Skip to Main content Skip to Navigation

Analysis and modeling methods for predicting functional robustness of integrated circuits during fast transient events

Abstract : Miniaturization of electronic circuits continues nowadays with the more recent technology nodes being applied to diverse fields of application such as automotive. Very dense and small integrated circuits are interesting for economic reasons, because they are cheaper to manufacture in mass and can pack more functionalities with elevated performances. The counterpart of size reduction is integrated circuits becoming more fragile electrically. In the automotive world, the new trend of fully autonomous driving is seeing tremendous progress recently. Autonomous vehicles must take decisions and perform critical actions such as braking or steering the wheel. Those decisions are taken by electronic modules, that have now very high responsibilities with regards of our safety. It is important to ensure that those modules will operate no matter the kind of disturbances they can be exposed to. The automotive world is a quite harsh environment for electronic systems. A major source of electrical stress is called the Electrostatic Discharge (ESD). It is a very sudden flow of electricity of large amplitude capable of destroying electronic components, or disturb them during their normal operation. This research focuses on functional failures where functionality can be temporarily lost after an ESD with various impact on the vehicle. To guarantee before manufacturing that a module and its components will perform their duty correctly, new analysis and prediction methods are required against soft-failures caused by electrostatic discharges. In this research, different approaches have been explored and proposed towards that goal. First, a modelling method for reproducing the ESD waveforms from the test generator up to the integrated circuit input is presented. It is based on a hierarchical approach where each element of the system is modelled individually, then added to the complete setup model. A practical case of functional failure at silicon-level is analyzed using simulation tools. To acquire more data on this fault, a testchip has been designed. It contains on-chip monitoring structures to measure voltage and current, and monitor function behavior directly at silicon-level. The last part of this research details different analysis methods developed for identifying efficiently functional weaknesses. The methods rely heavily on simulation tools, and prototypes have been implemented to prove the initial concepts. The first method models each function inside the chip individually, using behavioral models, then enables to connect the models together to deduce the full function's robustness. It enables hierarchical analysis of complex integrated circuit designs, to identify potential weak spots inside the circuit that could require more shielding or protection. The second method is focused on constructing equivalent electrical black box models of integrated circuit functions. The goal is to model the IC with a behavioral, black-box model capable of reproducing waveforms in powered conditions during the ESD. In summary, this research work has led to the development of several hardware and software prototypes. It has also highlighted important modelling challenges to solve in future works to achieve better functional robustness against electrostatic discharges.
Document type :
Complete list of metadata

Cited literature [81 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Friday, June 8, 2018 - 3:09:07 PM
Last modification on : Thursday, June 10, 2021 - 3:03:16 AM
Long-term archiving on: : Sunday, September 9, 2018 - 6:13:06 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01811079, version 1


Rémi Bèges. Analysis and modeling methods for predicting functional robustness of integrated circuits during fast transient events. Electromagnetism. Université Paul Sabatier - Toulouse III, 2017. English. ⟨NNT : 2017TOU30046⟩. ⟨tel-01811079⟩



Record views


Files downloads