It sometimes feels that the future of everything is related back to technology. Drones? Technology. Autonomous Vehicles? Technology. Wearables? Technology. The Internet of Things? T-E-C-H-N-O-L-O-G-Y.
As a simulation engineer/manager, why should you care? Because eventually the system that you are working on, if not already, will likely incorporate electronic hardware that is the foundation of all technology. And, increasingly, the adoption of these technologies throughout all environments, not just home and office, can result in thermal, mechanical and reliability challenges. To respond to these, DfR Solutions created Sherlock Automated Design Analysis (ADA) software.
Finite element analysis (FEA) is typically the go-to-solution in resolving these challenges, but the electronic hardware development process has some unique challenges not seen in other areas. The first is that electrical engineers control an extensive amount of the hardware design process. Electrical engineers often know little about FEA. Some use electronic design automation (EDA) software that is not compatible with FEA (no .stp outputs, much less .cae or scripting).
This is where the Sherlock software first steps in and provides a powerful pre-processor for Abaqus users. Sherlock can read all EDA output files and translate them into intelligent 3D models mesh-ready for Abaqus. This is not a trivial process since EDA files are flat 2D plot files with almost zero mechanical information at the part level. Sherlock uses a unique combination of powerful parsing code, embedded and global libraries (materials, part, package, laminate, solder), part number translators, and standardized build modules (wire bonds, heat sinks, mount points, etc.) to provide an out-of-the-box solution.
An intuitive user interface, consisting of easy to use fill-in fields and drop-down commands, allows the user complete control over model complexity. Users can start with a simplified model, consisting of bricks on a flat plane, with simulation times under 10 minutes. As knowledge is gained about the product and the user desires to better represent real-world geometry, Sherlock is able to ramp up detail to the point where every feature (ball, trace, via, lead, corner) is represented in high detail. Important aspects of electronic hardware can be captured in Sherlock, including complex daughter card/mother board configurations, flex and rigid-flex, and even housing. The result is that most users are able to go from raw EDA to a Python script ready for input into Abaqus in under 90 minutes.
The user, whether a beginner from the design team or extensive Abaqus expert, can now perform a range of simulation activities. This is because the Sherlock Python script has not only provided model dimensions, parts, and material properties, but also tied contacts, boundary conditions and environmental loads. Abaqus is now able to do what it does best, performing a wide range of implicit and explicit simulations designed to capture the temperature rise (thermal) and mechanical response (strain, creep, fatigue, displacement) under a range of conditions (temperature cycling, harmonic vibration, sinusoidal vibration, mechanical shock, bending).
Once the Abaqus simulations are complete, we experience the second challenge with electronic design: the sheer number of unique parts and failure modes. The average electronic hardware has between 500 and 5000 parts that could fail. Each part, could have between 1 and 10 different mechanisms that could cause failure. The complexity of it all often results in organizations defaulting to ‘rules-of-the-thumb’ (e.g., temperature rise cannot exceed 20C, board-level strain cannot exceed 500 ustrain) rather than harvesting the true value of the simulation activity.
Now, Abaqus simulation results (Thermal, mechanical, or both) are imported into Sherlock’s rigorously validated damage models. To access this potent capability, the user has to literally do nothing other than import the results from their Abaqus simulation (thermal, mechanical, or both) and leverage Sherlock’s position as one of the most powerful Abaqus postprocessors on the market today. Sherlock automatically identifies every part in the model, recognizes it’s technology and packaging, determines the failure modes most relevant to that part, and then makes a prediction for every unique combination of failure mode and electronic part.
The result is a complete life-curve for your technology. True tradeoff analysis (this material is 8% cheaper, but will raise warranty costs by 5%) can now start at the very beginning of product development process. First-pass success can now be assured, driving down time to market (especially true for automotive with their 6-12-week thermal cycle requirements). And minor improvements and modifications can now be validated through software rather than a complicated and costly physical validation testing.
Because of this powerful synergy between Sherlock and Abaqus, we have forged a deep and close relationship with Dassault Systèmes SIMULIA over the last several years. Our partnership goal is to drive deeper integration between the two software tools and to educate the larger technology world of the value of iterative Sherlock-Abaqus analyses during the product development process.
For more information, visit DfR Solutions website.
