Development of a CMOS pixel sensor for embedded space dosimeter with low weight and minimal power dissipation

Size: px

Start display at page:

Download "Development of a CMOS pixel sensor for embedded space dosimeter with low weight and minimal power dissipation"

Laurel Stevens
5 years ago
Views:

1 Development of a CMOS pixel sensor for embedded space dosimeter with low weight and minimal power dissipation Yang ZHOU Ph.D. thesis defense 23 th of September 2014

2 Outline Challenges in space radiation counters Space radiation environment Past generation devices & the main challenges How CMOS pixel sensors work and potentially cope with the challenges New concept realization Requirements translation from measurement to sensor design Sensor design Concept validation Tests of the prototype Embedded smart digital process performances Conclusion Thesis contributions 1

Part 1: challenges in space radiation counters Earth orbit radiation environment (radiation belt region) Why monitoring this radiation environment? Recommendations for spacecraft design (eg.

3 Part 1: challenges in space radiation counters Earth orbit radiation environment (radiation belt region) Why monitoring this radiation environment? Recommendations for spacecraft design (eg. surface damage, solar cell damage) Improvement of mission planning (avoid the high radiation region) Safety concerns for human space habitation and exploration (< 25 Rems per mission) Various particle species: electrons, protons, X rays, various heavy ion species (~1%); What should be measured? electrons & protons High flux density: E - : 10 4 several 10 7 particles/cm 2 /s P + : particles/cm 2 /s Large energy range: E - : 100 kev 10 MeV; P + : 100 kev 400 MeV; 2

Part 1: challenges in space radiation counters Earth orbit radiation environment (radiation belt region) Monitors in current use: Scientific payloads: Good particle measurement capability Large mass

4 Part 1: challenges in space radiation counters Earth orbit radiation environment (radiation belt region) Monitors in current use: Scientific payloads: Good particle measurement capability Large mass (>> 1kg) and power requirements (>> 1W) Small support instruments: Limited functionality (dose rate) and offer little or no particle identification Why small scale monitors? Spatial resoultion of this region requires a lot of monitors working on different orbits Particles species, energy & intensity vary with different orbits 3

5 Part 1: challenges in space radiation counters Earth orbit radiation environment (radiation belt region) Why particle identification is important? Radiation effects on biology and spacecraft devices strongly depend on particle species and energy Particle Energy Typical Effects Proton electron 100 kev 1 MeV 1 10 MeV MeV > 50 MeV kev >100 kev >1 MeV Surface damage; Surface material & solar cell damage; Radiation damage (both ionizing and nonionizing); Background in sensors; Single-event effects; Nuclear interaction-caused background; Surface electrostatic charging; Deep-dielectric charging Background in sensors Solar cell damage; Radiation damage (ionizing); A small & accurate instrument: suitable for widespread use on satellites in Earth orbit could open new prospects for radiation monitoring in space 4

Part 1: challenges in space radiation counters Past generation small-scale monitors in space Geiger counter: (gaseous detector) on the

Scintillating Fibre Detector (SFD): on the EQUATOR-S satellite in 1997 Compactness (397 g, 332 cm 3 and 105 mw) Dose rate measurement No

6 Part 1: challenges in space radiation counters Past generation small-scale monitors in space Geiger counter: (gaseous detector) on the Explorer-1 satellite in 1958 First radiation detector used in space; No particle species identification Very low count rate The Scintillating Fibre Detector (SFD): on the EQUATOR-S satellite in 1997 Compactness (397 g, 332 cm 3 and 105 mw) Dose rate measurement No single-particle pulse measurement The Standard Radiation Environment Monitor (SREM): Reliable operation: 7 flight missions (STRV-1c, PROBA 1, Integral, GIOVE B, etc.) since 2000 to 2010 Single-particle pulse analysis; good spatial resolution of the radiation belts Flux limitation: /cm 2 /s << the highest flux rate in the radiation belt; The particle species and energy identification is probabilistic only 2.6 kg, 2400 cm 3 and 2.5 W 5

7 Part 1: challenges in space radiation counters Main challenges Main challenges: High flux measurement: ~ 10 7 Part./cm 2 /s Single particle species and energy identification; Small-scale device: mass, power; Why pixelated detectors? Higher flux cope capability; Lower operation frequency required (less power); May provide additional method for particle identification (cluster analysis) Merging impacts Individual impacts Non-segmented Segmented 6

Part 1: challenges in space radiation counters Pixel sensors Charge Coupled Devices

Hybrid pixel sensors: Excellent radiation tolerance High complex and laborintensive

radiation tolerance (> 100 krad) Low fabrication cost Highly integrated: on-chip

8 Part 1: challenges in space radiation counters Pixel sensors Charge Coupled Devices (CCD): Uneconomical Complexity of using No processing inside the sensor possible Hybrid pixel sensors: Excellent radiation tolerance High complex and laborintensive chip bonding process ~ 300 µm-thick sensing volume CMOS Pixel Sensors (CPS): Good radiation tolerance (> 100 krad) Low fabrication cost Highly integrated: on-chip signal processing ~10 15 µm-thick sensing volume: less electron scattering, better energy resolution 7

Part 1: challenges in space radiation counters CMOS pixel sensor working principles and advantages Basic advantages for particle detection: Very sensitive: excellent noise performance (could be ~10 e

9 Part 1: challenges in space radiation counters CMOS pixel sensor working principles and advantages Basic advantages for particle detection: Very sensitive: excellent noise performance (could be ~10 e -, SNR > 20 for MIP) High detection efficiency: sensitive; 100% fill-factor; almost dead time free; Why CPS? High flux cope capability: high granularity (10 10 µm 2 possible); fast readout speed (~10 k frames/s) Good particle identification potential: very thin less deposited energy fluctuation Small-scale device: on-chip signal processing Radiation tolerance: >100k Rad good enough based on the ESA project requirement less e - scattering Higher measurable precision 8

Part 1: challenges in space radiation counters On-going project with CMOS pixel sensor Highly Miniaturized Radiation Monitor (HMRM): A CMOS pixel Sensor based detector Identify

Highly miniaturized: 15 cm 3, 52 g; HMRM designed by STFC (Rutherford Appleton Laboratory), ESA & Imperial College London, currently undergoing integration on the TechDemoSat-1

10 Part 1: challenges in space radiation counters On-going project with CMOS pixel sensor Highly Miniaturized Radiation Monitor (HMRM): A CMOS pixel Sensor based detector Identify particle fluxes: up to 10 8 particles/cm 2 /s (expected pile-up probability < 5%); Particle identification principle: based on the particles' different penetration capability; Highly miniaturized: 15 cm 3, 52 g; HMRM designed by STFC (Rutherford Appleton Laboratory), ESA & Imperial College London, currently undergoing integration on the TechDemoSat-1 spacecraft Sensor main specifications: pixel array; 20 µm pixel pitch; Sensitive area: 1 mm 2 ; Frame rate: 10, 000 fps; Column-parallel 3-bit ADC; Digital output; Data processing: in a FPGA 9

architecture 1. How to perform the complex treatment without compromising sensitivity?

11 Part 1: challenges in space radiation counters Proposed new concept: cluster analysis This study: further exploiting the full potential of a single CMOS monolithic pixel sensor Shield Difficulties: HMRM architecture 1. How to perform the complex treatment without compromising sensitivity? (Sensitivity & saturation; identify strategy & measurement precision; speed & power dissipation ) 2. The embedding of a smart digital process 10

12 Part 2: New concept realization

Challenge 1: high flux Sensitive area, granularity & operation speed Sensitive area (drives the flux estimation speed and accuracy): 10 mm 2 sensitive area 10 3 part.

13 Challenge 1: high flux Sensitive area, granularity & operation speed Sensitive area (drives the flux estimation speed and accuracy): 10 mm 2 sensitive area 10 3 part./cm 2 /s: 1 s with 10 % relative uncertainty; Higher flux: better accuracy or within a shorter time; Granularity & operation speed (drive the hits pile-up): < 0.1% for 10 7 part./cm 2 /s; assuming electrons trigger < 3 3 pixels 20 µm pitch size, 100 µs/frame 50 µm pitch size, 20 µs/frame? More specifications: o Pixel signal range; o ADC bits; o Embedded algorithm 4 steps of simulation 11

14 Challenge 2: particle identification Requirements from measurement to sensor design Step 1: signal generated: Step 2&3: pixel matrix response Limit the particle incident angle Reduce the deposited energy fluctuation Signal over pixels; Digitization: 3-bit; Step 4: cluster analysis Particle classification: by cluster ADC counts Interval containing: 68%, 95% of the values Cluster ADC counts Particle species >71 p + < 2 MeV protons 70 [21,70] p + : [2 MeV, 30 MeV] electrons 50 MeV Monte Carlo simulation results (obtained using GEANT4) 20 8 [9,20] p + : [30 MeV, 50 MeV] [1,8] e - & p + > 50 MeV 12

15 Challenge 3: highly integrated and smart detection system Specifications for each block of the sensor design Pixel: Speed: Sensing diode fast reset and recovery; Low noise: sensitive to MIP; linear response: for downstream digitization Signal range: up to ~ 4400 e - ADC: Column-level; 3-bit resolution Tunable threshold: for detection efficiency Proposed CMOS Pixel Sensor architecture developed through MIMOSA IPHC Intelligent digital processing: Cluster separation and counting 13

16 Challenge 3: highly integrated and smart detection system Specifications for each block of the sensor design General principles: No temperature sensitive blocks (eg. Logarithm amplifier/adc) Low system power dissipation and less complexity o Rolling shutter mode: one row at a time; o Switch off amplifiers/adcs when not using o Add one extra bit for the ADC output to simplify the digital processing block Proposed CMOS Pixel Sensor architecture developed through MIMOSA IPHC The sensor: Rolling-shutter mode: <315 ns/row Functionalities: particle identification and counting Output: Low data rate 14

17 Challenge 3: highly integrated and smart detection system Design of the pixel: active reset & in-pixel offset cancellation Pixel schematic Key issues for pixel design: Speed Low noise linear response Signal range Signal processing chain in pixel: Active reset: speed Pre-amplify stage: low noise Double sampling: low noise Source follower: speed Timing diagram: 240 ns (< 315 ns) for one pixel 15

Challenge 3: highly integrated and smart detection system Design of the pixel: key points Signal transmission: CS stage SF stage To column The Common Source

transistors: should remain in linear region for a fast reset Linear response in the useful signal range Pixel schematic NMOS transistors only: PMOS competes

18 Challenge 3: highly integrated and smart detection system Design of the pixel: key points Signal transmission: CS stage SF stage To column The Common Source (CS) amplifier simple architecture (low noise) Key issues for pixel design: Noise and speed Capacitor value: trade-off between noise and operation speed Reset transistors: should remain in linear region for a fast reset Linear response in the useful signal range Pixel schematic NMOS transistors only: PMOS competes for charge collection. Pre-amplification stage: trade off between gain& linearity in the range SF stage: stable gain in the signal range match with CS output; 16

19 Challenge 3: highly integrated and smart detection system Design of the ADC: Successive approximation ADC Chosen architecture: SAR ADC Low power dissipation: only one comparator with 3 or 4 comparisons are enough for the whole conversion Expected transfer function of the ADC Specifications: Tunable threshold: adjust for detection efficiency 240 ns/sample: match with pixel signal processing Proposed ADC architecture: based on the SAR logic 17

20 Challenge 3: highly integrated and smart detection system Design of the ADC: sample & hold block 4 capacitors, an output buffer & 6 switches Timing diagram of the sample and hold circuit Key design issues: Signal subtraction: obtain the signal amplitude from pixel double sampling Speed: followed by capacitor arrays Free from offset The architecture with 3 steps operation: Operation Step 1&2: signal subtraction Operation Step 3: offset cancellation Small value C H1 : fast sampling (speed) Unit gain buffer: fast readout (speed) 18

Challenge 3: highly integrated and smart detection system Design of the ADC: first comparison & power efficient design Timing diagram for the first comparison Based on charge redistribution principle

21 Challenge 3: highly integrated and smart detection system Design of the ADC: first comparison & power efficient design Timing diagram for the first comparison Based on charge redistribution principle Key design issues: Power efficient design: An additional comparison before typical SAR cycle Turn-off ADC (signal value < threshold) Tunable threshold & expected transfer function: 4 external references Appropriate timing 19

Challenge 3: highly integrated and smart detection system Design of the ADC: high precision comparator Offset cancellation Tracking Latching Key design issues: High precision:

22 Challenge 3: highly integrated and smart detection system Design of the ADC: high precision comparator Offset cancellation Tracking Latching Key design issues: High precision: Output Offset Storage (OOS) technique, One additional mode than typical tracking + latching Low kickback noise: Two buffers are added in the differential inputs of the amplifier 20

23 Challenge 3: highly integrated and smart detection system Design of the ADC: amplifier & latch in the comparator Amplifier: Dynamic latch: 21 Key design issues: Amplifier: First stage: gain; second stage: speed & kickback noise reduction Dynamic Latch: good speed, no-static power dissipation

by row Correct functionalities Shutter readout row by row Sum of rows 0 0 0 0

0 1 7 2 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 5 0 0 0 0 20 Clusterization out

24 Challenge 3: highly integrated and smart detection system Design of the digital processing stage: targets Key design issues: Real-time processing row by row Correct functionalities Shutter readout row by row Sum of rows Clusterization out Sum ADC counts in each cluster Separation Counting Cluster trimming driven by zeros 22

25 Challenge 3: highly integrated and smart detection system Design of the digital processing stage: principle Reading row 2: 240 ns Reading row 4: 240 ns Reading row 3: 240 ns Step 1 Step 2 Step 3 23

26 Challenge 3: highly integrated and smart detection system Design of the digital processing stage: schematic 24

27 Challenge 3: highly integrated and smart detection system Design of the digital processing stage: layout Layout of the digital processing stage (synthesized in a 0.35 µm process) Realization: Algorithm described using Verilog; Size: mm 2 Control timing clock frequency: 25 MHz Power dissipation: 56.7 mw (3.3 power supply) 25

3 V) o Operation speed: ~ 65, 000 frames/s o Output rate: 80 bps Layout of the propose sensor: designed in a

28 Challenge 3: highly integrated and smart detection system The full size sensor o Core area: 16.8 mm 2 o Power dissipation: ~100 mw (Vdd = 3.3 V) o Operation speed: ~ 65, 000 frames/s o Output rate: 80 bps Layout of the propose sensor: designed in a 0.35 µm process, with 14 µm thick, high resistivity (1k ohm-cm ) epitaxial layer, and named as COMETH* *: COunter for Monitoring the Energy and Type of charged particles in High flux 26

29 Part 3: concept validation

CVF (charge to voltage conversion factor); CCE (charge collection efficiency); Electrons source: validating the simulated sensor

30 Part 3: concept validation Tests of the prototype Pixel matrix: illuminated with 3 types of radiations Monochromatic X-rays: calibration. CVF (charge to voltage conversion factor); CCE (charge collection efficiency); Electrons source: validating the simulated sensor response with electrons Verify the CCE obtained in X-ray tests; Reduced scale prototype with pixels & 32 column ADCs Laser illumination: validating the pixel design Linearity; Range; Speed; Protons illumination test: lack of appropriate source ADCs 27

31 Part 3: concept validation 55 Fe radioactive source: 5.9 kev X-rays, 1640 e - CCE (seed pixel): 115/364 = (31.6±0.5)% CCE (3 3 pixels): 323/364 = (88.7±0.5)% Satisfactory high CCE for a pixel sensor with 50 µm pitch size Calibration peak: 1640 e - CCE (5 5 pixels): 343/364 = (94.3±0.5)% Seed pixel calibration peak: 364 ADC unit; 4.5 e - /ADC unit; CVF: ~33 µv/e - 28

simulated sensor response to electrons: small triggered cluster size Threshold (determines the

32 Part 3: concept validation 90 Sr source: e - energy spectrum end point being at 2.3 MeV Seed pixel MPV (Most probably value): 371±2 e - ~ % Threshold Confirmed the simulated sensor response to electrons: small triggered cluster size Threshold (determines the detection efficiency): 120 e - for 99.95% Most of the clusters contain: 2 to 4 pixels > threshold; 9 cluster 29

33 Part 3: concept validation Estimated from CDS value variation without any source Events For the smallest signal in this application: electrons ENC: 30 e - ; SNR (for electrons): 13 (MPV) 30

34 Part 3: concept validation Pulsed laser illumination: 1603 nm infrared Laser illumination test: Linear response range: 0 to 4600 e - ; (satisfy the expectation 0 to 4400 e - ) Sensing diode fast reset and recovery: confirmed though no remaining signal observed after large intensity illumination. 31

Part 3: concept validation ADCs test: External voltage biases replacing the pixel outputs Single ADC transfer function Threshold tunable to meet the detection efficiency 1LSB = 700 e - = 700 e - CVF

35 Part 3: concept validation ADCs test: External voltage biases replacing the pixel outputs Single ADC transfer function Threshold tunable to meet the detection efficiency 1LSB = 700 e - = 700 e - CVF = 23.1 mv Noise: INL and DNL for a single ADC: < ±0.12 LSB Temporal noise (rms): 0.02 LSB; Fix pattern noise between 32 columns (rms): 0.21 LSB (need to be improved in next version) Power efficient design: With signal: 759 µw/adc; Without signal: 532 µw/adc (most of the cases); 32

Part 3: concept validation Digital processing stage: reconstruction performances Deposited energy reconstruction performances: Based on the reconstructed cluster ADC counts, with 10% standard

36 Part 3: concept validation Digital processing stage: reconstruction performances Deposited energy reconstruction performances: Based on the reconstructed cluster ADC counts, with 10% standard deviation Layout of the digital processing stage Cluster reconstruction performances: (checked by the full simulation with the layout): Success cases: Normal convex clusters; Tricky clusters with one dead pixel; Most of two dead pixels cases; Failure cases: (marginal) Two adjacent dead pixels in a cluster center; One or both of the two adjacent dead pixels locates in the middle of the last row of a cluster Reconstructed energy (kev) Deposited energy (kev) 33

37 Part 3: concept validation Summary of the concept validation Performances have been validated to ensure the expected capability: High flux capability: (Fully checked) Electron triggered clusters size are small Designed operation speed; sensing diode fast reset and efficient recovery Particle identification: (Mostly checked) Tested sensor responses with electrons and protons match with the simulations ( cluster size, cluster ADC counts); Circuits function as the design specifications: pixels; ADC; Digital processing stage 34

38 Part 4: conclusions

Part 4: conclusions Conclusions COMETH COMETH: Handle with high flux: up to 10 8 particles/cm 2 /s for electrons (hits pile-up < 5%); Single particle identification capability: Electrons & protons >

39 Part 4: conclusions Conclusions COMETH COMETH: Handle with high flux: up to 10 8 particles/cm 2 /s for electrons (hits pile-up < 5%); Single particle identification capability: Electrons & protons > 50 MeV Protons: [30 MeV, 50 MeV] Protons: [2 MeV, 30 MeV] Protons < 2 MeV Highly integrated, miniaturized, low power/mass/data rate, smart detection system: no signal treatment power aside the sensor required Sensor area < 20 mm 2 ; ~ 100 mw power dissipation; 80 bps output data rate: no data transmission stress for the satellite; Differentiation beyond this energy could achieved by a strategy of exploiting several sensors equipped with various shields. A very compact monitor could be foreseen High level output information: directly output particles species/energies/flux information; 35

40 Part 4: conclusions Thesis contributions Contributions of this thesis: Developed & validated a new concept for single particles identification with CPS: Pulse measurment Multiple sensors Cluster analysis Measurable energy range Low High High Measurable flux rate Low High High Monitor archiecture Simple complex Simplest Monitor scale (mass, power) Large medium small Developed a CPS architecture from particle tracking to energy measurement: By sacrificing the hit position information The embedding of a smart digital process: Significantly reduced the complexity and scale of a monitor 36

41 Publications & communications during this study Publications: Y. Zhou, J. Baudot, Ch. Hu-Guo, Y. Hu, K. Jasskelainen and M. Winter. COMETH: a CMOS pixel sensor for highly miniaturized High-flux radiation monitor, Proceedings of Science (PoS) Y. Zhou, J. Baudot, C. Duverger, Ch. Hu-Guo, Y. Hu and M. Winter, CMOS Pixel Sensor for a Space Radiation Monitor with very low cost, power and msss, 2012 JINST 7 C Communications: Y. Zhou, J. Baudot, Ch. Hu-Guo, Y. Hu, K. Jaaskelainen and M. Winter. COMETH: a CMOS pixel sensor for a highly miniaturized high-flux radiation monitor, Oral presentation at the Technology and Instrumentation in Particle Physics (TIPP) conference 2014, 2-6 June, Amsterdam, Holland. Y. Zhou, J. Baudot, C. Duverger, Ch. Hu-Guo, Y. Hu and M. Winter, Development of a CMOS Pixel Sensor for Space Radiation Monitor, Poster at l école IN2P3 de microélectronique 2013, June Porquerolles, France. Y. Zhou, J. Baudot, C. Duverger, Ch. Hu-Guo, Y. Hu and M. Winter. CMOS Pixel Sensor for a Space Radiation Monitor with very low cost, power and mass, Oral presentation at the 14TH International Workshop on Radiation Imaging Detectors (IWORID2012), 1-5 July Figueira da Foz, Coimbra, Portugal. 37

Radiation dose in the radiation belts The highest recommended limit for radiation exposures is for astronauts-25,000 millirems per Space Shuttle

44 Radiation dose in the radiation belts The highest recommended limit for radiation exposures is for astronauts-25,000 millirems per Space Shuttle mission. 85% protons in space <10 MeV 0.7 mm aluminum can shield e - < 0.5 MeV and p < 10 MeV 2 mm aluminum can shield e - < 1.5 MeV and p < 20 MeV 44

Background Introduction Detector options Conclusion Those effects which do exist but be ignored (1/2) X ray 100 10 1 wavelength (A) Only hard X-ray: 0.

45 Background Introduction Detector options Conclusion Those effects which do exist but be ignored (1/2) X ray wavelength (A) Only hard X-ray: 0.1 A 1 A (12 kev 120 kev) can penetrate the shielding GOES X-ray satellite data (35,800 km above the Earth) Considering the efficiency of 62.5 kev X-ray in the sensor is lower than 1%, that means every 100 seconds, X-ray deposit 62.5 kev energy in the sensor. That is marginal. 45

Background Introduction Detector options Conclusion Those effects which do exist but be ignored (2/2) Proton nuclear reaction (rare) depends on the energy and range.

46 Background Introduction Detector options Conclusion Those effects which do exist but be ignored (2/2) Proton nuclear reaction (rare) depends on the energy and range. Still do not have exact data to support, but the probability would be less than 1% Various heavy ion species (~1%) *J. E. Mazur. An Overview of the Space Radiation Environment. Crosslink, Volume 4, Number 2 (Summer 2003) 46 Nuclear reaction *A.B. Rosenfeld, et al,. A New Silicon Detector for Microdosimetry Applications in Proton Therapy, IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 47, NO. 4, AUGUST 2000

47 Model for signal over pixels *PSF: Point Spread Function; was generated from the test results of MIMOSA 5, 17, 18, 22, 24 and LUCY with the same collecting diode size; 47

48 Dead time: 100 ns/ 240 ns/64 = 6.5%

49 Pixel-level offset cancellation 49

50 Normal pixel noise Noise pixel with RTS noise 50

51 51 Pedestal

52 Challenge 1: high flux Sensitive area, granularity & speed Probability of misjudgment by hits pile-up Flux estimation: 10 mm 2 sensitive area Frame time (speed) Pixel pitch Cluster size in pixels P(N 2) 100 µs 20 µm % % 50 µm % % 50 µs 50 µm % % 20 µs 50 µm % % % 10 3 part./cm 2 /s 1 s 100 s 10 7 part./cm 2 /s 0.1 ms 10 ms 7

53 Challenge 2: single particle identification Sensor response simulations: strategy performances Step 4: Check both the high-flux cope and particle identification (by the deposited energy reconstruction) capability Reconstruction capability : Flux: /cm 2 /s; (not an absolute limit of this architecture) Energy: a relative value of 10% standard deviation; Number of particles reconstructed Number of particles hit the sensor/frame Electrons and protons (50/50) with mixed energies (1 MeV to 100 MeV) Conclusion of the simulations: High flux counting capability: Reconstruct flux (particles/frame) Small cluster size triggered by electrons: 2 to 3 pixels in average;. Reconstructed energy (kev) Reconstruct Energy Deposited energy (kev) Based on the various cluster ADC counts Single particle identification capability: particles could be discriminated by their triggered cluster ADC counts Protons triggered cluster size is inversely proportional to its energy. (low energy protons possibly trigger 9 9 pixels) 9

54 Accumulated charges_cometh 54

55 Column FPN channel charge injection: 55

56 Part 4: conclusions Solutions to reduce the ADCs column FPN Column FPN (rms value) ~ 147 e - Idea 1: compensating the offset error by additional references: Idea 2: a switch multiplexer and 8 references for all the columns Column FPN (rms value) could be reduced to: Idea 1: ~ 7.5 e - ; Idea 2: ~ 15 e - ; Employs a 6-bit trimming DAC, each ADC has its own trimming register Architecture 2: especially attractive to low resolution, column ADCs for small size sensors 56

57 Offset cancellation in S/H circuit 57

58 Challenge 2: single particle identification Logarithm amplifier example Logarithm amplifier example: Id = Is (e (Vd/Vt) -1) Ir = Id Ir = Vin/R Vin/R = Id Vin/R = Is e (Vd/Vt) Vout = -Vt In(Vin/IsR) Vt: thermal voltage 58

59 Digital processing stage: estimated power Signal column mw Block mw Memory_temp FSM1_N Adder_Column Adder* SUM_Ctroller FSM2_N System mw Counter* Accumulator* Whole System = ( ) mw 59

60 Timing for the digital processing stage Clock_ADC, Clock1, Clock1_delay1/2/3: pulse width 120 ns, period 240 ns, frequency 4.17M Hz; Clock2: 6 times speed of clock1; pulse width 20 ns, period 40 ns, frequency 25M Hz; Chosen depends on the largest cluster size 9 N (Column Row); Clock3: Flexible Not scaled 60

61 Cluster reconstruction examples Cluster merges Failure cases 61

COMETH: a CMOS pixel sensor for a highly miniaturized high-flux radiation monitor

COMETH: a CMOS pixel sensor for a highly miniaturized high-flux radiation monitor Yang Zhou 1, Jérôme Baudot, Christine Hu-Guo, Yann Yu, Kimmo Jaaskelainen and Marc Winter IPHC/CNRS, Université de Strasbourg