2. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Vaqtec feedthoughs with PEEK base 2 Air side Vacuum side 3 prototypes mounted on DN40 CF flanges delivered to Frascati 04/12

3. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 HV test program for new feedthroughs 3 Measurement of leakage current at 4 kV, pin-to-pin and pin-to-ground Performed December 2015 Endurance testing of 2 feedthroughs: all pins at 2kV for ~1 month In progress Not much time to validate solutions: Test as best we can in available time Additional considerations: Accelerate aging by application of thermal cycles? Concerns about hygroscopicity? PEEK hygroscopicity very low but not zero (0.1% water absorption) Difficult to measure – need control over relative humidity

4. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 HV leakage tests at 4 kV 4 Performed in Frascati in December by G. Corradi, D. Tagnani, S. Ceravolo HV on 1 central pin and 1 outer pin Pins in 2nd ring connected to nanoamperometer, current measured to ground Connector body grounded Pins in outer ring tested individually Typical leakage current = 20 pA Test conditions Test voltage 4kV Temperature 20-25 C Relative humidity 20%

5. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 HV leakage tests vs. humidity 5 Bell jar with gradual evaporation of water from dish, hygrometer to measure RH Female cable plug inserted Tested at 4 kV over several few days Kept at 2 kV during Christmas closure

6. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Long term HV tests 6 Attempt first to reproduce failures at Frascati: LAV04 and LAV08 flanges from experiment 3 HV boards and cables = can test 3 feedthroughs Empty vacuum side connectors attached for safety Flanges grounded to HV chassis

7. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Long term HV tests 7 After failures reproduced, compare performance of new prototypes: Prototypes #1 and #2 mounted on 40mm CF flanges, grounded Compare time-to-failure with LAV4.05 feedthrough (never used in NA62)

8. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Long term HV tests: Summary 8 ConnectorHistoryTest conditionResult LAV4.02+03+04Multiple fails in 2015100 hrs at 1500V 3 connectors in series 13 ch tripped (10 uA) 9 ch drawing (1 uA) Failure reproduced LAV8.01Multiple fails in 2015120 hrs at 1500V7 ch tripped (10 uA) 6 ch drawing (1 uA) Failure reproduced LAV8.00No fails in 2015160 hrs at 1500V1 ch tripped (10 uA) Partly reproduced LAV4.05Never used280 hrs at 1500VNo problems observed Prototype #2New design (PEEK)280 hrs at 1500VNo problems observed Prototype #1New design (PEEK)235 hrs at 1500VNo problems observed Feedthroughs that failed during run also fail in lab tests Tests of new feedthroughs not yet long enough to cause failure of control case (an original feedthrough that was never used)

9. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Summary of vacuum tests 9 Prototype #1Originally supplied with visible defect in weld to flange Used for electrical tests and repaired after Christmas closure After repair, tested at Vaqtec: Cycled between 0C and 100C before testing Vaqtec leak measurement: 1 × 10 −10 mbar l/s Returned to Frascati to confirm test results Frascati leak measurement: 1 × 10 −12 mbar l/s Prototype #2Satisfactory test at Frascati on delivery in December First leak test: 3 × 10 −9 mbar l/s Unchanged after 1 cycle to 70C (4 hrs) After 2 cycles to 80C leak rate 3 × 10 −8 mbar l/s Prototype #3Supplied in December with suspected defect in weld to flange Repaired at Vaqtec and tested to 1 × 10 −10 mbar l/s Test confirmed at Vacqtec/Ecovide after 2 cycles to 75C Tested at Frascati mid-January: major leak inside feedthough body Returned to Vaqtec again for examination Problem discovered with weld between parts of feedthrough PEEK assembly tested with vacuum putty to 4 × 10 −10

10. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Diagnosis of prototype #3 10 Connector disassembled and problem spotted on weld between connector halves Weld was probably critical and failed after thermal cycles PEEK capsule cannot be re-tested in a new assembly: knife-edge seal We think the problem is understood, but need careful QC for welds during production

11. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Mechanical design for new HV flange 11 Feedthroughs individally mounted on custom DN40 flanges Defective connectors can be replaced without welding Mechanical design nearly final, parts ordered Cost ~€15k/11 flanges

12. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Mechanical design for new HV flange 12 Feedthrough flanges slightly smaller than standard CF, with O-ring seal Flange thickness optimized for access to bayonet from both sides No centering ring for large O-ring Helps to recover space for HV cables

13. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Cables and plug connectors 13 HV cables removed and brought to SPIE 30/11-04/12 SPIE set up test equipment in December Automated tests for continuity, channel map, HV leakage Plug connectors delivered by RS 2 nd week of January Connector replacement completed by SPIE by 01/02 We now have better cable diagnostics than ever before: 7 inversions found on Radiall side – can be fixed by inversion on MIL side 2 cables with 1 pin missed and translation error propagated to remaining pins – can be fixed on MIL side with some effort 2 cables with continuity problems on single channels traced to Radiall side – requires replacement of Radiall connector, deferred Test procedure verified using our equipment Monday 08/02 3 sample cables tested 1 continuity problem found by SPIE found also by me 1 cable found to be good by SPIE found to be good by me 1 continuity problem found by SPIE not found by me, but not found by SPIE when re-tested – intermittent problem

14. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Schedule 14 04/10 (done)Uncable LAV HV, bring cables to SPIE 05/01 (done 08/01)Plug connectors delivered, SPIE starts work on cables 08/01 (done 29/01) Last date to order flanges for feedthrough delivery 01/03 Delayed by one month, will try to recover 01/02 (done?) New HV cables ready Last few inversions and corrections to do this week 15-19/02Frascati technicians re-install install HV cables 01/03 Feedthroughs ready at Vaqtec, installation in flanges Will probably slip a week or two 15/03Flanges ready for installation and testing 01/04Begin to evacuate blue tube. All work completed Largely on schedule, but have probably incurred a small delay (~ 1 week)

15. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Addendum: Spare LAV-FEE boards 15 Comprehensive survey in May 2015 5 problem boards + 3 usable (at least half) Situation evolved during run – need new survey Currently 5 uninstalled boards, with good diagnostics 2 have problems with individual channels – substitute ToT mezzanines? 3 with major problems – to be used for spares Some pieces may have already been used

16. LAV HV repair status – M. Moulson (Frascati) – Photon Veto WG – CERN – 09 February 2016 Addendum: Primary HV power supplies 16 Meeting with G. Di Maio, CAEN, 16 December ROAL does not like the MOSFET used to ruggedize the PS for LAV2 2 alternatives suggested, 1 similar to the MOSFET used for LAV2 Decide to modify all PSs Decide to upgrade 50% of PSs with 1 of the 2 suggested MOSFETs, 50% with the other Upgrade work is being performed by ROAL ROAL modules should be available 22-26 Feb CAEN personnel will install ROAL modules in A4531s at CERN Work will be done in ECN3 Updates from Riccardo