A coverage prediction that looks excellent on paper can fail quickly when a meter is installed in a basement, a sensor sits behind process equipment, or a gateway antenna is mounted below a roofline. To calculate LoRaWAN link budget accurately, teams need to account for the complete radio path, not just a gateway's advertised range. The result is a practical estimate of whether a device has enough signal margin to communicate consistently at its planned location.
For municipalities, utilities, industrial operators, and system integrators, link budget is an early design control. It helps determine gateway density, antenna placement, device configuration, and whether a proposed network can support its intended service level before hardware reaches the field.
What a LoRaWAN link budget measures
Link budget is the difference, expressed in decibels, between the power available at the receiver and the minimum signal level that receiver can decode. In simple terms, it starts with transmitted power, adds useful antenna gain, then subtracts every loss between the end device and gateway.
A larger budget generally supports a longer or more obstructed radio path. It does not guarantee coverage. Buildings, terrain, vegetation, interference, antenna polarization, and installation quality can all consume margin that a basic calculation does not capture.
The received-power equation is:
`Received power (dBm) = Transmit power + transmit antenna gain - transmit losses - path loss - additional losses + receive antenna gain - receive losses`
Once received power is known, calculate margin with:
`Link margin (dB) = Received power (dBm) - receiver sensitivity (dBm)`
Because receiver sensitivity is a negative number, subtracting it increases the result. For example, a received signal of -120 dBm at a receiver with -132 dBm sensitivity has 12 dB of link margin.
The inputs needed to calculate LoRaWAN link budget
The calculation is only as credible as its inputs. Manufacturer data sheets are a useful starting point, but installation-specific losses and the actual regional channel plan matter just as much.
Transmit power and effective radiated power
Start with the radio's conducted transmit power in dBm. Many LoRaWAN end devices operate at values such as 14 dBm or 20 dBm, depending on the region, channel, hardware, and configured data rate.
Then calculate effective isotropic radiated power, or EIRP. This includes the transmitting antenna gain and subtracts losses from cable, connectors, surge protection, and any other RF components between the radio and antenna.
`EIRP = conducted transmit power + antenna gain - feeder losses`
A 20 dBm device connected to a 2 dBi antenna through 1 dB of loss produces 21 dBm EIRP. Regulatory limits apply to the full EIRP, not simply the radio's conducted power. US deployments using the 902-928 MHz band must be designed around the applicable FCC requirements, channel configuration, and equipment certification.
Receive antenna gain and feeder loss
Gateway antenna systems deserve the same scrutiny as the device side. A 6 dBi outdoor antenna may improve the path substantially over a low-gain antenna, but only if its radiation pattern suits the site and its feeder losses are controlled.
Long coaxial runs can remove much of the expected benefit. Lower-loss cable, a short feeder path, properly rated lightning protection, and weatherproofed connectors are often more valuable than selecting an antenna with a slightly higher gain rating. Also consider antenna pattern: high-gain omnidirectional antennas concentrate energy closer to the horizon and may create weaker coverage above or below the gateway elevation.
Path loss
Free-space path loss estimates signal reduction over distance with a clear line of sight. A common equation is:
`FSPL (dB) = 32.44 + 20 log10(distance in km) + 20 log10(frequency in MHz)`
At 915 MHz across 10 km, free-space path loss is approximately 111.7 dB. That is useful as a baseline, not as a field prediction. Urban canyons, dense vegetation, metal-clad facilities, reinforced concrete, and terrain obstructions introduce additional loss. A device installed inside a utility vault or electrical room may face 10 to 30 dB more attenuation than an outdoor device at the same distance.
Receiver sensitivity
LoRa modulation can achieve very high sensitivity, especially at slower data rates. Sensitivity varies with spreading factor, bandwidth, coding rate, noise figure, and receiver implementation. A gateway or end-device data sheet may show sensitivity near -132 dBm at a given configuration, while a faster configuration may require a much stronger signal.
This creates a central LoRaWAN trade-off. Higher spreading factors can improve link budget, but they increase airtime and reduce network capacity. Designing every device for SF12 may make isolated difficult links work, yet it can create congestion and duty-cycle limitations in a growing deployment. Adaptive Data Rate should be used to move stable devices toward the fastest reliable setting, while preserving lower-rate options for devices that genuinely need them.
A worked LoRaWAN link budget example
Consider a 915 MHz uplink from an outdoor monitoring device to a gateway 10 km away. The end device transmits at 20 dBm through a 2 dBi antenna with 1 dB of loss. Its EIRP is therefore 21 dBm.
At the gateway, assume a 6 dBi antenna with 1.5 dB of feeder and protection loss, for an effective receive-side gain of 4.5 dB. Free-space path loss is 111.7 dB. With no other losses, the predicted received power is:
`21 dBm - 111.7 dB + 4.5 dB = -86.2 dBm`
If the selected LoRaWAN configuration has a receiver sensitivity of -132 dBm, the theoretical link margin is 45.8 dB. That number may appear generous, but it is not the margin available to the deployed system. Add a 12 dB building-entry loss, 15 dB of foliage or local clutter, and a 10 dB fade allowance, and the remaining margin falls to 8.8 dB.
That may be acceptable for a noncritical outdoor sensor, depending on packet frequency and required delivery performance. It may not be sufficient for a meter installed in a below-grade enclosure or for an operational alarm that needs dependable uplink and downlink connectivity. The right response could be a better gateway site, an elevated antenna, an additional gateway, or a different sensor installation position. Raising transmit power alone is rarely the cleanest answer.
Plan for fade margin, not best-case range
Fade margin is the reserve that remains after expected losses. It protects the link against seasonal foliage, rain-related environmental changes, moving vehicles, local interference, equipment variation, and the imperfect repeatability of field installations.
There is no universal requirement, but 10 to 20 dB is a sensible planning range for many professional deployments. Dense urban, industrial, indoor, and difficult terrain environments often need more. A rural, elevated, near-line-of-sight link may need less, provided it is validated through site testing.
Do not treat fade margin as an arbitrary correction that can be added after an optimistic design. Use it to force a realistic conversation about installation conditions. A link that only works at the highest spreading factor with 2 dB of predicted margin is a risk, not coverage.
Check both uplink and downlink paths
LoRaWAN networks are often uplink-focused, but downlink performance matters for acknowledgments, configuration changes, multicast, clock synchronization, and device management. The end device and gateway may have different transmit powers, antenna systems, and receiver sensitivities. Calculate each direction separately.
Downlink can be the weaker path when a gateway has a constrained antenna location or when an end device is buried inside an enclosure. It also has operational limits: gateways must share downlink capacity across devices, and regional regulations can constrain transmission behavior. A strong uplink does not automatically mean a dependable bidirectional service.
Turn the calculation into a deployment decision
A link budget is most useful when paired with a propagation model, a site survey, and representative field tests. Model coverage to identify likely weak areas, then verify it using the actual gateway, antenna height, device enclosure, and device configuration planned for production. Record RSSI, SNR, spreading factor, packet success rate, and gateway diversity rather than relying on a single received packet.
Gateway placement usually has more influence than small changes in device hardware. Height, clear exposure, separation from noise sources, proper grounding, and an antenna selected for the coverage objective can improve network results across thousands of endpoints. In multi-gateway deployments, overlapping coverage also adds resilience and reduces dependence on one marginal path.
LoRaWorld supports this engineering approach with vetted gateway and antenna infrastructure from established LoRaWAN manufacturers, helping deployment teams align hardware choices with real coverage requirements.
Before approving a network design, test the locations that are hardest to serve, not only the locations that make the coverage map look good. Those difficult endpoints will determine whether the link budget is a planning exercise or a dependable operating network.