Rigid-flex architecture for a wearable sensor

A wearable sensor architecture should begin with the locations of the sensing surface, battery, radio, and mechanical supports. Rigid-flex becomes useful when these functions need different orientations or positions but must remain part of one connected assembly. The first decision is where the product needs mechanical freedom, not how tightly the board can be folded.

Texas Instruments’ wearable skin-temperature reference design discusses mechanical mounting, thermal paths, and the signal chain. These are useful starting points for a board architecture review. A mechanically convenient sensor location still needs measurement validation in the actual product rather than an assumption that the component specification describes the entire assembly.

Map the functions before choosing the board outline

Consider a fictional wrist-mounted temperature logger. A sensing island must face the intended contact surface, while a battery and processing island occupy a thicker cavity. The radio needs an RF review in the final enclosure. Draw these locations in the mechanical model before placing the connecting flexible span.

Separate sensing requirements from packaging assumptions

In the example, moving the sensing island away from the main electronics might solve a packaging conflict. It does not, by itself, prove a required response time or measurement accuracy. Write a test question for the proposed location: which assembly configuration, contact condition, operating state, and reference measurement will be used?

Keep the electrical and mechanical revisions attached to that result. If a later enclosure change alters the sensor support, identify which earlier conclusions need review. This is more useful than recording a single “sensor tested” check without its physical configuration.

Plan the assembly route through the enclosure

Start with the flat board and walk through installation. Identify which island enters first, where tools must reach, and whether a flexible span passes a sharp edge or a fastener. Then reverse the sequence to examine service access. The installed shape alone can conceal a difficult or uncontrolled installation step.

For the logger, a useful prototype exercise is to record three states: flat handling, partial insertion, and fully retained assembly. Review the board outline and clearances at each state. Use a mechanically representative prototype and document its limitations; a soft paper mock-up cannot establish the reliability of the actual copper and laminate construction.

Define what the sensing island measures

For the fictional logger, distinguish the temperature at the contact surface from the temperature of the sensor package and the surrounding air. The team should write which of these quantities the product intends to report. Moving the package closer to the contact surface is a geometric change; deciding whether the resulting signal represents the desired quantity is a measurement question. Keep these decisions connected in the design review without treating them as interchangeable.

Create a cross-section through the enclosure, contact interface, sensor package, board, and any backing support. Annotate every intentional contact and every possible air gap. Record the nominal dimensions and the tolerances that could open or compress the interface. This drawing gives the mechanical and electronics teams a common object to review. A top-view PCB screenshot cannot reveal whether the sensor is actually touching the intended thermal path after assembly.

Plan a comparison between two clearly documented mounting concepts before committing to the outline. Hold the sensor, firmware, sampling method, and reference arrangement constant as far as practical. Change only the mounting feature under investigation, then record both the transient response and the settled reading under the chosen conditions. This is a proposed engineering experiment, not a claim that either concept will achieve a particular accuracy. If several variables must change together, document that limitation when interpreting the result.

Build a power-state worksheet before dividing the islands

The largest component does not necessarily define the most important operating state. A short radio burst, a long sleep interval, and a measurement conversion have different current profiles. List these states against the components that participate in them. Then identify which supply and return paths cross the flexible region. This helps determine whether a proposed island split creates an awkward shared path between a sensitive measurement circuit and a switched load.

Consider an illustrative one-second measurement cycle with three non-overlapping states: a 20 ms transmit interval at 12 mA, a 10 ms acquisition interval at 2 mA, and a 970 ms sleep interval at 0.02 mA. The average current for this simplified cycle is (12 × 0.020 + 2 × 0.010 + 0.02 × 0.970) / 1 = 0.2794 mA. These are invented worksheet inputs, not specifications for a selected component or a battery-life promise.

The average does not replace a peak-current review. The supply path still has to support the instantaneous transmit state, and the real product may add indicators, memory writes, regulators, leakage, retries, or overlapping operations. Measure the assembled system in representative firmware modes and update the worksheet. For architecture selection, the important output is a traceable list of loads, timing assumptions, and paths rather than one attractive lifetime number.

Include charging as a separate physical configuration if the product supports it. Describe where the device sits, which surfaces touch the fixture, and whether measurements are expected during charging. Do not assume the normal worn configuration and the charging configuration share the same thermal surroundings. An operating restriction, such as not reporting a measurement during a particular mode, belongs in the product requirements and firmware review rather than an undocumented layout assumption.

Separate retention loads from the flexible connection

Draw the load path for each substantial part. In the logger concept, a housing feature might retain the battery while a separate feature locates the sensor island. The flexible span then connects the supported regions. Identify what happens if the enclosure is opened, the strap is pulled, or the assembly is dropped. The question is where the force travels, including temporary installation states, not simply whether all components fit inside the closed housing.

Classify each span by its actual motion. A folded board locked inside a rigid enclosure has a different requirement from a span that follows a strap every time the wearer moves. Count service operations separately from normal use and describe the motion envelope. Avoid assigning a universal minimum bend radius from a generic illustration; submit the intended construction, flex layers, motion, and available space for a project-specific manufacturability review.

Also mark surfaces that could rub against the span. Nominal clearance in a CAD model does not establish clearance at tolerance extremes or under movement. A useful review overlays the smallest enclosure opening, the largest relevant board outline, and the expected assembly variation. If the space is insufficient, resolve the geometry or retention concept before treating a protective material as an automatic cure. Any added layer can also change thickness and the way the span moves.

Keep test access available through the build sequence

An integrated board can be electrically accessible while flat and difficult to reach after folding. Decide which checks happen before insertion, after retention, and after the housing closes. For example, programming pads may be available only in the flat state, whereas the sensing interface requires the final enclosure. Write those dependencies into the assembly sequence so a passed early test is not mistaken for evidence about the finished configuration.

Do not prescribe acceptance numbers before the product requirements exist. Instead, assign an owner to each open criterion and identify the evidence needed to close it. The RF review, for example, needs the final nearby materials and operating configuration; a connection observed on an open bench is only a result for that bench arrangement. This avoids turning an exploratory prototype demonstration into an unsupported performance statement.

Release a coordinated architecture package

The final architecture package should connect the region drawing, enclosure cross-section, motion description, power-state worksheet, and proposed verification sequence. Use consistent revision identifiers so a supplier can tell which mechanical model belongs with the electrical data. Highlight unresolved assumptions explicitly, including contact geometry, available bend space, and test access. This gives the next review a concrete set of decisions rather than requiring another engineer to reconstruct the design intent from disconnected screenshots.

Compare the complete integration tradeoff

Compare the rigid-flex concept with separate boards and an interconnect. Include assembly steps, available volume, test access, replacement strategy, and the consequences of a defect in one island. The integrated option should answer an identified project problem. It is not automatically the best architecture for every wearable.

See our sensor and medical-electronics application overview and rigid-flex design guide. Submit your enclosure concept and region drawing to discuss the board architecture. Product performance and any medical-device qualification require their own project-specific evidence.