This thesis explores the energy-delay space of eight widely referred flip-flops in a 0.13µm CMOS technology. The main goal has been to find the smallest set of flip-flop topologies to be included in a “high performance” flip-flop cell library covering a wide range of power-performance targets. Based on the comparison results, transmission gate-based flip-flops show the best powerperformance trade-offs with a total delay (clock-to-output + setup time) down to 105ps. For higher performance, the pulse-triggered flip-flops are the fastest (80ps) alternatives suitable to be included in a flip-flop cell library. However, pulse-triggered flip-flops consume significantly larger power (about 2.5x) compared to other fast but fully dynamic flip-flops such as TSPC and dynamic TG-based flip-flops.

In this work, an existing generator for layout of Static Random Access Memory (SRAM) is improved. The tool is completed with a block decoder, which was missing when the thesis started. A feature of generating schematic files is also added. The schematics are important to get a better overview, to test parts properly, and enable Layout versus Schematics (LVS) checks. The main focus of this thesis work is to implement and evaluate a new SRAM cell, called Zero Aware Asymmetric SRAM cell. This cell saves major power when zeros are stored. Furthermore the pull-up circuit is modified to be less power consuming. Other parts are also modified to fit the new memory cell. Several minor flaws are corrected in the layout generator. It does still not produce a complete memory without manual interventions, but it does at least create all parts with one command. Several tests, including Design Rule Checks (DRC) and LVS checks, are carried out both on minor and larger parts. Development of documentation that makes it easier for users and developers to use and understand the tool is initiated.

With an increasing number of transistors integrated on a single die, the need for global on-chip interconnectivity is growing. Long interconnects, in turn, have very large capacitances which consume a large share of a chip’s total power budget. Power consumption can be lowered in several ways, mainly by reduction of switching activity, reduction of total capacitance and by using low voltage swing. In this project, the issue is addressed by proposing a new encoding based on Pulse Width Modulation (PWM). The implementation of this encoding will both lower the switching activity and decrease the capacitance between nearby wires. Hence, the total effective capacitance will be reduced considerably. Schematic level implementation of a robust transmitter and receiver circuit was carried out in CMOS090, designed for speeds up to 100 MHz. On a 10 mm wire, this implementation would give a 40% decrease in power dissipation compared to a parallel bus having the same metal footprint. The proposed encoding can be efficiently applied for global interconnects in sub-micron systems-on-chip (SoC).

In this work, an existing RAM generator is analysed and evaluated. Some of the aspects that were considered in the evaluation are the optimization of the basic SRAM cell, how the RAM generator can be ported to newer technologys, automating the simulation process and the creation of the workflow for the energy model. One of the main focus of this thesis work is to optimize the basic SRAM cell. The SRAM cell which is used in the RAM generator is not optimized for area nor power. A compact layout is suggested which saves a lot of area and power. The technology that is used to create the RAM generator is old and a suitable way to port it to newer technology has also been found. To create an energy model one has to simulate a lot of memories with a lot of data. This cannot be done in the traditional way of simulating circuits using the GUI. Hence an automation procedure has been suggested which can be made to work to create energy models by simulating the memories comprehensively. Finally, basic ground work has been initiated by creating a workflow for the creation of the energy model.

Since the invention of wireless telegraphy by Marconi in 1897, wireless technology has not only been enhanced, but also has become an integral part of our everyday lives. The first wireless mobile phone appeared around 1980. It was based on first generation analog technology that involved the use of Frequency Division Multiple Access (FDMA) techniques. Ten years later, second generation (2G) mobiles were dependent on Time Division Multiple Access (TDMA) techniques and Code Division Multiple Access (CDMA) techniques. Nowadays, third generation (3G) mobile systems depend on CDMA techniques to satisfy the need for faster, and more capacious data transmission in mobile wireless networks. Wideband CDMA (WCDMA) has become the major 3G air interface in the world. WCDMA employs convolutional encoding to encode voice and MPEG4 applications in the baseband transmitter at a maximum frequency of 2<i>Mbps</i>. To decode convolutional codes, Andrew Viterbi invented the Viterbi Decoder (VD) in 1967. In 2G mobile terminals, the VD consumes approximately one third of the power consumption of a baseband mobile transceiver. Thus, in 3G mobile systems, it is essential to reduce the power consumption of the VD. Conceptually, the Register Exchange (RE) method is simpler and faster than the Trace Back (TB) method for implementing the VD. However, in the RE method, each bit in the memory must be read and rewritten for each bit of information that is decoded. Therefore, the RE method is not appropriate for decoders with long constraint lengths. Although researchers have focused on implementing and optimizing the TB method, the RE method is focused on and modified in this thesis to reduce the RE method's power consumption. This thesis proposes a novel modified RE method by adopting a <i>pointer</i> concept for implementing the survivor memory unit (SMU) of the VD. A pointer is assigned to each register or memory location. The contents of thepointer which points to one register is altered to point to a second register, instead of copying the contents of the first register to the second. When the pointer concept is applied to the RE's SMU implementation (modified RE), there is no need to copy the contents of the SMU and rewrite them, but one row of memory is still needed for each state of the VD. Thus, the VDs in CDMA systems require 256 rows of memory. Applying the pointer concept reduces the VD's power consumption by 20 percent as estimated by the VHDL synthesis tool and by the new power reduction estimation that is introduced in this work. The coding gain for the modified RE method is 2. 6<i>dB</i> at an SNR of approximately 10-3. Furthermore, a novel zero-memory implementation for the modified RE method is proposed. If the initial state of the convolutional encoder is known, the entire SMU of the modified RE VD is reduced to only one row. Because the decoded data is generated in the required order, even this row of memory is dispensable. The zero-memory architecture is called the MemoryLess Viterbi Decoder (MLVD), and reduces the power consumption by approximately 50 percent. A prototype of the MLVD with a one third convolutional code rate and a constraint length of nine is mapped into a Xilinx 2V6000 chip, operating at 25 <i>MHz</i> with a decoding throughput of more than 3<i>Mbps</i> and a latency of two data bits. The other problem of the VD which is addressed in this thesis is the Add Compare Select Unit (ACSU) which is composed of 128 butterfly ACS modules. The ACSU's high parallelism has been previously solved by using a bit serial implementation. The 8-bit First Input First Output (FIFO) register, needed for the storage of each path metric (PM), is at the heart of the single bit serial ACS butterfly module. A new, simply controlled shift register is designed at the circuit level and integrated into the ACS module. A chip for the new module is also fabricated.

The explosive growth of battery operated devices has made low-power design a priority in recent years. Moreover, embedded SRAM units have become an important block in modern SoCs. The increasing number of transistor count in the SRAM units and the surging leakage current of the MOS transistors in the scaled technologies have made the SRAM unit a power hungry block from both dynamic and static perspectives. Owing to high bitline voltage swing during write operation, the write power consumption is dominated the dynamic power consumption. The static power consumption is mainly due to the leakage current associated with the SRAM cells distributed in the array. Moreover, as supply voltage decreases to tackle the power consumption, the data stability of the SRAM cells have become a major concern in recent years. <br /><br /> To reduce the write power consumption, several schemes such as row based sense amplifying cell (SAC) and hierarchical bitline sense amplification (HBLSA) have been proposed. However, these schemes impose architectural limitations on the design in terms of the number of words on a row. Beside, the effectiveness of these methods is limited to the dynamic power consumption. Conventionally, reduction of the cell supply voltage and exploiting the body effect has been suggested to reduce the cell leakage current. However, variation of the supply voltage of the cell associates with a higher dynamic power consumption and reduced cell data stability. Conventionally qualified by Static Noise Margin (SNM), the ability of the cell to retain the data is reduced under a lower supply voltage conditions. <br /><br /> In this thesis, we revisit the concept of data stability from the dynamic perspective. A new criteria for the data stability of the SRAM cell is defined. The new criteria suggests that the access time and non-access time (recovery time) of the cell can influence the data stability in a SRAM cell. The speed vs. stability trade-off opens new opportunities for aggressive power reduction for low-power applications. Experimental results of a test chip implemented in a 130 <em>nm</em> CMOS technology confirmed the concept and opened a ground for introduction of a new operational mode for the SRAM cells. <br /><br /> We introduced a new architecture; Segmented Virtual Grounding (SVGND) to reduce the dynamic and static power reduction in SRAM units at the same time. Thanks to the new concept for the data stability in SRAM cells, we introduced the new operational mode of Accessed Retention Mode (AR-Mode) to the SRAM cell. In this mode, the accessed SRAM cell can retain the data, however, it does not discharge the bitline. The new architecture outperforms the recently reported low-power schemes in terms of dynamic power consumption, thanks to the exclusive discharge of the bitline and the cell virtual ground. In addition, the architecture reduces the leakage current significantly since it uses the back body biasing in both load and drive transistors. <br /><br /> A 40Kb SRAM unit based on SVGND architecture is implemented in a 130 <em>nm</em> CMOS technology. Experimental results exhibit a remarkable static and dynamic power reduction compared to the conventional and previously reported low-power schemes as expect from the simulation results.

Advances in technology and the introduction of high speed processors have increased the demand for fast, compact and commercial methods for transferring large amounts of data. The next generation of the communication access network will use optical fiber as a media for data transmission to the subscriber. In optical data or chip-to-chip data communication, the continuous received data needs to be converted to discrete data. For the conversion, a synchronous clock and data are required. A clock and data recovery (CDR) circuit recovers the phase information from the data and generates the in-phase clock and data. In this dissertation, two clock and data recovery circuits for Giga-bits per second (Gbps) serial data communication are designed and fabricated in 180nm and 90nm CMOS technology. The primary objective was to reduce the circuit power dissipation for multi-channel data communication applications. The power saving is achieved using low swing voltage signaling scheme. Furthermore, a novel low input swing Alexander phase detector is introduced. The proposed phase detector reduces the power consumption at the transmitter and receiver blocks. The circuit demonstrates a low power dissipation of 340µW/Gbps in 90nm CMOS technology. The CDR is able to recover the input signal swing of 35mVp. The peak-to-peak jitter is 21ps and RMS jitter is 2.5ps. Total core area excluding pads is approximately 0.01mm2.

With rapid technology scaling, flip-flops are becoming more susceptible to metastability due to tighter timing budgets and the more prominent effects of process, temperature, and voltage variation that can result in frequent setup and hold time violations. This thesis presents a detailed methodology and analysis on the design of metastable-hardened, high-performance, and low-power flip-flops. The design of metastable-hardened flip-flops is focused on optimizing the value of τ mainly due to its exponential relationship with the metastability window δ and the mean-time-between-failure (MTBF). Through small-signal modeling, τ is determined to be a function of the load capacitance and the transconductance in the cross-coupled inverter pair for a given flip-flop architecture. In most cases, the reduction of τ comes at the expense of increased delay and power. Hence, two new design metrics, the metastability-delay-product (MDP) and the metastability-power-delay-product (MPDP), are proposed to analyze the tradeoffs between delay, power and τ. Post-layout simulation results have shown that the proposed optimum MPDP design can reduce the metastability window δ by at least an order of magnitude depending on the value of the settling time and the flip-flop architecture. In this work, we have proposed two new flip-flop designs: the pre-discharge flip-flop (PDFF) and the sense-amplifier-transmission-gate (SATG) based flip-flop. Both flip-flop architectures facilitate the usage in both single and dual-supply systems as reduced clock-swing flip-flop and level-converting flip-flop. With a cross-coupled inverter in the master-stage that increases the overall transconductance and a small load transistor associated with the critical node, the architecture of both the PDFF and the SATG is very attractive for the design of metastable-hardened, high-performance, and low-power flip-flops. The amount of overhead in delay, power, and area is all less than 10% under the optimum MPDP design scheme when compared to the traditional optimum PDP design. In designing for metastable-hardened and soft-error tolerant flip-flops, the main methodology is to improve the metastability performance in the master-stage while applying the soft-error tolerant cell in the slave-stage for protection against soft-error. The proposed flip-flops, PDFF-SE and SATG-SE, both utilize a cross-coupled inverter on the critical path in the master-stage and generate the required differential signals to facilitate the usage of the Quatro soft-error tolerant cell in the slave-stage.

Guarded evaluation is a power reduction technique that involves identifying sub-circuits (within a larger circuit) whose inputs can be held constant (guarded) at specific times during circuit operation, thereby reducing switching activity and lowering dynamic power. The concept is rooted in the property that under certain conditions, some signals within digital designs are not "observable" at design outputs, making the circuitry that generates such signals a candidate for guarding. Guarded evaluation has been demonstrated successfully for custom ASICs; in this work, we apply the technique to FPGAs. In ASICs, guarded evaluation entails adding additional hardware to the design, increasing silicon area and cost. Here, we apply the technique in a way that imposes minimal area overhead by leveraging existing unused circuitry within the FPGA. The LUT functionality is modified to incorporate the guards and reduce toggle rates. The primary challenge in guarded evaluation is in determining the specific conditions under which a sub-circuit's inputs can be held constant without impacting the larger circuit's functional correctness. We propose a simple solution to this problem based on discovering gating inputs using "non-inverting paths" and trimming inputs using "partial non-inverting paths" in the circuit's AND-Inverter graph representation. Experimental results show that guarded evaluation can reduce switching activity by as much as 32% for FPGAs with 6-LUT architectures and 25% for 4-LUT architectures, on average, and can reduce power consumption in the FPGA interconnect by 29% for 6-LUTs and 27% for 4-LUTs. A clustered architecture with four LUTs to a cluster and ten LUTs to a cluster produced the best power reduction results. We implement guarded evaluation at various stages of the FPGA CAD flow and analyze the reductions. We implement the algorithm as post technology mapping, post packing and post placement optimizations. Guarded Evaluation as a post technology mapping algorithm inserted the most number of guards and hence achieved the highest activity and interconnect reduction. However, guarding signals come with a cost of increased fanout and stress on routing resources. Packing and placement provides the algorithm with additional information of the circuit which is leveraged to insert high quality guards with minimal impact on routing. Experimental results show that post-packing and post-placement methods have comparable reductions to post-mapping with considerably lesser impact on the critical path delay and routability of the circuit.

Ultra-low-power circuits are becoming more desirable due to growing portable device markets and they are also becoming more interesting and applicable today in biomedical, pharmacy and sensor networking applications because of the nano-metric scaling and CMOS reliability improvements. In this thesis, three main achievements are presented in ultra-low-power adders. First, a new majority function algorithm for carry and the sum generation is presented. Then with this algorithm and implied new architecture, we achieved a circuit with 75mV supply voltage operation. Last but not least, a 64 bit current-mode majority-function adder based on the new architecture and algorithm is successfully tested at 75mV supply voltage. The circuit consumed 4.5nW or 3.8pJ in one of the worst conditions.