Designing a Wearable GPS Tracker Board

A wearable tracker for a child or a pet is one of the harder small boards to design. It has to receive a satellite signal that arrives at roughly minus one hundred and thirty decibels, transmit through a cellular network with a power amplifier running at over one watt, survive being carried against a body, and run for days on a battery no larger than a coin.

This article covers the board-level decisions that make such a device work: the two antennas and the ground they need, the module placement, the charging and battery arrangement, and the mechanical and encapsulation choices that decide whether it survives real use.

What a Wearable Tracker Has to Do

The receiver has to acquire and track satellites, which means it needs a clear view of the sky and a low-noise signal path from the antenna to the receiving front end. The transmitter has to reach a base station, which means it needs an efficient antenna and a power amplifier whose harmonics do not disturb the receiver sitting millimetres away.

Both functions compete for the same small volume. The two radios work in different bands and are active at almost the same time, so the isolation between them is a layout problem as much as a filtering problem. The rest of the design, including the charging circuit, the microcontroller and the accelerometer, has to fit in the space left over without contributing noise to either radio.

The GPS Antenna and Its Ground Plane

A ceramic patch antenna is the usual choice for a small tracker. It is a resonant structure that needs a defined ground plane beneath it, and its performance depends on the size and the shape of that plane as much as on the antenna itself. An antenna specified for a given gain will not achieve that gain on a ground plane that is much smaller than the reference used for the measurement.

The practical consequence is that the ground plane under the antenna must be treated as part of the antenna design. It should be continuous, it should extend symmetrically away from the patch where the geometry allows, and it should not be cut by slots or crowded with components. The feed point should be kept short and routed away from any switching node, because the receiver has to detect a signal that a nearby converter can easily obscure.

Placing a Ceramic Antenna

The antenna belongs at one end of the board, on the top layer, with no copper underneath it and no components closer than the clearance the supplier specifies. That clearance is not a suggestion: placing a component or a battery connector inside it detunes the antenna and reduces the radiated power, and the effect is not recoverable by adjusting the matching network.

The matching network is a small pi network between the feed and the receiver, and it exists to compensate for the difference between the antenna as measured and the antenna as installed. It should be placed as close to the feed as possible, with the components on the top layer and short connections, because the purpose of the network is to correct the impedance at that point rather than somewhere along the trace.

The GSM Module and Its Antenna

The cellular side is the louder of the two. A transmit burst draws a large current in short pulses, which produces both a conducted disturbance on the supply and a radiated field from the antenna, and the receive path of the module is sensitive during the receive window. The module supplier specifies the supply decoupling and the bulk capacitance required to support the transmit burst, and those capacitors have to be placed as the supplier describes rather than wherever space permits.

The cellular antenna is usually a flexible structure or a metal frame element, and like the patch it needs a defined clearance area. The two antennas should be placed so that they do not face each other and so that the transmit antenna is as far from the patch as the board allows. Where the enclosure is small enough that both antennas are within a few centimetres of each other, the filtering on the receive path and the grounding of the transmit path become the two parameters that decide whether the product meets its sensitivity specification.

Charging, Battery and Connector

The charging connector is one of the few openings a sealed wearable has, and it deserves attention for that reason alone. A small connector with a defined insertion depth and a robust anchor to the board survives use; a connector held only by its solder joints will eventually lift, and the pad damage from that failure usually cannot be repaired on a thin board.

The battery charging path should be routed with its own ground return, and the charger should be placed so that its switching node is not near the receiver front end. For the battery, the constraint is the peak current during a transmit burst rather than the average consumption, and the trace width and current calculation has to be run on the burst current instead of the average, because a trace sized for the average will impose a voltage drop that resets the module at the beginning of every transmission.

Encapsulation, Flexibility and Washing

A device worn against the body is exposed to moisture and, if it is designed to be carried in clothing, to washing. Conformal coating protects the assembly from humidity and condensation, but it is not a substitute for a sealed enclosure where liquid immersion is possible. Where a fabric-mounted module is intended to be removed before washing, the mechanical design should make that obvious, because a user who forgets will not be impressed by a return.

Encapsulation of a GPS tracker board brings its own constraints. A potting compound adds mass and can detune an antenna if it flows into the clearance area, and a rigid compound on a thin board introduces mechanical stress as the assembly flexes. Where the board has to tolerate bending, a thin substrate with the components placed away from the bend line and a compliant coating is a better answer than a stiff encapsulant. The properties of potting and dispensing materials should be chosen from the cured state, including the dielectric constant, rather than from the uncured viscosity alone.

Test and Verification

A radio product cannot be verified by continuity testing. The meaningful tests are conducted and radiated: a sensitivity measurement on the receive path, a power and spectrum measurement on the transmit path, and a comparison of each against the same measurement taken with the antennas disconnected and connected. Comparing the two states is what identifies a board-level problem such as an inadequate ground plane or a detuned matching network, as opposed to a module problem.

Two practical tests catch most layout errors early. A near-field scan around the board while the transmitter is operating shows where the field is concentrated and whether the receive path is inside it. And a thermal check during a long transmit burst shows whether the supply drop at the module is acceptable, which is where a board with marginal copper reveals itself. Applying the same discipline used for compact processor core board design to the digital section keeps the rest of the board from contributing to either problem.

FAQ

Why does the antenna clearance area matter so much? Because a ceramic antenna is a resonant structure whose resonance depends on its surroundings. Copper or a component inside the clearance area shifts the resonance and reduces efficiency, and no matching network can restore what the detuning has removed.

Can a tracker share one antenna for GPS and cellular? It can, with a diplexer and careful layout, but the isolation between the two paths has to be designed. Separate antennas remain the simpler and usually better answer where the enclosure has room for both.

How should the supply for a cellular module be sized? From the transmit burst current rather than the average. The module draws a large current for a short time, and the supply network has to hold its voltage within the module limits throughout that burst.