PXL Ladder assembly problems 03212013 Report HFT Hardware

PXL Ladder assembly problems 03/21/2013 Report HFT Hardware Meeting Giacomo, Leo, Michal, Xiangming

Assembly work flow chart (Engineering Run) Probe tested sensors Electrically tested driver boards Electrically tested low mass cables Dimensionally checked composite backer • Ladder characterization • Reworking and troubleshooting • Quality assessment • Initial validation Ladder wire bonding Full functionality test Wire bond encapsulation Frame mode test 1 day without problems Ladder assembly Sector assembly Frame mode test Sector metrology Full functionality test HFT PXL 3/6/2021 GC - HFT PXL Ladder Assembly Problems 2

Summary of problems reported during ladder assembly • Single sensor flaws – – BIAS register not programmable (1 x) Clamping voltage flaw (2 x, U 1 sensors) Broken LVDS differential line (1 x) Sensor not responding to internal and external Start (1 x) ü Not affecting the general ladder behavior: acceptable for the engineering run • Short between power and ground lines on ladder – – – Wrong wire bonding (change in U 2 pads, bonding errors) Temporary shorts burned by flowing current (2 x) Reduced adhesive layer between sensor and flat cable Flaw in the sensor VDDA regulators (2 x) Flaw in the sensor Memory Blocks (2 x) Flaw in the sensor analog power net (1 x) ü Can prevent from operating the whole ladder • Shorts on LVDS outputs (1 x) • Cracks in sensors (6 x on 4 ladders) 3/6/2021 GC - HFT PXL Ladder Assembly Problems 3

Power-GND shorts • Vd-GND – Power. On_Reset pads connected to power and ground • due to new bonding pad assignment in Ultimate 2 not propagated from sensor design to assembly – 131 from Digital. Power to Power. On_Reset – 132 from Ground to Power. On_Reset(probe) ü Diagram fixed, wrong wires removed – Temporary shorts (2 x) • Not associated with any other known issue, spotted by thermal camera before disappearing ü Burnt by pumping current through the power circuit, probably conductive residuals – Reduced adhesive layer between sensor and flat cable (1 x) • Spotted by visual inspection, potential cause of shorts between sensor bulk and cable traces ü Adhesive layer extended in both dimensions to increase the portion coming out from underneath the sensors 3/6/2021 GC - HFT PXL Ladder Assembly Problems 4

Power-GND shorts /2 – Internal short in VDDA regulators (2 x) • Thermal camera inspection localized hot spots in the regulators region ü Regulator digital ground pad disconnected, short fixed – Internal short in Memory blocks (2 x) • Thermal camera inspection localized a hot region across the border between digital section and pixel matrix ü The removal of the wires from the "digital power Memory" pads (112 -113, 161 -162) cured the short in both ladders • One of the flawed sensors showed high current during probe testing ü High current draw now added to rejection criteria; in contact with IPHC to investigate a possible connection with this issue • Va-GND short Thermal picture of L 018_chip 9 while injecting current in the power lines – Internal short between analog power net and ground (x 1) • Thermal camera inspection localized hot spots in the regulator region ü All analog power wires had to be removed to fix the short 3/6/2021 GC - HFT PXL Ladder Assembly Problems 5

Shorts on LVDS outputs • L 026 developed shorts on 10 different LVDS output pairs, distributed over 7 different chips • Never observed on any other ladder • No shorts between different outputs or with other lines • List of shorted LVDS pair is (chip#_output#): • 2_0, 2_1; 3_0, 3_1; 5_0; 6_1; 7_1; 8_0, 8_1; 9_0. • Electrical continuity tests and thermal camera inspection to localize the shorts • Hot spots have been localized underneath the sensor edge, in the digital section, by injecting a controlled amount of current through the LVDS lines – Not pointing to the actual short location in this case 3/6/2021 GC - HFT PXL Ladder Assembly Problems 6

LVDS/2 – Thermal picture of L 026 Chip 9_0 3/6/2021 GC - HFT PXL Ladder Assembly Problems 7

LVDS/3 - Other LVDS pair shorts in L 026 Chip 2_0 Chip 6_1 3/6/2021 Chip 3_1 Chip 7_1 GC - HFT PXL Ladder Assembly Problems 8

LVDS/4 - L 026 inspection and dissection • After excluding shorts in: – Sensors – Driver board – Exposed flat cable region • L 026 has been dissected, starting from the region spotted by thermal camera and proceeding along the shorted traces, trying first with the sensor edges, eventually removing one complete sensor (see next slide) • Conclusions: – Low impedance shorts beneath more than one sensor – Shorts are happening far from the sensor edges – Local shorts affecting two cable traces at most (~0. 5 mm distance) • Possible explanation: local lack of adhesive layer allows the sensor bulk to short trace pairs during the vacuum bag pressure cycle ü Long term solution: insulating coverlay to be applied on low mass cables 3/6/2021 GC - HFT PXL Ladder Assembly Problems 9

LVDS/5 - L 026 dissection pictures 3/6/2021 GC - HFT PXL Ladder Assembly Problems 10

Cracks on sensors during ladder assembly /1 After the assembly of ladders L 011 -L 018, ladders 011, 012, 014 and 015 contained sensors that had fractures: Fractures are generally along row or column boundaries, but can also present a pinhole along the crack • They must have developed after probe testing, during handling and ladder assembly • The whole ladder assembly procedure has been reviewed and monitored during the assembly of two dummy ladders made of bad probe tested sensors ü The fracturing was not repeated either in the dummy ladders or in the following 10 functional ladders assembled later • L 011 – chip 9 L 012 – chip 9 3/6/2021 GC - HFT PXL Ladder Assembly Problems 11

Cracks on sensors during ladder assembly /2 L 014 - 4 th chip L 015 - 4 th chip L 014 - 4 th chip • • Hints of cracks on 4 th chip of L 014 and L 015, in the same chip region The similarity of the two defects could indicate something in the corresponding region on the tool as cause of the problem. There is no damage on the other ladders in that region, therefore it was most likely a fragment or other dirt rather than a permanent defect of the tool On L 014, 4 th chip, the crack starts with a hole. Similar defects (hole/scratch + crack) are visible on 2 nd and 6 th chips in the same ladder ü Deeper cleaning procedures are now applied to sensors and tools at each assembly step •

Conclusions • The assembly of ladders and sectors for the PXL Engineering Run has been crucial to deal with a large number of unexpected issues • Thanks to this experience, we were able to refine the assembly procedure and to develop effective tools for troubleshooting ladders • The up-to-date ladder assembly statistics for the engineering run are: Quality assessment (03/20/2013) # % 15 55. 6 Lower quality operating ladders (>8 operating chips) 4 14. 8 Recoverable not operating ladders (need fix or workaround) 4 14. 8 Dead ladders (not recoverable) 4 14. 8 Fully operating ladders Total 27 100. 0 • We believe we are in good shape for the upcoming production of the final HFT PXL detector 3/6/2021 GC - HFT PXL Ladder Assembly Problems 13