Issue link: https://iconnect007.uberflip.com/i/1507356
12 DESIGN007 MAGAZINE I SEPTEMBER 2023 need two cable harnesses, one from board one to board two, and one from board two to board three. at's a grand total of five different item numbers I would need for each board. I would have three drawing numbers: a number for the schematic, a number for the board, and a number for the assembly. For three boards, that's nine drawings I must maintain, and then another two drawings per cable, and let's assume two cables. at's a total of 13 separate drawings and part numbers I have to maintain in my system. Now, if I take all of that and integrate it into a single rigid-flex design, it's much simpler to maintain. Before, when you had three differ- ent circuit boards and had to do a revision to one of the boards, you had to go through the process to make sure that the new revision of one board works 100% correctly with the old revision. I've seen that in my career multiple times. But with rigid-flex, it's all one design. It helps with packaging because you want to get smaller and smaller packaging. Again, no more harnesses, cables, or connec- tors, and it's much easier to deal with. Of course, material selection is much more complex, and there are a lot more manufacturing challenges, such as specifying rolled annealed copper for flexibility. We have to understand the stresses on copper. If I bend it, what are the bend radii and ratios that I have to deal with? I don't want to over-bend the copper and fracture it. ere's no woven reinforcement in the flex materials and you have a much higher amount of layer-to-layer misregistration, shi, and float during the lamination and bonding pro- cesses. I can't just take my part from a library that's good for a rigid board, throw it on flex, and expect that I'll get the same manufacturing producibility levels out of these boards. Again, these are all challenges you have to talk to your fabricator about as well. If we have a bunch of traces cross the flex section, we must be care- ful. Let's say I put one trace on top of the flex, and one on the bottom of the same flex core, so they're directly on top of each other. Now when I bend, one trace is put under compres- sive load, and I can have stress fractures and so on. We need to offset our traces from each other. Basically, I'm trying to get the students to understand the very high benefit you get from using rigid-flex, but also the gotchas, challenges, and pitfalls. It's not as simple as throwing a bunch of parts on a board and routing it. Shaughnessy: Your classes spend some time on rigid- flex tradeoffs. There seem to be quite a few. ere are multiple differ- ent factors that go into the break point of whether you should go rigid-flex, or dis- crete boards with flat cables, or traditional discrete boards. But I think the days of doing everything on discrete boards with traditional cable harnesses is fast going away. e fact is that cables are the single greatest point of failure in any modern system. Happy Holden: One thing that seemed clear to me is that rigid-flex, as a design alterna- tive, takes somebody with a lot of knowledge of all these different trade-offs. We need a complex tradeoff tool, where you can talk about each of your boards, each of your connecting harnesses and boundary I'm trying to get the students to understand the very high benefit you get from using rigid-flex, but also the gotchas, challenges, and pitfalls.