triroc!asic! - indico.cern.ch
TRANSCRIPT
TRIROC ASIC
TIPP 2014, Amsterdam 5 June 2014
Salleh AHMAD
The research leading to these results has received funding from the European Union Seventh Programme under grant agreement n° 602621
Christophe DE LA TAILLEa, Julien FLEURYb, Nathalie SEGUIN-‐MOREAUa,Ludovic RAUXa,
Stéphane CALLIERa, Gisele MARTIN CHASSARDa
Frederic DULUCQa, Damien THIENPONTa
aOMEGA/IN2P3/Ecole Polytechnique bWeeroc SAS
Weeroc and Omega
• Weeroc is a spin-‐off company from OMEGA lab • CEO: Julien Fleury • Weeroc addresses industrial needs for
microelectronics in Aerospace, Medical imaging, Scien[fic instrumenta[on, Homeland security, Nuclear industry …
• Weeroc and OMEGA (13 microelectronics engineers) provide : – off-‐the-‐shelf FE ASIC (the ROC chip family) – customer-‐specific ASICs – Services, Audit, Exper[se
Research Ins[tute
Educa[on University
Industry company
• OMEGA : formerly a microelectronics group of LAL ORSAY, became an independent lab (IN2P3/CNRS/Ecole Polytechnique) in June 2013. Located in Palaiseau, directed by Christophe de La Taille
Weeroc : www.weeroc.com OMEGA: omega.in2p3.fr
2
TRIMAGE European Project
• Cost effec[ve tri-‐modality (PET – MR – EEG) imaging tools for schizophrenia: • Find new biomarkers and define a suitable mul[modal paradigm with already available
PET, MR, EEG and PET/MR systems that provides clinical evidence on the feasibility of advanced schizophrenia diagnosis.
• Construct and test an op[mized cost-‐effec[ve trimodality imaging instrument (brain PET/MR/EEG) for diagnosis, monitoring and follow-‐up of schizophrenia disorders.
• Validate the trimodal imaging device built by this Consor[um with regard to the results and the clinical data obtained from objec[ve 1.
3
A European Collaborative project supported through the
Seventh Framework Programme for Research and Technological Development
The problem Schizophrenia affects about 7 per 1000 of the adult population but
because the disorder is chronic the overall incidence is higher, at around
1% of the population. The cost per person with psychotic disorders is
close to 20.000€ per year on average. The earlier the treatment is
initiated, the more effective it is, however the majority of people with
schizophrenia do not receive treatment, which has the effect of
prolonging their illness.
TRIMAGE aims to create a trimodal, cost-effective imaging tool consisting of PET/MR/EEG using cutting edge technology with performance beyond the state of the art. The tool is intended for broad distribution and will enable effective early diagnosis of schizophrenia and possibly other mental health disorders.
TRIMAGE Objectives The main TRIMAGE S&T objectives are:
x Find new biomarkers and define a suitable multimodal paradigm with
already available PET, MR, EEG and PET/MR systems that provides clinical
evidence on the feasibility of advanced schizophrenia diagnosis.
x Construct and test an optimized cost-effective trimodality imaging
instrument (brain PET/MR/EEG) for diagnosis, monitoring and follow-up
of schizophrenia disorders.
x Validate the trimodal imaging device built by this Consortium with
regard to the results and the clinical data obtained from objective 1.
The Consortium The TRIMAGE consortium brings together 11 multi-
disciplinary partners from 5 European countries, and is
based on high-level scientific expertise from Universities,
Research Centres and SMEs.
Acknowledgment The TRIMAGE project is supported by the European Commission through the
Seventh Framework Programme for Research & Development.
The 4 year project will run from 1st December 2013 until 30th November 2017.
An optimised TRImodality (PET/MR/EEG) imaging tool for schizophrenia
Project Partners Role in the project
University of Pisa (UNIPI) Coordinator & PET system development
Technological Educational Institute of
Athens (TEIA)
Dissemination & Monte Carlo simulations
Forschungszentrum Juelich GmbH (FZJ) Coil design & PET/MR/EEG integration
JARA BRAIN, RWTH (JRB) Clinical application
Technische Universitat Munich (TUM) Image quantification & clinical application
University of Zurich (PUK) Patient recruitment & clinical data analysis
Istituto Nazionale di Fisica Nucleare (INFN) PET system development & characterization
AdvanSiD (ASD) SiPMs and chip-scale package development
WeeROC (WRC) PET modules production & testing
Raytest GmbH (RAY) Mechanical parts design & market strategy
RS2D (RS2D) Design, assembly, test 1.5T MRI
Contact Scientific Coordinator Prof. Alberto DEL GUERRA
University of Pisa, Italy
Exploitation Manager Jean-Luc Lefaucheur
Raytest, France
Project website www.trimage.eu
Dissemination Manager Ms. Theodora Christopoulou
TEI Athens, Greece
Project Office Project Trimage Office
University of Pisa, Italy
TRIMAGE Project Partners
4
Project Partners Role in the project
University of Pisa (UNIPI) Coordinator & PET system development
Technological Educa[onal Ins[tute of Athens (TEIA)
Dissemina[on & Monte Carlo simula[ons
Forschungszentrum Juelich GmbH (FZJ)
Coil design & PET/MR/EEG integra[on
JARA BRAIN, RWTH (JRB) Clinical applica[on
Technische Universitat Munich (TUM)
Image quan[fica[on & clinical applica[on
University of Zurich (PUK) Pa[ent recruitment & clinical data analysis
Is[tuto Nazionale di Fisica Nucleare (INFN)
PET system development & characteriza[on
AdvanSiD (ASD) SiPMs and chip-‐scale package development
Weeroc (WRC) PET modules produc[on & tes[ng
Raytest GmbH (RAY) Mechanical parts design & market strategy
RS2D (RS2D) Design, assembly, test 1.5T MRI
The TRIMAGE system
FP7-HEALTH-2013-INNOVATION-1 TRIMAGE Part B - Stage 2
12
Fig. 4: (left) Arrangement of the 54 staggered LYSO crystals to form the PET ring; (center) schematic illustration of the staggered crystal-detector assembly; (right) example of a 4x4 array of SiPMs integrated on a common compact substrate with minimum footprint and SiPM-to-SiPM spacing produced by ASD.
1.2.4.3 Progress in MRI technology The MR system will be based on a superconducting magnet with a homogeneous magnetic field of 1.5 T, as in most clinical studies performed to date. It will have high stability (better than 0.1 ppm/h), high homogeneity (better than 5 ppm) and 30 cm diameter of homogeneous spherical volume. The 5 Gauss line is roughly at 320 cm axially and 220 cm transaxially from the isocentre of the scanner making the room requirements less stringent than for standard clinical MR scanners. The MR magnet has a bore of 60 cm, an asymmetric gradient coil covering 25 cm axially, that allows the introduction of the shoulders of the patient in the full diameter of the bore and an optimized RF coil (8-channels parallel receiver) for head studies with an active B1 field with a 20 cm length and a 22 cm diameter. A key feature of the magnet is that it is cryogen-free making the system much more compact and cost-effective compared to standard MRI scanners. Also, the system does not require any special safety measures for handling cryogenic gas leaks as is required in cryogen cooled magnets in standard MRI scanners. All these characteristics will facilitate better physiological measures since the patient’s arm will be accessible outside the magnet. It will then be easier to measure the input function for PET as well as requesting simple motor tasks, e.g., finger tapping. To date, such a system does not exist and – to the best of our knowledge – is not being planned.
1.2.4.4 Progress in multimodal imaging technology The ultra-compact design of the superconducting magnet and the small installation requirements are critical factors for the cost-effectiveness and the wide availability of a dedicated system for human brain studies. The main drawback of fMRI is its dependence on the low temporal resolution BOLD effect. This restriction can be overcome by the simultaneous combination of fMRI and electrophysiology (Neuner et al, 2010). On the other hand, EEG is capable of measuring neuronal function at a millisecond time scale (Michel and Murray, 2012). Moreover, it is clear that PET assessment is also paramount in order to complete the pathophysiological frame of the disorder. It is not intended here to develop an EEG system from the ground up. Rather, building on the experience gained at the FZJ, a commercially available system that has already been tested in the hybrid PET/MR environment will be integrated into the PET/MR system proposed here. System integration will take care of synchronisation of the three modalities and the display of the data from them. As noted earlier, the integration of all of three modalities, PET, MR and EEG is per se a demanded step forward in this field. A design of the proposed integrated instrument is presented in Fig. 5.
Fig. 5: Dimensional outline (left) and artistic view (right) of the dedicated brain PET/MR/EEG system (the EEG cap is not shown).
5
PET ring : one module
6 Gamma Camera Module
PET ring : Gamma Camera Module
• 256 channels per module • 54 modules per ring, 14k channels – 64-‐channel ASIC, 216 ASIC per PET ring – Front-‐end board with SIPM on one side and FEE on the other side
7
Features: • 64-‐channel SiPM readout : posi[ve & nega[ve polarity
inputs • 8-‐bit Input DAC for SiPM HV tuning • Time Stamp and ADC charge outputs • 64-‐channel trigger outputs • Power Pulsing : Analog, ADC & Digital • Event rate : 50k events/s • Process: AMS SiGe 0.35µm
Time Stamp
ADC Charge
8-‐bit input DAC
High Gain shaper
Low Gain shaper
Triroc Design
8
Analog Design
– High Bandwidth Common base pre-‐amp (4GHz GBW) • accep[ng both polarity inputs
– 8-‐bit input DAC – 2 shapers for charge measurements
• High gain : up to 100 pe, Low Gain : up to 2000 pe – 2 triggers: Time and Charge (events valida[on) – 10-‐bit DACs for threshold + 6-‐bit trimming in each channel – Wilkinson ADC, Track & Hold/Peak sensing 9
Digital and backend
• Reject unwanted events: – Valida[on with charge trigger
• Ini[ate data conversion: – 10-‐bit ADC – TDC coarse and fine [me
• Readout – 80 MHz LVDS – 4x { DataOut + TransmitOn } – Global Counter (Top level)
• External Pins: – Reset Channel – External Channel Hit – Start Conversion 10
Input DAC linearity : • Bit : 8 • Range : 1 -‐ 3 V • LSB : 8.18 mV
Simula[ons -‐ Input DAC
11
Simula[ons – Trigger Threshold DAC
Time Trigger Threshold DAC linearity : • Bit : 10 • Range : 1.446 – 2.348 V • LSB : 0.89 mV
Charge Trigger Threshold DAC linearity : • Bit : 10 • Range : 0.998 – 1.889 V • LSB : 0.87 mV
Time Trigger Threshold DAC Trimming : • Bit : 6 • Range : Time Trigger DAC -‐ 40 mV • LSB : 0.6 mV
12
Pre-‐amp simula[ons • Input : 20 pe
Posi[ve input
Nega[ve input
Posi[ve input
Nega[ve input
Linearity for both inputs up to 30 pe
40 pe
40 pe
Simula[ons – Pre-‐amplifier
13
High gain shaper linearity up to 100 pe
High gain Shaper simula[ons • Input charge : 20 pe • Shaper peaking [me : 10 ns
Shaper peak detec[on
Shaper Track & Hold
Shaper peak detec[on
Shaper Track & Hold
Posi[ve input
Nega[ve input
Simula[ons – High Gain Shaper
14
Low gain shaper linearity up to 2000 pe
Low gain Shaper simula[ons • Input : 1000 pe • Shaper peaking [me : 20 ns
2000 pe
Shaper peak detec[on
Shaper Track & Hold
Posi[ve input
Shaper peak detec[on
Shaper Track & Hold
Nega[ve input
Simula[ons – Low Gain Shaper
15
Power consump[on
• Bias + common bloc : 11.815 mA * 3.3V = 39mW • 1-‐channel : 0.6mW
• 64-‐channel : 125.76 mA * 3.3V = 415mW • 1-‐channel : 6.5mW
• Digital : ~16 mA *3.3V = 53mW (es[ma[on) • 1-‐channel : 0.8mW
• Total for 1-‐channel : 7.9mW
VDD = 3.3V, without Output Buffer power consump[on
16
Conclusion & Status
• Design of dual polarity inputs ASIC at Weeroc/Omega – Versa[le ASIC for any SiPM on market – Suitable for PET/PET-‐ToF protoyping – Low external component counts – Shares a lot of characteris[c with Pe[roc2 (see C. de la Taille talk)
• Tapeout in March 2014 • ASIC currently in packaging : – BGA 353 balls : 12 x 12 mm – Compact front-‐end board in PET ring
• Test board s[ll in design. Fab scheduled in June 2014 • Tests scheduled for July 2014 • Front-‐End board in Sept 2014 17