Connected Device PCB: Design for Wireless Products

What Makes a Connected Board Different

A connected device is one that talks to something else, whether over Wi-Fi, Bluetooth, Zigbee, Thread, LoRa or a cellular network. The circuit board inside it is not fundamentally different from any other board, but three requirements shape it from the first layout decision: it must carry a radio without interfering with it, it must run from a battery for months or years, and it must fit inside a housing designed around the user rather than around the electronics.

Those three requirements interact. Making the board smaller brings the radio closer to the switching regulators that pollute it. Extending battery life means turning circuits off, which creates transient currents that disturb the supply. Fitting a curved housing usually means a flexible or rigid flex board, which changes the ground structure under the antenna.

The Functional Blocks

• Microcontroller or system on chip. The processor, the radio and often the memory in one package, which is the usual choice for a compact product because it removes the interface between the processor and the transceiver.
• Radio front end. The module, the matching network, the filter and the antenna, plus the keep-out area the antenna requires.
• Sensors and actuators. Temperature, humidity, motion, light, current or biometric sensors, and the drivers for motors, relays or displays.
• Power path. Battery, charger, protection, and one or more regulators, with a strong preference for switching regulators that stay efficient at light load.
• User interface. Buttons, LEDs, a buzzer or a small display, and the connector used for programming and test.

Radio Integration

Use a certified module where you can. A pre-certified module removes the radio design, the matching and, more importantly, the regulatory testing from the project. It costs more per unit than a chip-down design and it removes a long and expensive certification cycle. For most products below a few tens of thousands of units a year, the module is the cheaper route when the whole project is counted.

Respect the antenna keep-out. Every antenna datasheet defines an area under and around it that must contain no copper, no components, no battery and no metal. That area is often the single largest constraint on the layout, and it is better to place it first and design the rest of the board around it.

Keep the radio ground intact. The radio needs a solid ground plane directly beneath it. Splitting that plane, or routing a high current path through it, degrades the sensitivity and the radiated performance in ways that are hard to diagnose after the fact.

Keep the switching supply away. A switching regulator is a broadband noise source. Place it on the opposite side of the board from the radio, keep the switching loop small, and filter the supply to the radio separately.

Think about the enclosure. A metal housing shields the antenna, and a hand near the antenna detunes it. The antenna position is a mechanical decision as much as an electrical one.

Our notes on internet of things PCBA cover the wider system context, and our notes on PCB design and layout describe the grounding and routing practices that follow from these rules.

Low Power Design

A battery powered device spends most of its life asleep, so the average current rather than the peak current decides the battery life. Four things dominate that average.

Quiescent current. Every regulator, sensor and interface has a standby current. Adding ten components that each draw a microampere adds ten microamperes to the average, which may be more than the radio uses when it is asleep.

Pull-up resistors. A pull-up that is permanently powered and pulled low by a switch is a continuous drain. Choose the largest value the interface tolerates, or switch the pull-up from a GPIO.

Leakage through unpowered sections. Where a section is powered down, current can still flow through a signal line into the unpowered device. Level shifters and series resistors with the right bias prevent it.

Duty cycle. The radio is the largest single consumer when it transmits. The protocol and the reporting interval matter more than any component choice, and the design should make it easy to change them.

Miniaturisation and Board Construction

A rigid board is simplest and cheapest and is used whenever the housing allows it. A flexible board fits a curved or folding product and removes the connector between two rigid sections. A rigid flex board combines both, with the components on the rigid parts and the flexible parts used for the fold.

Miniaturisation also means fine pitch. A system on chip with a fine pitch land grid array needs a small track width and spacing, a thin solder mask dam and often a via in pad construction. Each of those raises the fabrication requirement, and the design should stay one step inside the fabricator capability rather than at its limit. Our notes on flex PCB assembly describe the handling and support that flexible and rigid flex assemblies need.

Manufacturing and Test

The board is fabricated like any other, with the finish chosen so that the fine pitch pads solder reliably; electroless nickel immersion gold is common for that reason. Assembly places the radio module and the fine pitch processor with the same process as any other surface mount build, and the module is usually placed last or protected from the reflow of the other side.

Test is where a connected product differs. Every board is functionally tested, and the radio is tested too: the transmit power, the receive sensitivity and the antenna performance should be verified at least on a sample, because a matching component that is out of tolerance or a keep-out that was violated produces a product that works in the lab and fails in the field. The firmware is loaded and verified, the identity and the calibration data are written, and the result is recorded against the serial number. Our notes on PCB assembly describe the build, and our notes on PCB manufacturing cover the fabrication requirements for the fine geometry.

Where These Boards Are Used

• Smart home. Thermostats, locks, lighting, sensors and hubs, usually mains powered with a radio module and a small display.
• Wearables and health trackers. Battery powered, extremely compact, often rigid flex with a sensor on a flexible tail against the skin.
• Industrial IoT. Predictive maintenance, environmental monitoring and asset tracking, where the radio has to coexist with motors and drives and the enclosure is often metal.
• Automotive and electric vehicles. Telematics, remote monitoring and battery management, with the radio qualification added to the automotive requirements.
• Asset tracking. Cellular or LoRa devices that sleep for years and transmit occasionally, where the sleep current is the whole design.

What a Connected Board Costs

Cost follows the complexity rather than the connectivity. Indicative ranges for small quantities are around 5 to 15 US dollars for a simple two layer rigid board with a module, 15 to 40 dollars for a wearable flexible board, and 50 to 120 dollars for an advanced multilayer board with sensors and a high performance processor. Volume production reduces all three substantially.

The module, the sensors and the fine geometry drive the price, not the laminate. Where the design needs a flexible or rigid flex construction, the fabrication cost rises sharply at low volume and falls as the quantity grows, so the break-even between a rigid flex board and two rigid boards with a connector moves with the volume.

FAQ

Do I need a module or a chip-down radio design? A module below a few tens of thousands of units a year, because it includes the certification. Chip-down becomes economic at high volume where the certification cost is spread over many units.

How much does a connected device PCB cost? From around 5 to 15 US dollars for a simple two layer rigid board in small quantity, and considerably more for a compact flexible or multilayer design.

Why does the battery not last as long as the calculation said? Usually the quiescent current of the regulators and the sensors, or a pull-up that is permanently powered, rather than the radio itself.

Can a flexible board carry an antenna? Yes, but the ground structure changes as the board bends, so the antenna is usually placed on a rigid section and the flexible part is used for the fold and the sensor tail.

Conclusion

A connected device PCB is designed around three constraints: the radio needs space, an undisturbed ground and distance from noise, the battery needs every microampere accounted for, and the housing decides the board shape. Place the antenna first, keep the switching supplies away from it, audit the sleep current component by component, and verify the radio performance on real boards rather than assuming the layout worked.