Proposition of MKD & MKBH modification with the goal to reduce maximum voltage (and risk of erratic) Viliam Senaj, 05/05/2017

Outline Motivation Proposed solutions Benefits and drawbacks of the proposed solutions for MKD and MKBH R2E related consequences Cost estimation Conclusion

Motivation Risk of MKD and MKBH erratic increases for energies above 6.5 TeV (26.8 kV for MKD, 24.7 kV for MKBH) Erratic in our case results from a sudden discharge of a charge (electrons) accumulated on a surface of HV insulator of GTO stack into a HV conductor Nowadays we try to reduce risk of insulator surface charging by avoiding the local Efield non-homogeneities (dust, hairs, fibres … ) and electron emitters (metal particles, stripes…): cleaning of generator and especially GTO stack insulator A possible mitigation would be applying of high resistivity coating on an insulator surface to avoid charge accumulation – development is ongoing Recently developed mitigation measure – new GTO stack geometry with reduced E- field on insulator surfaces (1.5 MV/m = ½ of ionisation limit in air) will still require clean environment to avoid pollution by electron emitting particles (metal dust and stripes found in past); Internal source of dust was recently identified (including metallic particles) – Ross Relay: friction and spring vibrating in its coil stray magnetic field (50 Hz); increased inductance of stack with effect to T_rise New geometry GTO stack will not be ready for 7.x TeV operation (due to R2E)

Proposed modification to MKD Replacement of the main capacitor by ~ 25 % higher value and such reducing max voltage by ~ 12% for the same peak current Significantly reduced sparking and SEB risk at 7 TeV 7.5 TeV ready solution; no need to add 11th GTO (stack and TT modifications)…

Modification to MKBH Replacement of the main capacitor (9.5 uF) by ~ 25 % higher value and such reducing max voltage by ~ 12% for the same peak current Significantly reduced sparking and SEB risk at 7 TeV 7.5 TeV ready solution; No need to replace GTO by ABB ones (R2E)

Consequences to MKD 7 TeV voltage reduced from 28.7 kV to 25.6 kV (~ 6.2 TeV today) Increased T_rise – partially compensated by faster GTO commutation with new TT - 2.83 µs instead of today 2.63 µs (based on preliminary Cu TT measurement) Expected minimum abort gap duration ~ 2.9 µs (todays real need ~ 2.8 µs) AG prolongation by 100 – 200 ns to keep the same margin as today would be necessary

Consequences to MKBH 7 TeV voltage reduced from 26.6 kV to 23.1 kV (~ 6.2 TeV today) Waveform delay by ~ 1.3 us Reduced R_max2 (increased damping) from ~ 80% to ~ 70% (potentially favourable to avoid overlapping in case of modified MKB re-triggering strategy)

R2E related consequences Today 6.5 TeV PTM = IXGN100N170 7 TeV Modified capacitor PTM = IXGN200N170 7.5 TeV 2.68 kV/GTO (MKD) 2.47 kV/GTO (MKBH) 1.17 kV/IGBT (MKD+MKB) 2.87 kV/GTO (MKD); 2.66 kV/GTO (MKBH) 2.56 kV/GTO (MKD); 2.31 kV/GTO (MKBH) 1.14 kV/IGBT (MKD+MKB) 2.75 kV/GTO (MKD); 2.48kV/GTO (MKBH) ABB SEBc-s [cm2] 2e-10 5e-9 ~2e-11 4e-10 Dynex SEBc-s [cm2] ~3e-8 1e-7 ~5e-9 PTM IGBT IXGN100N170 SEBc-s [cm2] 1e-9 HEH fluence estimation [HEH.cm-2.y] ~5e4 Failure probability MKD (600 GTO) [y-1] 6e-3 1.5e-1 6e-4 1.2e-2 MKD (360 IGBT) [y-1] 9e-2 2e-2 MKBH (80 GTO) [y-1] 1.2e-1 4e-1 MKB (120 IGBT) [y-1] 3e-2 Total AD (MKD GTO + IGBT) [y-1] 0.1 0.24 0.03 0.04 Total SD (MKB GTO + IGBT) [y-1] 0.15 0.43 0.13

Cost estimation - MKD TT replacement and PTM modification - already part of consolidation program Cu TT more demanding than Al; additional cost ~50 kChf Testing of Cu TT same as Al; assembly more demanding ~30 men weeks 90 capacitors (35 generators x2 + 20 spares) x ~ 1500 Chf = ~135 kChf Modifications of the GTO stack architecture (only main insulator and top flange - less expensive) – already part of consolidation program ~- 50 kChf Manpower (capacitor replacement; MKD’s will have to pass LS2 maintenance anyway - including modification of one return bar of capacitor planned since a while – bars are ready) <20 men weeks Potential 7.x TeV op - no need for additional modifications (11 GTO stacks; new TT with 1 more secondary, new principal capacitors… ~400 kChf) Short term Total ~ 135 kChf and ~50 men weeks Long term Total savings of ~-400k and 100s of men weeks

Cost estimation - MKBH PTM modification - already part of consolidation program 14 capacitors (10 generators + 6 spares) x ~ 2.5 kChf = ~40 kChf Manpower (capacitor replacement) ~2 men weeks Future savings: no need for GTO replacement (by ABB) for operation at 7.x TeV: 150 GTO (10 generators + 5 spare stacks) = ~-160 kChf In case of > 7TeV operation request main capacitors would most likely need to be replaced anyway

Conclusion Reduction of the maximum voltage of MKD and MKBH generators by increasing of the principal capacitors value is cost effective way of reliability improvement from several points of view: Significant reduction of erratic triggering risk at 6.5 TeV and above without sacrifying the kick strength Significant reduction of radiation related failure risks at 7 TeV and higher Allows potential future cost saving by avoiding the need for future modifications of MKD and MKBH for potential 7.x TeV operation request Due to increased rise time of MKD current - moderate prolongation of abort gap duration would be necessary (100 - 200 ns to keep the same margin as today) Preliminary results on 2 prototypes of MKD trigger transformer shows significantly better performance of prototype No.1 – (copper made; different topology); it is however, more demanding on production and assembly To be decided if the switching speed gain (~100ns) justifies higher cost (~50k?) and higher assembly manpower (outsourcing as well?) request for Cu TT

MKD capacitors geometry

Influence of trigger transformer to T_del

Influence of trigger transformer to T_rise

Influence of trigger transformer to T_thresh