Issue link: https://iconnect007.uberflip.com/i/746712
November 2016 • The PCB Magazine 57 UNIQUE IMPLEMENTATION OF A RIGID-FLEX CIRCUIT to PTHs (plated through holes) the expansion of the materials in the z-axis with temperature rise becomes a real concern. As the material 'swells' in the z-axis it creates strain on the plat- ed barrel—with enough strain the barrel cracks, resulting in an electrical open. It is important the CTE match as closely as possible. The most common flex materials have a pretty good CTE match in the 'x' or 'planar' direction, but not so for the 'z'. As we can see in Table 1, once we rise above the Tg (glass transition temperature) of acrylic adhesive at 40°C, our z-expansion rises to 4x the rate of the other materials in our stack. This is the critical reason we moved to a rigid-flex type 4 construction with cutback coverlay and acrylic bonding films. These layers were stopped in the flex/rigid transition zone. This eliminat- ed acrylic adhesive in the via area. The cutback coverlay and acrylic bonding films can be seen in Figures 7 and 8. A note on the Tg value regarding acrylic ad- hesive. There can be concern on the part of de- signers when they see a Tg value of 40°C. This is really of no concern as long as proper design considerations are followed in our stack-up. Acrylic adhesive is uniquely qualified for use in FPCs. Its properties are both thermo-set and thermoplastic. It is exactly these properties that allow the multiple lamination cycles required when building up a multilayer with several sub- composites. FPCs of this construction are often rated for a MOT (maximum operating tempera- ture) of 105°C. Many flexes of this construction are operating at elevated MOTs of 150°C for ex- tended time with no detrimental effect. Bend Radius Considerations IPC-6013 recommendations state that a flex circuit should have a minimum bend radius of 10-times its thickness. As far as guidelines go, this one is pretty safe, however it fails to account for initial part thickness or material type. A very thin flex can go significantly tighter, even fold- ing over on itself, whereas a thicker circuit may require a more relaxed bend. In our application we had to bend a very thick circuit of 0.965 mm (0.038")—if we did not use an unbonded/loose- leaf design. Our calculated minimum allowed radius per IPC-6013 would have been 9.65 mm (0.38"). In real-life situations radii are not per- fect and we often require a bit more flex length for a comfortable installation. We had an un- bonded flex section length of 15.25 mm to form our flex to 90 degrees. Our 'real world' finished radius requirement was determined to be 2.54 mm (0.1") for best fit. This makes our planned design 3.8 times tighter than allowed by indus- try design standards without unbonded/loose- leaf/bookbinder. In a typical instance where we must go be- low the minimum recommended bend radius, we would incorporate a 'loose-leaf' approach. Individual layers would be separated within the flex stack-up (Figure 5). By separating the layers into multiple sub-composites within the stack- up, we can recalculate our minimum bend ra- dius based on the thickness of the individual layer. This can be quite effective; however, in our situation we still had a flex region of only 15.25 mm. Even unbonded layers would not al- low for enough give within the stack to allow for a stress-free bend. The Bookbinder By making each of our unbonded sub-com- posites slightly longer than the one below it, we can allow room for our sub-composite to bend without applying undue stress to itself or Table 1: Common CTE values for flex materials (z-axis). Table 2: Progressive lengths of sub-composites.