The Myriad Challenges of Testing Automotive Electronics

The SAE (Society of Automotive Engineers — http://www.sae.org/) predicts that by 2010, 40% of the sell price of an average car will be the electronic content. In terms of dollars, the semiconductor content alone was about $20 Billion in 2008 (Source Databeans Inc. Market Research — http://www.databeans.net/). While this will drop understandably in 2009, this is only temporary as vehicles can no longer be made without electronics. Emissions, fuel economy, navigation, and safety are all controlled by electronics.

So, as test engineers, how do we address the testing complexity of next generation ECUs (Electronic Control Unit) before it is installed in your favorite sedan? What test systems — and I do emphasize systems as no one ATE platform can do it all any more — are right for the task? In this short article, I will try and make you aware of applications and test strategies. The bottom line is I hope you think very carefully on the questions you need to ask your System Integrator and/or ATE vendors to help you make the right choices.

Product Development

Of course, all new designs are verified and tested prior to production release. What is becoming more important in the testing strategy in the lab is what is known as HILs, or Hardware in the Loop. Basically, the premise is to test an ECU to see how it reacts when something goes wrong — lost signal, shorts to Ground or V_bat, and bad data. The need for Fault Insertions switching systems is being answered by the ATE industry. Several companies provide hardware to create/inject the faults I mention here. As safety has become more of the realm of ECUs in the vehicle, one can see where this type of testing is critical for new designs.

Process Testing

In terms of verifying the quality of the assembly process, the Automotive industry is somewhat conservative in implementing denser technology. According to iNEMI, or the International Electronics Manufacturing Initiative (http://www.inemi.org/), the Automotive industry is slower than other market niches to adapt to smaller geometry (test pad sizes, microvias, etc) in terms of PCB layout. The primary reason is the environment — temperature ranges of —40 to +250 degrees Centigrade (Exhaust Sensors can see up to 580 degrees C) and multiple G force shock potential demands a mechanical design that is more robust than, say a home computer. Component densities per square centimeter will likely remain lower than other consumer products as well. Also, as most clock rates will be under 150 MHz for the foreseeable future, test pads will stay mechanically and electrically viable. Add the issues of product liability when a safety systems fails promotes a go slow/careful philosophy. So it appears that In—circuit testing is still viable in the assembly line for several more years.

AOI (Automated Optical Inspection) is also important. For those ECUs that have limited electric access, AOI is the lowest cost way to inspect for process faults. As Automotive ECUs usually have a large percentage of higher power components (Relays, Coils, etc.), camera height and the ability to focus on both large and small components on the same ECU will be important.