Designing Flex and Rigid-Flex PCBs for Makers: KiCad Workflow, Bend Radius, and Fab House Requirements

Every PCB design guide on this site — KiCad schematic-to-Gerber, multi-layer stack-ups, JLCPCB ordering — assumes a rigid FR4 board. Flexible and rigid-flex PCBs are a genuinely different fabrication process, not just a material swap, and they solve a problem rigid boards can't: a circuit that needs to bend, fold into a tight enclosure, or survive repeated flexing in a wearable or a moving mechanism. This guide covers what actually changes when you design for flex fabrication in KiCad, how bend radius constraints drive your layout, and what to check before sending a flex design to a fab house.

Flex vs Rigid-Flex: What You're Actually Ordering

A pure flex PCB is a circuit built entirely on a flexible substrate (almost always polyimide, the same family of material as Kapton tape) with copper traces and, often, a coverlay film instead of solder mask. It bends throughout its length and is typically used as a cable replacement — connecting two rigid boards, or wrapping around a mechanism — rather than hosting dense component placement itself. A rigid-flex PCB combines rigid FR4 sections (where your components live) with flexible polyimide sections connecting them, laminated together as one continuous board with no connector or cable in between. Rigid-flex is what you want when you need populated circuit boards in two or three locations of an enclosure with a permanent, vibration-proof connection between them instead of a ribbon cable or wire harness that can work loose.

Designing in KiCad

KiCad supports flex and rigid-flex design through its standard PCB editor with a few workflow differences from a rigid board:

• Stack-up definition. In the board setup dialog, define separate layer stacks for the rigid and flex regions if you're doing rigid-flex — this tells your fab house which layers are polyimide-based flex core and which are FR4, and where the transition happens. Most fabs want this expressed both in your Gerber/ODB++ export and in a written stack-up drawing, since the visual layer boundary isn't always unambiguous from the copper layers alone.
• Board outline layers for flex zones. Draw a separate outline on a dedicated mechanical or user layer marking exactly where the flex region starts and stops — this is what tells the fab where to apply coverlay instead of solder mask and where the stiffener (if any) goes.
• Bend line annotation. Mark intended bend lines explicitly, usually on a fab or comment layer, especially if the flex section needs to fold to a specific angle in the final assembly rather than just being generally flexible. Fab houses use this to advise on trace routing angle relative to the bend.
• Coverlay vs solder mask. Flex regions typically use a polyimide coverlay film (adhesive-laminated) instead of liquid photoimageable solder mask, because standard solder mask cracks under repeated flexing. Specify this in your fab notes; most PCB CAD tools don't have a distinct "coverlay" layer type, so this is usually communicated via the same solder mask layer with a note in your fab documentation.

Bend Radius: The Constraint That Drives Everything Else

The single most important design rule for flex is minimum bend radius, and it's driven by the flex stack thickness: the general guideline is a minimum bend radius of roughly 10 times the total flex circuit thickness for a static (bend-once, stay-put) application, and closer to 100 times the thickness for a dynamic application that flexes repeatedly over its service life — a phone hinge or a robot joint cable, for example. Violating this doesn't just stress the material cosmetically; it fatigue-cracks copper traces and delaminates layers, often after enough flex cycles that the failure shows up in the field long after it looked fine on the bench.

Design ElementRigid Board PracticeFlex Region Practice Trace routing direction at a bendNot applicableRoute traces perpendicular to the bend line, not parallel or diagonal, to minimize strain on each trace Copper fill / ground planesPour freely for EMI and thermal performanceAvoid solid copper pours in flex zones — use a hatched or cross-hatched fill, or avoid it entirely, since solid copper resists bending and cracks Via placementAnywhere convenientNever place vias inside a flex or bend region — only in rigid sections or fully static flex areas Component placementAnywhere on the boardNever place components in a region that bends; components belong only in rigid sections or unbent static flex areas Trace width/spacingPer current and fab minimumsSlightly wider traces and rounded (not sharp right-angle) corners reduce stress concentration at bend points

Stiffeners and Connector Regions

Anywhere a connector, a through-hole component, or a stress-sensitive solder joint needs to sit on a flex circuit, add a stiffener — a laminated backing of FR4, polyimide, or stainless steel bonded to the underside of that specific area. This gives the flex enough local rigidity to survive plugging and unplugging a connector or reflow soldering without flexing at exactly the point where a joint would crack. Specify stiffener material and location clearly in your fab drawing; it's a common miscommunication point between designers used to rigid boards and their first flex order.

Ordering and Cost Expectations

Flex and rigid-flex fabrication is meaningfully more expensive per board and has longer lead times than an equivalent rigid design — expect several times the cost of a comparable rigid board in low quantities, and expect fab houses to require more design review back-and-forth before committing to a run, since flex tolerances and lamination are less forgiving of ambiguous documentation than standard FR4. Not every fab that handles rigid PCBs (including budget prototype houses many makers default to) supports flex or rigid-flex at all, so confirm capability and get a design-rule-check file from the specific fab before finalizing your layout — flex-capable fabs typically publish their own minimum bend radius, trace width, and stiffener rules that are stricter than generic guidelines and should override them.

Flex and rigid-flex open up mechanical possibilities a rigid board plus a ribbon cable can't match — a connection that survives constant motion, or a board that folds into an enclosure shape no flat PCB could fit. The tradeoff is a fabrication process with real material physics constraints (bend radius, trace orientation, no components or vias in the flexing area) that have to be designed around from the first layout pass, not fixed after the fact — get the stack-up and bend geometry right before you route a single trace, because reworking a flex layout after the fact is far more disruptive than on a rigid board.