Over and over again, we still see application notes and reference designs that Advocate the use of a ferrite bead in a PDN to support a high-speed digital component with fast I/Os. These components are often used in two possible instances:

1. As an isolating element between two rails powered by the same voltage regulator module
2. As a noise reduction element (filter) with the goal of reducing noise from transients on the PDN

The appropriateness of the first instance is debatable, as it depends on the power spectrum required at the output of each of the isolated rails, the resonant frequency of the ferrite bead, and its resistance at resonance, all of which demand simulation and testing to fully qualify. In the other case, where the ferrite is placed in series with the power regulator output pin or the VDD supply pin on a digital integrated circuit, it is almost always bad and creates more noise when I/Os demand current at fast edge rates.

I discussed these points with Heidi Barnes of Keysight, an expert in signal integrity and power integrity. Heidi's take aligns with what is described above: the addition of a ferrite in a PDN for a digital processor generally creates more problems than it will solve.

Watch the clip below with Heidi Barnes or watch the full Altium OnTrack podcast episode here.

The Takeaway

The persistent reliance on outdated PCB manufacturing formats like Gerber, despite the availability of intelligent standards like IPC-2581 and ODB++. Resistance to change, lack of awareness, and industry inertia are cited as key barriers, with many designers and manufacturers sticking to what’s familiar—even if it introduces errors and inefficiencies. Adoption of IPC-2581 remains low, but the speaker emphasizes that when large OEMs mandate it, the supply chain adapts. The takeaway is clear: intelligent data formats reduce defects, improve communication, and streamline fabrication—yet progress will only come when enough pressure is applied from customers or forward-thinking vendors.

Transcript

Zach: I'm going to admit I don't use IPC 2581, but part of the reason for that is because all the fabricators and assemblers that I work with—none of them use it. The fab guys want Gerbers, and I think part of that is because 95% of their customers are giving them Gerbers. So they just kind of throw up their hands and say, "Ah, screw it. Give me the Gerbers." The assembly guys, on the other hand, like the ODB++ format. So I always just generate both directly at the same time to make sure that they are consistent, because you brought up the issue of one set saying one thing and the other set saying another—you at least want to make sure they're consistent. So I generate them at the same time and then hand them off.

Those are the formats that I use, and I think a lot of designers are like that. But it seems like the issue here is that either the design community doesn't know enough about this and the value, or manufacturers don't know enough about this and the value. Or, everybody is still stuck in the past because that's what they've been doing for the last 15 years, and there's no impetus to change. So why am I going to stop generating Gerbers?

Dana: Oh yeah—humans hate change, which is normal. Actually, the stats we have show IPC 2581 is used by around 5% of users. ODB++, according to them, is around 20%. So after 20-some odd years of two intelligent data formats, three-quarters of the world still does it the old way. In Asia, it's funny—you talk to a board vendor and say, "Send us your Gerber file." You want to send them IPC 2581? "Yeah, that's Gerber." They call everything Gerber files. It doesn't matter what the format is—CU, whatever—that's all they know. It’s just "Gerber." No, it’s not Gerber at all.

When we did IPC 2581 at the OEM level, we needed a couple of large OEMs, and we needed CAD and CAM software people involved; otherwise, it’s not going to go anywhere. You've got to have a few people make it happen. Now you're seeing large companies—there are already many—requiring IPC 2581 from vendors. As a fabricator, I can't make anyone do it because I’m not their customer. But if a large company comes in and tells me, the fabricator, "In 12 months, I’m not sending you Gerbers anymore—you’re going to take IPC 2581," everyone figures it out.

I'm going through that right now with a company. You can tell your vendors you're not going to send Gerbers anymore, and guess what? It’ll work. Trust me—I've been through this on both sides many, many times. The issue is that people aren’t feeling the pain. The user doesn’t think they're feeling the pain. It doesn’t cost them enough money when it fails. But in reality, it actually does cost money, and someone will figure it out eventually—as long as there’s enough push, maybe from customers or vendors like you've said, to move to this more modern format.

That’s because it requires less human typing when you get the data on both ends. And that’s a defect reduction—flat out. You can’t trust humans. And I’m sure it would be very nice for people who have to manage this whole process to maybe shorten that email chain significantly. I mean, whenever I have to deal with a fab order, I feel like there’s just going to be a stream of emails for the next week. Questions like: "Ordering parts—can you clarify why you didn’t include this layer? We see this in the design file but we don’t see it in the Gerbers…" On and on and on—it’s always something.

The worst part is, companies—especially in Asia—are very customer-centric, so they don’t want to irritate their customer. So they don’t ask nine-tenths of the questions they really have. They just make an assumption, and you never hear about it until you get a board that doesn’t work. "Why didn’t you ask a question?" And some designers, to be honest, hate having questions coming back because they’re busy on the next revision. They don’t want to work on it anymore, so they kind of push people: "Don’t ask me questions."

When I ran my groups between the various companies we worked for, I figured we saw probably over a million packages come in over the years. And our rule was: ask the question. I’ll take the abusive phone call. We want to get them a board that works—not just one that’s manufacturable, but one that actually works. It’s okay—we’ll take the abuse. People fix things after it becomes painful enough. So yeah, that’s kind of an interesting problem.

---

---

---

---