23 Link Budget and Coverage Planning

A link budget is the radio range ledger for an IoT deployment. It starts with transmit power, adds antenna gains, subtracts cable, path, obstacle, and reserve losses, then compares the predicted received power with receiver sensitivity. The useful output is not a single optimistic range number. It is the remaining margin and the assumptions that created it.

Use the record before choosing gateway spacing, antenna placement, enclosure location, or battery targets. If the ledger has no named environment, no reserve margin, or no field measurement plan, the design is still a guess.

The ledger matters because every term answers a different review question. Transmit power and antenna gain describe how much useful signal leaves the device. Cable loss, connector loss, and antenna orientation describe how much of that launch power actually reaches the path. Path loss, material attenuation, Fresnel clearance, foliage, and body blocking describe what the site does to the signal. Receiver sensitivity and the required packet success target describe how weak the signal can be before the application stops meeting its requirement.

A simple budget also makes trade-offs visible. Suppose a device launches 14 dBm, gains 2 dB from its antenna path, loses 96 dB to distance, and loses another 15 dB to the site. The predicted received power is -95 dBm. If the deployed receiver mode is sensitive to -118 dBm and the design requires 10 dB of reserve, the remaining release margin is 13 dB. That is a reviewable decision. If any input changes, the team can recompute the same ledger instead of debating a vague range claim.

Read Figure 23.1 as the order in which this chapter’s ledger is filled in.

+

Each step adds one line to the record. Transmit power opens it. Antenna gains at both ends are added, then path, cable and wall losses are subtracted. The running total is compared with receiver sensitivity, which is where a design first looks either promising or short. A fade margin, typically 10 to 20 dB, is then held back, so the answer survives a bad day rather than a good one. The two endings are the point of Figure 23.1. A calculated pass still needs field validation. A fail names possible repairs: more height, more gain, permitted power, or more gateways.

A useful link budget is traceable. Each number should have a source, a location, and an owner: radio configuration, antenna datasheet, cable length, path model, measured RSSI, or field survey. Keep the calculation simple enough that a field engineer can update it when gateway height, antenna alignment, enclosure material, or sensor location changes.

Build the record in the same order the signal experiences the path. Start with the transmitter and antenna system, then the open-air or log-distance path estimate, then installation losses, then receiver sensitivity, then reserve. Add a short evidence note beside each uncertain term. For example, wall loss may come from a site survey, a conservative design allowance, or an early RSSI walk test. The review should be able to tell which terms are measured, which are data-sheet values, and which are deliberate safety allowances.

After installation, keep the budget alive. Field measurements should not replace the record; they should calibrate it. If measured RSSI is weaker than predicted, update the loss assumption and check whether similar paths are also at risk. If measured RSSI is stronger than predicted, keep the conservative reserve unless the team has enough representative samples to reduce it. That habit turns the link budget into an operating document for gateway moves, enclosure changes, seasonal foliage, and radio-mode updates.

Use Figure 23.2 to trace every gain, loss, and reserve behind an LPWAN coverage claim.

+

Begin with send power and antenna gain. Then subtract loss in the cable, plugs, path, walls, and site. Compare the result with receiver sensitivity at the chosen data rate. Maximum coupling loss is the total loss that the pair can bear, not a range claim. The link still needs margin for fades, model error, weather, antenna angle, and age. Keep all terms in decibels so gains and losses add in one unit. The record should name each source, site type, receiver setting, spare margin, and field test that backs or changes the model.

Where The Antenna-Gain Line Actually Comes From

The ledger's antenna-gain line is not a number to look up once and forget. It is a choice between antenna types that trade directionality for reach, and that choice decides how much of the launched power actually lands where the deployment needs it. An omnidirectional antenna radiates in all directions around one plane, spreading a fixed amount of power over the widest area and giving typically the least gain of the common choices. A Yagi antenna lines up several parallel elements to concentrate that same power into a narrower beam, trading coverage angle for a substantial gain increase. A parabolic reflector antenna goes further, using a dish or mesh reflector built larger than the wavelength it focuses to concentrate power into a tight beam with typically the highest gain of the three, at the cost of needing the far end to stay inside that narrow beam. Actual gain depends on frequency, size and efficiency.

Selecting between them is a directionality, gain, and bandwidth decision made together, not a single spec copied from a datasheet: how tightly can the deployment predict where the far end will be, how much gain does the remaining margin actually need, and does the antenna's usable bandwidth cover the intended band. The same decision should also record polarization. A receiver's orientation is rarely known in advance -- a handheld device, a vehicle-mounted sensor, or a tag on a rotating asset can present any polarization to the antenna. Some access points answer this by propagating both vertical and horizontal polarization, or by using a configurable or diversity antenna, so an unknown receiver orientation does not silently turn into an unrecorded polarization-mismatch loss. Treat polarization the same way the ledger already treats cable loss: name it in the record, or budget an explicit mismatch allowance for it.

Read the steps in Figure 23.3 in order, because one of them is far larger than the rest.

+

The example signal starts at 14 dBm of transmit power and gains 2 dBi from the antenna. Cable loss takes 1 dB back, which looks trivial. Then path loss removes 105 dB, and the scale of the drawing makes the point: nothing you can add at the transmitter matches that single loss. The fade margin, typically 10 to 20 dB, is held back after it, not before. The final box asks whether what remains still sits above receiver sensitivity. That is why Figure 23.3 is read step by step. A budget that passes on path loss alone can still fail once installation losses are added.

Once the ledger fields are named, the fastest way to see how they trade off is to move them. Try the range calculator below to estimate wireless range from link budget, path loss, receiver sensitivity, and fade margin, and to compare that estimate across common IoT protocols under the same deployment assumptions.

The budget is arithmetic, but the risk is in the assumptions. Free-space path loss is a baseline, not a site guarantee. Receiver sensitivity depends on data rate, bandwidth, coding, and required error behavior. Antenna gain depends on orientation and mounting. A clean design keeps those assumptions visible instead of hiding them inside one range claim.

dB arithmetic is useful because gains and losses can be added in one chain, but it can also hide weak evidence. A quoted 6 dBi antenna gain only helps when the antenna pattern points toward the receiver and is mounted away from detuning surfaces. A receiver sensitivity number only applies to the bandwidth, coding, spreading, data rate, packet error target, and interference assumptions used to measure it. A path-loss number only applies to the environment model and distance convention recorded with it.

Margin is therefore not spare decoration at the end of the spreadsheet. It is the buffer between an assumed world and the deployed world. Fade, multipath, human movement, wet foliage, antenna tilt, component tolerance, enclosure changes, and measurement error all consume that buffer. When the reserve is too small, the system may work during commissioning and fail later under normal site variation. When the reserve is visible, engineers can decide whether the right fix is more gateway density, a different mounting point, a slower radio mode, or a field survey.

For review, separate nominal margin from release margin. Nominal margin is predicted received power minus receiver sensitivity. Release margin subtracts the reserves the team promised to protect: fade allowance, installation uncertainty, site-change allowance, and measurement uncertainty. That second number is the one that should drive approval because it says what headroom remains after known risk has been paid for.

Figure 23.4 works the chapter’s own numbers in three stages, so check each stage against the arithmetic you would do yourself.

+

Stage one adds and subtracts in dB: 14 dBm of transmit power, plus 2 dB of antenna path gain, minus 96 dB of path loss and 15 dB of site loss, giving -95 dBm at the receiver. Stage two compares that with a receiver sensitivity of -118 dBm, and the gap is 23 dB. Stage three is the one a review must not skip. Subtract the 10 dB reserve the team promised, and 13 dB is what actually remains. Figure 23.4 separates those three results because only the last one is margin you may spend.

Use Figure 23.5 to distinguish raw link margin from the reserve left after deployment risks are charged.

+

Read received power against the receiver threshold: their difference is the raw margin, with stronger signals above and weaker signals closer to failure. Then subtract explicit allowances for fading, body or wall loss, antenna variation, installation error, and model uncertainty. For example, -105 dBm - (-120 dBm) = 15 dB raw margin; reserving 10 dB leaves 5 dB decision margin. The remaining value, not the raw gap, supports release. Field percentiles and outage behaviour must still confirm that the chosen reserve covers the required service condition.

Compare the urban and rural paths in Figure 23.6 to see why identical radios can have very different range.

+

The urban path in the illustration adds buildings, lower antennas, reflections, interference, and penetration loss, consuming link budget quickly. The rural path offers clearer geometry and fewer obstructions, so the same transmitter, receiver, and settings can cover roughly five to fifteen times the distance in this illustration. That ratio is an observed class, not a conversion rule for every site. Gateway height, terrain, foliage, indoor placement, and traffic load still change coverage. A coverage plan should use separate environment models and repeated field passes, then place gateways from measured margin and outage evidence rather than one advertised radius.