Rigid-flex DFM comes down to one boundary: the line where the rigid section ends and the flex begins.
Almost every failure in a rigid-flex build starts within a few millimetres of that boundary. Cracked traces, barrel fractures in vias, solder joints that open after the ribbon is folded into the enclosure. The layout decisions around that transition determine most of the outcome, and the assembly process decides the rest.
The rules below cover both halves: what the design has to get right, and what happens to a rigid-flex panel once it reaches an SMT line.
Bend Radius Rules for Static and Dynamic Designs
Bend radius is set by total flex thickness and by how many times the ribbon will move.
Static bending means the flex is folded once during installation and stays there. Common minimum ratios, expressed as radius against total flex thickness:
- Single-layer flex: 6× thickness or greater
- Double-sided flex: 12× or greater
- Multilayer flex: 20× or greater
Dynamic bending means repeated motion in service, as in a hinge or a moving sensor head. The working figure rises to 100× or greater, and rolled annealed copper becomes mandatory rather than preferred, because electrodeposited copper fatigues far sooner under repeated flexing.
Add margin rather than designing at the limit. IPC-2223 is the reference standard for these calculations, and a 20% to 30% cushion absorbs the thickness tolerance the fabricator actually delivers.
Venture Electronics builds rigid-flex with 1 to 6 flex layers against 4 to 30 rigid layers, so the layer count on the flex side is usually the first thing to settle when a bend radius looks tight.
What Must Stay Out of the Bend Zone
The bend zone is a keepout, and the list of what it excludes is longer than most layouts assume:
- Vias and plated through holes. Barrel cracking under bending stress is the classic rigid-flex failure and it is not detectable by visual inspection.
- Component pads and solder joints. A joint inside a flexing region fatigues regardless of how well it was formed.
- Sudden trace width changes. Any discontinuity concentrates strain at that point.
- Coverlay-to-soldermask transitions. The material change is itself a stress riser, so it belongs in a non-bend area.
- Stiffener edges. A stiffener that ends inside or immediately beside a bend zone creates exactly the concentration point the bend was meant to avoid.
Mark the bend line, direction, angle, radius and installed shape on the fabrication drawing. A bend zone that exists only in the mechanical model does not exist to the people building the board.
Stack-Up Symmetry and Routing Across the Bend
The flex stack-up should be symmetrical around its neutral axis. Copper weights and dielectric thicknesses matched above and below keep bending strain balanced instead of concentrating it on one side.
Use low-flow or no-flow prepreg at the transition. Resin squeezing into the flex area stiffens the region that was supposed to move.
Routing rules inside the flex section:
- Run conductors perpendicular to the bend axis wherever the layout allows
- Use curved or gently angled corners, never sharp turns, inside the flex region
- Stagger traces on adjacent layers rather than stacking them directly on top of one another
- Use crosshatched copper for planes in the flex area, and confirm the pattern with the fabricator when impedance matters
Anchor features earn their place at the transition. Teardrop pads, anchoring spurs and a bead of flexible epoxy strain relief at the rigid-to-flex boundary all reduce the chance of a lifted pad during handling. Details on construction options sit on therigid-flex PCB design page.
Stiffener Selection, Placement, and Overlap
Stiffeners exist to give the flex a rigid area where it needs one: under connectors, under mounting hardware, and under any dense component cluster.
Material follows the job. FR-4 for component mounting and general support, polyimide for thin ZIF connector areas, stainless steel where mechanical loading is high or the area is very thin.
Placement rules that matter at assembly:
- Stiffener edges overlap the coverlay by 0.75mm (30 mil) or more, so the transition is supported rather than abrupt
- Corners are rounded, not square, to spread stress across the edge
- Stiffeners never overlap a via array unless that combination has been tested
- For ZIF connectors, the stiffener builds total local thickness to 0.30mm ±0.05mm, which is what the connector's insertion force is specified against
Getting the ZIF stack-up wrong is a quiet failure. The board assembles normally, and the connector either will not seat or loses retention in the field.
Component Placement Near the Rigid-to-Flex Transition
Components belong on rigid sections or on stiffened areas, never on unsupported flex.
Keep a clearance band back from the transition boundary itself. The exact figure depends on construction and is confirmed by the fabricator, but a keepout starting in the 1mm to 3mm range is a workable starting assumption for layout.
Two further constraints come from the assembly side:
- Prefer low-profile, low-mass packages near the transition. A tall or heavy part sitting close to a flexing region loads its own joints every time the ribbon moves.
- Height is governed by the carrier, not by the line. No fixed ceiling applies to component height, but a rigid-flex pallet is machined around the panel, so a tall part sitting over a support rib needs its relief cut into the pallet. Those clearances are confirmed while the carrier is being designed, not after the parts are placed.
Why Rigid-Flex Panels Need Carriers Through SMT
Here is where rigid-flex assembly separates from everything above. A rigid-flex panel cannot lie flat on a conveyor, and nothing in the layout changes that.
The flex sections have no self-supporting stiffness. Under squeegee pressure they dip; in the oven they sag; at placement the vision system reads an inconsistent height. Every one of those produces a defect that looks like a printing or placement problem and is not.
The answer is a machined carrier that holds the panel rigid through the whole process. Two details decide whether it works:
- Double-sided flex needs top and bottom configurations, because the same pallet cannot support both passes.
- Locating features follow the array, so the carrier and the panel outline are designed together rather than in sequence.
Venture Electronics machines these carriers in house from composite stone and ESD-safe polymers, and its engineering team reviews theflex and rigid-flex assembly array layout and perimeter connections during the DFM stage rather than after the panel is fixed.
Carriers are tooling and tooling has lead time. Raised during design review, the carrier is built in parallel. Raised after a failed first build, it adds weeks.
Assembly Steps That Behave Differently on Rigid-Flex
Four steps on the line work differently once a flex section is involved. Each one has a rigid-board equivalent that does not transfer.
Moisture Control Before Reflow
Polyimide and the adhesives in the flex stack absorb moisture readily. Trapped moisture flashes to steam at reflow temperature and delaminates coverlay from copper.
- Boards are stored sealed with desiccant from the moment they arrive
- Baking follows J-STD-033 once exposure limits are passed
- The damage is not recoverable, so this step is not one to compress when a schedule slips
This is the most frequently skipped step on rigid-flex builds.
Reflow Profiling Across Two Thermal Masses
A rigid-flex panel carries two very different masses in the same oven pass. The rigid section heats slowly. The flex ribbon heats fast.
A profile tuned to the rigid area overshoots the flex, and a profile tuned to the flex leaves the rigid section short of full wetting.
- Reflow ovens holding ±1°C accuracy give room to run a gentler ramp that satisfies both
- Profiling is done on a populated panel, not a bare one, since component mass changes the result
Depaneling a Rigid-Flex Array
V-scoring across a flex section is not an option. Separation has to be planned around the rigid areas.
- Break tabs run through the rigid sections, using routed tabs rather than score lines
- Flex regions are routed or die cut, never snapped
- The assembly line takes boards from 50mm × 50mm to 774mm × 710mm, so small rigid-flex parts have to be arrayed regardless of the separation method
Inspection Limits on Flex Sections
Automated optical inspection assumes a consistent board plane. Any residual distortion in a flex area shifts height readings and produces calls that are not real defects.
Hidden joints under BGA and QFN packages on the rigid sections cannot be judged optically at all, so X-ray remains a standard step on boards carrying those packages.
Review the Bend Zone and the Carrier in the Same Pass
Rigid-flex is one of the few build types where the layout and the tooling have to be agreed together. The bend zone, the stiffener stack-up, the array outline and the carrier all reference the same geometry, and settling them in separate conversations is what creates rework.
Send the stack-up, mechanical drawing and array with yourPCB assembly enquiry so the bend zone keepouts and the carrier design are confirmed against the same line that will build the boards.
FAQs About Rigid-Flex PCB Assembly DFM
Q1: What is the minimum bend radius for a rigid-flex PCB?
For static bends, roughly 6× total flex thickness for single-layer, 12× for double-sided and 20× for multilayer. Dynamic applications rise to 100× or more and require rolled annealed copper rather than electrodeposited.
Q2: Can vias be placed in the bend area?
No. Plated barrels crack under bending stress and the failure is invisible to optical inspection, so vias, plated holes and pads are excluded from the bend zone and moved into rigid or stiffened areas.
Q3: How far from the rigid-to-flex transition should components sit?
A keepout starting in the 1mm to 3mm range is a practical layout assumption, with the working figure confirmed against the actual construction duringPCB fabrication review. Components belong on rigid or stiffened areas, never on unsupported flex.
Q4: Which stiffener material should be used?
FR-4 for component and connector support, polyimide for thin ZIF areas, stainless steel where mechanical load is high. Edges overlap the coverlay by 0.75mm or more and corners are rounded rather than square.
Q5: Why does a rigid-flex panel need a carrier during SMT?
Flex sections have no stiffness of their own, so the panel dips under the squeegee and sags in the oven. A machined carrier holds it flat, and double-sided flex requires separate top and bottom configurations.
Q6: Does a rigid-flex assembly need baking before reflow?
Yes, whenever moisture barrier exposure limits have been passed. Polyimide and flex adhesives absorb moisture readily, and trapped moisture delaminates coverlay from copper at reflow temperature.
Q7: How are rigid-flex panels depaneled?
Through the rigid sections using routed tabs, with flex regions routed or die cut rather than scored. The array outline is best agreed duringDFM and DFA review, since it also defines the carrier geometry.
