University of Ioannina
Self-healing integrated circuits/systems in semiconductor nanometer technologies
Abstract
dc:descriptionThe evolution of CMOS technology over the years has allowed the presence of billions of transistors in an integrated circuit. As the size of the transistors escalates, important reliability issues, both at the transistor level and at the circuits and systems level, have emerged, due to aging phenomena such as bias-temperature instability (BTI), hot-carrier injection (HCI), dielectric breakdown and electromigration or due to ever-changing environmental conditions. One approach to cope with reliability issues is the self-healing concept. That is, developing circuits and systems that will "sense" their aging status as well as changes that the environment sets and react appropriately to continue operating reliably under any conditions. The present thesis focuses on the analysis of the mechanisms that influence the reliable operation of VLSI circuits and systems and the development of innovative methods that can offer them self-healing so that they function reliably and uninterrupted throughout their life. The target of this dissertation is to develop embedded techniques in the field of SRAM memories for early, on-line, excess aging prediction and oncoming failure diagnosis, aiming to maintain, by proper actions, the reliable operation and prolong the lifetime of the integrated circuit/system where the SRAM belongs. These techniques will allow the sensing of the SRAM status and thus, the prediction of upcoming failures on its sub-circuits. Next, our goal is to present SRAM operation adjusting techniques or repairing options for the self-healing of the corresponding sub-circuits in order to maintain the memory’s reliable operation. Two new self-healing methods are proposed for the SRAM Memory Cells and Sense Amplifiers (SAs) and one new method for the SRAM Decoders. Firstly, a scheme that addresses individually the SRAM SAs and the SRAM Memory Cells for aging monitoring on their transistors is presented and corresponding self-healing options are suggested in order to maintain the reliability of the SRAM. The monitoring scheme is based on the use of a small Differential Ring Oscillator (DRO) and the duty cycle of the DRO signal is used for the discrimination of aged memory cells or SAs. Next, a unified approach of the above scheme is presented for the aging monitoring of both memory cells and SAs. An alteration of this technique is also proposed with the use of a reconfigurable DRO (rDRO) for the needs of the self-healing on SRAM SAs. Finally, an aging monitoring technique for the SRAM Decoders, along with an adjusting technique to succeed self-healing are presented. The monitoring scheme suggests the addition of a simple, low cost embedded circuit in order to early diagnose the Decoder’s aging and properly react to ensure reliability and prolong the lifetime of the SRAM. The simulations performed on the above schemes validate their ability for early (before thepresence of failures) self-healing of overaged cells, SAs and Decoders respectively, while offering low cost in silicon area as well as the ability to avoid aging of the circuits implemented when the SRAM is operating in normal mode.
Degree
thesis:*- Grantor dc:publisher
- University of Ioannina
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Dounavi, Eleni-Maria
- Δούναβη, Ελένη-Μαρία
Subjects
dc:subject × 16- Αυτο-ίαση
- Ανίχνευση γήρανσης
- Μνήμες SRAM
- Αξιοπιστία
- Πρόβλεψη αστοχιών
- Self-healing
- Aging monitoring
- SRAM memories
- Reliability
- Failure prediction
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Νανοτεχνολογία
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
- Nano-Technology
Rights
dc:rights- Statement dc:rights
-
- BY_NC_ND
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/47477