Supervisor

Topic title

Description

Damian Panter

Investigation of Pulse-Width Modulation Methods for Signal Processing

Information can be represented in several ways in the context of an electrical circuit. Next to the well-known voltage-mode and current-mode signals, time-mode signals are an interesting alternative. Pulse-width Modulation (PWM) is a popular method to generate time-mode signals and shall be investigated for this seminar. The aspects of interests are:

• PWM Sampling Theorem and its implications

• Different PWM carrier signals and their benefits

• Alternative PWM and Voltage-to-Time Conversion methods

Damian Panter

Survey of State-of-the-Art Time-Mode Signal Processing

Information can be represented in several ways in the context of an electrical circuit. Next to the well-known voltage-mode and current-mode signals, time-mode signals are an interesting alternative. For this seminar, operations and processing on time-mode signals shall be examined. These are the relevant aspects:

• Overview of time-mode signal processing: benefits, drawbacks & SotA applications

• Operations on time-mode signals and their circuit implementations

• Conversion of and reconstruction from time-mode signals

Nihal Deshpande

Comparison of the different methods to calculate threshold voltages for MOSFET devices

Characterizing MOSFET devices requires calculating the threshold voltage, a crucial parameter for defining the operating point. However, various methods exist, leading to confusion among new designers due to a lack of uniformity. This study aims to evaluate and compare different methods for calculating threshold voltage, examining their implementation reasons and impact on defining operating points for SPICE simulations.

Nihal Deshpande

Use of SINDy for Parameter Estimation in Advanced Modeling

Advanced modeling methodologies require optimizing hundreds of parameters to fit observations. Current methods, such as Evolutionary Algorithms and Swarm Algorithms, have limitations, including hyperparameter optimization and computational pitfalls. This study explores Sparse Identification for Nonlinear Dynamics (SINDy), a data-driven algorithm that uses multiple transcendental functions for curve fitting and identifies the best combination of functions and variables to fit the data. The goal is to evaluate SINDy as a potential candidate for parameter estimation frameworks and compare its performance with standard metaheuristic methods for black-box optimization problems.

Nihal Deshpande

Thermal Network Models to predict Self-Heating Effect for cryogenic temperatures

Electronic devices generate heat during operation, causing a self-heating (SH) effect that elevates temperature above ambient. This effect significantly impacts device reliability and lifetime, particularly at cryogenic temperatures, where it can lead to thermal stresses, noises, and deviations from expected behavior. Current thermal networks and methodologies for modeling SH need to be adapted to account for non-linear thermal resistances and capacitances at cryogenic temperatures. This study aims to evaluate and analyze existing thermal networks and methodologies, and propose a suitable approach for optimally modeling SH at cryogenic temperatures.

Investigation of the temperature dependent band-gap used in Compact Models

Compact models simulate device behavior by combining physics and empirical expressions for SPICE simulations. The core physics relies on standard semiconductor physics, with energy band-gap being a critical component. Different compact models, such as BSIM and HiSIM, use various methods to calculate the temperature dependence of the band-gap, resulting in discrepancies that require additional tuning parameters to match device behavior. This study aims to review the literature on band-gap models within different compact modeling families, evaluating their development, correlation with measurement data, and impact on compact models.

In modern handheld mobile devices, the matching condition of the planar inverted-F antenna is strongly influenced by the surrounding environment. Hence, real-time impedance tuning is necessary to fulfill the requirement on power efficiency and linearity of the transmitter. In this work, the student is encouraged to perform a survey on various topologies which realize the closed-loop antenna tuning and compare them based on their advantages/disadvantages.

In recent years, near-field wireless power transfer has facilitated the use of higher frequencies in the multi-MHz range for both high and low power applications. Common inverter structures employed include Class-D and Class-E. Given the need for high efficiency and minimal distortion within the overall system, selecting the right switches is crucial. The objective is to explore the types of switches currently in use and identify the most promising semiconductor technologies for future systems.

Electrically-control reconfigurable filters are highly required in multi-functional multi-standard communication and radar systems due to their potential to achieve frequency agility, circuit-size reduction and design simplicity. In this seminar, you should study the circuit design and direct synthesize approach for bandpass filter with tunable bandwidth and center frequency.

Impurities within the SiO2 bulk oxide and at the Si-SiO2 interface of MOS devices create defects/traps that can be electronically active and influence the device characteristics. This seminar should describe the characteristics of the most important (electronically active) defects/traps within the SiO2 oxide and their interaction with charge carriers.

Computing in Memory (CIM) using RRAM for AI application

• Basic structure of the analog CIM using RRAM
• Basic structure of the digital CIM using RRAM and DRAM or SRAM
• Method of the MVM realization using RRAM for both analog and digital type (multiplication, convolution, etc.)

Energy-Efficient Direct RF-modulation transmitter in V-band

Conventional Gilbert-Cell based mixers are not suitable for ultra low-power battery-operated radio devices due to its high power consumption and hence it is desirable to perform modulation (OOK/ ASK) directly at the power amplifier end for such application devices and  hence eliminates the mixer usage. In this seminar you will perform a survey of such CMOS-based transmitter architectures operating in the V-band (40-75 GHz) and compare their functional performance in terms of operating power gain, output power, efficiency and bandwidth .

Electrical Impedance Spectroscopy (EIS) is a widely utilized non-invasive method in biology for analyzing cellular responses. The technique allows for the measurement of ions, plant tissues and water inside plants in a wide spectrum of frequencies. The study will focus on the electrical and physical behaviors and limitations of EIS on plants.

The continued scaling of CMOS technologies has driven supply voltages steadily lower, exacerbating voltage headroom constraints. In such environments, multistage operational amplifier (op-amp) topologies become essential to maintain performance. Ring amplifiers, which scale efficiently with advanced CMOS processes, have emerged as a promising solution for switched-capacitor circuits. This seminar should investigate the fundamental principles of ring amplifier design and its practical application as residue amplifiers in SAR-ADCs.

As the semiconductor landscape evolves, heterogeneous integration (HI) emerges as a key player in addressing the demands for enhanced performance, functionality, and yield while reducing power consumption and area usage. The anticipated rise of the chiplet market promises exciting innovations and opportunities. However, with these advancements come unforeseen security challenges that could jeopardize the integrity of chiplet and HI systems. This seminar should explore and discuss, which security threads may occure for chiplets and hetero-integrated devices (e.g. laser probing, hardware trojans, chiplet intellectual property…).