From Detection to Safety: Fault Simulation for Functional Safety

In the early days of computer-aided engineering, the primary goal of fault simulation was straightforward: find manufacturing defects. Tools like the LASAR platform created a “fault dictionary”—a cheat sheet to quickly diagnose failing chips on the production line. The focus was on yield and quality. Engineers used simplified models, like “stuck-at-0” or “stuck-at-1,” to represent physical flaws, allowing them to test vast designs efficiently. The process was all about detection and diagnosis, closing the loop between design and manufacturing.

The Paradigm Shift: From Manufacturing to Operational Safety

Today, the story is different. With the rise of self-driving cars, advanced medical devices, and robust data centers, we’ve entered the era of functional safety. The question is no longer just “Can we find the defect?” but “What happens when a fault occurs while the system is running?\” In safety-critical applications, a fault must not lead to harm. This shift, mandated by standards like ISO 26262 for automotive, redefines the role of faults. They are no longer just unwanted manufacturing errors; they are an expected part of the operational environment that the system must tolerate, mitigate, or respond to safely.

The Core of Safety Verification: FMEDA and Beyond

At the heart of functional safety is a framework called FMEDA (Failure Modes, Effects, and Diagnostic Analysis). Think of FMEDA as a detailed safety plan. It breaks down a system, predicts how each part could fail, and assigns safety mechanisms to catch those failures. However, a plan on paper is not enough. This is where advanced verification comes in. Fault simulation and fault emulation are used to inject these predicted failures into the actual electronic design to see if the safety mechanisms work as intended in real-time.

The New Complexity: System-Level Behavior

Modern systems, especially in cars, are incredibly complex networks. A single vehicle might have dozens of electronic control units (ECUs) from different suppliers, all communicating. A fault in a sensor’s ECU must be detected, and the braking ECU must react appropriately—perhaps by engaging backup systems or alerting the driver—all within milliseconds. Verifying this chain reaction is a massive challenge. The number of potential fault scenarios can reach into the billions, and running them all in traditional software simulation could take months.

The Essential Role of Fault Emulation

This is where fault emulation becomes a game-changer. Emulation uses specialized hardware to run the electronic design at speeds much closer to real-time. While simulation is perfect for analyzing localized faults in detail, fault emulation is essential for executing long, system-wide tests with real software, like booting an operating system. The industry now uses a hybrid approach: simulation for deep, localized analysis and emulation for fast, full-system validation. Crucially, both methods feed into a single, unified fault campaign and database, giving engineers consistent and fast results.

A Unified Solution in Action

A complete functional safety verification solution connects all these pieces: the safety intent (FMEDA), the design code, software workloads, and verification results. Key technologies include fast simulation and emulation engines, tools to manage and reduce the massive number of faults, and automated analysis to calculate safety metrics.

Benchmark tests show the dramatic impact. For a typical RISC-V processor design, booting Linux to test faults could take over four months in simulation. Using fault emulation, the same task was completed in just over seven hours. This acceleration is not just about speed; it enables engineers to achieve the high diagnostic coverage required by ISO 26262 with confidence, turning weeks of waiting into days of productive iteration.

Conclusion: A Transformative Journey

The journey of fault simulation is one of transformation, not replacement. Its core principles remain vital, but its purpose has expanded from the factory floor to the open road. It has evolved from a tool for finding defects to a foundational pillar for proving functional safety. By combining the detail of simulation with the speed of emulation within a unified, software-driven flow, engineers can now practically verify that complex electronic systems will behave safely under real-world fault conditions. As technology advances, this methodology will remain essential for building the trustworthy systems of tomorrow.