STATUS OF MEDIPIX-3, PLANS FOR TIMEPIX-2
X. Llopart
-2- X.Llopart
Medipix 3 - reminder
• Medipix3 builds on the success of Medipix2 as a single photon counting imaging chip
• Added Features– Analogue charge summing to keep all charge information– Spectroscopic mode with 8 threshold levels– Continuous readout mode (no dead time)– Increased counter depth increasing dynamic range– Increased readout speed– Increased radiation hardness from 130nm CMOS process
-3- X.Llopart
Status
• First engineering run (12 wafers of 100 chips) delivered early this year
• Wafer probing complete
• One wafer diced and bonded to PCBs
• Using the IC Tester data transfers speeds of 1.6 Gb/s were achieved
• Initial readout system working at low speed (USB from Prague)
• Electrical characterisation is almost completed
• Medipix3 readout integration in Pixelman is underway
• First bump bonded assemblies with wafers expected soon (~1 month)
17.3
mm
14.1 mm
-4- X.Llopart
Electrical Characterisation Status
• Almost everything works as designed:– Complex pixel functionality: Charge summing, 8 independent thresholds,
programmable counter depth– Expected ENC and minimum threshold has been measured as expected
in all the different pixel modes
• Need to confirm the electrical characterisation with Medipix3 Si assemblies
• We found 2 issues with the chip– Cas and FBK DACs are not able to supply the designed nominal voltage– Continuous Read Write is not working correctly
-5- X.Llopart
The “analog switches” (problem)
Digital
Vin
R=100K
Vout
!EnablePixelCom
EnablePixelCom
EnablePixelCom
VSSA
Cascode DAC (Vcas) x65536 Pixels
75pA
75pA5A
5A
10A FS
Due to radiation and temperature an increase in the leakage current of NMOS transistors is observed
-6- X.Llopart
Temperature Measurements Setup
• The Medipix3 has an on-chip temperature sensor in the chip periphery
• At CERN we have available the Climatic Chamber ESPEC EGNX12-6CWL (-70 °C to +180 °C)
• Measurements covered the -60 °C to 60 °C range
y = 1.0555x + 24.173R² = 0.9999
-60
-40
-20
0
20
40
60
80
100
-60 -40 -20 0 20 40 60Tem
p M
edip
ix3
[ºC]
Temp In Climate Chamber [ºC]
-7- X.Llopart
Vcas and Vfbk temperature sensitivity
• Vcas is pulled low since the current leakage path goes to GND• Vfbk is pulled high• The consequence of this is effect is that the operational biasing of
the front end is not stable in temperature and it must be controlled carefully
0
0.2
0.4
0.6
0.8
1
1.2
1.4
-60 -40 -20 0 20 40 60
DAC
Out
put [
V]
Temp In Climate Chamber [ºC]
Cas @ 0xF8
Fbk @ 0x00
-8- X.Llopart
Temperature effects in one pixel
• The chip is operational in all the studied temperature range• But, we observe variations on the noise, threshold position and gain
0 100 200 3000
1000
2000
3000
4000
5000
Pixel at 60CPixel at 20CPixel at -40C
THL DAC
Cou
nts
-9- X.Llopart
Pixel to pixel gain variation
• Gain variation (0.086 e-/THL DAC step) is due to the Medipix3 shaper stage
y = 0.0861x + 31.816R² = 0.9552
0
5
10
15
20
25
30
35
40
45
-60 -40 -20 0 20 40 60
Gai
n [e
-/TH
L D
AC st
ep]
Temp In Climate Chamber [ºC]
-10- X.Llopart
Overall pixel noise variation
• Noise is increasing at low temperature probably due the change in front end bias voltages Vcas and Vfbk
0
20
40
60
80
100
120
140
160
-60 -40 -20 0 20 40 60
ENC
[e-]
Temp In Climate Chamber [ºC]
-11- X.Llopart
Overall threshold position shift
• Significant effect visible in the threshold position due to the variation on Vcas
• When the shaper’s pol-zero cancellation circuit is activated the measured temperature sensitivity is 16.7 e-/°C and 3.7 e-/°C when it is switched off.
y = 0.0167x + 8.453R² = 0.9744
0
2
4
6
8
10
12
-60 -40 -20 0 20 40 60
Title
Title
-12- X.Llopart
400Mrad Irradiation
• Used a calibrated X-ray machine (Seifert RP149)• Beam profile is smaller than the Medipix3 → Two runs:
On the Pixel Matrix 60Mrad
Threshold Variation
Gain Variation
Noise Increase
On the Periphery 400Mrad
Check DACs
E-fuses
Logic functionality
60 MRad
460 MRad
400 MRad
-13- X.Llopart
Performance after 460MRad
Threshold Noise
Yie
ld a
rtifa
ct
b 190.733061= σ 35.304781= Fit 0.000026=
0 2 4 6 8 10 12 14 16 18 200
200
400
600
800
THL [ke-]
Cou
nts
σ=1.72 ke-
µ=9.3 ke-
b1 8.798851= σ1 1.591887= Fit1 0.000247=
0 20 40 60 80 100 120 140 160 180 2000
2000
4000
6000
8000
10000
noise[e-]
Cou
nts
σ=12.9 e-
µ=71.6 e-
Threshold can be re-tuned using 5 bit equalisation
-14- X.Llopart
0 15 30 45 60 75 90 105 120 135 1500
1000
2000
3000
4000
5000
Pixel Non-IrradiatedPixel Irradiated at 460 MRad
THL [DAC step]
Pix
el c
ount
s
Performance after 460MRad
Yie
ld a
rtifa
ct
Qin=2ke-
0 100 20040
60
80
100
Row Number
Noi
se [
e-]
Row
[0:2
55]
After 460MRad there is essentially no gain variation observed
-15- X.Llopart
DAC Measurement
0.5
0.525
0.55
0.575
0.6
0.625
0.65
0.675
0.7
0 10 20 30 40 50 60 70 80 90 100 110
Volt
age
[v]
Hours (1hour→ 4Mrad)
Band Gap
Accumulated dose of 396 Mrad
0.3
0.325
0.35
0.375
0.4
0.425
0.45
0.475
0.5
0 10 20 30 40 50 60 70 80 90 100 110
Volt
age
[v]
Hours (1hour→ 4Mrad)
Accumulated dose of 396 Mrad
9mV
NMOS DAC (Preamp)
0.95
0.975
1
1.025
1.05
1.075
1.1
1.125
1.15
0 10 20 30 40 50 60 70 80 90 100 110
Volt
age
[v]
Hours (1hour→ 4Mrad)
Accumulated dose of 396 Mrad
33mV
PMOS DAC
For this technology there worst threshold shifts due to radiations happen ~3Mrad
-16- X.Llopart
3MRad irradiation
3 MRad
Worst effect at 3MRad, After this the effect of further radiation is much
less.
Irradiated a small area of the chip up to 3Mrad to try and keep Cas DAC
working by only damaging a limited number of pixels
-17- X.Llopart
General behavior after 3 Mrad
0 3 6 9 12 15 18 21 24 27 300
200
400
600
THL [ke-]
Cou
nts
Threshold Noise
b1 4.928843= 1 0.991838= Fit1 3508.489764=
0 20 40 60 80 100 120 140 160 180 2000
5000
10000
15000
noise[e-]
Cou
nts
Yie
ld a
rtifa
ct
-18- X.Llopart
Overall threshold position variation
0 100 200180
200
220
240
260
280
300
Row Number
TH
L [
DA
C s
teps
]
Main contributor is the protection diode. In order to minimize the pixel to pixel threshold variation Ikrum is set to 16nA which indicated that the leakage of this diode at 3Mrad is ~5 to 10 nA. This is the worst radiation operation point.
Row
[0:2
55]
-19- X.Llopart
3MRad Noise and Gain Result
Yie
ld a
rtifa
ct
0 100 20040
60
80
100
Row Number
Noi
se [
e-]
0 15 30 45 60 75 90 105 120 135 1500
1000
2000
3000
4000
5000
Pixel Non-IrradiatedPixel Irradiated at 3 MRad
THL [DAC step]
Pix
el c
ount
s
Qin=2ke-
Row
[0:2
55]
In this worst case situation the noise is
still less than 100e- and the gain variation is still
minimal
-20- X.Llopart
Wafer Probing Summary
Name # AA Bs Cs Ds E F # Chips Where
VK7DG2H 0 100 CERN/diced
VT7DCWH 1 27 10 11 45 3 4 100 CERN
VT7DEVH 2 71 17 2 2 6 2 100 Bump bonding
VV7DFAH 3 33 18 6 32 7 4 100 CERN
VN7DE0H 4 57 18 2 15 8 0 100 Bump bonding
VU7DCVH 5 18 5 1 63 10 3 100 CERN
VW7DF9H 6 36 18 4 30 6 6 100 CERN
VQ7DCZH 7 61 18 2 8 6 5 100 Bump bonding
VS7DCXH 8 6 0 1 84 9 0 100 CERN
VU7DDCH 9 47 21 2 16 12 2 100 CERN
VL7DFJH 10 33 19 7 36 5 0 100 CERN
VW7DDAH 11 48 22 5 15 9 1 100 CERN
437 166 43 346 81 27 1100
39.7% 15.1% 3.9% 31.5% 7.4% 2.5% 100.0%
-21- X.Llopart
Best and worst wafer
-22- X.Llopart
Conclusions
• The electrical characterisation of the chip is basically completed
• By monitoring the chip temperature and the radiation effects through Vcas line the chip is practically insensitive to radiation large temperature fluctuations
• Only remaining issue is to understand the failure mechanism of the CRW mode
• Yield results are very encouraging given the complexity of such chip
• First Si assemblies will be available shortly and will be use to confirm electrical measurements
-23- X.Llopart
Towards Timepix2
• In the Medipix3 collaboration there is a growing interest in Timepix2
• It might be possible to fund this development with the Medipix3 funds
• Main specs:– Pixel to measure TOT and Arrival time information simultaneously– <2ns time resolution– Triggered readout– Sparse and very fast readout
• Many building blocks used in Medipix3 can be reused for a new chip
• Start of design depends on available experienced man-power