Issue link: https://iconnect007.uberflip.com/i/1007258
60 FLEX007 MAGAZINE I JULY 2018 ing with flex. Several decisions must be made: RA or ED copper, adhesive-based materials, or adhesiveless materials, copper thickness, dielectric thickness, coverlay or flexible sol- der mask, what type of stiffener, polyimide or FR-4? Skill and knowledge is required to bal- ance those decisions with the end use of the circuit, available materials, and cost. There were a few stories about "the flex that didn't flex" when a multilayer stack-up became so thick there was no way to bend the circuit without cracking the copper. This seems to be a common occurrence—it has happened to me in the past! As a side note, most were resolved using unbonded layers in the stack-up. Another common message was that material lead time seems to be longer than expected, with more questions about the stack-up than anticipated. It is true there are a lot of vari- ables in inventory, preferences, and capabili- ties between fabricators. The piece of advice given most often was, "Work with your fab- ricator during the design and to understand their capabilities." Great advice. Conductor Routing Conductor routing practices was another category that stood out in the conversations about lessons learned. Nearly everyone has a story about cracked traces and the learning curve they went through to be confident in the flex design and performance. A flexible circuit is a hybrid of mechanical and electrical design. This introduces a lot of variables. I'll share one story that stood out. The appli- cation required a double-sided circuit that was expected to be flexed during installation and test, but not over the life of the product. The first design used solid copper for shielding and was manufactured with adhesive-based mate- rials. It cracked in the bend area during instal- lation. Several new ideas were implemented for the second revision. The traces were rerouted perpendicular to the bend area, materials were changed to adhesiveless, and cross- hatch shielding was added. These are all great options for improving flexibility. The second revision cracked in the same location. For the third revision, traces were routed to just one side of the bend area, and all copper was removed in that area. In addi- tion, polymide stiffeners were added to help more specifically direct where the bend was occurring. Even though all the best practices were employed in this design, the third revi- sion cracked also. The problem was resolved when they realized that the circuit was not just absorbing the stress from the known bend area, but as the unit was working stress was being applied in another axis. A slight redesign of the unit eliminated the cracking. This had to be a painful and frustrating experience for all involved, but it also was a good lesson in ways to improve the flexibility in any design. I received a lot of real-world advice for conductor routing. A few of the key items included: avoid abrupt changes in conductor size and direction, route conductors uniformly and perpendicular to the bend area, add radius to all inside corners, make pad patterns big- ger to add stress relief, and add anchoring tie points to the solder pads to reduce the oppor- tunity for pad lifting during assembly. Improving Flexibility Another common topic of discussion was the learning curve for options to improve flex- ibility. The previous example provides many tips and tricks pertaining to conductor routing. Wisdom was shared for additional options to consider, especially relevant for dynamically flexing and applications and when tighter than There were a few stories about "the flex that didn't flex" when a multilayer stack-up became so thick there was no way to bend the circuit without cracking the copper.