1. 1 U. Goerlach, LECC Colmar Sept 2002  Design Considerations and a some History  Technology Choices  First Prototypes on Ceramics and Results  New Developments in Rigid-Flex and Full-Flex Technology  Preparation of Industrial Production and Testing (FHIT)  Conclusions  Design Considerations and a some History  Technology Choices  First Prototypes on Ceramics and Results  New Developments in Rigid-Flex and Full-Flex Technology  Preparation of Industrial Production and Testing (FHIT)  Conclusions Front-End Hybrids for the CMS Silicon Tracker Front-End Hybrids for the CMS Silicon Tracker Laboratories: IReS & LEPSI, Strasbourg UCL-Louvain & RWTH Aachen III Contributing Aachen I, Bari, CERN, IC-London, Karlsruhe, Pisa, Vienna U. Goerlach, LECC Colmar Sept 2002

2. 2  Start in 2000  Definition of functionality  Modularity: hybrid geometries and shapes  Technology choices  Option for fully industrial production  2000-2002  Prototyping in thick-film technology ca. 180 FE-hybrids  Many design changes to adapt to detector-module integration  Electrical system tests  Development of an automatic test-station (FHIT) for industrial production  Identification of industrial partners  2002  Revision of initial technology choice  Start of industrial production planned Design Considerations and some History N.Marinelli, Tuesday: Hybrids need some additional work

3. 3 U. Goerlach, LECC Colmar Sept 2002  Electrical functionality (2002):  Analogue read-out chips 4 or 6 APV25  Power lines, grounding, decoupling  Auxiliary chips: MUX, PLL and DCU  Measurements (with DCU-chip) of :  Supply voltages,  Temperatures on hybrid and detector (with internal and external thermistors)  Detector bias return current  No HV on hybrid  Connector(s) to interconnect-cards and opto-hybrid for analogue read- out Definition of FE Hybrid (I) Detector bias return NAIS 40 pin NAIS 30 pin SAMTEC 20 pin 47 mm 60 mm 25+3 mm APVAPV MUXMUX PLLPLL DCUDCU

4. 4 U. Goerlach, LECC Colmar Sept 2002  Kapton cable  1 or 2 connectors (NAIS and SAMTEC)  Bending radius of 1-1.5 mm (180° turn)  SMD components  Minimal height  R and C of type 0402 and 0603  LPCC (since 2002): MUX, PLL, DCU  Naked die ASICs to bond  4 or 6 APVs, alignment  30  m  (MUX-PLL)  (DCU)  Glob-top cover for bonds? (Radiation hardness ?)  Multi-layer board  Mechanical parameters  Two geometries  Heat transport to frame (  3 Watt)  Support of pitch-adapter  Thickness limitations, less than one mm without components  No connector on the hybrid  Rigid and flat within 100  m  Operation at -10° to -20 °C  Radiation hardness  Electrical parameters (2002)  4 metal layers  Via: Ø 100/300  m  Line width: 120  m  Separations (line/via): 180/90  m  Bias line resistance: 20-50 m  Definition of FE Hybrid

5. 5 U. Goerlach, LECC Colmar Sept 2002 Modularity and Geometries TIB TID TOB TEC TIB TID TECTOB  Need industrial production! (Version 2002)

6. 6 U. Goerlach, LECC Colmar Sept 2002

7. 7 InterConnect Bus InterConnect Cards Module frame Cooling pipe Patch panel 15 cm 110 cm TOB-rod Opto- hybrids

8. 8 U. Goerlach, LECC Colmar Sept 2002 Loaded TOB Rod with Si-Detector Modules Used for double sided modules

9. 9 U. Goerlach, LECC Colmar Sept 2002 TIB Shell loaded with Si-Detector Modules

10. 10 U. Goerlach, LECC Colmar Sept 2002 Technology choices (I)  Industrial availability and price  Xo budget of tracker  Radiation safety (no Ag) Technology Candidates (Very approximate figures!)

11. 11 U. Goerlach, LECC Colmar Sept 2002  Thermo-mechanical properties  Heat (max. 3 Watt) has to be transported to detector frame  CTE compatibility with module frame material:  Hybrid will be glued on frame (carbon fibre or graphite)  Deformation  Lift-off  Many simulations and tests to validate concept and substrate-material  Suitable for automatic mounting on Gantry  Flatness  Rigidity  Integration of cable and connector(s) ? Technology choices (II) Material CTE ppm/ o C Thermal Conductivity (W/mK) Xo (mm) Al2O37.024.075.5 FR4/G1012-160.2-0.3194.0 Carbon Fibre < 1.0200-400 || 1  250.0 Graphite E779 7.465. Polyimide45.00.2286.0 Cu17.0390.014.3 Au14.0318.03.35

12. 12 U. Goerlach, LECC Colmar Sept 2002  First Circuits produced at CERN Workshop then  Populated with SMD and ASICs and then  Tested in Strasbourg    APVs need careful bias decoupling on V125  Market survey MS2991 (Spring 2000)  Small via diameters and spacing at industrial technical limit (120  m)  Very few candidates Prototyping in Thick-Film Technology on Ceramic Substrates (2000-2001) MUX_PLL 47 mm 28 mm OptohybridInterconnect card Kapton cable DCU Combined MUX-PLL chip (naked die) Small pitch: Min.: 115  m

13. 13 U. Goerlach, LECC Colmar Sept 2002 Read-Out Chip (APV) Stability on Hybrid Oscillation on V125 power line (across 4.7  F) ca. 80 kHz,  200 mV VSS V125 V250 100 nF VSS V125 V250 4.7-10  F Common bias lines VSS V125 V250 100 nF VSS V125 V250 100 nF VSS V125 V250 100 nF VSS V125 V250 100 nF VSS V125 V250 100 nF Decoupling of APV-power rails: Power line resistances on hybrid from connector to APV: 20-40 m , including 15 m  in cable M.Raymond, IC London

14. 14 U. Goerlach, LECC Colmar Sept 2002  Split V125 line for preamplifier and inverter  Re-generate V125 from V250 by individual 100  resistor  Solution has been tested up to common resistances of up to 5  on the V125 line  Robust solution  New split power pads (bonding!): Solution to APV (In-)Stability M.Raymond, IC London

15. 15 U. Goerlach, LECC Colmar Sept 2002 Dorazil and MIPOT Ceramic Hybrids 150 hybrids 5 hybrids Main Difficulties: Small feature size not suitable for mass production Soldering of kapton cable Main Difficulty: At technical limit of company After a market survey (MS2991) only two companies agreed to produce hybrids for CMS Tracker M200 Milestone at reasonable price

16. 16 U. Goerlach, LECC Colmar Sept 2002  Good APV performance preserved on FE-hybrid!  300 e - Merging all APVs pedestals calibrations Uniformity: Noise: APV 1+2 APV 3+4 APV 5+6 calibration APV Read-Out on Hybrid

17. 17 U. Goerlach, LECC Colmar Sept 2002 Frank Hartmann IEKP - Universität Karlsruhe (TH) Irradiation Test of Ceramic FE-Hybrid Proton current: 2 µA " 26 MeV protons " Irradiation of 120x120mm 2 up to 10 14 p/ cm 2 takes 20 min Irradiation of Hybrid During irradiation: Clock, Trigger (10Hz), Error Buffer Readout (1Hz), Reset Irradiation with 2 µA protons up to 2.7 10 14 p/ cm² (~ 30years LHC) First verification at cyclotron with over 5m cables and provisional readout ( Hybrid is still activated (2µSv) ) Hybrid still alive! Irradiation of some SMD components and glue has been performed by the Vienna group All final components have to be qualified with irradiation !

18. 18 U. Goerlach, LECC Colmar Sept 2002 Summary on Ceramic Hybrids  In total about 180 ceramic hybrids were produced  Yield about 80% (we did not repair everything)  About half will be used for prototype Si-detector modules until December  Assembly procedure  Performance in test beam  Different electronic and electrical and mechanical system tests for the configurations of TIB TOB and TEC.  Temperature cycles  Irradiation tests  General concept of FE-hybrid validated  Several changes requested (  re-design of layout)  Re-design with larger feature size necessary  Possible with new encapsulated control chips in 0.5mm pitch housings  Now other potentially significant cheaper technologies available  New R&D and industry survey necessary at very late stage of project

19. 19 U. Goerlach, LECC Colmar Sept 2002 Temperature Cycling of FE-Hybrids

20. 20 U. Goerlach, LECC Colmar Sept 2002  With new LPCC Control and Service ASICS  MUX, PLL, DCU  Multi-layer board in “advanced” FR4 printed circuit board technology  Could be very cheap in large quantities  First circuits in January 2002  Boards are correct  Great difficulties to solder cable  Only a few working proto-types  Next step was cable integration:  Rigid-Flex hybrid New Technological Choices Prototyping I FR4 board with connector

21. 21 U. Goerlach, LECC Colmar Sept 2002  Flex-Rigid(FR4) Hybrid:  Different companies with slightly different technologies and quality  2 bottom layers on rigid FR4  2 upper layers on polyimide  Tolerances in thickness up to 100 micron  Non-Flatness in the order of 100 micron before SMD montage, and we have seen more  Extend FR4 part under PA  More prototype boards available soon Promising path, but quite a bit more expensive than a simple FR4 board Polyimide copper Coverlay FR4 FR4- Noflow New Technological Choices Prototyping II Thickness

22. 22 U. Goerlach, LECC Colmar Sept 2002 Coverlay 25 Pi + 25 glue  Flex-Flex (all Kapton) laminated on  Carbon-fibre substrate  FR4-substrate with thermal heat conducts  Mechanical properties  Final rigidity?  Final flatness < 100  m?  Proto-types available soon Polyimide 25 micron Copper 20 micron Lamination 25 micron CICOREL New Technological Choices Prototyping III Carbon fibre or FR4

23. 23 U. Goerlach, LECC Colmar Sept 2002  Large quantities (more than 15000) can only be produced reliably in industry  Numbers will help to achieve uniformity throughout production  Industry will also be charged with the final acceptance test before delivery  Preferably only one manufacturer or consortium, delivering final product  Technical specifications have to be well defined before tendering  Careful evaluation and system tests mandatory!  Qualification of manufacturer by proto-type runs  Quality assurance during production:  In depth test/characterisation of random samples  Some temperature or other cycles in industry before final acceptance test  Relay on industrial standard during mass production  Duration of production will be about one year N.M.: Final results of system test expected in spring 2003 Industrial Production of FE-Hybrids

24. 24 U. Goerlach, LECC Colmar Sept 2002 Front-End Hybrid Industrial Tester (I) Task: Simple acceptance test of hybrid in factory Components: Mechanical structure Transition board (FEHC) FHIT: electronic circuit including switching matrices Active component, a connection to ARC and fast Controllers ARC Read-out system Power supplies PC Barcode reader Software ARC board

25. 25 U. Goerlach, LECC Colmar Sept 2002 Front-End Hybrid Industrial Tester (II) http://www.fynu.ucl.ac.be/themes/he/cms/activities/tracker/hybrids.html Test sequence:  Power supply control  Barcode scanning, recognition of hybrid type  Continuity test  Electrical test, including I2C scan  Functionality test (read-out of APVs)  Log file creation + error file + hybrid identification file  XML (CMS database) Response (simplified for operator!): green or red light Block diagram

26. 26 U. Goerlach, LECC Colmar Sept 2002 APV current consumption distribution I_250 I_125

27. 27 U. Goerlach, LECC Colmar Sept 2002 DCU ADC calibration Signal amplitude as a function of MUX resistors being switched on

28. 28 U. Goerlach, LECC Colmar Sept 2002  Successful development of FE-hybrids for the CMS-tracker  In total we have produced over 200 hybrids in different technologies  Proto-typing of FE-hybrids in thick film on ceramic  Required performance achieved!  Many modifications implemented to help system integration!  Need for revised layout using ASICs with larger footprint  Now open to new (cheaper) technologies  New technologies (FR4, flex-rigid, full-flex) are being explored  Proto-types in industry in progress  Will determine final technical specification and the choice of substrate  Full industrial production foreseen (duration is about one year)  Industrial tester has been developed for acceptance test  Quality assurance:  Relies on large scale industrial production standards  Extensive characterisation of samples during production Summary and Conclusions

29. 29 U. Goerlach, LECC Colmar Sept 2002 DCU  Digital Test: OK  -1 LSB < DNL < 1 LSB  Transient noise RMS ~ 1/4 LSB  Power dissipation < 40mW  ADC Gain vs. X-ray dose: -0.4 %/Mrad  No evident changes in INL and transient noise RMS during and after X-ray irradiation Gain ~ 2.2 LSB/mV INL < 1LSB

30. 30 U. Goerlach, LECC Colmar Sept 2002 Study of frames deformations and cooling efficiency RWTH Aachen I reflecting surface grid CCD >100  m dummy hybrid at correct temp.