At the conclusion of our webinar, Designing Flex and Rigid-Flex PCBs to Prevent Failure, we had several questions submitted to our presenter, Zachary Walker, Product Manager of Flex and Rigid-Flex Circuits at Epec. We have compiled these questions into a readable format on our blog.
Q&A From Our Live Flexible PCB Webinar
Quick Links:
- In the past, I had a part that delaminated. Can it be fixed once it starts or workable in some way?
- At what point does a flex board become considered “dynamic” flex vs. “static” flex? If the end user only flexes the board a few times (i.e. while installing), can you design with multiple layers in mind?
- I've got a design that doesn't have copper fill, but I can't do the data changes without setting the project back with paperwork. Can you make the change for me?
- Is copper fill recommended for all layers, even the inner layers?
- You mentioned adding strain reliefs. Can you expand on [1] when that's necessary and [2] how you do it?
- What does IPC-2223 cover?
- How significant of a cost increase is there between 2- and 4-layer flex?
- My assembly doesn't allow for a bend radius to the numbers you listed. Is there any way to get around it?
- Why would gold FPC fingers start cracking?
- How is the minimum bend radius determined? Is there a large variation between construction types or stack-ups?
- In a 4-layer board, does one always have to put the flexible part to be attached to the rigid part on layers 2 and 3 or could I use other layers? Is there an advantage or disadvantage to using layers 2 and 3 as opposed to layers 1 and 2?
- In a 4-layer board can I have flex circuits come out in the same place and place one flex on layers 1 and 2 and the other flex on layers 3 and 4? Or is this not done?
- Are there significant delamination issues “microvia-ing” from rigid to flex layers? Mitigation/possible?
Watch the Recording Below:
Question: In the past, I had a part that delaminated. Can it be fixed once it starts or workable in some way?
Answer: Unfortunately, parts that have started delaminating aren't really fixable. Trust me, I've tried for my personal projects. Once they start going, it's not something that can really be fixed. Sometimes the delamination can be stopped, but you still run a risk later on that the delamination continues and perhaps gets worse even. So, ultimately if a design starts to delaminate, most of our customers will opt to get rid of or replace the boards. Usually, by the time they start to delaminate, the only good use for the board is perhaps for something like a fit check or something like that. Electrically, they're either not going to function or they will function but again you don't want to trust a delaminating board.
Question: At what point does a flex board become considered “dynamic” flex vs. “static” flex? If the end user only flexes the board a few times (i.e. while installing), can you design with multiple layers in mind?
Answer: The difference between static and dynamic is a bit of a messy question. IPC-2223 itself, to my knowledge and a search, does not seem to delineate the two with a hard and fast bend count. The way we like best to advise customers is that a static bend is best defined by boards that are meant to bend to install, and perhaps only a few more times over the lifetime, so something on the order of 10 or fewer bends. Whereas dynamic bends tend to approach the more “infinite” number of bends, as they’re meant more for repeating bending over the lifetime of the flex.
Question: I've got a design that doesn't have copper fill, but I can't do the data changes without setting the project back with paperwork. Can you make the change for me?
Answer: So, most of the changes that we require of you can be done with permission. So, those added non-functional pads or copper fill. Now, some changes, we are going to either have to look at on a case-by-case or maybe have a customer do, such as via movements, things that could require layout shifts. Ultimately, some things like copper fill, line width changes to a small degree, or things like that, can be done by us but again it's kind of case-by-case. So, I don't want to make a promise one way or another on that. We can advise on it, and we can inform you if you have those issues in the design, whether or not we can change them, or if we believe you should do so.
Question: Is copper fill recommended for all layers, even the inner layers?
Answer: Yes, copper fill is especially recommended for inner layers. Copper fill serves to keep a uniform thickness and good point of adhesion for lamination processes, which is needed for inner layers to laminate properly. Without copper fill, boards may end up uneven on the surface, or worse, delaminating later.
Question: You mentioned adding strain reliefs. Can you expand on [1] when that's necessary and [2] how you do it?
Answer: Strain reliefs are necessary when a bend is occurring close to the rigid-flex interface. If a bend is done too close to the transition, it can sometimes sharply crease which can exceed minimum bend requirements and cause circuits to deform and lose ductility. A second reason is that if a bend occurs in that region, sometimes the rigid prepreg will have extruded out of the rigid portion partially during lamination, which can form sharp, jagged edges that may puncture or cut the board if it is bent towards that interface. Regarding the process of applying strain relief, it’s done by mixing a two-part epoxy that is then poured into a syringe. Once inside the syringe, it is usually pushed out mechanically or pneumatically and guided by an operator along the edge of the interface. After it is fully applied, it is then baked to fully cure it.
Question: What does IPC-2223 cover?
Answer: IPC-2223 is the design standard for flex and rigid-flex. It's kind of the guideline on designs that was put out by IPC. It's going to cover things like cover lay requirements just via placements, like, we mentioned earlier, notation issues or notations on copper, things of that nature. In a way 6013 is the builder's handbook but IPC-2223 is the designer's handbook.
Question: How significant of a cost increase is there between 2- and 4-layer flex?
Answer: Unfortunately, that becomes a difficult question to answer accurately, as it is material, size, and copper weight dependent. Typically, we recommend submitting a request for a quote to help determine it for any parts in question, as we can rapidly get the quotes over to you.
Question: My assembly doesn't allow for a bend radius to the numbers you listed. Is there any way to get around it?
Answer: The best way to kind of fix bend radius when allowed by a design is to just ultimately reduce that flex thickness, so that way you can get your bend radius down to what you need it to be. This can be done by cutting copper weights down and reducing dielectric thicknesses. If you're unsure if it can be done or if you're unsure because your design has to function in a certain way, then it's always best to reach out to us because we will be able to help with that design to assist and find ways where maybe we can help reduce that minimum bend radius, and get it to where it can suit your design.
Question: Why would gold FPC fingers start cracking?
Answer: One possible explanation is that the FPC fingers received a bend during a process which may have resulted in the crack forming due to the gold. If the gold is done by ENIG, this involves an electroless nickel underlayer, and that layer would be brittle enough that should it bend, it would crack and also crack the copper underneath it and the gold above it. We recommend that any area receiving ENIG should never be bent, as it may result in the formation of cracks and intermittent opens.
Question: How is the minimum bend radius determined? Is there a large variation between construction types or stack-ups?
Answer: Minimum bend radius is determined by taking the thickness of the flex stack-up and multiplying by numbers provided in IPC-2223 to achieve a minimum bend radius as shown in the presentation. These numbers can vary based on static or dynamic bends and layer counts. There can be variations, such as airgap constructions where the minimum bend radius is pertinent to the sections rather than the whole flex summation, or if a plane is crosshatched it would very slightly reduce the minimum bend radius. Otherwise, the guidelines pertain to a large amount of flex applications.
Question: In a 4-layer board, does one always have to put the flexible part to be attached to the rigid part on layers 2 and 3 or could I use other layers? Is there an advantage or disadvantage to using layers 2 and 3 as opposed to layers 1 and 2?
Answer: Our preferred location is 2 & 3 for rigid-flex designs. This will facilitate manufacturing, allow for a balanced construction that eliminates warp and twist, results in a more flexible flex section, and eliminate added material and processing costs that it may incur if it was on layers 1 and 2.
Question: In a 4-layer board can I have flex circuits come out in the same place and place one flex on layers 1 and 2 and the other flex on layers 3 and 4? Or is this not done?
Answer: This appears to be referring to an ‘airgap’ construction, which is unfortunately only available in rigid-flex constructions and not in flex constructions.
Question: Are there significant delamination issues “microvia-ing” from rigid to flex layers? Mitigation/possible?
Answer: Unfortunately, we do not support blind micro-vias which terminate at a flex layer due to significant manufacturing issues/limitations and significant reliability issues.
