Most electronic products start as a flat board: components on a stiff laminate, traces in straight lines, and every board-to-board connection made through a connector or a separate cable. That works until the product has to fold, twist, survive vibration, or fit a curved housing, and the connector becomes the weakest link. Rigid-flex PCB is the alternative. Instead of a rigid board with a flex cable bolted to it, the two become one laminated structure. When a rigid-flex PCB manufacturer quotes that structure, the conversation always returns to the same points: how many layers, where the bend zones sit, and whether the flex section is bent once during assembly or flexed repeatedly in service.
A rigid-flex PCB and a traditional rigid PCB differ in three things that matter. Structure: a rigid-flex board combines FR-4 rigid sections with polyimide flexible sections in a single laminate, while a traditional PCB is FR-4 only and stays flat. Interconnect: on a rigid-flex board the flexible section replaces the board-to-board connectors and wire harnesses a rigid design depends on.
Design rules: the bend zone imposes bend radius, copper-type, and component-placement limits that a flat board never has to satisfy. Rigid-flex costs more per board and takes longer to fabricate, and it earns that cost when the product must fold, curve, move, or run in a vibration-heavy environment. A traditional PCB is cheaper, faster, and simpler, and it remains right for flat enclosures with few connectors.
The sections below cover materials, stackup, bend rules, cost, and selection criteria.
What separates a rigid-flex PCB from a traditional board?
The difference is structural, not cosmetic. A traditional PCB is a single flat laminate that carries components and passes signals through copper and plated holes, and it stays rigid for the life of the product. A rigid-flex PCB bonds rigid FR-4 sections to thin polyimide sections in one continuous board, so the flexible section carries signal or power across a fold or hinge with no separate connector at that joint. Cost, lead time, design complexity, and reliability all follow from that one fact.
| Parameter | Traditional rigid PCB | Rigid-flex PCB |
|---|---|---|
| Substrate | FR-4 or high-Tg FR-4 throughout | FR-4 rigid zones bonded to polyimide flex zones |
| Form factor | Flat, non-bendable | Rigid mounting zones plus 3D fold, twist, and curve routing |
| Board-to-board link | Connectors, headers, wire harnesses | Built-in flex layer, no external interconnect |
| Component mounting | Anywhere on the board | Rigid zones only |
| Flex-zone copper | Not applicable | Rolled and annealed for dynamic zones, electrodeposited acceptable for static |
| Main mechanical risk | Solder joint and connector fatigue | Stress at the rigid-to-flex transition and along the bend line |
| Typical use | Motherboards, power supplies, control boards | Wearables, camera modules, medical devices, automotive sensing, robotics |
Rigid-flex does not win because it is newer. It wins in a narrow set of mechanical situations, mostly when the product is too small for a connector or when vibration would eventually fatigue a solder joint.
The four main types of PCBs
- Rigid PCB—solid FR-4 laminate, no bend. Flat enclosures, high layer counts.
- Flex PCB—thin polyimide throughout, bendable. Display tails, sensor leads, and cable replacement.
- Rigid-flex PCB—FR-4 zones plus polyimide zones in one stack. Several mounting areas joined across a hinge or curve.
- HDI PCB—microvia and blind or buried via structures. Fine-pitch density where layer count must stay low.
Rigid-flex and HDI are not competing categories. Many rigid-flex builds add microvias in the rigid zones, so one board can be both.
Flex PCB versus rigid-flex PCB: where the line sits
A pure flex PCB has no rigid section, so it cannot carry a connector, a shield, or a heavier component without added stiffener material. A rigid-flex PCB has FR-4 areas that take mounting hardware, connectors, and dense parts, while the polyimide areas route between them.
The distinction is commercial as well as technical. Flex and rigid-flex run through different lamination sequences, tooling, and pricing. A request that reaches a fabricator labeled only as “flex” can come back as a board with no rigid mounting capability at all. State which construction the design needs and where the flex zones sit. For the flat, single-zone case, a rigid PCB manufacturer runs the standard process without flex-specific cost.
Materials and stackup: what actually differs
Material choice is where rigid-flex earns its higher price. A traditional PCB uses one dielectric family, normally FR-4, across the whole board. A rigid-flex board combines two: an FR-4 or high-Tg core in the rigid zones and a polyimide film in the flex zones, joined by an adhesive system that must survive both. The two do not expand or flex at the same rate, and that mismatch is why the process is harder to control.
| Element | Rigid zones | Flex zones |
|---|---|---|
| Base material | FR-4 or high-Tg FR-4 | Polyimide film |
| Conductor | Electrodeposited copper, standard foil weights | Rolled annealed copper in dynamic sections |
| Bonding | Prepreg during lamination | Adhesive film, or adhesiveless cast polyimide |
| Stiffener | Not required | Polyimide, FR-4, or stainless steel where a connector sits |
| Continuous flexing | Not applicable | Yes, subject to bend radius rules |
| Thermal behavior | Standard laminate response | Polyimide tolerates higher processing temperatures than FR-4. |
Rolled annealed copper is the detail that matters most in moving designs. Its grain structure follows the bend axis, and it tolerates repeated flexing far better than electrodeposited copper, which is why fabricators specify RA foil wherever a section will move in service rather than just during assembly. Published capability for a production rigid-flex line includes adhesiveless polyimide base material with RA copper, builds from 2 to 20 layers, minimum core thickness of 0.051 mm, minimum trace and space of 0.089 mm, 0.18 mm smallest mechanical drill, 0.127 mm smallest laser drill, up to 3 oz finished copper, and 0.1 mm flex zone stiffeners in polyimide, FR-4, or stainless steel (published rigid-flex capability data, checked 10 October 2026).
Rigid-flex PCB stackup options
| Stackup | Typical use | Main risk to check |
|---|---|---|
| Single flex layer between rigid zones | Wearables, display tails, simple sensors | Bend radius, coverlay quality, connector support |
| Two flex layers | Medical handhelds, compact sensors, IoT devices | Plane continuity and impedance across the transition |
| Multi-layer flex with rigid zones | Camera modules, RF, dense electronics | Lamination alignment, via reliability, yield |
| Rigid-flex with HDI microvias | Compact industrial and automotive electronics | Microvia reliability, laser drilling quality, panel cost |
Layer balance matters more here than on a flat board. An asymmetric stackup lets the finished part twist, and twisting after reflow is what cracks copper in a flex zone. Symmetrical copper distribution and consistent dielectric thickness keep the board flat through reflow. Where signals run at controlled impedance, the rigid-to-flex transition needs its layer count and impedance profile planned in advance rather than adjusted afterward.
Bend radius and flex-zone design guidelines
Minimum bend radius decides whether a rigid-flex board works or fails, and it is set by two things: the total thickness of the flex zone and whether the section is bent once or repeatedly. The widely used starting point is a multiple of flex-zone thickness—roughly 6x for a single-layer static bend, 10x to 15x for a double-layer static bend, and 20x to 40x or more once the section moves continuously in service. Layer count, copper weight, and material push the required ratio further, so treat any published figure as a starting point, not a specification.
Static flex versus dynamic flex
| Condition | Definition | Design consequence |
|---|---|---|
| Static flex (bend-to-install) | Bent once or a few times during final assembly, then fixed | Electrodeposited copper is usually acceptable; static radius ratio |
| Dynamic flex | Flexed repeatedly in use, such as a hinge or folding product | Rolled annealed copper required; dynamic radius ratio; cycle-life target agreed in writing |
Designing a dynamic application with static-flex margins is a common and expensive mistake. Vendors publish different ratios because the ratio depends on layer count, copper weight, and application, so treat any published figure as a starting point. The method is consistent: measure from the inside surface of the bend, add a 20 to 30 percent margin over the calculated minimum, and confirm the cycle count with the fabricator before release. Material also stiffens in cold conditions, so a radius that passes at room temperature may not pass in a cold enclosure.
Flex-zone layout rules
The bend zone is the highest-stress area on the board and the one most often designed wrong.
- Keep components, vias, test pads, plated holes, and solder joints out of bend areas. Most fabricators ask for roughly 1.5 mm clearance from the rigid-to-flex transition for surface-mount parts.
- Keep vias at least 1.0 to 1.5 mm clear of the transition, and never place a through-hole via in a bend zone.
- Route traces along the bend axis, not across it, and use curves or gentle angles instead of sharp 90-degree corners.
- Offset traces on opposite layers of a double-sided flex zone rather than stacking them directly above each other.
- Avoid solid copper pours in dynamic bend zones; hatched or voided copper flexes better.
- Use no-flow prepreg at the boundary so resin does not creep into the flex area.
- Terminate stiffeners well back from the bend line, and extend coverlay past the outermost flex-zone trace.
Manufacturing and reliability differences
Rigid-flex boards cost more and take longer to build because the process is less forgiving, not because the materials are exotic. A rigid board runs one lamination cycle on a mature, highly automated flow. A rigid-flex board runs multiple lamination and bonding cycles, needs rigid and flex layers to align accurately within the panel, and needs its delicate flex zones protected through drilling, plating, and handling. The reward is a board with fewer interconnect parts, which matters most in vibration.
Reliability is where rigid-flex justifies itself. Every connector removed is a potential failure point removed with it: no mating halves to wear, no solder joints to fatigue under vibration, and no harness to chafe. That is why rigid-flex shows up in medical handheld devices, camera modules, automotive sensing and battery systems, and aerospace electronics, where interconnect fatigue rather than component capability tends to set the field failure rate.
Yield risk concentrates in four places. Material compatibility is the first: FR-4 and polyimide expand at different rates, and a poor match surfaces later as delamination. Lamination is the second, since each bonding cycle needs precise alignment and controlled adhesive flow. Third, drilling and plating, since holes and vias pass through both zone types. Fourth, handling: thin polyimide is easily damaged and needs clean handling and soft tooling.
Standards that define the build
Design and acceptance are governed by two different documents, and specifying only one is a common source of rework. IPC-2223 is the sectional design standard for flexible and rigid-flexible printed boards, covering material selection, mechanical and physical properties including flexible areas, performing bend, stiffeners, and strain-relief fillet guidelines, and electrical properties such as impedance control. IPC-6013 is the qualification and performance specification for flexible and rigid-flex boards, from single-sided flex through multilayer rigid-flex, including stiffeners, plated through-holes, microvias, and blind or buried vias.
The two work together: IPC-2223 is the design language, and IPC-6013 is the acceptance language. Performance classes run from Class 1 for general electronics through Class 2 for dedicated service electronics to Class 3 for high-reliability medical and aerospace work. State the class on the purchase order rather than writing “IPC compliant,” and state which standards the board is designed and qualified to. Class 3 adds stricter flex-cycle validation, thermal stress, and delamination requirements and should be qualified as Class 3 from the first DFM review, not at final inspection.
Rigid flex PCB cost: bare board versus total system cost
A rigid-flex board costs more than a comparable rigid board, and comparing the two on bare board price is the most common way to make the wrong decision. The purchase order covers one board; the assembly covers connectors, harnesses, the labor to crimp and route them, the enclosure volume they occupy, and the field failures they eventually cause. Rigid-flex usually wins on that total and usually loses on the line item. Judge it on the second number.
| Cost driver | Direction on a rigid-flex build |
|---|---|
| Base materials | Higher—polyimide film, RA copper, separate adhesives, stiffeners |
| Lamination | Higher—additional bonding cycles and process time |
| Panel utilization | Lower—irregular outlines and flex zones generate more waste |
| Engineering and DFM | Higher—bend radius, stackup, and transition rules need review. |
| Testing | Higher where dynamic flex is specified—bend testing and microsection |
| Assembly | Lower—fewer connectors, no harness to crimp, fewer joints to inspect |
| Enclosure and shipping | Lower—smaller assembled footprint and less wiring weight |
None of these is a fixed percentage, and the spread between suppliers is wide enough that any published ratio is directional only. What moves the number most is the design itself: layer count, bend radius ratio, dynamic versus static flex, copper weight, finished surface, and full versus partial panel. Two boards with identical schematics can land far apart if one has a 6x static bend and the other a 40x dynamic bend with RA copper.
For a quote that reflects the real part, send the fabricator the Gerber files with layer stackup and flex zone definitions marked, the quantity, surface finish, copper weight, stiffener specification, IPC class, and the bend-cycle target. If those are missing from a quote, the number on it is a guess.
How to choose between the two
Choose a traditional rigid PCB unless something in the enclosure physically requires a bend, a curve, or a moving joint. Rigid-flex is right when the product folds, hinges, wraps a curved surface, bridges two separated mounting areas, or lives in a vibration environment where a connector is a liability. If none of those apply, the extra cost and lead time buy nothing.
Work through these before deciding:
- Does the board bend after assembly, and if so, how often in service?
- Are components or connectors physically separated, so something has to span the gap?
- How much enclosure volume would connectors and cable routing consume?
- What vibration, thermal cycling, and chemical exposure will the assembly see?
- Is the connector count high enough that solder joints dominate the failure estimate?
- Can the production volume absorb a longer lead time and a higher per-board price?
| Application | Usual choice | Reason |
|---|---|---|
| Motherboards, power supplies, industrial control | Rigid | Flat, high layer count, no bend required |
| Bulk consumer electronics, fixed enclosures | Rigid | Cost per unit dominates. |
| Wearables and smart devices with curved housings | Rigid-flex | The board must follow the housing shape. |
| Medical handheld and imaging devices | Rigid-flex | Space, weight, and chemical resistance in one structure |
| Camera modules and optical assemblies | Rigid-flex or flex | Short interconnect across a tight or moving joint |
| Automotive ADAS, battery, and sensor systems | Rigid-flex | Vibration resistance and 3D packaging |
| Robotics and industrial motion systems | Rigid-flex | Repeated flexing with connector elimination |
| Consumer foldables and hinges | Rigid-flex, dynamic flex rules | Cycle life is the governing requirement. |
Before releasing a rigid-flex design, walk the DFM checklist once more: bend zones defined before parts are placed, nothing mounted in the bend area, radius ratio chosen for the actual cycle count, RA copper in every dynamic section, stiffeners clear of the bend line, coverlay extended past the outer traces, and the IPC class written into the purchase order.
FAQ
Can a flex zone be bent more than once if the bend radius is large enough?
No. A large radius reduces strain on each bend but does not change the fact that the copper is being stressed. Repeated cycles need dynamic-flex design rules—rolled annealed copper, a radius from the dynamic ratio, and a cycle-life target confirmed against your actual cycle count.
Will rigid-flex replace traditional rigid PCBs?
Not broadly. It costs more per board, takes longer to fabricate, and adds design rules that only pay off where the product genuinely needs to bend, curve, or move. Most products using rigid-flex today are hybrid in practice: a rigid-flex board inside one module, with traditional rigid boards doing the rest.
What should a purchase order specify for a rigid-flex board?
At minimum: the IPC class, the standards the board is designed and qualified to, whether the flex is static or dynamic, the bend-cycle target, and the flex zone boundaries in the fabrication data. Naming only “IPC compliant” leaves acceptance criteria undefined and is a frequent source of disagreement at inspection.
