretary, ABP – IR magnets (cold and warm), but low-beta quadrupoles) Stephane Fartoukh (ABP - Magnet evaluation activity leader, ABP - MB)Massimo Giovannozzi Yannis Papaphilippou (ABP - SSS)Jean Bernard Jeanneret Frank Schmidt (ABP – Low-beta quadrupoles )Stephane Fartoukh Jean Bernard Jeanneret (Aperture) Davide Tommasini (MB)Jose Carlo Pereira Lopes Michele Modena (SSS)Theodor Tortschanoff Nuria Catalan Lasheras (MS and DS SS)Ranko Ostojic Ranko Ostojic (IR)Karl-Hubert Mess Suitbert Ramberger (Resistive Magnets)Willi Kalbreier Stephane Sanfilippo (Field Quality) Elena Wildner (Geometry)Walter Scandale Andrzej Siemko (Quench and Protection)Pierre Pugnat Karl-Hubert Mess (Electrical Engineering)Ranko Ostojic Dominique Missiaen (Survey)Patrick Winkes coordinators experts ABP