Local quality control of LHC electrical interconnections during the - PowerPoint PPT Presentation

Local quality control of LHC electrical interconnections during the 2012 shutdown 1 LHC Splice Review, C. Scheuerlein, 18.10.2010

Outline LHC splice quality control (QC) sequence during 2012 o shutdown QC of existing main interconnection splices • QC of disconnected main busbar cables before splice assembly • QC of newly connected main interconnection splices (without • shunts) QC of shunt solder contacts; feasibility studies and development • of a new QC test QC of entire splices with shunts • QC of line N connections • QC of line M connections • QC of splice insulation • Other quality assurance (QA) activities o C. Scheuerlein, 18.10.2010 2

QC of existing (“old”) main interconnection splices In the LHC two families of main interconnection splices o can be distinguished: About 9500 “old” splices produced during installation: • Most “old” splices have gaps and/or large steps between the Cu profiles, and are not conform with visual acceptance criteria (see IEG-C-BR- 001 rev C “Contrôle Strongly distorted connection visuel des brasures”). (QBQI.25R5-M1-QRL-lyra) About 500 “new” splices produced during 2009 shutdown, • which are conform with IEG-C-BR-001 acceptance criteria. Three possible consolidation strategies: o 1) Open and redo almost all “old” splices before adding shunts. 2) Do not open any “old” splices and apply shunts as is. Gap between U-piece and 3) Repair the worst cases before applying shunts. busbar tongue (QBQI.33L2- M3-QRL) In order to avoid opening all “old” splices, it is planned o to repair only the worst cases. Therefore, the geometrical, as well as the electrical resistance acceptance criteria are less strict for “old” splices. Step between U-piece and C. Scheuerlein, 18.10.2010 3 busbar (QQBI.32R6-M2)

The R-8/R-16 test Whenever possible QC is based on nondestructive test results obtained o in-situ in the LHC. QC of main interconnection splices is based on room temperature (RT) resistance measurements (first suggested in 2009 by Howie Pfeffer and Bob Flora from Fermilab). The so-called R-8/R-16 test [1] has been introduced during the 2009 o shutdown in order to detect splices with a high excess resistance in the normal conducting state. This was a breakthrough in the QC of the main interconnections splices. In case of a complete transverse gap o between Cu splice profiles and bus bar stabiliser, the R-8/R-16 result is nearly proportional to the defect size (insulated cable length inside the busbar stabiliser). The estimated R-8/R-16 random error is o ±1 µ Ω , and the systematic error is about +10 %. R-16 measurement with test splices in SM18 [1] “ Production and Quality Assurance of Main Busbar Interconnection Splices during the LHC 2008-2009 C. Scheuerlein, 18.10.2010 4 Shutdown ”, IEEE Trans. Appl. Supercond., accepted.

Distribution of additional R-8 values measured for “old” splices produced before 2009 The (biased) distribution of R-8 excess resistance values measured in o 2009 for “old” splices (produced before 2009) may be helpful for determining an R- 8 acceptance threshold value for “old” splices. The excess resistance is defined as the measured resistance minus an o average resistance value for good splices (5.6 µ Ω and 9.3 µ Ω for dipole and quadrupole splices, respectively). An R- 8 acceptance threshold value for “old” splices could be +5 µΩ o excess resistance (redo “old” splices when R-8>10.6 µ Ω and R-8>14.3 µ Ω for dipole and quadrupole splices, respectively). C. Scheuerlein, 18.10.2010 5

Alternative nondestructive tests: US transmission test US transducers mounted to M3 splice Before the introduction of the R-8/R-16 o test, the US transmission test [2] was the most important QC tool in order to detect splices in which no solder foil had been added and/or for which the solder melting temperature had not been reached. Both defect cases will also cause high R-8 o excess values. Therefore, during the 2012 Standard US test result (result “OK”) shutdown the US test is not needed anymore, and it will not be used as a routine QC test [2] F . Caspers, T. Kroyer, J. Kulka, J.- P . Tock, L.R. Williams, “ Ultrasound C. Scheuerlein, 18.10.2010 Additional US test between U-piece 6 diagnostics of the superconducting cable connections between the main and busbar nose (result “not OK”) ring magnets of LHC ”, LHC -Project-Report-1032, 2007.

Alternative nondestructive tests: Gamma ray imaging Comparatively o large volume defects are detectable in the radiographs: Lack of SnAg alloy o Lack of SnAg in busbar ends. in the busbar extremities. Un-molten SnAg o foil at the splice extremities. Gamma ray o imaging is an important tool for trouble shooting. Unmolten SnAg foil (additional R-8=31 µΩ ). Radiographs courtesy J.-M. Dalin. 7

Gamma radiography as a QC tool? The spatial resolution of gamma radiography (for the LHC splices in the order of o 0.1 mm ) is insufficient to exclude the presence of defects. Gamma radiography is NOT a QC tool. During the 2012 shutdown gamma radiography will only be used occasionally, for o instance in order to study if through improved splice heating equipment the loss of SnAg from the busbar ends can be reduced. QBBI-B16L7-M1-COR-cC QBBI-B16L7-M1-COR-cL QQBI-16L7-M1-COR-cL QQBI-16L7-M1-COR-cC 8

QC of main busbar cables before splice assembly At RT it is impossible to control the o busbar cables and the cable to cable contacts inside a finished splice. Therefore, it is important to perform a Disconnection of cable SSS-221-up-M1 a visual control before the splices are assembled. Opening splices is a delicate task with o some associated risks. Every un-soldering/soldering cycle invariably causes a cable degradation. If the splice disconnection is performed as planned, the degradation is small and acceptable [3]. Disconnected cables are much more vulnerable than a busbar or a o complete splice. We expect a variety of cable defects to show up when a large number of o splices will be opened. In most cases it will not be clear if a defect has been introduced during the o first interconnection, or if it occurred during or after splice disconnection (previous 1.9 K QC result is lost). [3] “ Temperature induced degradation of Nb-Ti/Cu composite superconductors ”, 2010 J. Phys.: Conf. Ser. 234 022031 C. Scheuerlein, 18.10.2010 9

Examples of damaged cables found in 2009 by the personnel performing the interconnection work NCR 1002739: Entire cable squeezed between U-piece and wedge NCR 990852, 2 strands cut NCR 992513: Cable overheated (>580 °C) NCR 1002739: Entire cable was squeezed Half of the strands of one M3 cable of 10 between U-piece and wedge MB2446 were found cut in SMI2.

QC of newly produced main interconnection splices (before application of shunts) Acceptance criteria for splices that will be o produced during the 2012 shutdown are more strict than for the existing “old” splices. Visual inspection according to standard o QBBI.A27R3-M3-corridor Gap between U-piece and busbar pocedure: IEG-C-BR- 001 rev C “Contrôle stabiliser (QBBI.A27R3-M3- visuel des brasures” corridor “new” before repair) No macroscopic (visible) gaps • No steps between the Cu profiles >1 mm • QC is mainly based on in-situ R-8 results o (measurement of R-16 is done as a cross check of R-8 results). Gap between U-piece and busbar tongue (QBQI.33L2-M3-QRL) Step between U-piece and busbar C. Scheuerlein, 18.10.2010 11 (QQBI.32R6-M2)

Distribution of additional R-8 values measured for “new” splices produced in 2009 The average values for “new” dipole (M3) o and quadrupole (M1 and M2) splices are 5.6 µΩ and 9.3 µΩ, respectively. Based on the R- 8 distribution of “new” o splices, additional resistance acceptance threshold values of 2 µΩ and 3 µΩ are suggested for dipole and quadrupole splices, respectively. Redo “new” splices if: o R-8 dipole >7.6 µΩ and R-8 quad >12.3 µΩ . C. Scheuerlein, 18.10.2010 12