Choosing an Antenna for 2.4 GHz IoT: Gain, Pattern and Placement

The antenna is the one component in your 2.4 GHz product that physics will not let you hide. You can shield a noisy regulator and re-spin a bad matching network, but an antenna that was chosen for the wrong reasons will cost you range, certification time and, in the worst case, a mechanical redesign.

Gain is a trade, not a score

Datasheet gain numbers read like a leaderboard: 2 dBi, 5 dBi, 9 dBi. It is tempting to buy the biggest number. But antenna gain is not amplification — it is redistribution. A 9 dBi patch achieves its gain by narrowing the beam to roughly ±30°; those decibels are taken from every other direction.

For a fixed point-to-point link, that is exactly what you want. For an IoT device that could sit in any orientation — a sensor a customer velcros to a wall, a tracker in a moving vehicle — high gain means deep nulls, and deep nulls mean support tickets. The right specification for most mobile or arbitrarily-oriented devices is a low peak gain with a smooth pattern: 2 dBi and no hole deeper than −10 dB beats 6 dBi with a −25 dB null.

The ground plane is part of the antenna

Chip and PCB-trace antennas do not radiate by themselves; the return currents in your ground plane do half the work. Every 2.4 GHz chip-antenna datasheet assumes a specific board size and keep-out. Shrink the ground plane below about a quarter wavelength (~31 mm) and three things happen at once: the resonant frequency shifts up, the bandwidth collapses, and efficiency drops — sometimes by more than half.

Practical consequences:

Freeze the board outline before tuning the match. A matching network tuned on a 60 mm board is wrong on a 40 mm board.
Keep the keep-out honest. Copper pours, battery cables and mounting screws inside the keep-out detune the element.
Plastics matter. An enclosure wall 1 mm from the element shifts resonance down by tens of MHz. Tune in the final enclosure, not on the bench in free air.

External, patch or trace?

| Choice | When it wins | The catch | | ------------------------------------------------------------------ | ---------------------------------------------- | ---------------------------------------- | | PCB trace / chip | Cost, no assembly step | Efficiency lives and dies by your layout | | External whip | Best omni efficiency, easy certification story | Ingress, cost, someone will snap it off | | Patch (like our SL-ANT-5G8) | Fixed links, known orientation | Directional by design |

A selection checklist

Define orientation: fixed, constrained, or arbitrary?
Pick the pattern first, gain second.
Confirm the ground-plane and keep-out budget fits your enclosure.
Check VSWR across the band edges at temperature, not just at 2.44 GHz.
Ask for measured patterns — simulated farfields flatter everyone.

Every antenna we sell ships with measured E- and H-plane patterns from our range, because a pattern plot you can trust is the difference between a selection and a gamble.

Gain is a trade, not a score

The ground plane is part of the antenna

Practical consequences:

Freeze the board outline before tuning the match. A matching network tuned on a 60 mm board is wrong on a 40 mm board.

Keep the keep-out honest. Copper pours, battery cables and mounting screws inside the keep-out detune the element.

Plastics matter. An enclosure wall 1 mm from the element shifts resonance down by tens of MHz. Tune in the final enclosure, not on the bench in free air.

External, patch or trace?

A selection checklist

Define orientation: fixed, constrained, or arbitrary?

Pick the pattern first, gain second.

Confirm the ground-plane and keep-out budget fits your enclosure.

Check VSWR across the band edges at temperature, not just at 2.44 GHz.

Ask for measured patterns — simulated farfields flatter everyone.

Every antenna we sell ships with measured E- and H-plane patterns from our range, because a pattern plot you can trust is the difference between a selection and a gamble.