Speaker
Description
The scaling trend of highly integrated circuits makes them more and more sensitive to single event effects (SEE). It has now become necessary for circuit designers to figure out the sensitivity of their circuit and new technology during the design phase. The purpose of TRADCARE is to assess their reliability before manufacturing and therefore, reduce the testing costs. TRADCARE can also be used on existing technologies as preliminary analysis, to assess the testing need.
SEE occurrences are ruled by a stochastic process. Many prediction tools have grown up around of the Monte-Carlo approach which is commonly chosen to describe the particle transportation through the matter. Geant4 offers many interesting features both in technical (extended physiclist, low energy physics…) and practical (full custom, cross-platform, c++) aspects, and is the core of several prediction tools. However, the use of very detailed physical processes of particle transportation coupled with simplified approaches (i.e. parallelepiped sensitive volumes, single energy threshold value or critical charge) may only provide a rough estimation of SEE occurrences.
The spreading parasitic effects of a single particle strike results in an observable and measurable failure or defect at the circuit and system scale. Therefore, every physical mechanism involved in SEE should be thoroughly considered.
A multi-physics SEE prediction chain, including Geant4, was developed and implemented in the TRADCARE software. It is an engineering tool developed by TRAD, a company renowned worldwide for its unique expertise on radiation effects. This work presents:
• TRADCARE features and the characteristics of the implemented modules;
• the simulation flow with an emphasize on the Geant4 part;
• an application case on simple and complex circuits with the experimental comparison.
