Save Your Time and Effort in Optimizing a Differential Pair Via Transition

10 Jan 2024 • 5 minute read

Technology has become an essential part of our lives and is transforming our world. However, this leads to complex system design complexity and increased design simulation parameters required to ensure performance, functionality, and reliability concerns. Designers know optimizing a design with many simulation parameters can be highly challenging, as it requires significant computational resources, time, and cost. This can eventually become unsustainable. Consider a scenario where a design simulation has 10 controllable parameters, and every parameter has 10 possible values. To optimize this design, we must simulate all 10 billion simulations! It will take decades to complete using the traditional human-intensive flow (design, test, and refine). Realizing a design that works the first time is crucial, and accurate electromagnetic (EM) simulation is a critical step in this process but takes a lot of time.

Could we achieve a genuinely optimal design by replacing human involvement with a conventional optimization workflow?

It turns out that using reinforcement learning – as used in the Cadence Joint Data and Analytics (JedAI) Platform – we can very quickly build a machine learning (ML) model for the optimal design of a system of any size or complexity. The Cadence Optimality Intelligent System Explorer is a generative AI-driven multiphysics optimization software that uses advanced multidisciplinary analysis and optimization (MDAO) technology driven by artificial intelligence (AI).

Optimality Intelligent Explorer is fully integrated with Clarity 3D Solver and Sigrity X, our high-speed signal and power integrity (SI/PI) platform. By offloading system-level SI and PI simulation and analysis to Optimality, designers can arrive at an optimized design far more quickly – reducing the need for design respins and accelerating overall time-to-market. This blog focuses on the Clarity 3D solver part using a simple test case that showcases how to save time and effort in optimizing a differential pair via transition using Cadence Optimality Explorer. Here we have a PCB board with the 6-layer substrate, GND plane on top, bottom, and layer 3, Diff pair on layers 2 and 5, VDD plane on layer 4 Plated thru-hole (PTH) Vias, GND PTH Vias next to signal Vias.

Here, the design and analysis goals are to achieve the target differential impedance and optimize for low insertion loss (S₂₁) and return loss (S₁₁) at specific frequencies. Optimality is used to parameterize this layout to see what can be changed for a better solution; while comparing it with the manual process, we discovered that Optimality Intelligent System Explorer automates the design process and offers early design closure. As shown below, the manual process involves a repeated loop of creating an initial layout, defining ports, running simulations, documenting the results for each case, and exploring what to modify. This can be time-consuming and requires a lot of manual steps.

The application of Optimality Explorer with Clarity uses AI deep learning to find a solution more efficiently than brute force methods. Optimality Explorer automates the process and eliminates the need for manual interaction. The optimality flow involves defining the parameters to optimize and setting optimization goals, after which the tool automatically creates layouts and simulates them to achieve them. It enables multi-physics simulations to reduce the required simulation data or simulation time required to get to a design closure using the least number of simulations.

The Optimality approach is much faster and can be run overnight without manual interaction, making it a more efficient way to explore different design options. It can be used in all design phases, such as Spacing, the position of the return path via, cutouts in the planes, drill size, and anti-pads are some of the values that can be explored depending on what is feasible in the manufacturing process.

How it Works/Steps: How to Set Up an Optimality Run?

Optimality is a powerful tool that automates the design process. It intelligently picks random values within your set boundary conditions by defining the parameters and variables for optimization. The objective function goal and function type can also be defined for optimization. Once the simulation is completed, you can observe the graphical and convergence plots for the best result. If you're unsatisfied with the outcome, you can resume the simulation and continue optimizing until you achieve your desired results.

Utilizing this tool can speed up finding a superior solution. Further, the template feature allows the automatic addition of design variables to the parameter list, making optimizing routes and maintaining consistency across all design layers easy. Even in the low-frequency range, you can improve significantly by using this method to find a better solution. The complexity of positioning vias around a return path, via around your signal, can be significant, especially if you have micro via and buried via. The signals' transition from the core to the bottom layers can be complicated. Using this tool's features, you can quickly sort and load the parameters, compare the best and worst case, and see the significant difference between them. With the help of this tool, early adopters have witnessed the optimization of via structure to improve their results.

“As an early adopter of the Cadence Optimality Intelligent System Explorer, we stressed its performance on rigid-flex PCB with multiple via structures and transmission lines. The Optimality Explorer’s AI-driven optimization allowed us to uncover novel designs and methodologies that we would not have achieved otherwise. Optimality Explorer adds intelligence to the powerful Clarity 3D Solver, letting us meet our performance target with accelerated efficiency.”

-Kyle Chen, Principal Hardware Engineer at Microsoft

Conclusion

The industry needs a paradigm shift from the good old-fashioned flow that uses design, test, and refine loops to generative AI-driven technology to obtain an optimized design while meeting time constraints. Many SoC designers and early users of Optimality Intelligent System Explorer now proclaim the end of conventional human-intensive optimization processes. Optimality Explorer enables design teams to optimize their designs, resulting in faster detection and mitigation of thermal issues early in the design process, significantly reducing the time it takes to reach a truly optimized design iteration. It enables design engineers to explore 3D electromagnetic (EM) and high-speed signal and power integrity results and zero in on the optimal design. It helps design teams accelerate electronic systems' analysis and optimization realization without compromising accuracy. It empowers them to analyze and optimize 3D electromagnetic (EM) and high-speed signal and power integrity results easily.

Learn More

The blog is an excerpt from the presentation by Martin Biehl, AE Director, and Lokaja Bonagiri, Sr AE at CadenceLive’23 Europe. Please register to view the complete presentation.