LHC Electronics irradiation tests in the CNGS side gallery September 2009

Transcription

1 LHC Electronics irradiation tests in the CNGS side gallery September 2009 D.Kramer for the RADWG Thanks to B.Todd, E.Gousiou, E.Calvo, E.Effinger, R.Denz, K.Roeed, J.Lendaro, M.Brugger, EN/MEF, DG/SCR, Fluka team and the RadMon team (T.Wijnands, A.Nyul, C.Pignard) 30/9/09 report to LMC 1

2 Outline Test facility overview Planning of slots Calibration of test positions Comparison with Fluka User results CRYO BIC/PIC BLM BPM QPS HTS CL heaters not enough data WorldFip not enough data TE/EPC no data yet SURVEY no data yet 30/9/09 report to LMC 2

3 Testing area in the TSG4 side gallery of CNGS Radiation fields similar to the ones expected in LHC Area available after the displacement of the CNGS electronics to the ventilation chamber 30/9/09 report to LMC 3

4 HTS CL heaters QPS 2 Test areas High & Low flux and dose rate with multiple calibrated locations Target chamber Hottest test area in TSG46: ~ 1.5 Gy(SiO 2 )/week ~ (1MeV n eq.)cm -2 /week RadMon Survey New FLUKA simulations are now available thanks to K.Roeed BIC/PIC Target chamber WorldFip Hottest test area in TSG45 : ~ 12Gy(SiO 2 )/week ~ (1MeV n eq.)cm -2 /week This is expected in the LHC arcs* or UJ56/14/16 in 1 nominal year (10Gy/y, 3e11cm -2 ) BLM CRYO TE/EPC 30/9/09 *Alongside a dipole report to LMC 4

5 Calibrated radiation field data for all test locations accessible via dedicated Java GUI Hottest test area TSG46: ~ 1.5 Gy(SiO 2 )/week ~ (1MeV n eq.)cm -2 /week Hottest test area TSG45 : ~ 12Gy(SiO 2 )/week ~ (1MeV n eq.)cm -2 /week 30/9/09 report to LMC 5

6 Like it is the case for FLUKA, Safety Factor of 2 should be used for all the CNGS measurements (high gradients, small detector size) Calibrated radiation field data for all test locations accessible via dedicated Java GUI Hottest test area TSG46: ~ 1.5 Gy(SiO 2 )/week ~ (1MeV n eq.)cm -2 /week By assuming all the devices to be installed in the worst locations, the estimations are more conservative than 2 times (for SEE related failures) Example for the BLM location 1month time Hottest test area TSG45 : ~ 12Gy(SiO 2 )/week ~ (1MeV n eq.)cm -2 /week 30/9/09 report to LMC 6

7 Planning for the user tests Additional measurements: Activation of water samples TLD (Li6 & Li7) Au & Bi activation foils BLMI sensitivity to thermal neutrons Temporary procedure for handling irradiated electronics was put in place for CNGS. Valid until 30/11/09. Radioactive electronics WORKSHOP has to be used afterwards. All users requesting beam time in 2009 received a slot Only the vacuum group didn t receive the turbo pumps -> no tests possible Slot # End in week TSG45-1 TSG45-2 TSG46-3 TSG WorldFip BLM/BPM CRYO QPS 2 29 WorldFip BLM/BPM CRYO QPS 3 33 WorldFip BLM/BPM CRYO QPS 4 38 WorldFip + BLM/BPM BIC/PIC + QPS TE/EPC CRYO 5 41 TE/EPC Survey + BIC/PIC QPS BLM 6 47 TE/EPC Survey BIC/PIC QPS 30/9/09 report to LMC 7

8 USER test results CRYO BIC/PIC BLM BPM QPS HTS CL not enough data WorldFip not enough data TE/EPC no data yet SURVEY no data yet The presented radiation levels are for nominal LHC beam Estimations for the 2009/10 operation are added when necessary 30/9/09 report to LMC 8

9 Tunnel equipment in CNGS: Tunnel Electronics CRYO Tunnel electronics (~9500 channels in LHC tunnel) have received till now a cumulated dose of: TID (Gy) ~125 NIEL (n/cm 2 ) ~4e12..and the tests are still ongoing. No Single Event Errors LHC years in 90% of the cases (apart from DS) >10 o Output accuracy still within specs ( <0.3% ) Resistance measurements of 5 conditioners: 51 Resistance (Ω) 48 0 TID (Gy) 120 -Temperature reading (18 Channels) - Helium level reading (12 Channels) - Digital inputs (24 Channels) - Cold mass electrical heater DC supply (4 Channels) - FIP communication (8 agents) - Power supply (2 cards) 30/9/09 report to LMC 9

10 Protected Areas Electronics Cumulative Effects Failures CRYO Safety factor 2x to be applied o QRL electrical Heater AC supply o Insulated Temperature reading (~45 channels in protected areas): (~2400 channels in protected areas): TID (Gy) 5 NIEL (n/ cm 2 ) LHC years in the worst case (UJ14, UJ16, UJ56) 7e LHC years assuming 09/10 conditions in the worst case (UJ14, UJ16, UJ56) 500 TID (Gy) 70 NIEL (n/ cm 2 ) LHC years in the worst case (UJ14, UJ16, UJ56) 2e12 7 LHC years assuming 09/10 conditions in the worst case (UJ14, UJ16, UJ56) 7000 Failing component: solid state relay. Planning of moving electronics or shielding of protected areas. Failing component: DC-DC converter. Planning of scheduled replacements. o Same results for several channels and reproduced in two different CNGS locations. 30/9/09 report to LMC 10

11 Protected Areas Electronics Single Event Upsets CRYO o Insulated Temperature Conditioners (~2400 channels in protected areas): Cross Section (cm 2 ) 2e-9 LHC SEU/ hr for the total channels if all in UJ56 Failing component: ISO 150 Digital Isolator. 6 Foreseen SEU rate in 2009/10 For the total channels if ALL in UJ56 2/month Mitigation technique for LHC (in progress): soft reset automatically forced by the control system; No influence on proper operation of the machine. o Same results for 12 channels and reproduced in two different CNGS locations. 30/9/09 report to LMC 11

12 USER test results CRYO BIC/PIC BLM BPM QPS HTS CL 30/9/09 report to LMC

13 CPLDs Used in the Interlock Systems (5V) 500nm (3.3V) 350nm Beam Interlock System XC95288 x 300 XC95288XL x 34 UA, UJ, RR UA, SR, TZ (5V) 500nm Power Interlock Controllers XC95144 x 36 UA, UJ, RR 95288/144 (5V) failure = maintenance required = not machine critical 95288XL (3.3V) failure = can compromise safety* N.B. Whole BIS is redundant = *needs two identical failures during a mission 95288XL is used in the control VME chassis TE/MPE/MI has installed a Radiation Test Bench in CNGS XC95144 x 32 Almost = # in LHC XC95288XL x 32 30/9/09 report to LMC 13

14 XILINX CPLD test setup in CNGS for BIS and PIC Dedicated test setup produced only for the radiation tests One test card (5V) failed after 77Gy Setup remains in CNGS to verify total dose resistance of the CPLDs No destructive Latchup observed so far (to be confirmed) 30/9/09 report to LMC 14

15 XC9500XL (3.3V) [ UAs, SR, TZ76 etc ] Failure (SEE induced) cross-section per device = 2.8e-10 cm 2 In UA87 one expects high energy hadron fluence less than (conservative!) 1e8 cm -2 /y The resulting failure rate for 1 device = SEEs/y If all the 36 CPLDs were in UA87 (not considering the shielding installed last week), we would get 0.9 failures/y, out of which 20% produce loss of redundancy (5.5y MTBF) 80% lead to a false dump (1.4y MTBF) We have to keep in mind that this is a low voltage device and might be sensitive to thermal neutrons. In areas like UAs their fluence could be up to 10xhigher than hadrons>20mev (Fluka analysis to be performed). If this is true, the 0.9 failures/y is a conservative but not exaggerated estimation. Tests in thermal neutron beam are being considered 30/9/09 report to LMC 15

16 XC9500 (5V) The device has a very low error cross-section! Failure (SEE) cross-section = 3.8e-13 cm2 We can assume conservatively 1e9cm -2 /y for the high energy hadron fluence in the RRs. We get then 3.8e-4 SEEs/y/device which makes 0.14 SEEs/y for the 366 devices, i.e. 7.2y MTBF In 10% of cases the SEE produces a False dump In 90% of cases the SEE produces a Monitoring problem These figures are of the same order of magnitude as the Electrical Reliability of the devices (according to Xilinx) In both cases above we don t care about the total dose, one should not reach 1Gy/y for the concerned locations (except the locations below) Relocation from UJ56/UJ14/16 is in preparation (cables placed, new location identified) 30/9/09 report to LMC 16

17 USER test results CRYO BIC/PIC BLM BPM QPS HTS CL 30/9/09 report to LMC

18 Quick report on the early QPS measurements WorldFip problems Field-bus coupler type DQAMGS (new QPS layer) Errors produced by SEUs in the WorldFip part Using latest MicroFip TM and FIELDRIVE TM (line driver) version Data are still readable but no longer updated -> *Access required prior to next fill Cross section ~ 2.8x10-10 cm 2 Assuming 4x10 10 cm -2 /y (10Gy/y) in the arcs and 450 devices one gets: 14 errors*/day Hardware solution was found and being successfully (so far) tested in CNGS Accelerated SW compatible MicroFip TM and FIELDRIVE TM development would be much appreciated not only by the QPS team for any future developments This year s tests in CGNS confirmed the radiation tolerance of the detection systems of the new QPS layer including the power supplies. Detailed evaluation still to be done. 30/9/09 report to LMC 18

19 USER test results CRYO BIC/PIC BLM BPM QPS HTS CL 30/9/09 report to LMC 19

20 BLM- System: Installation overview Arc Installation In all the arcs from 12L 33L, 12R 34R Expected Radiation levels: worst case Dose = 1-10 Gy/y (1MeV eq. neutrons = 3* *10 11 cm -2 /y) 360 * BLECF Card (Analogue data acquisition cards) 360 * Hand made power supplies Straight Section Installation RR13, RR17, UA23, UA27, UJ33, UA43, UA47, RR53, RR57, UJ63, UJ67, RR73, RR77, UA83, UA87 Expected Radiation levels: worst case Dose = <0.7Gy/y (1MeV eq. neutrons = 1*10 9-1*10 10 cm -2 /y) 309 * BLECF Card (Analogue data acquisition cards) 74 * 5V Power supplies (Haltec-USR515-5A, of-the-shelf) 37 * 2.5V Power supplies (Haltec-USR515S-2A, of-the-shelf) 30/9/09 report to LMC 20

21 BLM- System: test conclusions Power supplies 5V: 19 years of nominal LHC operation possible Power supplies 2.5V: 20 years of nominal LHC operation possible BLECF (Tunnel acquisition card): more than 20 years LHC operation possible 1 SEE induced failure (false dump) observed Cross section = 2.2x10-13 cm 2, this means for 600 cards assuming all in the arcs 70 days MTBF (fail safe)

22 USER test results CRYO BIC/PIC BLM BPM QPS HTS CL 30/9/09 report to LMC 22

23 BPM electronics Tested up to 123 Gy (>10 nominal years) 1 Chassis similar to the installed on LHC. 1 Power supply 545x in LHC 2 WBTNs 1 Intensity card 1 ufip controller card Eva Calvo (BI/QP) 2160x in LHC

24 BPM electronics Tested up to 123 Gy (>10 nominal years) Chassis similar to the one installed in LHC. 1 Power supply OK 545x in LHC 2 WBTNs 1 Intensity card 1 ufip controller card OK failed SEEs Eva Calvo (BI/QP) 2160x in LHC 330x in LHC

25 BPM electronics Tested up to 123 Gy (>10 nominal years) Intensity card failed very soon Aim of the intensity cards is to detect obstacles in the machine during its initial operation (low intensity beam) 3x10 10 cm -2 (1MeV n. eq.) means ~ 0.5 nominal LHC years Most probably not an issue for 2009/10 Measurement corrupted after the event (SEL?, card to be inspected) ufip controller card sensitive to SEEs sending the sensitivity settings of the WBTNs, launch calibrations, or monitoring the power supplies Return to the default state after SEE (calib. off, high sens.) SEE Cross section = 4x10-12 cm 2 thus 2.2days MTBF Worst case: 250um offset for 60s for 1 channel can be compensated by orbit feedback -> not an issue for operation 30/9/09 report to LMC 25

26 Summary (detailed report to be written after the end of tests) CRYO tunnel cards perform very well no concerns Some CRYO non-tunnel cards sensitive to total dose, replacements after 0.5y of nominal beam UJ14/16/56 not critical for operation Some CRYO non-tunnel cards sensitive to SEUs, software mitigation in progress (6 SEUs/h for nominal beam) BIC VME chassis uses sensitive CPLDs but only a small number and not in the most critical areas (5.5y MTBF) very little concern BIC/PIC installations use low cross-section CPLDs which always fail in a safe way (7.2y MTBF) no concerns No cumulative effect problems expected for the BLM system Marginal sensitivity of tunnel card to SEE means 70 days MTBF (fail safe) Very sensitive latest WorldFip chip could compromise the new QPS layer hardware solution looks very promising (450 cards in LHC) If uncorrected, 14 errors/day in nominal conditions (no issue for safety) BPM cards showed no radiation related problems, intensity cards don t stand high intensity (needed for low intensity), ufip card sensitive to SEEs (2.2days MTBF) but handled by the orbit feedback CNGS side gallery proved to be a very useful facility!! 26

27 Reserve slides CNGS Fluka simulations 30/9/09 report to LMC 27

28 Simulated radiation fields of the selected most critical non-tunnel locations LHC Area Beam conditions Hadrons>20MeV [cm -2 /y] 1MeV n. eq. fluence [cm -2 /y] UJ56 09/10 4x x10 7 4x UJ56 nominal 4x x x x UJ14/16 09/10 4x x10 7 4x UJ14/16 nominal 4x x x x Dose [Gy(air)/y] RR13/17 nominal 3x x x x RR77/73 09/ x RR77/73 nominal 1x x10 8 5x x UJ76 09/ x UJ76 nominal x UA87 2x10 15 protons on TED/y* x /9/09 *Assuming 2 shots per week on TED report to LMC 28

29 BLM- System: Status CNGS Result power supplies PS1, PS2, PS3 and BLECF Device Date Hadrons>20MeV [1e11*cm-2] 1MeV eq. neutrons [1e11*cm-2] Dose [Gy] Observation PS 1 (5V) (Haltec- USR515-5A) PS2 (5V) (Haltec- USR515-5A) PS3 (2.5V) (Haltec- USR515S-2A) BLECF_ : Voltage starts to increase : Voltage increased by 20% (1V) : Voltage dropped to 0V : Voltage starts to increase : Voltage increased by 11% (0.55V) : Voltage dropped to 0V : Voltage starts to increase : Voltage increase of 10% (0.25V) : Voltage dropped to 0V : Card stopped : Card reset -> card works again : Card still working RadMon data for the BLM crate positioned in TSG45. Beam intensity data was scaled with the calibration data from 2009.