RF exposure limits explained: the FCC and ICNIRP numbers in plain English

People who start measuring the radio signals around them soon ask how their numbers compare with the official limits. This guide explains who sets those limits, what they are designed to protect against, how the frequency formulas work, and how a professional checks compliance. It works through the arithmetic at the frequencies you're most likely to meet. It also explains, step by step, why a reading from a software radio, Consentiva's included, isn't an exposure measurement, and what you'd need to turn a reading into a fraction of a limit.

Who sets the limits

Several bodies publish radio-frequency (RF) exposure limits. They overlap a great deal, but the numbers and the fine print differ, so it helps to know which document someone is quoting.

The FCC (United States)

In the United States the Federal Communications Commission sets the legally binding limits for transmitters it regulates. The maximum permissible exposure (MPE) limits are in 47 CFR § 1.1310, which you can read on the official eCFR. The same section sets the specific absorption rate (SAR) limits used for phones and other devices held close to the body. Related rules say when a licensee has to evaluate a transmitter (47 CFR § 1.1307) and how portable and mobile devices are assessed (47 CFR §§ 2.1093 and 2.1091).

The FCC's limits are based largely on earlier recommendations from the National Council on Radiation Protection and Measurements (NCRP) and the Institute of Electrical and Electronics Engineers (IEEE). The FCC's technical guidance on how to evaluate compliance is OET Bulletin 65, published by its Office of Engineering and Technology.

ICNIRP (international)

The International Commission on Non-Ionizing Radiation Protection is an independent scientific body. It publishes guidelines rather than law, but many countries adopt them, in full or in part, as their national limits. Its 1998 guidelines were in force for over two decades and are still the basis of some national rules. In 2020 it published revised guidelines covering 100 kHz to 300 GHz. Both documents are available from icnirp.org.

IEEE

The IEEE publishes its own standard, IEEE C95.1-2019, on safety levels for human exposure to electric, magnetic and electromagnetic fields from 0 Hz to 300 GHz. You'll see it cited in engineering work. It isn't the FCC rule, though, so in the US the FCC table is the one that decides legal compliance.

What the limits are designed to protect against

At the frequencies used for radio, TV, mobile phones and Wi-Fi, the established effect of strong RF fields is heating of body tissue. At the lowest frequencies ICNIRP covers, up to about 10 MHz, its guidelines also consider nerve stimulation. The limits are set at levels well below the thresholds at which those established effects occur, with safety margins built in. The general-public limits carry a larger margin than the occupational ones.

This guide describes what the limits are and how they are applied. It doesn't make claims about health beyond what the regulators themselves state. If you have health questions, the regulators' own publications and your doctor are the right sources.

General population versus occupational exposure

Both the FCC and ICNIRP have two tiers of limits.

• General population / uncontrolled exposure (FCC) or general public (ICNIRP). This applies to anyone who may be exposed without knowing it or without being able to control it: residents, passers-by, shoppers, children. These are the stricter limits.
• Occupational / controlled exposure (FCC) or occupational (ICNIRP). This applies to people exposed as part of their job who are fully aware of the potential for exposure and can exercise control over it, for example tower climbers and broadcast engineers who have had RF safety training. These limits are higher.

In the FCC tables, the occupational power-density limits are five times the general-population limits everywhere above 30 MHz. The averaging time is shorter too, as the next section explains. ICNIRP keeps roughly the same five-to-one relationship between its occupational and general-public whole-body levels.

A rooftop that's off limits to the public, with signs and locked access, can be managed under the occupational limits for trained workers. The pavement below has to meet the general-population limits. That's why RF compliance work often consists of drawing boundaries: where the public may stand, where only trained workers may go, and where nobody goes while the transmitter is on.

Averaging time: why the limits are averages

The FCC limits are time averages, not instantaneous ceilings. For the general population the averaging time is 30 minutes; for occupational exposure it is 6 minutes. Because the protected effect is heating, and body temperature responds slowly, what matters is the energy absorbed over a period rather than the peak at any instant.

Two hypothetical examples show how the averaging works:

1. General population at 1.9 GHz. The limit is 1.0 mW/cm² averaged over 30 minutes. Suppose someone stood in a spot at 3.0 mW/cm² for 10 minutes and then at 0 for the remaining 20 minutes. The average is (3.0 × 10 + 0 × 20) ÷ 30 = 1.0 mW/cm², exactly at the limit.
2. Occupational at 2 GHz. The limit is 5 mW/cm² averaged over 6 minutes. Ten mW/cm² for 3 minutes followed by 3 minutes at zero averages (10 × 3 + 0 × 3) ÷ 6 = 5 mW/cm², again exactly at the limit.

In practice, compliance work for the public doesn't rely on people moving on. Engineers usually assume continuous exposure at the worst spot a member of the public can reach, which is the conservative case. The averaging time matters more for workers whose time near an antenna can be controlled.

ICNIRP 2020 also averages over time. Its whole-body reference levels for the general public use a 30-minute average. It also has separate local-exposure limits with a 6-minute average, aimed at heating of a part of the body rather than the whole body. ICNIRP 1998 used a 6-minute average for most frequencies.

How the limits are expressed: power density, field strength and SAR

The limits come in two layers.

• Basic restrictions (ICNIRP's term) are limits on what happens inside the body, mainly the specific absorption rate (SAR), in watts per kilogram. SAR is the rate at which RF energy is absorbed by tissue. It can't be measured directly on a living person, so it's assessed in models and test phantoms.
• Reference levels (ICNIRP) or maximum permissible exposure (FCC) are the field limits you can measure in the air: electric field strength in volts per metre (V/m), magnetic field strength in amperes per metre (A/m), and power density. They are derived so that meeting them keeps the SAR below the basic restriction.

Above 30 MHz, at a distance from the antenna, power density is the usual headline figure. It is expressed in milliwatts per square centimetre (mW/cm²) in the FCC rules and in watts per square metre (W/m²) in the ICNIRP guidelines. You'll also see microwatts per square centimetre (µW/cm²) on consumer RF meters. The conversions:

• 1 mW/cm² = 10 W/m² = 1,000 µW/cm².
• 1 W/m² = 0.1 mW/cm² = 100 µW/cm².
• 1 µW/cm² = 0.01 W/m² = 10 mW/m².

In the far field, power density and electric field strength are linked by the impedance of free space, about 377 Ω: S = E² ÷ 377, with E in V/m and S in W/m². So 1 V/m corresponds to 2.65 mW/m², or 0.265 µW/cm². Our guide to RF units covers these conversions in depth.

The FCC limits, frequency by frequency

Here are the FCC limits from 47 CFR § 1.1310, Table 1. In the formulas, f is the frequency in MHz and the power density is in mW/cm².

General population / uncontrolled exposure (30-minute average)

* At these low frequencies the power-density figure is a "plane-wave equivalent". Near an antenna, the electric and magnetic field limits are what really apply, because the simple relationship between field and power density doesn't hold close in.

Occupational / controlled exposure (6-minute average)

Below 30 MHz the occupational table continues with its own field-strength and power-density values. If you work at those frequencies, read the rule itself.

Why the limit changes with frequency

The shape of the table follows how strongly the human body absorbs RF energy. A body absorbs most efficiently when its size is comparable to the wavelength, a little like an antenna. For an adult that whole-body resonance lies at the low end of the VHF range, depending on height and whether the person is grounded; smaller bodies resonate at higher frequencies. So the limit is lowest, 0.2 mW/cm², across 30–300 MHz. Above 300 MHz the energy is absorbed less deeply and less efficiently by the whole body, so the limit rises in proportion to frequency until it levels off at 1.5 GHz.

Notice that the formulas join up at the boundaries. At 300 MHz, f/1500 gives 300 ÷ 1500 = 0.2 mW/cm², matching the flat 0.2 below it. At 1,500 MHz it gives 1,500 ÷ 1,500 = 1.0, matching the flat 1.0 above it. At 30 MHz, 180/f² gives 180 ÷ 900 = 0.2, and at 1.34 MHz it gives 180 ÷ 1.7956 ≈ 100. If you calculate a value and it doesn't fit between its neighbours, you've probably used the wrong row or the wrong units.

Worked examples: the FCC formula at five common frequencies

For each frequency: find the row, apply the formula in mW/cm², then convert to W/m² (multiply by 10) and µW/cm² (multiply by 1,000). The equivalent far-field electric field is E = √(377 × S), with S in W/m².

100 MHz (FM broadcast)

100 MHz falls in the 30–300 MHz row, where the limit is flat. General population: 0.2 mW/cm² = 2 W/m² = 200 µW/cm². The equivalent field is √(377 × 2) ≈ 27.5 V/m, which is the value printed in the table. Occupational: 1.0 mW/cm² (10 W/m²).

600 MHz (low-band cellular)

The 600 MHz band is used for mobile service in the US. It falls in the 300–1,500 MHz row. General population: 600 ÷ 1500 = 0.4 mW/cm² = 4 W/m² = 400 µW/cm². Equivalent field: √(377 × 4) ≈ 38.8 V/m. Occupational: 600 ÷ 300 = 2.0 mW/cm² (20 W/m²).

900 MHz (cellular)

Same row. General population: 900 ÷ 1500 = 0.6 mW/cm² = 6 W/m² = 600 µW/cm². Equivalent field: √(377 × 6) ≈ 47.6 V/m. Occupational: 900 ÷ 300 = 3.0 mW/cm² (30 W/m²).

1.9 GHz (PCS cellular)

1,900 MHz is above 1,500 MHz, so the limit is flat. General population: 1.0 mW/cm² = 10 W/m² = 1,000 µW/cm². Equivalent field: √(377 × 10) ≈ 61.4 V/m. Occupational: 5 mW/cm² (50 W/m²).

3.7 GHz (C-band 5G)

Also in the flat top row. General population: 1.0 mW/cm² = 10 W/m²; equivalent field ≈ 61.4 V/m. Occupational: 5 mW/cm². For what's transmitted in this band and others, see 5G frequency bands explained.

Summary

One more, below 30 MHz: at 10 MHz the general-population formula gives 180 ÷ 10² = 1.8 mW/cm² as a plane-wave equivalent. At that frequency you'd normally assess the electric and magnetic field limits (824 ÷ 10 = 82.4 V/m and 2.19 ÷ 10 = 0.219 A/m) rather than power density.

The ICNIRP reference levels, and how they compare

ICNIRP 2020 sets whole-body reference levels for the general public in W/m², averaged over 30 minutes. In the range most surveys cover:

The ICNIRP formula uses a different divisor because it works in W/m², not mW/cm². It also moves the boundaries to 400 MHz and 2 GHz. Again the pieces join up: at 400 MHz, 400 ÷ 200 = 2 W/m², and at 2,000 MHz, 2,000 ÷ 200 = 10 W/m².

The same general-public power densities appear in ICNIRP's 1998 guidelines for this range. The 2020 revision changed other things, among them the whole-body averaging time, the treatment of local exposure, and the approach above 6 GHz, where it uses absorbed power density at the body surface as the basic restriction. If a document quotes "ICNIRP", check which year it means.

Side by side at the five example frequencies

Between 300 MHz and 2 GHz the ICNIRP general-public level is lower than the FCC limit. From 400 MHz to 1.5 GHz it is exactly 75 % of it: the FCC's f/1500 mW/cm² is f/150 W/m², and (f/200) ÷ (f/150) = 0.75. Just either side of that span the ratio lies between 75 % and 100 %. From 30 to 300 MHz and from 2 to 100 GHz the two are equal. Some countries set national limits lower than either. A figure that's "x % of the limit" therefore depends on which limit you mean, so always say which.

Multiple sources: the sum-of-fractions rule

Real places have many transmitters at once: FM and TV, several mobile operators on several bands, Wi-Fi, paging, two-way radio. Each has its own limit, so you can't simply add the power densities and compare the total with one number. Instead you work out each signal's fraction of its own limit and add the fractions. Compliance requires the total to be no more than 1 (100 %).

Total fraction = S₁/L₁ + S₂/L₂ + S₃/L₃ + …

When you have field strengths instead of power densities, use the squares: (E₁/E_L1)² + (E₂/E_L2)² + …, because power density goes with the square of the field.

Worked example (hypothetical levels)

Suppose a calibrated, frequency-selective measurement at one spot found these time-averaged levels:

Against ICNIRP 2020, the same levels in W/m² (0.02, 0.03, 0.05 and 0.01) give fractions of 0.02 ÷ 2, 0.03 ÷ 4.5, 0.05 ÷ 9.5 and 0.01 ÷ 10, which sum to about 0.023 (2.3 %). Notice that the FM signal, the smallest in absolute terms after C-band, contributes the largest fraction, because its limit is the lowest.

The FCC also has a rule about responsibility. Where several transmitters contribute to a spot that exceeds the limit, each licensee whose contribution is more than a small percentage of its own limit shares responsibility for fixing it. For the exact threshold, read 47 CFR § 1.1307 rather than relying on a summary.

Near field and far field

Close to an antenna, the electric and magnetic fields don't yet have the fixed relationship of a travelling wave. This matters because the simple formulas, S = E² ÷ 377 and the inverse-square law, only hold in the far field.

• Reactive near field. Very close to the antenna, roughly within λ ÷ 2π of it, energy is stored and exchanged as well as radiated. The E and H fields must be measured separately. At 100 MHz (λ ≈ 3.0 m) that's about 0.48 m; at 1.9 GHz (λ ≈ 15.8 cm) about 2.5 cm.
• Radiating near field (Fresnel region). Further out, the wave is radiating but the beam hasn't fully formed. Power density doesn't fall off as 1/r², and the pattern varies with distance.
• Far field (Fraunhofer region). For an antenna whose largest dimension is D, the far field is usually taken to start at about 2D² ÷ λ. Here the beam shape is stable and power density falls off as 1/r².

Worked example: where the far field starts

Take a panel antenna 1.3 m tall, a size you might see on a rooftop cell site (illustrative only):

• At 900 MHz, λ = 0.333 m, so 2D² ÷ λ = 2 × 1.3² ÷ 0.333 ≈ 10.1 m.
• At 1.9 GHz, λ = 0.158 m, so 2D² ÷ λ = 2 × 1.69 ÷ 0.158 ≈ 21.4 m.

So for the same antenna size, the far field starts further away at higher frequencies. Inside that distance, a calculation using the far-field formula isn't reliable. OET Bulletin 65 gives separate methods for the near field of aperture and other antennas. Measurements there need probes designed for it, and at lower frequencies separate E-field and H-field probes.

Predicting power density from a transmitter

Before anyone measures, engineers usually predict. In the far field and in the main beam, power density from a transmitter with effective isotropic radiated power (EIRP) at distance R is:

S = EIRP ÷ (4πR²) (S in W/m², EIRP in W, R in m)

OET Bulletin 65 also gives a more conservative version for ground-level estimates that allows for a reflected wave adding in phase with the direct one. It multiplies the result by 2.56 (a field reflection factor of 1.6, squared), giving S = 0.64 × EIRP ÷ (πR²).

Worked example (hypothetical transmitter)

Take a 900 MHz transmitter with an EIRP of 1,000 W in its main beam.

1. At 100 m: S = 1,000 ÷ (4π × 100²) = 1,000 ÷ 125,664 ≈ 0.0080 W/m² = 0.80 µW/cm². The FCC limit at 900 MHz is 600 µW/cm², so this is about 0.13 %.
2. With the 2.56 reflection factor: 0.80 × 2.56 ≈ 2.04 µW/cm², about 0.34 %.
3. At 30 m, with reflection: 2.56 × 1,000 ÷ (4π × 30²) ≈ 0.226 W/m² = 22.6 µW/cm², about 3.8 %.
4. At 10 m, with reflection: 2.56 × 1,000 ÷ (4π × 10²) ≈ 2.04 W/m² = 204 µW/cm², about 34 %. That's right at the edge of the far field for the 1.3 m panel above, so a careful evaluation would use near-field methods or measurement here.

Two lessons follow. First, the inverse-square law is steep: going from 100 m to 10 m multiplies the power density by 100. Second, the formula assumes you're in the main beam. Cell-site panels are usually aimed towards the horizon with a slight downtilt. Someone directly below, or behind the panel, may be well outside the main beam. The hypothetical high fractions at short range are why the areas right in front of rooftop antennas are often restricted, while the street is usually evaluated as a far-field case.

SAR: the limits for phones and other body-worn devices

For a device used close to the body, field limits in the air aren't a useful test, because you're in its near field. Instead the limit is on the specific absorption rate, the RF power absorbed per kilogram of tissue.

The European Union's recommendation for the general public uses the same 2 W/kg over 10 g local limit as ICNIRP, as do many countries that follow ICNIRP. The FCC's 1.6 W/kg over 1 g and the 2 W/kg over 10 g aren't directly comparable. Averaging over a smaller mass captures more of the hotspot, so the same phone generally gives a higher figure over 1 g than over 10 g.

How phone SAR is tested

Under the FCC rules, a device designed to be used within 20 cm of the body is a "portable" device and is assessed by SAR (47 CFR § 2.1093). One used further away is a "mobile" device and is assessed against the MPE limits (47 CFR § 2.1091). Phones are tested in standard head and body-worn positions, in a phantom filled with tissue-simulating liquid, transmitting at their highest certified power on each band. The reported SAR is the highest value found across those tests.

The FCC points out that a phone's SAR figure is a compliance test result at maximum power, not a measure of the exposure in everyday use, and that comparing SAR values between models doesn't tell you which one exposes you less in practice. Phones adjust their power to what the network needs. To look up a model, find its FCC ID (usually in the settings or on the label) and search it on the FCC's equipment authorisation database; the SAR test report is among the published documents.

A spectrum survey can't measure SAR. SAR concerns absorption inside the body very close to the device, which only lab testing can assess.

How compliance is actually evaluated

When an engineer evaluates a site, the process generally looks like this:

1. Gather the transmitter data. Frequencies, transmitter powers, antenna models and patterns, heights, azimuths and tilts, for every operator at the site and for nearby sites that contribute.
2. Predict. Use the OET Bulletin 65 formulas, or modelling software, to estimate power density at places people can reach, at the transmitters' maximum power. If every predicted spot is far below the limits, that may be enough.
3. Measure where needed. Use a broadband meter with a calibrated isotropic probe, which responds equally to fields from all directions and polarisations, or a calibrated frequency-selective instrument when it matters which signal is responsible. The equipment has a current calibration certificate, and the engineer records its stated uncertainty.
4. Average in space and time. The fields near the ground vary from spot to spot because of reflections. Guidance describes spatial averaging over the area a body would occupy, rather than taking a single point. Results are compared on the appropriate time-average.
5. Scale to worst case. A network's output varies with traffic. Frequency-selective measurements can be scaled up to the transmitters' maximum power, so the result covers the busiest moment, not just the moment of measurement.
6. Sum the fractions across all signals and compare with 100 %.
7. Mitigate and document. If any area could exceed the limits, the remedy may be signs, barriers, locked access, antenna changes or power reductions. The result is a written report.

The people who do this are RF safety engineers and consultants with the training and equipment for it. Their reports are what licensees, landlords and regulators rely on.

Why a receiver reading isn't an exposure measurement

A spectrum analyzer measures the power arriving at its input connector. An exposure limit is about the field in the space a person occupies, averaged over time and over the body. Getting from one to the other takes a chain of steps, and with a consumer software radio such as the HackRF One, every link in that chain is either missing or uncertain.

1. Absolute level. The HackRF reports levels relative to its own analog-to-digital converter's full scale (dBFS), not in dBm. The relationship between the two depends on frequency, on every gain setting, and on the individual unit. Without calibration against a known source, you don't know the absolute power at all.
2. Antenna gain. To go from power to field strength you need the antenna's gain, or its antenna factor, at that exact frequency, from a calibration sheet. A telescopic whip or a small wideband antenna has a gain that varies by many dB across 1 MHz–6 GHz, and nobody has measured it for your particular antenna.
3. Direction and polarisation. An exposure probe is isotropic: it sums the field from every direction and in all three polarisations. A single antenna picks up some directions and polarisations well and others poorly. A signal arriving cross-polarised can read many dB low.
4. Your body and surroundings. Holding the phone and radio near your body changes what the antenna receives. Reflections make the field vary over distances of a fraction of a wavelength, which is why standards average over space rather than trusting one point.
5. Bandwidth. A spectrum analyzer shows power per resolution bandwidth (RBW) bin, not the total power of a wide signal. A 5 MHz-wide signal measured in 100 kHz bins reads about 17 dB (10 × log₁₀ 50) lower per bin than its total channel power. The units guide explains how to integrate across a channel.
6. Time. Limits are 30-minute or 6-minute averages, scaled to maximum transmitter power. A walking survey captures moments. A sweep that catches a burst sees the peak; one that misses it sees nothing.
7. Summing. Compliance is the sum of fractions across every signal present. A survey sees the signals in the range and bandwidth it sweeps; Consentiva's Full plan reaches 6 GHz and the Free and Standard plans 2.7 GHz.
8. Overload. A receiver driven near its maximum input can create false signals (intermodulation) and misread true ones, and too much input power can damage it. Close to a strong transmitter, a sensitive front end with its gain turned up can be pushed into error long before the damage limit is reached.

Each of these can shift the result by several dB, and together they make a percentage-of-limit figure from an uncalibrated SDR survey meaningless. Our comparison of RF meters and spectrum analyzers goes further into which instrument answers which question.

A reading that looks strong on a spectrum analyzer is usually a tiny fraction of an exposure limit. For example, −60 dBm received at 900 MHz on an ideal isotropic antenna corresponds to about 0.000011 µW/cm². That's around two millionths of one per cent of the FCC general-population limit at that frequency. A strong-looking trace isn't, by itself, a sign of high exposure, and a weak one isn't proof of low exposure either. Only a proper evaluation tells you that.

Converting a calibrated reading to a fraction of the limit

If you do have calibrated equipment, the arithmetic is straightforward. Here are two worked routes for a 1.9 GHz signal. The numbers are examples, not typical values.

Route 1: a calibrated isotropic field probe

A calibrated broadband meter with an isotropic probe reads 2.5 V/m, and you know the field is at 1.9 GHz.

1. Power density: S = E² ÷ 377 = 6.25 ÷ 377 ≈ 0.0166 W/m².
2. In µW/cm²: 0.0166 × 100 ≈ 1.66 µW/cm².
3. FCC general-population limit at 1.9 GHz: 1,000 µW/cm². Fraction: 1.66 ÷ 1,000 ≈ 0.17 %.
4. ICNIRP 2020 general public at 1.9 GHz: 9.5 W/m². Fraction: 0.0166 ÷ 9.5 ≈ 0.17 %.

Route 2: a calibrated receiver and a calibrated antenna

A calibrated receiver reads −20 dBm at 1.9 GHz in a measurement bandwidth that covers the whole channel. The antenna's gain at 1.9 GHz is 0 dBi, and for now we'll ignore cable loss.

1. Power: −20 dBm = 10⁻² mW = 10⁻⁵ W.
2. Wavelength: λ = 3×10⁸ ÷ 1.9×10⁹ ≈ 0.158 m.
3. Effective area of a 0 dBi antenna: A = λ² ÷ (4π) ≈ 0.158² ÷ 12.57 ≈ 0.00198 m².
4. Power density: S = P ÷ A = 10⁻⁵ ÷ 0.00198 ≈ 0.00505 W/m² = 0.505 µW/cm².
5. Fraction of the FCC limit: 0.505 ÷ 1,000 ≈ 0.050 %.
6. Field strength, as a check: E = √(377 × 0.00505) ≈ 1.38 V/m.

The antenna-factor route gives the same answer. In a 50 Ω system, −20 dBm is −20 + 107 = 87 dBµV. The antenna factor of a 0 dBi antenna at 1,900 MHz is 20 × log₁₀(1900) − 29.77 ≈ 35.8 dB/m. So E = 87 + 35.8 = 122.8 dBµV/m, which is 10^{(122.8 − 120)/20} ≈ 1.38 V/m.

What happens when the inputs are wrong

• Cable loss. If 2 dB is lost in the cable, the power at the antenna was −18 dBm, not −20. Every figure scales by 10^0.2 ≈ 1.58, so S becomes about 0.80 µW/cm².
• Unknown antenna gain. If the real gain could be anywhere within ±5 dB of the assumed 0 dBi, the power density could be anywhere from about 0.16 to 1.6 µW/cm², a factor of ten between the extremes. That's typical of what an uncalibrated antenna does to a result.
• Several signals. Repeat the calculation for each signal and add the fractions, as in the sum-of-fractions example above.
• Peak instead of average. If the reading was a peak hold on a bursty signal, the time average could be much lower. Exposure limits use the average.

What Consentiva measures, and what it doesn't

Consentiva is a walking spectrum survey. A HackRF One connected to an Android phone sweeps the spectrum while you walk, and each sweep is recorded with its GPS position. Over repeated walks, Consentiva learns a baseline for each block, about 150 m, separately for night, morning, afternoon and evening. It then flags signals that are stronger than that block's normal as Unusual.

• Units. Levels are in uncalibrated dBFS, unless you enter a calibration offset in the settings, in which case they're shown in dBm. Even in dBm, a reading is power at the receiver input, not field strength or power density.
• Comparisons are relative. An Unusual flag means "this many dB above what's normal here at this time of day". Relative comparisons stay meaningful if you keep the gain settings identical between walks, because the unknown calibration then cancels out.
• No exposure claims. Consentiva doesn't measure exposure and doesn't compare readings with the FCC or ICNIRP limits. An Unusual flag isn't evidence of any limit being exceeded.

What a survey is good at is where and when: which blocks and which hours have the strongest signals, which frequencies they're on, and whether something new has appeared. Its reports and tamper-evident evidence packages record that history in a form you can share. See how to read a Consentiva report or open the sample report.

Scenario walk-throughs

A cell antenna on the building across the street

You notice new panels on a rooftop opposite your window and want to know if they meet the limits.

1. Identify what's there. Our guide to finding cell towers near you covers the FCC's public databases and what you can see from the street.
2. A survey before and after the installation will show new signals and which bands they're on. That tells you something changed. It doesn't tell you whether the limits are met.
3. Ask the site owner or the operator for the RF compliance evaluation for the site. For a window across the street, the distance and the direction of the main beam are the key inputs to that evaluation.
4. If you still have concerns, the FCC is the regulator for compliance with its rules, and an independent RF engineer can measure at your window.

A Wi-Fi router in a bedroom

Under the FCC's Part 15 rules, a 2.4 GHz Wi-Fi transmitter using digital modulation is limited to 1 W conducted output with an antenna gain of up to 6 dBi, which is 4 W EIRP. As an upper bound, put 4 W EIRP into the far-field formula at 1 m: S = 4 ÷ (4π × 1²) ≈ 0.32 W/m² = 32 µW/cm². That's about 3.2 % of the FCC general-population limit. At 3 m it's about 3.5 µW/cm², or 0.35 %. This is a worst case: many devices run below the legal maximum, and Wi-Fi transmits in bursts, so the time average is lower still.

A survey will show the router's channel lighting up whenever it's busy. That's a useful way to identify which device is which, but the brightness of the trace doesn't translate into a percentage of a limit.

A phone

For a phone held against your head, the relevant limit is SAR, not power density. Look up the model's SAR test report by its FCC ID. A survey can show when your phone transmits and on which band, but not how much energy your body absorbs.

A workplace near transmitters

If your job takes you onto rooftops, towers or into transmitter rooms, the occupational limits and your employer's RF safety programme apply. Look for signage, ask for the site's RF safety information and training, and follow the access rules. In the US, the Occupational Safety and Health Administration (OSHA) also has a role in workplace safety. A survey app isn't a personal RF safety monitor.

Something new and strong appears on your street

If Consentiva flags an Unusual signal that persists, the first question is what it is, not whether it's dangerous. It could be a new cell carrier, a construction crew's radios, a neighbour's new equipment, or interference. Our guides to surveying radio signals and to documenting interference for an FCC complaint cover the next steps.

What to do if you're concerned

If you want to know whether a particular place meets the limits, these are the steps that lead to a real answer:

1. Identify the source. Work out which transmitters are nearby and who operates them. A survey helps here: it shows which frequencies are strongest, where, and at what times. The FCC's licensing and antenna structure databases, on fcc.gov, can tell you who holds a licence or registered a tower.
2. Ask for the compliance information. Licensees are responsible for complying with the FCC limits. Ask the operator, or the owner of the building or land where the antenna sits, for the RF exposure evaluation for the site and for any access restrictions in place.
3. Ask the regulator. In the US, the FCC handles questions about compliance with its RF exposure rules. Elsewhere, ask the national radio or health regulator which limits apply and how they're enforced.
4. Hire a qualified RF engineer if you need a figure you can rely on, for a lease, a dispute or a complaint. Ask for calibrated, isotropic measurement equipment with a current calibration certificate, a statement of measurement uncertainty, the limit used (FCC or ICNIRP, general or occupational), spatial and time averaging as the guidance describes, and results as a fraction of the limit for each signal and in total.
5. Use your survey to target the work. A Consentiva survey can't replace that measurement, but it can make it cheaper and more focused. It shows the engineer where and when the strongest signals were, and on which bands.

Common mistakes

• Treating dBFS or dBm as exposure. A receiver reading is power at a connector, not a power density. Without calibration and an antenna factor, it can't be turned into a fraction of a limit.
• Mixing up the units. mW/cm² and W/m² differ by a factor of ten. Putting a W/m² measurement against a mW/cm² limit makes a reading look ten times smaller (or larger) than it is.
• Using the wrong formula. The FCC's f/1500 gives mW/cm²; ICNIRP's f/200 gives W/m². Each only applies within its own frequency range.
• Comparing a peak with an average limit. Limits are time averages. A peak-hold trace overstates a bursty signal.
• Comparing one point with a spatially averaged limit. The field varies over short distances. One spot can be well above or below the area average.
• Measuring one bin of a wide signal. The power in one RBW bin is only part of the channel's power.
• Forgetting the other signals. Compliance is the sum of fractions across every signal present, including ones outside the range you measured.
• Using far-field formulas in the near field. Close to an antenna, S = E² ÷ 377 and the inverse-square law don't hold.
• Quoting "the limit" without saying whose. FCC and ICNIRP differ between 300 MHz and 2 GHz, and general and occupational limits differ by a factor of five.
• Treating an Unusual flag as a safety finding. It means stronger than normal for that block and time, nothing more.

Glossary

• Antenna factor (AF). The ratio, in dB/m, that converts the voltage at an antenna's output into the field strength around it. It comes from the antenna's calibration.
• Averaging time. The period over which exposure is averaged for comparison with a limit: 30 minutes for the FCC general population, 6 minutes for occupational exposure.
• Basic restriction. ICNIRP's term for a limit on a quantity inside the body, such as SAR.
• dBFS. Decibels relative to the full scale of a receiver's analog-to-digital converter. Not an absolute unit.
• EIRP. Effective isotropic radiated power: the power an ideal isotropic antenna would need to produce the same field as the real antenna in its main beam.
• Far field. The region beyond about 2D² ÷ λ from an antenna, where power density falls with the square of distance and S = E² ÷ 377.
• General population / uncontrolled. The FCC tier for people who may not know they're exposed or can't control it.
• Isotropic probe. A field sensor that responds equally to fields from all directions and polarisations.
• MPE. Maximum permissible exposure: the FCC's field and power-density limits in 47 CFR § 1.1310.
• Near field. The region close to an antenna where the electric and magnetic fields must be assessed separately.
• OET Bulletin 65. The FCC's technical guidance on evaluating compliance with its RF exposure limits.
• Occupational / controlled. The FCC tier for workers who are aware of the potential for exposure and can control it.
• Power density (S). RF power flowing through a unit area, in mW/cm², W/m² or µW/cm².
• Reference level. ICNIRP's term for a measurable field limit derived from the basic restrictions.
• SAR. Specific absorption rate: RF power absorbed per kilogram of tissue, in W/kg.
• Sum of fractions. Adding each signal's fraction of its own limit; the total must not exceed 1.

Sources: FCC, 47 CFR § 1.1310 (radio-frequency exposure limits), § 1.1307, § 2.1091 and § 2.1093; FCC Office of Engineering and Technology, OET Bulletin 65; ICNIRP, "Guidelines for limiting exposure to time-varying electric, magnetic and electromagnetic fields (up to 300 GHz)", Health Physics 74(4), 1998; ICNIRP, "Guidelines for limiting exposure to electromagnetic fields (100 kHz to 300 GHz)", Health Physics 118(5), 2020; IEEE C95.1-2019. This guide is general information, not legal, medical or safety advice. Always check the current text of the rules that apply to you.

Common questions

What is the FCC RF exposure limit for the general public?

It depends on frequency. Under 47 CFR 1.1310 it is 0.2 mW/cm² from 30 to 300 MHz, f/1500 mW/cm² from 300 to 1,500 MHz (f in MHz), and 1.0 mW/cm² from 1.5 to 100 GHz, averaged over 30 minutes.

What is the FCC limit at 900 MHz?

For the general population it is 900 ÷ 1500 = 0.6 mW/cm², which is 6 W/m² or 600 µW/cm². The occupational limit is 900 ÷ 300 = 3.0 mW/cm².

How do the FCC and ICNIRP limits differ?

They are equal from 30 to 300 MHz and from 2 to 100 GHz. Between 300 MHz and 2 GHz the ICNIRP 2020 general-public level is lower, for example 4.5 W/m² against the FCC's 6 W/m² at 900 MHz.

How do you combine exposure from several transmitters?

Work out each signal's measured level as a fraction of its own limit at its frequency, then add the fractions. The total must not exceed 1, or 100 per cent.

What is the SAR limit for mobile phones?

The FCC limit is 1.6 W/kg averaged over any 1 gram of tissue. ICNIRP and the EU use 2 W/kg averaged over 10 grams, so the two figures are not directly comparable.

Can a spectrum analyzer app tell me if I'm above the exposure limit?

No. An uncalibrated receiver measures power at its input in dBFS, and turning that into power density needs calibration, a calibrated antenna factor, isotropic measurement and time and spatial averaging. Consentiva compares signals with each block's normal, not with exposure limits.

Who should I contact if I'm worried about a nearby antenna?

Ask the operator or the property owner for the site's RF compliance evaluation, and contact the FCC with questions about compliance with its rules. For a figure you can rely on, hire a qualified RF engineer with calibrated equipment.