Issue link: https://iconnect007.uberflip.com/i/726276
18 The PCB Design Magazine • September 2016 The biggest challenges were: A) There was some bias against flex because it is considered too costly, impossible to modify and comes with a long lead time. Our first big use of flex was two impossibly complicated and expensive 30+ layer rigid-flex circuits that I de- signed for the first rover, Pathfinder. I still hear the same bias on each successive program, and on each program printed flex cables are an en- abling technology that allows them to meet the mission goals. B) I spent a great deal of time negotiating with instrument and electronic designers over pin-out designs that would enable efficient use of flex. For the 100 ohm differential stuff, it means talking them into broad-side instead of edge-coupled (not a huge deal for them since the electronics generally used wire between the connector and PWB. This gives thermal com- pliance between the PWB and the chassis, and doesn't overly constrain or stress the solder joints). We also segregated noisy stuff to one edge, and quiet to the other and placed shield line between. C) Controlled impedance. The flex cables used Dupont AP material and acrylic adhesive. To hit the 100 ohm differential I can reduce trace width, but I need to stop at some point to maintain the robustness of the trace (12 mils), increase the offset distance, but this needs to be kept as small as possible or the coupling will shift to through the shield layer and it eats up finite cable width resources (~24mils), and I can increase the distance between the trace layers and the shield layers, but this increases cable stiffness, increases needed twist capsule diam- eters, and static bend radii (~12 mils). D) I created flat patterns for each of the ca- bles by modeling them in their flight configura- tions, and then flattening each design using the sheet metal module in CV Cadds4. When it was all done, I think it came to- gether pretty well. Its original mission was sup- posed to be 90 sol (one Martian day). I don't recall exactly how long they ran them. I think it was over five years, and that they were still mostly operational when they decided to stop the operational funding. You will always need to check me on mission facts. I'm a design mer- rover exterior. From there the wiring goes to all of the actuators and instruments. For the rover's internal wiring, I developed a 4-layer printed flex cable construction (two conductors, two Faraday shields) with edge launched micro-d connectors. I think there were 50 flex cables in the front and rear cable tunnels, each about 1m in length. The only round wires exiting the rover were a couple of RF lines to the antenna, ~20 pyrotechnic lines, and a couple of others due to last-minute changes. This saved considerable mass, volume and, most importantly, it reduced our thermal leakage. The thermal leakage related directly to needed heater power and solar panel size, and operational constraints; for example, how long to heat up before we can do science?. In addition to this, I also created the rov- er wiring diagram, and the flex cable designs for the robotic arm (seven cables up to 3m in length), panoramic camera mast (seven cables up to 1m in length), high-gain antenna (HGA— three cables up to 1m in length), and the mo- bility system (six cables up to 1.8m in length). The mast used a COTS twist capsule, the arm and the HGA used its continuous flex in custom twist capsules, and the mobility had a one-time deployment of a rolling fold for the telescoping structure. JOHN CARDONE ON DESIGNING FLEX FOR SPACECRAFT