LoRaWAN beside Wi-Fi | OT Wireless

LoRaWAN moves a few bytes from a sensor across a whole site on very little power. It cannot carry a control loop or a safety connection. Its value to your Wi-Fi is the traffic it takes away.

What it is

LoRa is a long-range, low-power, low-data-rate modulation, a form of chirp spread spectrum. LoRaWAN is the network layer the LoRa Alliance specifies on top of it.

The network is a star of stars. Devices send over a single hop to gateways, which forward over another network, such as IP over Ethernet or cellular, to a central network server. The figures you will hear are 2 to 5 km in urban areas, 15 km in suburban areas and about 10 years of battery life¹. Treat them as claims to verify in your buildings.

The U.S. channel plan

The LoRa Alliance’s regional parameters divide the U.S. band into three groups:

64 uplink channels of 125 kHz, from 902.3 to 914.9 MHz in 200 kHz steps².
8 uplink channels of 500 kHz, from 903.0 to 914.2 MHz.
8 downlink channels of 500 kHz, from 923.3 to 927.5 MHz.

The rates are low. The slowest uplink, DR0, runs SF10 on 125 kHz at about 980 bit/s. DR4 reaches 12,500 bit/s on a 500 kHz channel, and the newest revision adds optional rates up to 15,625 bit/s. At DR0 a frame carries at most 11 bytes of application payload.

Dwell time, not duty cycle

In Europe’s 868 MHz band, end devices are held to a 1% duty cycle. The U.S. limits time on each frequency instead. Under 47 CFR 15.247, a frequency hopper with channels narrower than 250 kHz must use at least 50 hopping frequencies and occupy any one for no more than 0.4 seconds in a 20-second period³. Digitally modulated systems need a 6 dB bandwidth of at least 500 kHz and no more than 8 dBm in any 3 kHz.

LoRaWAN maps onto both. The regional parameters cap uplinks at 400 ms of dwell when the bandwidth is under 500 kHz, and set each data rate’s maximum payload from the maximum allowed transmission time. That is why the slowest rate carries so little. On the wide channels, the power-density rule holds a device to about +26 dBm conducted, and a hybrid of hopping and digital modulation to about 21 dBm.

How long a command waits

Every LoRaWAN device implements Class A. Classes B and C are options layered on top.

Class A. Each uplink is followed by two short receive windows, and a downlink at any other time waits for the device’s next uplink. By default the first window opens 1 s after the uplink and the second at 2 s. If the sensor reports once an hour, a command can wait an hour.
Class B. Gateways send a beacon every 128 s⁴, and devices open scheduled ping slots from about every 1 s to about every 128 s.
Class C. The receiver stays open nearly all the time except while transmitting, which costs the most power and gives the lowest latency.

Uplinks are another matter. A Class A device transmits on its own schedule with a random offset, ALOHA style. Classes B and C add receive windows, not transmit slots, so no class schedules the reading you need next.

Why it cannot carry control or safety traffic

Start with the strongest case for it. A 2020 simulation found LoRaWAN could serve event-triggered control of a slow water network, for systems that tolerate blackouts and delays of up to 10 s⁵. Small networks with low spreading factors and enough channels can serve applications that sample every second.

That is the ceiling. Industrial control loops may need response times of about 1 ms to 100 ms, while a 10-byte packet at SF7 spends about 40 ms on air. Because uplink access is contention-based, latency will not be deterministic. Reliability costs capacity too: confirmations are downlinks, and acknowledged traffic drains the network.

Put that beside a safety connection. A CIP Safety input connection at default settings faults, and puts its data in the safe state, if no valid packet arrives within 40 ms⁶. A single LoRaWAN uplink at the slowest U.S. rate is sized to a 400 ms dwell limit. Anything a PLC acts on every cycle stays on a network built for cycles.

Where it relieves your Wi-Fi

The fit is the long tail of slow sensors: readings that arrive every few minutes and can arrive late. Every one you put on LoRaWAN is one fewer client associated to the access points your vehicles depend on.

The band is separate as well. LoRaWAN lives in 902 to 928 MHz, so its airtime is not your 2.4, 5 or 6 GHz airtime. It shares that band with Wi-Fi HaLow, which the Wi-Fi Alliance places in the same 902 to 928 MHz⁷. It shares on unlicensed terms: spread spectrum systems there operate on a noninterference basis with government systems, which rank behind Part 18 ISM equipment.

The gateways need backhaul. If that backhaul is Wi-Fi, you have moved the sensors, not the traffic.

Field note. You have to know everything on the air so control and safety traffic are prioritized over large transfers and updates.

A sensor network you did not plan for is one more thing on the air you do not know about. Put LoRaWAN in the same inventory as everything else. The airtime guide covers the rest.

When to use it

Battery sensors that report every few minutes or less often, and can tolerate a late or lost reading.
Wide sites, yards and outbuildings where a gateway costs less than cable or more Wi-Fi.
Monitoring, never control.

What to ask a vendor

Which LoRaWAN version and regional parameters revision the devices and network server implement, and which data-rate option.
Which of the U.S. uplink channels each gateway listens on, and how devices learn that plan.
The device class, and the longest a downlink can wait in that class at your reporting interval.
Whether readings are confirmed, and what confirmations do to capacity at your device count.
How the gateways backhaul, and whether any of it rides your Wi-Fi.
How firmware is updated at these payload sizes.

LoRaWAN earns its place by taking the slow traffic off Wi-Fi. Ask nothing faster of it.