Issue link: https://iconnect007.uberflip.com/i/1359517
54 DESIGN007 MAGAZINE I APRIL 2021 When I ask that, I get answers like faster sig- naling speeds, calculating impedance or loss, opening eyes and avoiding inter-symbol inter- ference, controlling crosstalk, etc. ese are all good answers, but a bit on the periphery in my view. One astute signal integrity practitioner offered that we simulate for only two reasons: 1. To make design decisions (i.e., evaluate tradeoffs during design). 2. To verify a design before manufacturing (verification). e only question that designers really care about is, "Will it work and by how much?" is implies that the simulation should be able to produce tangible metrics that can be related to design success or failure. Fair enough. is is a good description for "why we simu- late," but so far, I've never heard anyone men- tion the most fundamental reason, in my opin- ion, for signal integrity (SI) or power integrity (PI) simulation: To predict the negative impact that the physical world has on the electrical world, and to mitigate or prevent the negative effects proactively. You can have hundreds of thousands of dol- lars' worth of VNAs, oscilloscopes, and simu- lation soware, but simulations will be invalid unless you carefully model PCB fabrication details. e bridge between a high-speed design and a high-speed design that works across several fabricators goes through the backbone of the PCB—the stackup— which touches every single high- speed signal. My chicken scratch in Figure 2 illustrates this. (Yes, I know how to use PowerPoint.) The Progressive Evolution Toward Simulation Over the years, I've tracked five different phases in the evolution toward simulating hardware perfor- mance, as illustrated in Figure 3. To do this, we need to do our best to "sim- ulate what we're going to build," as Figure 2 points out. A good simulator accelerates the rate at which you can experiment with alterna- tives, allowing engineers to optimize not only to signal quality, but also cost tradeoffs. But it's about more than having expensive simulation tools. In my work, quite a number of PCB stackups cross my desk, and depending Figure 2: The fundamental reason for signal-integ- rity or power-integrity simulation is to predict the negative impact that the physical world has on the electrical world, and to mitigate or prevent the neg- ative effects proactively. (Artwork by yours truly.) Figure 3: Most companies use some combination of these five approaches for resolving signal integrity problems.