RF PCB Manufacturing Capabilities

Rigid Flex RF PCB | Rogers Polyimide Hybrid Build

Rigid flex RF PCB technology merges high-frequency Rogers or PTFE laminates in rigid sections with flexible polyimide interconnects, enabling compact 3D RF assemblies for radar, avionics, and wearable medical devices while maintaining controlled impedance across the transition.

Cross-section of a rigid flex RF PCB showing Rogers and polyimide layers

What Is a Rigid Flex RF PCB?

A rigid flex RF PCB is a single unified board that combines rigid high-frequency laminate zones—typically Rogers RO4350B (Dk 3.48 ± 0.05) or RO3003 (Dk 3.00 ± 0.04)—with flexible polyimide interconnect sections in one lamination cycle. The rigid areas carry microstrip, stripline, or coplanar-waveguide transmission lines, while the polyimide flex zones fold to connect boards in three dimensions, eliminating RF connectors and cable assemblies that would otherwise degrade signal integrity above 6 GHz. This hybrid construction is standard in phased-array radar modules, surgical navigation devices, and satellite transponder assemblies where volume, weight, and insertion loss budgets are all tight.

Why Combine Rogers and Polyimide in One Board?

Folded rigid flex RF PCB with flex bridge connecting two rigid sections

Rigid-only RF boards need connectors or flex cables between sub-assemblies, each adding 0.1–0.3 dB of insertion loss and a potential impedance discontinuity. A rigid flex RF PCB removes those interfaces. The Rogers rigid sections provide the low Df (typically 0.0037 at 10 GHz for RO4350B) needed for filter, coupler, and antenna feed networks, while the polyimide flex sections (Dk ≈ 3.4, Df ≈ 0.008 at 1 GHz for standard Kapton-type films) handle digital, power, and lower-frequency signal routing where loss is less critical.

The combined structure also cuts assembly labor. A folded rigid flex RF PCB replaces what might otherwise be three separate boards, two flex cables, and four board-to-board connectors—reducing component count, solder joints, and field-failure modes in high-reliability programs governed by IPC-6013 Class 3 or equivalent.

Material Selection and Compatibility

Rigid Zone Materials

The rigid sections must balance RF performance with processability alongside polyimide. Thermoset hydrocarbon-ceramic laminates (Rogers RO4000 series, Isola Astra MT77) are preferred because their lamination temperatures (roughly 190–220 °C peak) are compatible with polyimide adhesive systems. Pure PTFE laminates such as RT/duroid 5880 (Dk 2.20) offer lower loss but require higher lamination pressures and sodium etch pre-treatment or plasma activation for reliable copper adhesion in the hybrid stackup.

Flex Zone Materials

Standard polyimide films for the flex sections are 1 mil (25 µm) or 2 mil (50 µm) thick, with or without adhesive. Adhesiveless polyimide (e.g., DuPont AP series) gives better dimensional stability and a thinner bend radius—critical when the flex must fold 180° inside a radar LRU chassis. For flexible RF interconnects carrying signals above 20 GHz, liquid crystal polymer (LCP) flex offers Df below 0.004 at 40 GHz, though at roughly 2–3× the material cost of polyimide.

Bonding and No-Flow Prepreg

The rigid-to-flex transition uses no-flow or low-flow prepreg (such as Isola 185HR or Panasonic R-1551) to prevent resin squeeze-out into the flex bend zone. The prepreg Dk should be characterized and accounted for in the impedance model; a mismatch of even 0.3 in Dk at the boundary can create a localized impedance bump visible on TDR above 10 GHz.

Material Zone Dk (10 GHz) Df (10 GHz) Tg / Max Temp
Rogers RO4350B Rigid RF 3.48 0.0037 280 °C
Rogers RO3003 Rigid RF 3.00 0.0013 500 °C (PTFE)
Isola Astra MT77 Rigid RF 3.00 0.0017 210 °C Tg
Polyimide (Kapton HN) Flex 3.40 0.0080 400 °C
LCP flex film Flex (mmWave) 2.90 0.0025 280 °C
No-flow prepreg (typical) Bond layer 3.8–4.2 0.010–0.018 170–180 °C Tg

Stackup Architecture for Rigid Flex RF PCBs

A representative 8-layer rigid flex RF PCB stackup might look like this, from top to bottom in the rigid section:

  1. L1 – RF signal (microstrip): 0.5 oz ED copper on RO4350B, 10 mil core.
  2. L2 – Ground plane: continuous copper pour, stitched with vias at λ/20 spacing.
  3. L3 – Signal/power: FR-4 or mid-Dk prepreg, 4 mil.
  4. L4 – Ground plane: separates digital from RF return paths.
  5. Flex core: 2 mil polyimide with 0.5 oz RA copper on both sides (L5/L6 in flex zone).
  6. L7 – Ground plane: mirrors L4 in rigid zone.
  7. L8 – Digital/control signal: FR-4 core, 6 mil.

The flex layers run continuously from one rigid island to the next. In the rigid sections, the flex core is laminated between rigid cores with no-flow prepreg; in the bend zone, the rigid material is routed away, leaving only the flex layers and their coverlay. This architecture allows blind and buried vias in the rigid zones while keeping the flex zone free of plated through-holes that would crack during bending.

Impedance Control Across the Transition

The biggest design challenge is maintaining 50 Ω (or 100 Ω differential) as a trace moves from the rigid Rogers section into the flex polyimide section. The dielectric constant changes—from 3.48 in RO4350B to roughly 3.4 in polyimide—which is a small shift, but the dielectric thickness also changes dramatically as rigid prepreg layers disappear. A microstrip trace that is 18.5 mil wide over 10 mil RO4350B may need to widen to 21 mil over a thinner flex stackup to hold 50 Ω. Model each zone independently in a 2D field solver (e.g., Altium’s impedance calculator or Polar Si9000) and use tapered transitions rather than abrupt width changes.

For designs involving HDI microvia interconnects in the rigid zones, ensure the via stub length is minimized through back-drilling or by using sequential lamination with buried microvias—stubs as short as 4 mil can cause noticeable resonance dips above 20 GHz.

Fabrication Process Considerations

Lamination Sequence

Rigid flex RF PCBs typically require two or three lamination cycles. The flex core is fabricated first (image, etch, AOI), then laminated to the rigid cores. Each additional lamination cycle adds registration tolerance—expect ±3 mil layer-to-layer in a well-controlled process, versus ±2 mil for a standard rigid board. This matters for tightly coupled stripline where the trace-to-ground spacing is only 4–5 mil.

Controlled Depth Routing

The rigid material must be routed away to expose the flex bend area. Depth-controlled routing accuracy of ±2 mil is standard; going tighter (±1 mil) may require laser routing at additional cost. Any nick into the flex copper or coverlay in this step creates a stress riser that will crack during bending.

Etch Compensation

Rogers laminates and polyimide flex etch at different rates. The rigid section may use 1 oz copper while the flex uses 0.5 oz rolled-annealed copper—each requiring different etch compensation factors to hit target trace widths. Specifying per-zone etch compensation in the fabrication notes is essential; a blanket compensation value will leave one zone over-etched and the other under-etched, shifting impedance by 3–5 Ω.

Buy vs. Build: Assembly Decision for Rigid Flex RF

Close-up of rigid-to-flex transition zone on an RF circuit board

Rigid flex RF PCBs blur the line between bare-board fabrication and system assembly. The board itself replaces cables and connectors, so the “build” option (fabricating rigid flex in-house or at a specialty shop) must be weighed against the “buy” option (using separate rigid RF boards connected by commercial flex cables or board-to-board connectors). Enter your volumes, connector count, and target frequency into the tool below to see where the crossover lies for your program.

Loading Buy vs Build Assembly Tool…

A typical result: at volumes above 200 units, a 6-layer rigid flex RF PCB replacing three connectors and two flex cables saves $8–$15 per assembly in connector and labor cost, even though the bare-board price is 40–60 % higher than equivalent separate rigid boards. Below 50 units, the NRE for rigid flex tooling (step-and-repeat panel design, depth-routing programs, multiple lamination fixtures) often makes the buy-separate approach cheaper. Factor in reliability: every eliminated connector removes a potential field-failure point, which in medical-grade RF assemblies or avionics programs can justify the premium even at low volumes.

Design Rules and Best Practices

  • Bend radius: minimum 6× the flex thickness for single-sided flex, 12× for double-sided. A 10 mil total flex thickness needs at least a 120 mil bend radius to survive 25+ flex cycles per IPC-2223.
  • No vias in the bend zone: plated holes crack. Route all inter-layer transitions into the rigid sections.
  • Stagger flex traces: rather than stacking traces directly over each other on both flex layers, stagger them to distribute bend stress.
  • Ground stitching at transitions: place ground vias at ≤ λ/20 spacing around the rigid-to-flex boundary to prevent slot-mode radiation from the ground plane discontinuity.
  • Coverlay openings: keep coverlay at least 10 mil back from SMD pads in the rigid zone; coverlay thickness affects local impedance.
  • Panelization: rigid flex panels are less efficient than standard rigid panels because of the routed flex tails. Discuss panel utilization strategies with your fabricator early—rotating or nesting flex tails can recover 10–15 % material yield.

Testing and Qualification

Beyond standard RF tests (network-analyzer S-parameter sweeps, TDR impedance profiling), rigid flex RF PCBs require mechanical qualification:

  • Flex-cycle testing: IPC-6013 Class 3 requires the flex zone to survive a minimum number of bend cycles at the specified radius without open circuits or impedance shift > 2 Ω.
  • Thermal cycling: –55 °C to +125 °C, 100 cycles minimum for mil-aero, verifying no delamination at the Rogers-to-prepreg or prepreg-to-polyimide interfaces.
  • Cross-section analysis: microsection at the rigid-to-flex transition to confirm no resin voids, copper cracking, or adhesive squeeze-out.

Frequently Asked Questions

Can FR-4 be used for the rigid sections of a rigid flex RF PCB?

FR-4 can serve as the rigid dielectric for digital and power layers, but it should not be used as the RF signal dielectric above about 3 GHz. Its Df of 0.020 at 10 GHz causes roughly 5× the insertion loss of RO4350B over a 50 mm trace. A common hybrid approach uses Rogers for the top RF layer and FR-4 for inner digital layers to balance cost and performance.

What frequency range suits rigid flex RF PCBs?

Rigid flex RF PCBs are routinely built for applications from UHF (300 MHz) through Ka-band (40 GHz). Above 40 GHz, the flex section loss in standard polyimide becomes significant, and designers typically switch to LCP flex or move the mmWave circuitry entirely onto the rigid section with wirebond or flip-chip transitions.

How many flex cycles can a rigid flex RF PCB survive?

For install-to-flex (bent once during assembly and never again), standard constructions are sufficient. For dynamic flex (repeated bending in service), single-sided flex zones with rolled-annealed copper and a bend radius ≥ 12× thickness can exceed 100,000 cycles. Double-sided flex zones are generally limited to install-to-flex or low-cycle applications.

What is the typical lead time for a rigid flex RF PCB prototype?

Expect 15–25 working days for a first-article rigid flex RF PCB with Rogers and polyimide, depending on layer count and lamination complexity. This is roughly 2× the lead time of a standard rigid RF board due to the additional lamination cycles, depth routing, and flex-zone AOI steps.

Ready for a quotation?

Send your requirements for a material, stackup and manufacturing review. Price and schedule are confirmed for your project.

Upload Gerber