2026-06-15 · wireless · BLE · WiFi · LoRa · IoT · RF

Wireless: range, data rate, power — pick two

Every wireless link is a negotiation with physics, and the physics is blunt: you want long range, high data rate and low power, and you may have any two. A radio that reaches kilometres on a coin cell moves almost no data. One that streams video drinks current and barely leaves the room. Choosing a wireless technology is choosing which corner of that triangle your product lives in — everything else is detail.

Why you only get two

Three levers set the outcome. Range comes from link budget: more transmit power, a lower data rate (more energy per bit and a narrower, more sensitive receiver), or a lower frequency — sub-GHz signals travel further and punch through walls better than 2.4 GHz. Data rate costs bandwidth and signal-to-noise, which means more power or less range (Shannon doesn’t negotiate). Power is what’s left to trade. Push any one lever and at least one other gives.

So the map of wireless has a shape: cheap, low-power radios cluster at *short range

  • modest rate* (BLE, Zigbee) or long range + tiny rate (LoRa). The top-right corner — long range and high rate — exists only with real power or a paid carrier behind it.

A range-versus-data-rate map of wireless technologies on logarithmic axes. The horizontal axis is range from ten metres to ten kilometres; the vertical axis is data rate from ten kilobits per second to one gigabit per second. Wi-Fi sits at short range and very high rate, marked high power. BLE and Zigbee/Thread sit at short range and low-to-moderate rate, marked low power. LoRaWAN sits at long range and very low rate, marked ultra-low power. Cellular NB-IoT and LTE-M sit at wide-area range and modest rate, marked carrier-powered. The long-range, high-rate corner is annotated as empty — it costs serious power or money. Figure 1 — The design space. Each technology claims a region; none reaches the long-range, high-rate corner cheaply. Read power as the third axis coming out of the page — the radios furthest right and lowest are the ones you can run for years on a battery.

The cast

TechnologyBandRangeData ratePowerTopology
BLE2.4 GHz10–100 m0.1–2 Mbpsvery lowstar / mesh, phone-hub
Wi-Fi (802.11)2.4 / 5 / 6 GHz30–100 m10s Mbps – Gbpshighstar (access point)
Zigbee / Thread (802.15.4)2.4 GHz10–100 m / hop250 kbpslowself-healing mesh
LoRaWANsub-GHz (868/915)2–15 km0.3–50 kbpsvery lowstar-of-stars
NB-IoT / LTE-Mlicensed cellularkm, nationwide0.1–1 Mbpslow–mediumcellular

Where each one wins — and doesn’t

BLE wins because the hub is already in everyone’s pocket: a phone pairs with it directly, no gateway to build. Coin-cell power and low cost make it the default for wearables, beacons, medical peripherals and phone-connected gadgets. It loses on range (a room, maybe a floor) and throughput, and it shares the crowded 2.4 GHz band with everything else.

Wi-Fi wins on raw throughput and the fact that it speaks IP straight to the internet — cameras, audio, firmware images, any device that moves real data and can afford it. It loses on power: association and a high-rate radio rule out a multi-year coin cell, so Wi-Fi wants mains or a fat battery. (802.11ah “HaLow” trades rate for sub-GHz range and lower power, but adoption is thin.)

Zigbee / Thread win on coverage through mesh: every mains-powered node repeats, so the network self-heals and reaches across a house at low power for the battery end-devices. Thread plus Matter is the home-automation backbone for lights, sensors and locks. They lose on throughput (250 kbps) and need a border router to reach the internet.

LoRaWAN wins where nothing else can go: kilometres of range on years of battery, on free spectrum with a gateway you own. It is the choice for farm, metering, asset and environmental sensors. The price is brutal data rate and latency — payloads of tens of bytes, sent seconds apart, capped by duty-cycle and airtime rules. Don’t design anything chatty, real-time, or OTA-heavy on it.

NB-IoT / LTE-M win when you need wide-area reach with no gateway to deploy — the carrier already covers the country. Trackers and smart-city nodes lean on them. The cost is a SIM and subscription, and more power than LoRa for the same idle sit.

The question that usually decides it

Before range or rate, ask what is the hub? Your phone → BLE. An access point you already run → Wi-Fi. A border router for a mesh → Thread/Zigbee. A gateway you deploy → LoRaWAN. The carrier’s network → cellular IoT. That single answer, crossed with your bandwidth and battery budget, picks the technology more often than any datasheet number.

Four wireless network topologies drawn as small node diagrams. Star: a central access point or phone with several devices hanging directly off it, labelled Wi-Fi and BLE, hub equals access point or phone. Mesh: a web of interconnected nodes with one acting as a border router, labelled Zigbee and Thread, self-healing, hub equals border router. Star-of-stars: many end nodes reporting to a couple of gateways that forward to a network server in the cloud, labelled LoRaWAN, hub equals a gateway you deploy. Cellular: a device talking to a tower that connects to the carrier cloud, labelled NB-IoT and LTE-M, hub equals the carrier. Figure 2 — Same question, four answers. The infrastructure a technology needs — a phone, an AP, a mesh router, a gateway, or a carrier — is often the real deciding factor, not the link itself.

A sorting diagram. A product requirement on the left feeds five buckets, each a wireless technology with the use cases it owns. BLE: phone-connected wearable, beacon, medical peripheral. Wi-Fi: camera, audio streaming, firmware images, anything above a megabit. Thread or Zigbee: whole-home sensors, lights and locks needing mesh reliability. LoRaWAN: battery sensor across a farm or city sending tiny data for years. Cellular NB-IoT or LTE-M: wide-area tracker with no gateway to deploy. Each bucket notes the deciding factor: phone hub, bandwidth, mesh coverage, range on a coin cell, or nationwide reach. Figure 3 — Sort the use case, not the buzzword. Bandwidth, the hub you can build, and the battery you can fit usually drop a product cleanly into one bucket.

Field notes

  • 2.4 GHz is a swamp. Wi-Fi, BLE, Zigbee and the microwave oven all share it; co-existence and channel planning matter. Sub-GHz (LoRa, 802.11ah, some Thread) is calmer and travels further for the same power.
  • Halve the datasheet range, at least. Quoted figures are line-of-sight ideal; walls, human bodies and 2.4 GHz absorption cut them hard. Budget a fraction.
  • Mind LoRa’s airtime. Duty-cycle and regional limits cap how often you may transmit — a chatty protocol or a firmware update over LoRa will hit a wall.
  • Energy per message, not peak current. A fast radio used briefly often beats a slow radio used long. Measure the joules to deliver one packet, including wake, associate and retry.
  • Count the hidden hub cost. “No carrier fee” LoRa still needs a gateway; “no gateway” cellular still needs a SIM and a subscription. Neither is free.