Ce travail vise à étudier l’apport des technologies d’intégration 3D à l’imagerie CMOS ultra-rapide. La gamme de vitesse d’acquisition considérée ici est du million au milliard d’images par seconde. Cependant au-delà d’une dizaine de milliers d’images par seconde, les architectures classiques de capteur d’images sont limitées par la bande passante des buffers de sortie. Pour atteindre des fréquences supérieures, une architecture d’imageur burst est utilisée où une séquence d’une centaine d’images est acquise et stockée dans le capteur. Les technologies d’intégration 3D ont connu un engouement depuis une dizaine d’années et sont considérées comme une solution complémentaire aux travaux menés sur les dispositifs (transistors, composants passifs) pour améliorer les performances des circuits intégrés. Notre choix s’est porté sur une technologie où les circuits intégrés sont directement empilés avant la mise en boitier (3D-SIC). La densité d’interconnexions entre les différents circuits est suffisante pour permettre l’implémentation d’interconnexions au niveau du pixel. L’intégration 3D offre d’intéressants avantages à l’imagerie intégrée car elle permet de déporter l’électronique de lecture sous le pixel. Elle permet ainsi de maximiser le facteur de remplissage du pixel tout en offrant une large place aux circuits de conditionnement du signal. Dans le cas de l’imagerie burst, cette technologie permet de consacrer une plus grande surface aux mémoires dédiées au stockage de la séquence d’image et ce au plus proche des pixels. Elle permet aussi de réaliser sur la puce la conversion analogique numérique des images acquises. / This work aims to study the inflows of the 3D integration technology to ultra-high speed CMOS imaging. The acquisition speed range considered here is between one million to one billion images per second. However above ten thousand images per second, classical image sensor architectures are limited by the data bandwidth of the output buffers. To reach higher acquisition frequencies, a burst architecture is used where a set of about one hundred images are acquired and stored on-chip. 3D integration technologies become popular more than ten years ago and are considered as a complementary solution to the technological improvements of the devices. We have chosen a technology where integrated circuits are stacked on the top of each other (3D-SIC). The interconnection density between the circuits is high enough to enable interconnections at the pixel level. The 3D integration offers some significant advantages because it allows deporting the readout electronic below the pixel. It thus increases the fill factor of the pixel while offering a wide area to the signal processing circuit. For burst imaging, this technology provides more room to the memory dedicated to the image storage while staying close to the pixel. It also allows implementing analog to digital converter on-chip.

This thesis deals with the design of a device that enables to monitor the behavior of the measured signal on the computer screen, using the principle of the digital oscilloscope. The control element of the device is the field programmable gate array (FPGA) on FITkit platform.  The FPGA configuration controls the input signal sampling and sends the received samples through the USB interface to the PC. The graphical application implemented in the computer tries to restore the signal and then displays it on the screen.

This paper explores the noise shaping and noise producing qualities of Delta-Sigma Modulators (DSM) and Pulse-Width Modulators (PWM). DSM has long been dominant in the Delta Sigma Analog-to-Digital Converter (DSADC) as a noise-shaped quantizer and time discretizer, while PWM, with a similar self oscillating structure, has seen use in Class D Power Amplifiers, performing a similar function. It has been shown that the PWM in Class D Amplifiers outperforms the DSM [1], but could this advantage be used in DSADC use-cases? LTSpice simulation and printed circuit board implementation and test are used to present data on four variations of these modulators: The DSM, PWM, the out-of-loop discretized PWM (OOLDP), and the cascaded modulator. A generic form of an Nth order loop filter is presented, where three orders of this generic topology are analyzed in simulation for each modulator, and two orders are used in physical testing.

A comprehensive and scalable solution for high-performance switched capacitor amplification is presented. Central to this discussion is the concept of ring amplification. A ring amplifier is a small modular amplifier derived from a ring oscillator that naturally embodies all the essential elements of scalability. It can amplify with accurate rail-to-rail output swing, drive large capacitive loads with extreme efficiency using slew-based charging, naturally scale in performance according to process trends, and is simple enough to be quickly constructed from only a handful of inverters, capacitors, and switches. In addition, the gain-enhancement technique of Split-CLS is introduced, and used to extend the efficacy of ring amplifiers in specific and other amplifiers in general. Four different pipelined ADC designs are presented which explore the practical implementation options and design considerations relevant to ring amplification and Split-CLS, and are used to establish ring amplification as a new paradigm for scalable amplification. / Graduation date: 2012 / Access restricted to the OSU Community, at author's request, from July 19, 2012 - July 19, 2013