Characterization of MAPS for the ALICE experiment

Transcription

1 Characterization of MAPS for the ALICE experiment Silvia Tedesco Relatrice: prof.ssa Stefania Beolè Università degli Studi di Torino 21 luglio 2017

2 Inner Tracking System (ITS) Upgrade Introduction Present ITS main limits: Silicon Drift Detector max readout rate: 1kHz Material budget of 1.1 % of X 0 per layer Resolution of 120 μm at p T = 500 MeV/c Upgrade main motivation: study of rare probes at very low p T (< 1 GeV/c). New ITS: 7 layers with new Monolithic Active Pixel Sensor called ALPIDE. Each layer is made up of staves: Space frame Cold plate Flex Printed Circuit (FPC) Power Bus 21st July 2017 Silvia Tedesco 2

3 Outline Flex printed circuit Visual analysis Studies of the transmission quality of a digital signal through the FPCs Half-Stave (HS) 0 Left Wire-bond quality Attenuation of a digital signal along the HS and FireFly cables 21st July 2017 Silvia Tedesco 3

4 Flex Printed Circuit FPC is connected directly to the chip with wire-bond. Main features: transmit clock and control signals transmit data out of the chips power the chips B0 Chips have a master/slave architecture. data lines control Top view A8 Stackup Status: production started, first 300 pieces arrived. 21st July 2017 Silvia Tedesco 4

5 FPC Visual inspection Best version of FPC from GS Swiss Differences between FPCs produced by three different industries: Rectangular pads too close due to track-pad interconnection Non symmetric pads Non centred vias 21st July 2017 Silvia Tedesco 5

6 Test setup Generator o ō Osc Agilent N5980A o ō i ī PC FPC FPC Setup for eye diagram s acquisition Setup for BER s acquisition Input: pseudo random bit sequence with different data rates: Data rates from 125 Mb/s to 3.1 Gb/s Signal amplitude: 400 mv 10 measurements of eye diagram parameters for each data rate. GOALs: 1. Effect of vias on differential lines 2. Edge effects on the FPC 3. Check the transmission quality on data lines 21st July 2017 Silvia Tedesco 6

7 Eye diagram and BER Eye diagram allows to study the transmission quality of a digital signal. A 1 UI Superimposing all the slices Parameters Eye height Eye width Jitter total (the deviation from true periodicity of a periodic signal) t Bit Error Rate (BER) gives information about errors in digital signal transmission. Acquisition time T = - ln(1- α) ν * BER α : confidence level ν : frequency BER : BER limit 21st July 2017 Silvia Tedesco 7

8 Examples of eye diagram Clock line (A8) Mb/s Mb/s Mb/s Mb/s Mb/s Mb/s Mb/s Mb/s 21st July 2017 Silvia Tedesco 8

9 1. Effect of vias on line The two lines studied are different only for the presence of vias: the clock line on M8 side has the vias. Max accepted jitter: 0.2 UI 400 Mb/s 1.2 Gb/s Vias have no effect on data transmission: fluctuations are due to experimental setup (same effect measured with external cables only). Same conclusions for eye width. 21st July 2017 Silvia Tedesco 9

10 2. Edge effect studies The two lines studied are different just because control line runs closer to the FPC edge. Max accepted jitter: 0.2 UI 400 Mb/s 1.2 Gb/s There are no edge effects on data transmission: fluctuations are due to experimental setup (same effect measured with external cables only). Same conclusions for eye width. 21st July 2017 Silvia Tedesco 10

11 3. Data lines Analysis of transmission quality on data lines for 1 and 2 FPCs. 400 Mb/s 1.2 Gb/s Eye height decreases with increasing the number of FPCs. BER for data, clk and ctrl lines < for all the data rates. BER = only at 3.1 Gb/s for 2 FPCs (very high data-rate). 21st July 2017 Silvia Tedesco 11

12 Half-Stave 0 Left HS is made up of seven HIC connected in series. GOALs: Wire-bond quality Attenuation of a digital signal along the HS and FireFly cables 21st July 2017 Silvia Tedesco 12

13 Wire-bond FPC is connected directly to the chip with wire-bond. Measure of #bond/pad for clock, control and data lines with Coordinate Measuring Machine by Mitutoyo. For each connection 3 wire-bonds are expected for safety reason. 21st July 2017 Silvia Tedesco 13

14 #Bond Wire-bond Good connections Bad connections 3,5 3 2,5 2 1,5 1 0,5 0 HIC in position 7 (side B0) #bonds #broken bonds Pad For few modules, 2 bonds/pad have been seen --> Lower quality due to not finalized production procedure. 21st July 2017 Silvia Tedesco 14

15 Data transmission B0 3 A8 2 RU 1 Oscilloscope 21st July 2017 Silvia Tedesco 15

16 Data transmission 3 A8 B0 2 FireFly c. RU 1 Oscilloscope Measurements with differential probes on control line of HS0L on 3 different points at the same time: On read-out unit (point 1) At the beginning of the HS (point 2) At the end of the HS (point 3) Only the module under test was powered on. Studies of attenuation along the stave and the FireFly cable. 21st July 2017 Silvia Tedesco 16

17 Oscilloscope signal Pos 7 Pos 6 Pos 5 Pos 4 Pos 3 Pos 2 Pos 1 RU CH1 CH3 CH2 Module in position 4 CH1: (800 ± 11) mv CH2: (870 ± 12) mv CH3: (875 ± 16) mv A8 Module in position 1 CH1: (825 ± 14) mv CH2: (890 ± 14) mv CH3: (675 ± 18) mv B0 21st July 2017 Silvia Tedesco 17

18 Attenuation along FireFly cables c2/c1: ratio of the signal amplitudes measured in c1 and c2. Slope = (-1.3 ± 7.7 ) 10-3 Z = 0.17 < 1.96 Slope = (-0.6 ± 5.6 ) 10-3 Z = 0.11 < 1.96 As expected the slopes are compatible with zero: the attenuation on the FireFly cable + FPC extension is the same for all the modules. 21st July 2017 Silvia Tedesco 18

19 Attenuation along HS c2-c3: difference between the amplitude measured in c2 and c3. Attenuation on side A8 is bigger than on side B0: there is a difference of impedance along the lines. R 2 A8 = (125 ± 1) Ω R 2 B0 = (123 ± 1) Ω R 3 A8 = (97 ± 1) Ω R 3 B0 = (95 ± 1) Ω 21st July 2017 Silvia Tedesco 19 r Ctrl Ctrl Tester for R 3 Tester for R 2 r

20 Conclusions Flex Printed Circuit Vias have no effect on data transmission There are no edge effect on data transmission Large operational margin at high data rates even with 2 FPCs Half-Stave 0 Left Wire-bond of good quality, 2 bonds/pad seen for few modules Attenuation along FireFly cables is the same for all HIC as expected Attenuation along HS is different for side A8 and B0 due to a different impedance A A8 = (36 ± 3) mv FPC A B0 = (28 ± 2) mv FPC A A8 > A B0 21st July 2017 Silvia Tedesco 20

21 Conclusions There results allowed to validate: The FPC layout by INFN TO The FPC produced by GS Swiss as components of the ALICE ITS upgrade To develop a test setup for data attenuation studies At present the HS on which I made the measurements is used to set up the first prototype of an Outer Barrel Stave. 21st July 2017 Silvia Tedesco 21

22 Thanks for the attention!!! 21st July 2017 Silvia Tedesco 22

23 Backup slides 21st July 2017 Silvia Tedesco 23

24 Pointing resolution Pointing resolution (mm) ALICE Current ITS, Z (Pb-Pb data, 2011) Upgraded ITS, Z Current ITS, rj (Pb-Pb data, 2011) Upgraded ITS, rj With the upgrade pointing resolution improved by a factor 3 in rϕ direction and a factor 5 in z direction p (GeV/c) T 21st July 2017 Silvia Tedesco 24

25 Eye width Clock lines (side A8 and B0) Control and clock (side A8) In both cases: Vias have no effect on data transmission. There are no edge effects on data transmission. 21st July 2017 Silvia Tedesco 25

26 Data lines Analysis of transmission quality on data lines for 1 and 2 FPCs. Eye height decreases with the number of FPCs. 21st July 2017 Silvia Tedesco 26

27 External cables contribution Analysis of data transmission along SMA cables of two different length: 1 and 2 metres. Fluctuation are due to experimental setup. 21st July 2017 Silvia Tedesco 27

28 Oscilloscope signal Signals for modules in position 6 and 7 side A8. There is a master malfunction in both cases: modules excluded from the measurement. 21st July 2017 Silvia Tedesco 28