Avoid Overbuilding your RF Printed Circuit Board


Today, many companies are overbuilding and “overmaterializing” their RF printed circuit boards. In this interview, James Hofer of Accurate Circuit Engineering (ACE) shares some strategies to avoid doing both, which will help lower the total cost of your PCB and improve the overall product quality. Hofer also discusses some of the challenges in the laminate supply chain.

Barry Matties: Why don't you start by talking about Accurate Circuit Engineering; what do you do and what is your specialty?

James Hofer: Accurate Circuit Engineering is a high-tech prototype facility. We specialize in low volume, high mix. Our specialty is RF and signal integrity products. We do a lot of good cavities, and we handle all specialty materials; we especially use a lot of Rogers, Taconic, and Arlon, which is now also Rogers. We stock R4000 series, because 4000 series is an advanced dielectric for signal integrity. We probably run somewhere between 15 and 30 jobs a week on R4000, which we stock because Rogers just recently was quoted as, "We're very proud to say our lead-time on the 4000 series is now only 35 days." That's more than a month.

We specialize in quick-turn. I can run 16-layer, high-frequency jobs with impedance on multiple layers in 24 hours. We can also go up to 20 days. We do everything from the AT test boards to gigantic antennas. Everybody who has a wave-guide analyzer or an amplifier is using boards for those, and our boards are just the kind of things to go in there.

Matties: Yes, RF is a booming area right now.

Hofer: RF satellite is back up and on top of everybody's minds right now, with all the advanced unmanned vehicles that everybody from civilian to military is using. All of that is RF. We're going faster and faster, higher and higher. All our smartphones are using RF-type and wireless technology. All of that requires test equipment, design equipment, prove-out equipment, and every one of those has a circuit board.

Matties: Your manufacturing focus is RF. How long have you had that focus?

Hofer: I brought that focus on board with Accurate about 12 years ago. I've been with Accurate 15 years now, and when I started we were just doing rigid boards, whatever came our way, and I wanted Accurate to move forward and have a specialty. When somebody says, “Who does RF boards? Oh, Accurate does RF boards,” we are right there. That has really worked out for us, because not only are we seeing a lot of activity in that area, but we were also very good at it. It has really propelled us to the next level.

Matties: We've done a few surveys recently—you may have seen—about supply chain and also the war on failure. In the RF arena, failure is just not even an option when you're dealing with military satellites, commercial aircraft, that sort of thing. How do you battle failure in your process?

Hofer: We start from the minute we get the customer design. We run it through DFM and DRC checks. We verify their calculations. We take a look at how their structure is and how their stack-up is and how they've laid everything out, to make sure that it's going to be robust enough in the finished product to withstand multiple heat cycles, vibration, etc.

Then during our own manufacturing process, because we're also MIL-certified and AS9100 certified, we have process checks and process documentation to ensure that when we're manufacturing the board, it's seeing only processes that have been vetted and are monitored in performance, tracked, and managed. Once the finished product is done, we put it through construction integrity testing, full electrical testing, and then full dimensional and measurement testing before we ship it out to the customer.

Matties: Interestingly, I've looked at this in the supply chain from fabricator to assembler, and it seems to me that no matter how stable your processes are, you could be using best practices and all that but inspections are the only line of defense.

Hofer: Nothing is 100%. Even computers are 99.9999%, and that's not a hundred. Inspection is really the key to prevent escapes, because you're right, in manufacturing, in life, mistakes are going to happen and people are going to make mistakes and mistakes make people. It's how you move forward from them that's important. The inspection steps—which is why we have so many inspection steps along the way of manufacturing and at final—are the only way you keep it from getting out to assembly and then getting out into a product. Even that, unfortunately, is not at 100%. Sometimes things escape. We see it on the news all the time.

Matties: Absolutely. Airbags, ignition switches, falling satellites, or whatever it happens to be.

Hofer: That's correct.

Matties: It's really interesting, and you mentioned materials. The other part we looked at was supply chain and how difficult it is to get materials. What sort of demands or impacts has that had on your business?


Hofer: We're a quick-turn manufacturer in prototyping, and in many good R&D processes, an engineer can run a little bit late, because you always think of a way to do it better or something that you didn't address, and you have to go back and fix it. We find an incredible amount of demand for RF products in five days or less. The biggest problem that we run up against is a lot of the advanced dielectric manufacturers have lead times that exceed 20 or 30 days.

Rogers is a big name, but there's also Taconic out there, there's Isola making high-frequency products, and actually we're very excited by some of the products that Isola is developing. You've got Nelco in that arena as well, and the issue comes when somebody needs a board in five days; if you have to wait 10 or 15 or 20 business days to get material, it throws the whole project out the door. We combat that by stocking a lot of that material and taking it upon ourselves to make sure that we keep a min/max stock on hand so we can eliminate that lead time.

Matties: That's good. Not everybody has that ability.

Hofer: No, and I think that sets us apart. We got an engineer who needs to put a satellite up next year, and he's got this last board in design and he says, “I need this one in three days.” No problem, just hold on three weeks while I get the material. It just doesn't go over well.

Matties: Isn't part of the problem that we see, though, is these OEMs and engineers are spec’ing that material? Are you seeing a shift where they're giving you second-source approvals?

Hofer: With customers that are repeat customers, we're finding that they understand that we offer help at the beginning when they're doing that stack-up. We can then say, “Okay, if you run with this material, you might have a lead time of 20 business days. Here's two options. This one has a lead time of 5–7 days, this one we stock, or this one has a lead time of 10 days. Here are the performance differentials.”

Typically, between high-performance material manufacturers, you can see Dks within 0.1 between differing materials and loss tangents between 0.001 between different materials, and sometimes your design can handle that. If it can and the lead-time is much less, it's nice to have that option.

Matties: In terms of bringing new customers into the business, one of the things that we're looking at is the process of qualifying a vendor. There was a day when audits were just mandatory. Are you seeing a lot of audits demanded in your situation?

Hofer: I think the industry is moving towards allowing the AS9000 certification and the MIL-certification to circumvent the audits. We still see some of the bigger OEMs maintaining an in-house auditor that goes out and does auditing. We've had a couple of smaller aerospace customers hire a third-party auditor to come in and audit, but I can say as an average, on-site audits are trending downward, because with the AS9100 certification, you've got audits every year. With the MIL-spec certification, we've got audits every two years, and those audits aren't just come in for a day like the old ISO audit. It's a three-day, two-man party that comes in and thoroughly vets the system. They're allowing your internal process controls and the certifications to suffice versus on-site audits.

Matties: Regarding the topic of lowering costs, there's obviously a big demand for lowering costs.

Hofer: There's always a demand for lowering costs.

Matties: What strategies do you bring, not necessarily to take the profit out of what you're doing, but to help your customers lower their costs?

Hofer: That's a very good question and difficult for a quick-turn prototype house, because one of the best ways to lower cost is have it built slower—increase your lead time. What we do is offer lower-cost alternatives. We help with the stack-up and the DFM at the beginning. We try to run some processes as lean as possible. It would be great if I could run my raw material very lean and just bring in material as I need it, but unfortunately the cost versus gains on lead times is not there.

What we try to do is offer design assistance and stack-up assistance to help use the most cost-effective performance materials and to make sure that the design is robust and gets you what you want the first run.

Matties: Don't overbuild it, and don't over-materialize it.

Hofer: I've had several people say, “Wow, why don't they just build the whole board out of Rogers?” Well, it's not cost-effective. Run some FR-4 in there. That's a fifth to a tenth of the price of the Rogers, where you don't need the Rogers. I've had other customers come up and say, “I've heard Rogers doesn't bond well with FR-4.” Well, as a manufacturer who has run literally every part number of Rogers that they've made, I've mixed 90% of them with FR-4 and they all work great.

I think there are also some misconceptions with some designers not understanding that they can mix and match materials to make the most cost-effective package.

3024.jpegAgain, that’s why we like to work with the designers and the layout folks at the beginning. You can ask us to help design cost-effectiveness into your package rather than send it to us and say, “Well, you asked for all Rogers. Gee, if you had talked to us a couple months ago, we could've suggested replacing a few of these layers with FR-4.” Or, “Oh, wow, you're using that material. This other manufacturer has a low-cost alternative that's almost exactly the same electrical properties as the one you're using, but it's at a third of the price or half the price.” That's why we like to work with them in the beginning.

Another thing that I like to do is go in and teach designers how printed circuit boards are manufactured. A lot of designers will design in aspects to their board that they think is the way they need it, not understanding that the cost of manufacturing this particular methodology is not cost-effective for them. They can do the same thing, for example, with somebody saying, “I need microvias because I know I can't overlay conventional vias.” Well, there's a back-drilling option, so if you can live with a 2-mil via stub, back-drilling is going to save you 8% on your board price.

A lot of people just don't know that, so when we say, “This is what you do when you manufacture a printed circuit board,” many people will respond, “Wow, I never knew you did it that way. Now when I go design a board, I'm going to think the way that you're manufacturing it.” It also helps that they understand the manufacturing process, because then they don't design stuff that doesn't manufacture, or that you can't do. It costs time and money when you have to go back and redesign your board.

Matties: Well, this is great stuff. I really appreciate your time today. Is there anything else that we should share with the industry that you think they should know?

Hofer: Keep moving forward with high-speed. The world is becoming a digital living space. Nothing's going to be tied down or hard-wired anymore. The wave of the future is all wireless, separate, and digital.

Matties: You mentioned high-speed. I just did an interview with Vario Optics, and they are embedding optics in circuit boards. Do you see that as something that you're going to look at to stay up with the speed and data transfer?

Hofer: Absolutely. A lot of printed circuit board laminates are made with woven glass. Using that glass as a fiber optic is the next logical step, and that's going to increase our speed ten times. I see that as definitely something we will be involved in.

Matties: How do you advance your process? Do you have a team of internal engineers, or do you go out to the supply base?

Hofer: We have a team of internal engineers, and then we work closely with other customers. I had a gentleman come by the booth talking about something that we've been discussing internally for a long time, and he's from a major OEM. First thing we said was, “We'd love to partner with you on this. We'll work something out so we can run it, you guys help us with testing it on your side, and we'll call it even.” I think those are the advances that are going to make the industry go forward.

Matties: James, thank you so much.

Hofer: Thank you, Barry. I appreciate it.