Malte Hildebrandt

50 pages
19 views

Please download to get full document.

View again

of 50
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
Drift Chamber System • hardware status Run 2008 • shutdown activities 2009. MEG Review Meeting Feb 2009. Malte Hildebrandt. Outline. • installation 2008 • MEG Run 2008: • characteristics of HV instabilities • tests with dc system in MEG
Transcript
Drift Chamber System • hardware status Run 2008 • shutdown activities 2009 MEG Review MeetingFeb 2009Malte HildebrandtMalte Hildebrandt MEG Review Meeting, 18.02.2009Outline• installation 2008• MEG Run 2008: • characteristics of HV instabilities • tests with dc system in MEG • tests in laboratory• Shutdown • preparations, tests • first test results • Summary / OutlookMalte Hildebrandt MEG Review Meeting, 18.02.2009Installation 2008• main changes/improvements in 2008 compared to installation in 2007: → topics were already noted in last years DC Status Talk at the Review Meeting as „needs to be done“ / „ will be improved“) • improved strain-relief of LV and signal cables at inside patch panel → no disconnected LV cable 2007: LV dc6u disconnected → 13 missing signal channels 2007: 42 missing connections ↔ crucial step during installation: closing end-cap • more detailed analysis of optical survey of dc system → geometrical alignment includes slope of wires along z-axis 2007: wire at constant x and y, parallel to z-axis (UCI group startet to implement slope) Malte Hildebrandt MEG Review Meeting, 18.02.2009Installation 2008• main changes/improvements in 2008 compared to installation in 2007: → topics were already noted in last years DC Status Talk at the Review Meeting as „needs to be done“ / „ will be improved“) • optimised target → new support spacers and 2007: target slightly misaligned new attachment screw-plate to correct position in space → slant angle adjusted to (20.5 ± 0.3)° 2007: (12.8 ± 0.5)°Malte Hildebrandt MEG Review Meeting, 18.02.2009Target20082007Malte Hildebrandt MEG Review Meeting, 18.02.2009Installation 2008• main changes/improvements in 2008 compared to installation in 2007: → topics were already noted in last years DC Status Talk at the Review Meeting as „needs to be done“ / „ will be improved“) • optimised target → new support spacers and 2007: target slightly misaligned new attachment screw-plate to correct position in space → slant angle adjusted to (20.5 ± 0.3)° 2007: (12.8 ± 0.5)° → identification marks on targetMalte Hildebrandt MEG Review Meeting, 18.02.2009Installation 2008• main changes/improvements in 2008 compared to installation in 2007: → topics were already noted in last years DC Status Talk at the Review Meeting as „needs to be done“ / „ will be improved“) • optimised target → new support spacers and 2007: target slightly misaligned new attachment screw-plate to correct position in space → slant angle adjusted to (20.5 ± 0.3)° 2007: (12.8 ± 0.5)° → identification marks on target → measurements of inclination: • conventional (sliding rule) 2007(2008): conventional • photogrammetric 2007: photogrammetric • optical survey Malte Hildebrandt MEG Review Meeting, 18.02.2009New DC HV Modules↔ S.Ritt↔ R.Schmidt• during Run 2008: modification of dc HV modules• observation (in 2007, June - July 2008): • several times “blocking” / “freezing” of complete communication: MSCB Submaster, HV nodes, LabView • in combination with HV trips of several HV modulesbut: cause and effect not clear…→ new HV modules: • watchdog: reboot of mController in case of missing toggle signal from mController (→ HV off) • new bus driver: „state machine“ decouples from the bus when not transmitting data → no blocking of bus line with „active high signal“ in case mController stucks • smaller capacitance at HV_output reduces trip propagation within HV module • geographical addressing of nodes within cratesMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Performance• summary: • many dc planes / modules suffered from frequent HV trips• consequently theses planes / modules could only be operated with reduced HV settings → huge impact on overall performance of dc system(→ talk by B.Molzon)Malte Hildebrandt MEG Review Meeting, 18.02.2009HV Performance• summary: • many dc planes / modules suffered from frequent HV trips• consequently theses planes / modules could only be operated with reduced HV settings → huge impact on overall performance of dc system(→ talk by B.Molzon)• phase 1: June – July • 30 / 32 planes >1800 V(remark: nominal 1850 V)m beam • 2 planes showed problems right from beginning• phase 2: Aug • dc system most of the time off, short period onp beam • 19 / 32 planes >1800 V • 6 / 32 planes 1700-1800 V • phase 3 : Sep – Dec • further deterioration of HV performancem beam • 11 / 32 planes >1800 V • 7 / 32 planes 1700-1800 V Malte Hildebrandt MEG Review Meeting, 18.02.2009Phase 1• phase 1: June – July • m beam, start-up phase, e.g. HV ramping • 2 dc planes showed problems right from beginning • dc10A: always <1300-1500 V • dc03A: once 1850 V, then setback down to 1300 V, recovery within weeks up to 1700 V • all other planes: air admixture to COBRA gas (outside module) necessary to achieve stable dc operation → cHelium ≈ 95-96 % (reading O2 sensors) (instead of „pure“ helium level around 99.0-99.5 %)Malte Hildebrandt MEG Review Meeting, 18.02.2009Air AdmixtureSat, 19.07., 02:0032ccm / 2000ccm(1.60%)cHelium=99%~96-97 %~95-96 %~95-96 %~94-95 %cHelium=95%cHe= ~93 %Wed, 16.07., 12:4024ccm / 2000ccm(1.20 %)Mon, 30.06., 08:1530ccm / 2000ccm(1.5 %)Tue, 01.07., 08:4555ccm / 2000ccm(2.75 %)Sun, 06.07., 23:0040ccm / 2000ccm(2.00 %)Sat, 12.07., 15:0032ccm / 2000ccm(1.60 %)Mon, 30.06., 17:4565ccm / 2000ccm(3.25 %) Malte Hildebrandt MEG Review Meeting, 18.02.2009Phase 1• phase 1: June – July • m beam, start-up phase, e.g. HV ramping • 2 dc planes showed problems right from beginning • dc10A: always <1300-1500 V • dc03A: once 1850 V, then setback down to 1300 V, recovery within weeks up to 1700 V • all other planes: air admixture to COBRA gas (outside module) necessary to achieve stable dc operation → cHelium ≈ 95-96 % (reading O2 sensors) (instead of „pure“ helium level around 99.0-99.5 %) → end of July: 30 / 32 planes operational with >1800V (remark: nominal 1850 V) • occasional problems with communication of HV modules / MSCB / LabView → modify HV modules during XEC runMalte Hildebrandt MEG Review Meeting, 18.02.2009Phase 2• phase 2: August • p beam, XEC run, Dalitz run • XEC run: • dc gas system running continuously • dc HV system off for 2½ weeks (modifications of HV modules) • 10 days dc HV at 800-1000 V • Dalitz run (e+ e- identification in Dalitz decay p0 → e+ e-g) • 5 days dc HV on nominal values → current load/plane: 0.7-1.0 mA(compared to 10-12 mA with m beam) • but: many HV trips, number of „weak“ planes increased → beginning of September: • 19 / 32 planes >1800 V • 6 / 32 planes 1700-1800 VMalte Hildebrandt MEG Review Meeting, 18.02.2009Phase 3• phase 3: September – December • m beam: MEG physics run • several tests with running system: • to understand reason of HV trips→ see separate transparency • to possibly stop or even recover deterioration • but still: many HV trips, number of „weak“ planes increased further → end of December: • 11 / 32 planes >1800 V • 7 / 32 planes 1700-1800 V • in parallel: new test setup in laboratory (HV pcb, potting of capacitors, HV)→ see separate transparencyMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Trips• characteristics of HV trips: • individual „treshold effect“ for affected planes • deterioration due to frequent trips (due to damage?) • no obvious correlation with beam off/on, magnetic field off/on or muon target/CW target tube ↔ exception for 2-3 planes: • beginning of run period: HV trip while beam blocker opened → improved during run time („training“ ?) • during run period: HV trip 10-20 min after beam blocker closedMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Trips• characteristics of HV trips: • beginning of run period: air admixture to COBRA gas necessary to achieve stable dc operation → cHelium ≈ 95-96 % • significant deterioration started after: • p beam time (XEC, Dalitz) • 2-3 months with dc + COBRA gas and HV → at beginning: same planes affected as in 2007 • further deterioration due to frequent HV trips during remaining run time (2008: May – Dec) even without any further p beam time • stable operation (with reduced HV settings) during second p beam time (5 days) 2007: cHelium ≈ 96 %2007:p at end of run2007: after 2-3 months with gas and HV2007: similar, but: shorter run time (Sep – Dec)2007: dc system off during p beam timeMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Trips• characteristics of HV trips: • beginning of run period: air admixture to COBRA gas necessary to achieve stable dc operation → cHelium ≈ 95-96 % • significant deterioration started after: • p beam time (XEC, Dalitz) • 2-3 months with dc + COBRA gas and HV → at beginning: same planes affected as in 2007 • further deterioration due to frequent HV trips during remaining run time (2008: May – Dec) even without any further p beam time • stable operation (with reduced HV settings) during second p beam time (5 days) 2007: cHelium ≈ 96 %2007:p at end of run2007: after 2-3 months with gas and HV2007: similar, but: shorter run time (Sep – Dec)2007: dc system off during p beam time→ deterioration due to p beam time ?Malte Hildebrandt MEG Review Meeting, 18.02.2009HV Trips• characteristics of HV trips: • beginning of run period: air admixture to COBRA gas necessary to achieve stable dc operation → cHelium ≈ 95-96 % • significant deterioration started after: • p beam time (XEC, Dalitz) • 2-3 months with dc + COBRA gas and HV → at beginning: same planes affected as in 2007 • further deterioration due to frequent HV trips during remaining run time (2008: May – Dec) even without any further p beam time • stable operation (with reduced HV settings) during second p beam time (5 days) 2007: cHelium ≈ 96 %2007:p at end of run2007: after 2-3 months with gas and HV2007: similar, but: shorter run time (Sep – Dec)2007: dc system off during p beam time→ deterioration due to p beam time ?→ deterioration due to helium environment ?Malte Hildebrandt MEG Review Meeting, 18.02.2009Tests during MEG Run• infrastructure / hardware: • independent Bertan HV power supplies • HV cables • trip test with oscilloscope (MEG and lab) → no improvement• variation of dp_dc regulation value (pdc-pCOBRA): • ↔ small leaks ?Malte Hildebrandt MEG Review Meeting, 18.02.2009Location of HV Problems• HV problems at top of U-branches: systematics or just by chance? 15th Oct 2008→ due to difference in ρof He and He/C2H6:• dc operation at slightly lower dp_dclevel of dp_dc measurementdp• dc operation at slightly higher dp_dcMalte Hildebrandt MEG Review Meeting, 18.02.2009Location of HV Problems• HV problems at top of U-branches: systematics or just by chance? 15th Oct 2008→ due to difference in ρof He and He/C2H6:• dc operation at slightly lower dp_dclevel of dp_dc measurementdp• dc operation at slightly higher dp_dcMalte Hildebrandt MEG Review Meeting, 18.02.2009Tests during MEG Run• infrastructure / hardware: • independent Bertan HV power supplies • HV cables • trip test with oscilloscope (MEG and lab) → no improvement• variation of dp_dc regulation value (pdc-pCOBRA): • ↔ small leaks ? • 0.2 Pa → 2.0 Pa → -10% current/plane due to breathing of dc modules → no improvement• increase ethane fraction in dc counting gas: • ↔ inside sensitive volume ? • He/C2H6: 50/50 → 45/55 → reduction of gas gain by nearly factor 2 → no improvementMalte Hildebrandt MEG Review Meeting, 18.02.2009Tests during MEG Run• increase air admixture to COBRA: • ↔ outside dc module • phase 1: cHelium ≈ 95-96% • cHelium ≈ 95-96% → ~40% → 0 % → only test that showed effect → but not that clear result as expected for obvious problem „outside of dc module“→ summary of tests with dc system during MEG run period: • no clear cause and effect • but: hint, that problem is connected to longterm exposure to helium (inside and/or outside of dc module) and - at least in some cases - is located outside of dc module! → ensure helium atmosphere during christmas holidays and shutdown → see separate transparencyMalte Hildebrandt MEG Review Meeting, 18.02.2009Tests in Laboratory• check of common aspects of construction/assemby • sequence of production and assemby → no hint from logbook • wire tension → no hint from logbook • HV pcb • sealing/potting of capacitors • sealing/potting of HV soldering spot2007: several times weak point→ new test setup: • HV test setup • pcb, potting material • helium environment • T ≈ 40-45° C • longterm test (>3 months)Malte Hildebrandt MEG Review Meeting, 18.02.2009HV Test Setup in Lab→ new test setup: • HV test setup • pcb, potting material • helium environment • T ≈ 40-45°C • longterm test (>3 months)Malte Hildebrandt MEG Review Meeting, 18.02.2009HV Test Setup in Lab• 6 test samples: • 2 samples concentrating on sealing of HV cable, no resistors, no capacitors (Fri, Nov 7th) • 2 samples with HV cable, resistors, capacitors (Fri, Nov 7th, Thu, Nov 13th) • 1 sample with pcb_left and pcb_right (Thu, Nov 13th) • 1 sample: pcb glued on Cu-plate (Thu, Nov 20th)• all tested in air at 2kV• status test: • cHelium > 99% (reading of three O2 sensors) • all HVs at 1990V • T = 40-45°C (since Mon, Nov 24th)→ update 17.02.2009: • no significant deterioration (still 2kV) → no conclusion for “aquarium” test(“aquarium” = dc test setup in laboratory during shutdown)Malte Hildebrandt MEG Review Meeting, 18.02.20092nd Pressure Control System• construction of 2nd pressure control system for laboratory → operate „aquarium“ independently from MEG pressure control systemreminder: • „aquarium“: • setup to operate two dc modules with He/C2H6 as counting gas • within helium atmosphereMalte Hildebrandt MEG Review Meeting, 18.02.2009Helium Cabin→ ensures helium environment for dc system during „waiting time“ in lab→ cHelium ≈ 95% (conditions like in COBRA)• closed volume (V = 5.7m3)• windows / frames removable → access to dc modules• operated with the MEG pcs• helium sensor in exhaust line of helium cabin (instead of dc exhaust line)• patch panel is interface/accessible → dc system can be operated like in COBRAMalte Hildebrandt MEG Review Meeting, 18.02.2009Test Tesults→ first results from dc module tests in laboratory • weak point: potting of HV soldering spot on pcbMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Connection→ weak point: potting of HV soldering spot on pcb hood readouthood Vernier patternanode decoupling capacitorspre-amplifier cardsHV connection to pcb + sealingMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Sealing→ nominal conditioncable isolationdielectricumpottingpcbHV linebraided shield→ observation after run periodseveral pottings show: • change of shape („flowed away“) • change of color (white → brown) • same observation in 2007/8 (badly applied? no!)Malte Hildebrandt MEG Review Meeting, 18.02.2009HV Soldering SpotMalte Hildebrandt MEG Review Meeting, 18.02.2009Test Tesults→ first results from dc module tests in laboratory • weak point: potting of HV soldering spot on pcb→ replace ThreeBond 1530 with epoxy EPO-TEK 302-3M Malte Hildebrandt MEG Review Meeting, 18.02.2009Test Tesults→ first results from dc module tests in laboratory • weak point: potting of HV soldering spot on pcb→ replace ThreeBond 1530 with epoxy EPO-TEK 302-3M • „circumstantial evidence lawsuit“ (no glow by eye, IR camera, ...): • signals on oscilloscope: positive anode signals negative cathode signals • change of gas mixture: no effect (or long delay) → discharge between anode channel and GND but: not towards cathode strip and not in sensitive volume maybe towards frame at the edge or on pcb itself→ open dc module to verify or falsify: • wires and cathode foil fine • edges of isolators fine → closer look at vias on pcbMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Viatop layerGND+HVbottom layer7 mmMalte Hildebrandt MEG Review Meeting, 18.02.2009dc01Aglueno glueno glueMalte Hildebrandt MEG Review Meeting, 18.02.2009He / C2H6Heairglue glue glue G10 isolatorglue G10 isolatorglue carbon framePCB Cross Sectionbottom layertop layerpcbpcbbottom layer+HV+HVGNDMalte Hildebrandt MEG Review Meeting, 18.02.2009PCB Cross Sectiondpcbpcb+HV+HVGNDglue glue glue G10 isolatorglue G10 isolatorglue carbon frameMalte Hildebrandt MEG Review Meeting, 18.02.2009PCB Cross Sectiondpcbpcb+HV+HVGNDglue glue glue G10 isolatorglue G10 isolatorglue carbon frameMalte Hildebrandt MEG Review Meeting, 18.02.2009PCB Cross Sectiondpcbpcb+HV+HVGNDglue glue glue G10 isolatorglue G10 isolatorglue carbon frameMalte Hildebrandt MEG Review Meeting, 18.02.2009PCB Cross Sectionpcbpcb+HV+HVGNDglue glue glue G10 isolatorglue G10 isolatorglue carbon frameMalte Hildebrandt MEG Review Meeting, 18.02.2009Depth of HV Vias dc01A→ example: depth of HV via on dc01A upstream downstream a0 0.73 ± 0.03 0.76 a1 0.47 0.80 a2 0.39 0.66 a3 0.14 0.83 a4 0.65 0.67 a5 0.57 0.46 a6 0.64 0.61 a7 0.00 0.47 a8 0.57 0.57ddesign = 0.80 ± 0.02location of discharge(identified by signals)remark: all numbers in mmMalte Hildebrandt MEG Review Meeting, 18.02.2009PCB Cross SectionHe / C2H6pcbpcb+HV+HVGNDglue glue G10 isolatorglue G10 isolatorglue carbon frameMalte Hildebrandt MEG Review Meeting, 18.02.2009Next Steps→ following activities will start immediately and run in parallel: motto:confirm„viahypothesis“,startnewconstruction/repairworktoproofsolution • operate „dc skeleton“ (2 anodes + middle cathode) in „aquarium“ → no hood cathode: observe discharge ? • prepare new sample for HV test box: • no glue on ring of via • glue only on ring of via • fill via completely with glue → confirm effect of different „via/glue conditions“ • start construction of anode frames with new pcb designMalte Hildebrandt MEG Review Meeting, 18.02.2009HV Print 2009pads for resistorsHV print 2009• traces for HV on middle layer → no HV traces on bottom layer → individual layers with „only HV“ or „only GND“ (3-layer → 4-layer pcb)• „blind vias“ → vias have only necessary depth to connect appropiate layers (like „blind hole“)print 2007outer edgeinner edge+HV tracesvias for +HVGNDprint 2009Malte Hildebrandt MEG Review Meeting, 18.02.2009Next Steps→ following activities will start immediately and run in parallel: motto:confirm„viahypothesis“,startnewconstruction/repairworktoproofsolution • operate „dc skeleton“ (2 anodes + middle cathode) in „aquarium“ → no hood cathode: observe discharge ? • prepare new sample for HV test box: • no glue on ring of via • glue only on ring of via • fill via completely with glue → confirm effect of different „via/glue conditions“ • start construction of anode frames with new pcb design → operate 1st and 2nd dc in „aquarium“ to confirm long term behaviour → chance to check for additional „hidden“ weak point (masked by „via problem“)Malte Hildebrandt MEG Review Meeting, 18.02.2009Time Schedule→ following activities will start immediately and run in parallel: motto:confirm„viahypothesis“,startnewconstruction/repairworktoproofsolution detector laboratoryFebruary – March • „dc skeleton“ in aquarium • „via test“ with pcb • test new dc‘s in aquariumApril – July • ongoing test in aquarium July – August • test of dc system in support structur in laboratorybeginning of September → dc system ready for installation in MEG experimentdetector workshop• construction of new dc‘s: anodes with new pcb recycled cathode + hood• ongoing constructionMalte Hildebrandt MEG Review Meeting, 18.02.2009New DC Preamplifier• goal: • improve signal/noise ratio→ improve R- and z-resolution for small signals→ improve efficiency• argument: • analysis of experimental data 2007 using low-pass digital filter:→ improvement in z-resolution and 10% higher efficiency• solution: • select low-noise operational amplifier• adapt schematics→ bandwidth reduced from 140MHz to 80MHz• status: • few prototypes produced and tested→ signal/noise ration improved by factor 1.5• to do: • direct comparison of old and new preamplifier with real signals→ dc in „aquarium“ with cosmics and 90SrMalte Hildebrandt MEG Review Meeting, 18.02.2009Summary / Outlook• many dc planes / modules suffered from frequent HV trips• consequently theses planes / modules could only be operated
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks