Beyond the Drawing Board: A Lesson Learned in Flex Design

Several recent blogs have been discussing “war stories” about flex and rigid-flex and the lessons learned when problem-solving and ultimately resolving the issue. We all have them. Flex is part art, part science, and even with the best of intentions and great attention to detail, there are occasionally “surprises” in performance. Most often they seem to defy logic until we stumble on the solution, and then of course everything is crystal clear in hindsight.

One of my favorite “war stories” might just be one of the first that I encountered, and I experienced it firsthand shortly after I started working in the industry. Let me set the stage. I had been working with flex for just a couple of years and thought I had learned a lot, but quickly realized what I had yet to learn. I am still thankful that I worked with such a knowledgeable group of engineers. We were asked to build a simple, rectangular, double-sided flexible circuit. There was nothing unusual or particularly note-worthy about the design. The drawing was calling for standard materials; adhesive based laminate with 1-ounce copper on both sides of 2 mil polyimide. One layer was a trace layer, the other a solid copper shield. Line width and space as well as hole size were all in the standard window for fabrication.

There was no particular concern about the design at all and the parts were built and shipped. But not too long after, we received the dreaded call. The unit was not working properly, and they had traced this back to a performance issue with the flex circuit. We started asking questions: How was this flexing? Was it flex to installor dynamically flexing? What was the bend radius? Was there anything violating standard recommendations? In this case, the flex was a dynamically flex application, but with relatively few flex cycles— and nothing in particular was pushing the limits for flexibility.

Unfortunately, after this review, there was no clear reason that this flex was not performing as expected. We did make a few recommendations to improve flexibility, however. For example, the solid copper shield was changed to a cross hatch pattern in the bend area, and the materials were changed to adhesiveless materials to reduce their overall thickness. We then proceeded with another round of prototypes.

Once again, the flexes were not performing as expected in end use and once again we started to review how the flex was being used and what could be putting stress on the conductors. Again, nothing seemed to be out of the ordinary and it was becoming frustrating for the end user and for us as fabricators. Without a clear root cause for the cracked conductors, we recommended additional steps that could be employed to improve flexibility in general. Panelization was adjusted to ensure that parts were fabricated with the grain direction of the copper. All copper was removed from the back side in the bend area, and polyimide stiffeners were added on either end of the bend area to help absorb stress and be sure that the bend was occurring exactly where intended.

An unlikely observation yields an “Aha!” moment

Frustratingly enough, once the flex was installed, there were still intermittent failures. This was definitely frustrating for all involved. We called a meeting at our customer’s site, or perhaps more accurately, we were summoned there to answer for this flex that wasn’t performing. For the first time, we all went to observe the flex in use, rather than reviewing prints, cross-sections, etc. Unfortunately, the culprit wasn’t immediately apparent upon observation. But after a lot of brainstorming and some time had passed, one of my team members noticed that the flex wasn’t just flexing in one direction. The way the box was configured, the flex was bending as we’d intended, but at the very end of the bend cycle, there was a little tweak—an almost undetectable twist occuring on one end. GOTCHA! That small, nearly imperceptible twist had been wreaking havoc for months. Once spotted, a simple adjustment was made, and the flex performed well from then on!

That was a very eye-opening experience. The obvious lesson here was to review all the directional forces acting on a flex, whether intended or unintended. Since that experience, I tend to ask more probing questions about how the flex is being bent, folded, or flexed, and when possible, I try to observe the flex in use. Sometimes minor details have a major impact.

I also will always remember this experience as one that taught me a progression of tips and tricks to help improve the flexibility of any circuit. Crosshatched copper, material selection, grain direction, selectively removing material, and strategically adding polyimide stiffeners, are all important options to keep in mind.

Just like all good war stories, this one was fraught with frustration and challenges as it was being experienced, but it also provided many lessons learned and even quite a bit of humor once all the frustration passed.  Do you have any war stories to be shared? I am sure others would enjoy both the stories—likely shared with a touch of humor—and more importantly, the lessons you learned.

Talk to an Altium expert today to learn more or continue reading about designing your rigid-flex PCB bend radius with Altium Designer^®.