Automotive Applications Demand Silicon Lifecycle Management

Every electrical engineer learns early in university studies that automobiles are a highly demanding environment for electronics. Temperature and humidity extremes, noise and vibration, electrical interference, exposure to alpha particles, and other factors all make it hard to design and manufacture chips that will operate properly under all conditions. These challenges are exacerbated as chips age, an important consideration since consumers keep their vehicles longer than many other types of products with electronic content.

On top of the environmental issues, automobiles have high reliability requirements. Owners expect their cars to be ready for a long drive at a moment’s notice. In remote locations or extreme weather conditions, lives may depend on this availability. The growing importance of Transportation as a Service (TaaS) is pushing availability requirements to 24/7/365.

Reliability is also critical while driving; many chip failure scenarios could result in serious damage, injuries, or death. This is even more true for autonomous vehicles, where faults in the field could cause behavior that the drivers cannot correct. As specified by the ISO 26262 safety standard, chips must be able to detect faults and take corrective action on their own.

For all these reasons, the development of chips for automotive applications is extremely hard. Proper pre-silicon design, verification, and validation is essential, but further attention to detail is needed during system bring-up, manufacturing, production test, and operation in the field. In recent years, Silicon Lifecycle Management (SLM) has emerged as a leading paradigm to make product development and deployment more deterministic. SLM consists of two primary steps:

• Deploying sensors, monitors, and structures embedded within silicon designs to gain insight on how devices are made and how they then perform in-field.
• Gathering and analyzing data at every opportunity throughout the lifecycle of silicon devices to provide powerful analytics that enable in-design, in-test and in-field observations and device improvements to be made.

The scope of SLM is broad, much of it powered by artificial intelligence (AI) and machine learning (ML). These techniques optimize power, performance, and area (PPA) better during the design phase, improve yield and reliability during manufacturing, and tune operating parameters for every individual device in the field. The following figure provides an overview of how SLM is applied throughout the complete silicon lifecycle.

The SLM vision can be summarized in a single phrase: monitor, transport, analyze, and act:

• Monitor – Embedded monitors and structures are integrated early in the design phase
• Transport – Data from the monitors is gathered and transported to a unified SLM database
• Analyze – The data from the embedded monitors is analyzed throughout the device lifecycle
• Act – Based on this analysis, insightful decisions can be made in real time at any lifecycle stage

SLM benefits the development and deployment of any chip, but it is particularly valuable for automotive applications. Many of the issues hitting the semiconductor industry in general are acute for vehicular electronics. The environmental conditions are exceptionally challenging, with a risk of faults that could compromise safety. Supply chain shortages have hit the automotive market especially hard, creating enormous pressure to maximize yield. Self-driving cars, connected vehicles, and over-the-air software updates generate security risks. The industry can no longer afford to be blind to what is happening inside automotive chips.

There are many ways in which SLM can improve the PPA, safety, security, and reliability for cars and vehicles of all kinds. For example, continual test throughout the lifecycle can detect many types of problems as soon as they occur. Traditional manufacturing test is complemented by power-on self-test when the chips are deployed in a vehicle and by in-field mission mode safety tests with advanced analytics. This creates a unified test and repair solution that sends critical warning messages when proper operation is compromised or likely to be compromised.

On-silicon monitors in field deployment provide fleet-level visibility into silicon health by collecting data from various sources inside the silicon. Aging effects and other vulnerabilities can be detected and reported early, well before failures occur. This enables predictive maintenance and replacement of chips likely to fail. The SLM analysis provides actionable insights based on purpose-built aging reliability, ML models, and real-time monitor data.

Clearly, SLM offers many advantages for automotive applications if it is applied in an effective and disciplined way throughout the lifecycle. This cannot be a manual effort; chip developers are already overwhelmed with their current tasks and responsibilities. Proper deployment of SLM requires an integrated, interconnected set of tools and intellectual property (IP) that share a unified database. Ideally, this should come from a single vendor so that the chip development team does not have to cobble together a suboptimal solution from disconnected pieces.

The Synopsys Silicon Lifecycle Management solution provides just such an integrated set of IP and automated tools, interconnected by a Unified Data Format (UDF) database. Its capabilities include:

• Embedded process, voltage, temperature (PVT) path margin monitor (PMM) IP
• Silicon-aware design flow improvement
• Yield improvement, production ramp, bring-up, and diagnostics
• High volume analytics screening and production control
• Adaptive high-bandwidth testing
• In-design and in-field performance optimization

SLM is a well-established approach for improving many aspects of chip development and deployment. Automotive applications are a natural beneficiary for SLM because they present so many daunting challenges. Synopsys has the track record and the experience to meet these challenges head-on and provide a better experience throughout the silicon lifecycle. For more on the Synopsys SLM solution, a white paper is available.