The senior director of Isola's Analytical Services Lab talks about working with engineers to solve common PCB design issues.
Peg Conn, senior director of the Analytical Services Lab
at Isola, surveys her domain.
EE Times on occasion sits down with technical professionals who are working behind the scenes to ensure the quality and reliability of the components and materials that go into the products engineers design. Here we interview Peg Conn, senior director of the Analytical Services Laboratory at Isola, a manufacturer of materials used in advanced multilayer printed circuit boards.
EE Times: Is running a busy analytical services lab that performs various electrical, thermal, and mechanical characterizations of resin components and laminate materials used in PCBs part of your grand career plan?
Peg Conn: I'm in this position more by luck than through any actual planning. I started out at Isola in manufacturing and then moved into the analytical lab as a technician. While working full time, I decided to go back to school part time to get a bachelor's degree in chemistry. The best part about college was that I was the only student who had no problem getting up early for an 8:00 a.m. chemistry class. Every day in the lab is an adventure -- some days more than others. I only wish I could spend more time in the lab and less time on paperwork.
EE Times: What's a typical day like in the lab?
Conn: Hectic. Just to give you an idea of how busy we are, we are open seven days a week from 5:00 a.m. until 1:30 a.m. the next day. At any given time, we test more than 100 samples, ranging from partially finished circuit boards to R&D samples from resin mixes to raw materials from a new supplier, and deadlines are tight. My boss once came up to me and told me that we had to find some way to test a material -- and we had 15 minutes to come up with a new test.
EE Times: How have things changed in the past 30 years?
Conn: When I started in this industry, the standard was all paper-based material with fiberglass barriers, and traces were a quarter to a half inch. Now they are 5 and 10 mils. Boards were two-sided. Now we're looking at 80 layers. And it is almost mind-boggling to consider the speed at which signals need to transverse a PCB.
EE Times: What is the most important thing engineers want to know when they come to your lab for testing?
Conn: Engineers are really interested in the curing process. If a circuit board is to perform thermally when it is fully populated, the base needs to be cured properly. It's such a big issue, because when you deal with polymers, you can easily and inadvertently be favoring a reaction that you don't want. Crosslinking of the polymers is science, but how you get there is really an art.
EE Times: Are there typical, common mistakes that you see PCB designers often make?
Conn: One of the biggest challenges engineers have when they design a very complex board is that they wind up with a lot of features, which all need to be filled in with resin. One problem we see frequently is called "glass stop," which means there is not enough resin, and the glass is actually hitting the features of the subassembly. That's a good place for delamination and a weak point for thermal excursions. Typically, engineers have no choice in the amount of copper they use, because it is required for performance. That can be overcome by using more prepreg plies, but today everyone is trying to minimize board thickness -- for cost or other reasons -- and board design becomes a delicate balancing act.
EE Times: How is the industry pushing the limits in materials advancements?
Conn: One of the biggest challenges today is that you still need glass fabric for reinforcement in the multilayer boards. Glass has a different dielectric constant as compared to the resin matrix. As an electric signal passes in between the glass fibers, it will affect the speed and, therefore, timing of a differential signal, resulting in skew. For differential pairs at high data rates, design engineers were adjusting signal traces to run on a 15-degree bias in order to mitigate the skew issues caused by the glass fabric. This is a cost prohibitive solution, due to yield issues, and perhaps not very effective at very high speeds. To address this issue, we have introduced GigaSync, which was engineered to eliminate the skew issues. This has required the lab to develop a new set of test processes. More fun in the lab.
EE Times: What do you predict will happen in the next 10 years?
Conn: I don't see the rate of change in this industry slowing down anytime in the near future. I think it is going to keep changing as fast as we can develop new applications for new materials.