In this thesis, a CMOS tunable second-order log-domain filter using pseudo-exponential approximation is proposed. MOSFETs in the circuit are working in the saturation region. This filter has higher frequency response than that of weak inversion CMOS filter. The circuit has been fabricated with 0.35um CMOS technology. It operates with a supply voltage 3V, internal capacitance C is 1pF, the biasing current varies from 2uA~10uA. The cutoff frequency can be turned from 2MHz~37.5MHz. The harmonic distortion is 0.93% and the power consumption is 772uW.

No description available.

The speed of fuzzy controllers implemented on dedicated hardware is adequate for control of any physical process, but too slow for todays high-complexity data networks. Defuzzification has been the bottleneck for fast implementations due to the large number of computationally expensive multiplication and division operations. In this thesis, we propose a high-speed fuzzy inferential system based on log-domain arithmetic, which only requires addition and subtraction operations. The system is implemented on a Xilinx Virtex-II FPGA with a processing speed of 67.6 MFLIPS having a maximum combinational path delay of 4.2 ns. It is a clear speedup compared to the reported fastest 50 MFLIPS implementation. A pipelined version of the controller is also implemented, which achieves a speed of 248.7 MFLIPS. Although a small approximation error is introduced, software simulation and hardware implementation on FPGA confirm high similarity of the outputs for control surfaces and a number of second-order plants. / Software Engineering and Intelligent Systems

Traditionally, the design of continuous time active filters usually has a trade offbetween low-voltage and high dynamic range. One way to solve this problem is companding technology. There are two methods for companding filters. The first method utilizes the exponential I-V characteristics of BJT in the saturation region. In order to reduce the cost andintegrate the analog and digital circuits, the other method was exploited using CMOS process. In this project, a new first-order low pass log-domain filter based on CMOS parasitic vertical BJTwill be proposed. This filter has higher frequency response than previous circuits. We will first employ Hspice to simulate the log-domain filter to ensure the correctness of the circuit and make it a reliable reference with the circuit layout. After summarizing all the simulations and analyses, the chip will be fabricated with 0.35um CMOS technology.

The cochlea is an organ which extracts frequency information from the input sound wave. It also produces nerve signals, which are further analysed by the brain and ultimately lead to perception of the sound. An existing model of the cochlea by Fragni`ere is first analysed by simulation. This passive model is found to have the properties that the living cochlea does in terms of the frequency response. An analog VLSI circuit implementation of this cochlear model in CMOS weak inversion is proposed, using log-domain filters in current domain. It is fabricated on a chip and a measurement of a basilar membrane section is performed. The measurement shows a reasonable agreement to the model. However, the circuit is found to have a problem related to transistor mismatch, causing different behaviour in identical circuit blocks. An active cochlear model is proposed to overcome this problem. The model incorporates the effect of the outer hair cells in the living cochlea, which controls the quality factor of the basilar membrane filters. The outer hair cells are incorporated as an extra voltage source in series with the basilar membrane resonator. Its value saturates as the input signal becomes larger, making the behaviour rather closer to that of a passive model. The simulation results show this nonlinear phenomenon, which is also seen in the living cochlea. The contribution of this thesis is summarised as follows: a) the first CMOS weak inversion current domain basilar membrane resonator is designed and fabricated, and b) the first active two-dimensional cochlear model for analog VLSI implementation is developed.

The cochlea is an organ which extracts frequency information from the input sound wave. It also produces nerve signals, which are further analysed by the brain and ultimately lead to perception of the sound. An existing model of the cochlea by Fragni`ere is first analysed by simulation. This passive model is found to have the properties that the living cochlea does in terms of the frequency response. An analog VLSI circuit implementation of this cochlear model in CMOS weak inversion is proposed, using log-domain filters in current domain. It is fabricated on a chip and a measurement of a basilar membrane section is performed. The measurement shows a reasonable agreement to the model. However, the circuit is found to have a problem related to transistor mismatch, causing different behaviour in identical circuit blocks. An active cochlear model is proposed to overcome this problem. The model incorporates the effect of the outer hair cells in the living cochlea, which controls the quality factor of the basilar membrane filters. The outer hair cells are incorporated as an extra voltage source in series with the basilar membrane resonator. Its value saturates as the input signal becomes larger, making the behaviour rather closer to that of a passive model. The simulation results show this nonlinear phenomenon, which is also seen in the living cochlea. The contribution of this thesis is summarised as follows: a) the first CMOS weak inversion current domain basilar membrane resonator is designed and fabricated, and b) the first active two-dimensional cochlear model for analog VLSI implementation is developed.

Τα φίλτρα λογαριθμικού πεδίου (log-domain filters) είναι μια οικογένεια φίλτρων συνεχούς χρόνου στην οποία γίνεται χρήση της μη-γραμμικής I/V χαρακτηριστικής των ενεργών στοιχείων (BJT τρανζίστορ, MOSFET τρανζίστορ σε ασθενή αναστροφή) για να επιτευχθεί γραμμική συμπεριφορά από είσοδο σε έξοδο. Επιπρόσθετα τα log-domain φίλτρα, προσφέρουν και άλλα πλεονεκτήματα, όπως την ηλεκτρονική ρύθμιση της συχνότητας αποκοπής, την λειτουργία σε υψηλές συχνότητες και την δυνατότητα λειτουργίας με χαμηλή τάση τροφοδοσίας. Στην παρούσα διατριβή προτείνεται μια καινούρια τεχνική σχεδιασμού log-domain φίλτρων, η οποία στηρίζεται στη κυματική μέθοδο (log-domain wave filters). Με αυτή την τεχνική, log-domain φίλτρα μπορούν να σχεδιαστούν είτε με την χρήση των κυματικών παραμέτρων σκέδασης (wave scattering parameters), είτε με την χρήση κυματικών τερματιστών θύρας (wave port terminator). Έτσι αρχικά, προτείνονται δομές log-domain κυματικών φίλτρων με χρήση των κυματικών παραμέτρων σκέδασης. Σε αυτήν την προσέγγιση, το log-domain κυματικό ισοδύναμο ενός επαγωγού σε σειρά κλάδο μετασχηματίζεται από το γραμμικό στο λογαριθμικό πεδίο. Τα log-domain κυματικά ισοδύναμα των υπολοίπων βασικών στοιχείων του παθητικού πρωτοτύπου φίλτρου προκύπτουν από το κυματικό ισοδύναμο του επαγωγού σε σειρά κλάδο με την εναλλαγή των εξόδων του ή/και την προσθήκη αντιστροφέων τάσης. Χρησιμοποιώντας μια διαφορετική προσέγγιση, προτείνονται δομές log-domain κυματικών φίλτρων με χρήση του log-domain κυματικού τερμαστιστή θύρας. Με την χρήση του σχεδιάζεται το κυματικό ισοδύναμο στο log-domain ενός πυκνωτή σε παράλληλο κλάδο, το οποίο χρησιμοποιείται ως το βασικό κυματικό ισοδύναμο δομικό στοιχείο. Τα κυματικά ισοδύναμα των υπολοίπων στοιχείων του παθητικού κυκλώματος προκύπτουν με χρήση των παραπάνω κυματικών ισοδυνάμων. Τέλος γίνεται μελέτη της επίδρασης των μη-ιδανικών χαρακτηριστικών του BJT τρανζίστορ και της επίδρασης των θερμοκρασιακών μεταβολών στην συμπεριφορά των προτεινόμενων φίλτρων, ενώ προτείνονται μέθοδοι τόσο για την επιμέρους όσο και την συνολική αντιστάθμισή τους. / Log-domain filters are a family of continuous time filters that instead of being designed using locally linearized components, directly exploit the non-linear nature of a BJT, or a MOSFET transistor operating in weak inversion, in order to obtain a system with overall linear performance. This way they exhibit high linearity even for large signal to bias ratios. In addition they have some other interesting features like tunability and the ability to work at high frequencies even with low voltage supplies. In the present dissertation a new method for the design of log-domain filters is proposed, based on the well-known wave method. According to this method log-domain filters can be designed using the wave scattering parameters or using the wave port terminator block. Structures of log-domain wave filters using wave scattering parameters are initially proposed. In this approach the log-domain wave equivalent of a series inductor is transformed from linear to log-domain. Log-domain wave equivalents of the other basic elements of a passive prototype are then derived by the interchange of its outputs or/and the addition of voltage inverters. These equivalents are then connected together forming the actual filter. Log-domain differential wave filters can be designed by the introduction of the log-domain differential wave equivalent of the series inductor. Using a different approach, log-domain wave filters have been designed, introducing the block of a log-domain wave port terminator. Using this block, the log-domain wave equivalent of a shunt capacitor is derived as the based structure element. The wave equivalents of other basic elements of the passive prototype are derived as before interchanging its outputs or/and adding voltage inverters. Finally the influence of the non-ideal characteristics of the BJT transistor on the frequency response of the derived filters is analyzed and efficient methods are proposed for their compensation.

Translinear circuits are circuits in which the exponential relationship between the output current and input voltage of a circuit element is exploited to realize various algebraic or differential equations. This thesis is concerned with a subclass of translinear circuits, in which the basic translinear element, called a multiple-input translinear element (MITE), has an output current that is exponentially related to a weighted sum of its input voltages. MITE networks can be used for the implementation of the same class of functions as traditional translinear circuits. The implementation of algebraic or (algebraic) differential equations using MITEs can be reduced to the implementation of the product-of-power-law (POPL) relationships, in which an output is given by the product of inputs raised to different powers. Hence, the synthesis of POPL relationships, and their optimization with respect to the relevant cost functions, is very important in the theory of MITE networks. In this thesis, different constraints on the topology of POPL networks that result in desirable system behavior are explored and different methods of synthesis, subject to these constraints, are developed. The constraints are usually conditions on certain matrices of the network, which characterize the weights in the relevant MITEs. Some of these constraints are related to the uniqueness of the operating point of the network and the stability of the network. Conditions that satisfy these constraints are developed in this work. The cost functions to be minimized are the number of MITEs and the number of input gates in each MITE. A complete solution to POPL network synthesis is presented here that minimizes the number of MITEs first and then minimizes the number of input gates to each MITE. A procedure for synthesizing POPL relationships optimally when the number of gates is minimal, i.e., 2, has also been developed here for the single--output case. A MITE structure that produces the maximum number of functions with minimal reconfigurability is developed for use in MITE field--programmable analog arrays. The extension of these constraints to the synthesis of linear filters is also explored, the constraint here being that the filter network should have a unique operating point in the presence of nonidealities. Synthesis examples presented here include nonlinear functions like the arctangent and the gaussian function which find application in analog implementations of particle filters. Synthesis of dynamical systems is presented here using the examples of a Lorenz system and a sinusoidal oscillator. The procedures developed here provide a structured way to automate the synthesis of nonlinear algebraic functions and differential equations using MITEs.

Οι εφαρμογές της σύγχρονης τεχνολογίας επιτάσσουν τη χρήση συσκευών με όσο το δυνατόν μικρότερες διαστάσεις, χαμηλή τάση τροφοδοσίας, χαμηλή κατανάλωση ισχύος και ταυτόχρονα υψηλές επιδόσεις. Το αντικείμενο της εργασίας αυτής, αφορά στη σχεδίαση αναλογικών ολοκληρωμένων φίλτρων, χαμηλής τάσης τροφοδοσίας, για υλοποίηση μεγάλων σταθερών χρόνου, στο πεδίο του λογαρίθμου. Προς αυτή την κατεύθυνση, μελετώντας και σχεδιάζοντας δομές αναλογικών φίλτρων στο πεδίο του λογαρίθμου, επιτεύχθηκε η σχεδίαση φίλτρων δεύτερης τάξης με μεγάλες σταθερές χρόνου, διατηρώντας τις φυσικές διαστάσεις των κυκλωμάτων σε εξαιρετικά χαμηλά επίπεδα. Αρχικά, παρουσιάζονται κάποια εισαγωγικά στοιχεία για την σχεδίαση ολοκληρωμένων κυκλωμάτων σε περιβάλλον χαμηλής τάσης τροφοδοσίας. Γίνεται εισαγωγή στην ιδέα των λογαριθμικών φίλτρων και αναλύονται οι βασικές αρχές σχεδίασης. Παρουσιάζονται βασικά χαρακτηριστικά των κυκλωμάτων στο πεδίο του λογαρίθμου, καθώς και ανάλυση των τελεστών και των διαγωγών που αποτελούν τη βάση της σχεδίασης στο λογαριθμικό πεδίο. Επιπλέον, παρουσιάζονται οι υλοποιήσεις των ολοκληρωτών των φίλτρων στο πεδίο του λογαρίθμου. Στη συνέχεια, γίνεται τοπολογική εξομοίωση 2ης τάξης βαθυπερατών φίλτρων στο πεδίο του λογαρίθμου. Σχεδιάζονται φίλτρα με την κλασική μέθοδο υλοποίησης, κάνοντας χρήση ισοδύναμων των παθητικών στοιχείων στο λογαριθμικό πεδίο, αλλά και φίλτρα υλοποιημένα με διάγραμμα ροής (SFG). Παρουσιάζονται τα πρώτα αποτελέσματα των εξομοιώσεων που πραγματοποιήθηκαν με το λογισμικό Cadence και το γραφικό περιβάλλον που διαθέτει για την σχεδίαση αναλογικών ηλεκτρονικών κυκλωμάτων (Virtuoso Analog Environment). Προτείνονται, δύο κυκλώματα τα οποία πραγματοποιούν πολλαπλασιασμό της χωρητικότητας των πυκνωτών, επιτυγχάνοντας έτσι μεγάλες σταθερές χρόνου, και η υλοποίηση νέων ολοκληρωτών που κάνουν χρήση των πολλαπλασιαστών. Δημιουργούνται έτσι οι βάσεις για την υλοποίηση φίλτρων με εξαιρετικά μικρές διαστάσεις, των οποίων η σχεδίαση, η εξομοίωση και η φυσική σχεδίαση (layout design) παρουσιάζονται, αναλύονται και συγκρίνονται. / The technological evolution and market requirements have led to an increasing demand of low - power portable devices, featuring the reduced size of the devises and high efficiency. This M.Sc project deals with the design of analog integrated, Log - Domain filters, for low - voltage implementation, with large time - constants. In this direction, the design of a second order, low - pass filter, with the above features, and with the occupied silicon area maintained at very low levels, was achieved. In Chapters 1 and 2, an introduction to the design of integrated circuits in low voltage environment is presented. There is an introduction to the idea and the basic principles of Log - Domain filters. The key characteristics of circuits in a large signal operation point of view, and an analysis of the operators and the exponential transconductor cells are, also presented. Furthermore, the basic Log - Domain integrators has been analyzed. A topologic analysis of second order Log - Domain filters is given in Chapter 3. Filters has been initially designed firstly with the classic implementation, using Log - Domain equivalent of passive elements. In a second step, the filter has been realizes by employing the signal flow diagram (SFG) representation. These filters were simulated with the Analog Design environment of the Cadence software. the obtained simulation results confirmed the correct operation of the circuit. Two implementations for realizing the Log - Domain equivalent of a capacitor multiplier are introduced. In addition, implementations of new Log - Domain integrators, that use the capacitor multipliers, are given in Chapter 4. Using these implementations, Log - Domain filters, with reduced total area and large time - constants, are designed, simulated and characterized in Chapter 5. Finally, the layout design of a second - order has been performed in Chapter 6 and the provided post - layout simulation results show that the performance of the filter was close to that of the filter realized in schematic level.

<p>Neuromorphic computing endeavors to develop computational systems capable of emulating the brain’s capacity to execute intricate tasks concurrently and with remarkable energy efficiency. By utilizing new bioinspired computing architectures, these systems have the potential to revolutionize high-performance computing and enable local, low-energy computing for sensors and robots. Organic and soft materials are particularly attractive for neuromorphic computing as they offer biocompatibility, low-energy switching, and excellent tunability at a relatively low cost. Additionally, organic materials provide physical flexibility, large-area fabrication, and printability.</p><p>This doctoral dissertation showcases the research conducted in fabricating a comprehensive spiking organic neuron, which serves as the fundamental constituent of a circuit system for neuromorphic computing. The major contribution of this dissertation is the development of the organic, flexible neuron composed of spiking synapses and somas utilizing ultra-low voltage organic field-effect transistors (OFETs) for information processing. The synaptic and somatic circuits are implemented using physically flexible and biocompatible organic electronics necessary to realize the Polymer Neuromorphic Circuitry. An Axon-Hillock (AH) somatic circuit was fabricated and analyzed, followed by the adaptation of a log-domain integrator (LDI) synaptic circuit and the fabrication and analysis of a differential-pair integrator (DPI). Finally, a spiking organic neuron was formed by combining two LDI synaptic circuits and one AH synaptic circuit, and its characteristics were thoroughly examined. This is the first demonstration of the fabrication of an entire neuron using solid-state organic materials over a flexible substrate with integrated complementary OFETs and capacitors.</p>