Issue link: https://iconnect007.uberflip.com/i/1267313
70 DESIGN007 MAGAZINE I JULY 2020 load "B" and so on comes from the PDN. Generally, the PDN consists of either power planes or power routes with ground planes, decoupling capacitors, and sometimes other devices as the main ingredients. I have done a longer ver- sion of this class at PCB West that was 3.5 hours and took the audience through every- thing I consider important for power delivery. For the PCEA grand opening, I'm going to share an hour of good tips on key features that need low inductance, how to maintain that, and other things regarding PDNs that they need to understand first and foremost. For example, one of the most important elements of good PDN design is maintaining a low inductance through the PDN at a broad range of frequen- cies at which the energy gets delivered. One of the myths engineers believe is that the princi- pal energy of the circuit is the clock frequency, but it isn't; it's driven by the rise and fall times of the IC outputs. They run at a much higher frequency than the clock. Dack: Will you tell the audience some horror stories about how this misconception can play out in design? I love the horror stories you sometimes share because they illustrate what can go wrong when improper assumptions are made at critical points in the design process. What do designers miss with regard to this topic? Hartley: Designers need to care about induc- tance because high inductance at any of the frequencies of concern can cause major prob- lems—and there is a broadband of frequen- cies that the PDN has to deliver. It starts at the clock and extends out to a very high-fre- quency level based on the rise time. All of that energy has to be delivered through the inductance of the power bus. If the induc- tance—or impedance—is high at any of those frequencies, you get large voltage drops that lead to switching noise, which leads to sig- nal integrity and EMI problems, as well as others. It is important to understand how to design the PDN to ensure these things do not happen. Dack: That sounds outstanding. This is you speaking about physics in action. It sounds like you are going to provide the audience with visual representations of stackups and show how the conductors work to distribute power. Hartley: Yes. With everything I present, I try to bring a simple but realistic view of the phys- ics to people so that they can understand why things happen, including a behind-the-scenes approach on how this stuff works and why people need to know it. I do not approach the subject matter like a professor in college who launches into high-level math. Instead, we'll talk about subjects like the proper location of decoupling capacitors. For example, with a BGA part, do the caps belong under the BGA, or do they go on the same side of the board, next to the BGA? The answer to a question like this is, "It depends." It depends on the board stackup, the BGA, the power, and a number of factors. Design- ers have to possess that knowledge when they deal with these elements so that they won't do it incorrectly. Without consider- ing these elements, they will end up creating some amount of power starvation, resulting in inductive losses and switching noise that will lead to other problems. It is very impor- tant to know these tips, and I have chosen the important ones to go through during the one-hour class. Rick Hartley One of the myths engineers believe is that the principal energy of the circuit is the clock frequency, but it isn't; it's driven by the rise and fall times of the IC outputs.