How to Architect a Connected Device: From Antenna to Cloud — Part 2

This three-part series explores why and how device OEMs, systems integrators, and their customers should take a holistic view of the full connectivity stack: antenna, module, network, cloud, and data flow. This installment focuses on impedance matching and antenna efficiency, which are two of the most overlooked and underappreciated aspects of device design.

Waste Not, Want Not

Part 1 discussed when and how to minimize uplink and downlink data to help maximize the battery life of IoT devices. But as important as that is, minimizing payload size isn’t the only key consideration. If the RF path is inefficient, every byte worth sending still uses more power than it needs to.

To avoid that waste, pay close attention to impedance matching when choosing and integrating the antenna. Impedance mismatch wastes battery power and service life because the device’s power amplifier (PA) typically is designed and characterized into a 50 ohm (Ω) load. If the antenna and matching network don’t present ~50 Ω at the module’s RF pin, some of the transmitted power reflects toward the PA instead of radiating (quantified by VSWR/return loss/S11).

To compensate, the PA must be driven harder to achieve the target on-air power and satisfy the network’s link budget/power control commands. That draws disproportionately more current from the battery for the same effective radiated power (ERP). It also can trip thermal foldback or force retransmissions that require even more battery power.

Follow these four best practices to maximize performance and battery life:

• Base the match on the board rather than the datasheet. Module vendors publish a reference matching network, but it’s tuned for their reference antenna and layout. Any changes to the antenna, trace length, enclosure (plastic vs. metal), or nearby ground pour cause the optimal component values to shift. Skipping board-specific tuning is one of the most common reasons a certified module underperforms once soldered into a real enclosure.
• Measure it; don’t assume it. Conduct a VSWR/return-loss sweep across the full operating band (not just the center channel) during EVT, measure antenna efficiency, plus an OTA (TRP/TIS) check once the antenna is in its final enclosure. Plastic housings, batteries, and nearby metal such as meter cans and junction boxes all detune antennas in the field in ways bench testing with an open board won’t catch.
• Minimize the hidden multiplier of retransmissions. A marginal link doesn’t just waste power on reflection. Weaker ERP means more retries, longer time on air, and in extreme cases dropped registration/re-attach cycles, all of which cost far more energy than the few percent lost to a slightly off match. Good matching pays for itself many times over specifically because it prevents this cascade.
• Don’t fall for a cheap fixed match. An example is devices that will be installed in a wide variety of enclosure types and indoors and outdoors. A fixed L-match tuned once for a “typical” install is cheapest but will struggle to deliver that performance in other installations. In a 10,000-unit smart meter deployment, for example, the bulk of the support costs come from just 5% of installations: the ones in challenging locations such as in a cast-iron pit under a metal lid. Characterize the full distribution of installs, not just the typical one. Antenna tuner ICs or SKU-specific matching cost a little more but can save money in the long run by eliminating the expense of truck rolls and other work to make a device meet performance KPIs.

Falling Down the Stairs

Many IoT systems use LPWAN technologies such as NB-IoT and LTE-M because they meet business requirements such as affordability, long battery life, and the ability to remain in service for five years or longer. An often overlooked aspect is that their link margin degrades in steps rather than smoothly, which directly affects battery life.

For example, if an antenna is 3 dB less efficient than it should be, the network will command repetition (2–128×), which means energy per message jumps by a multiple, not a percentage. That also undermines all of the time and effort spent minimizing data payloads. These caveats highlight why battery life vs. antenna quality is a step function, not a slope.

Finally, when comparing antennas, keep in mind that their datasheets lead with peak gain (directional, measured on a reference ground plane). What really matters is total radiated efficiency and TRP/TIS. A 2 dBi antenna at 35% efficiency loses to a 0 dBi model at 65%.

Get in touch for orders or any queries: sales@rfdesign.co.za / +27 21 555 8400

Courtesy of Taoglas