What is the difference between EU and US SAR limits?

The European Union, based on Council Recommendation 1999/519/EC and the ICNIRP guidelines, sets localised head and trunk SAR at 2 W/kg averaged over 10 g of tissue for 6 minutes. The FCC, in 47 CFR section 2.1093, adopts 1.6 W/kg averaged over 1 g. The US limit is numerically lower and the averaging mass smaller, which makes it, for the same physical phenomenon, more demanding than a glance at the numbers suggests. A product sized for 2 W/kg is never automatically compliant with 1.6 W/kg, and the conversion must be verified for each configuration.

When should a product be evaluated for SAR rather than MPE or power density?

SAR applies to equipment used within 20 cm of the body (phones, smart watches, hearables, body-worn medical devices, smart glasses) across 100 kHz to 6 GHz. Above roughly 6 GHz and for the 5G FR2 mmWave bands (24 to 100 GHz), penetration of the field into tissue becomes negligible and the evaluation quantity switches to surface power density. The FCC formalises this split in KDB 941225 (mmWave) and 447498. ICNIRP 2020 and IEEE C95.1-2019 in parallel introduce time-averaged power density as the reference metric above 6 GHz.

Why does IEC 62209 have several parts?

The series was split to track increasing product complexity. IEC 62209-1 covers handheld phones used near the head in the 300 MHz to 6 GHz band. IEC 62209-2 covers body-worn equipment. IEC 62209-3 introduces methods for multi-source and multi-band simultaneous transmitters, the typical scenario for 5G smartphones combining cellular, Wi-Fi and BLE. IEC 62209-4 defines derived methods to reduce the test scope on product variants already qualified, relying on experimentally validated numerical simulation.

What is the SAM phantom and why is it used?

The SAM (Specific Anthropomorphic Mannequin) is a low-loss dielectric shell shaped on the adult male 90th percentile per the US military report AEL-TR-91-029. Filled with a tissue-equivalent liquid whose permittivity and conductivity mimic the electromagnetic properties of the human head at the test frequency, it provides a reproducible reference between labs. For body tests, a flat phantom filled with the same kind of liquid is used, the product being placed against the flat face at the separation distance declared by the manufacturer.

What is the typical duration of a multi-band SAR campaign?

A complete campaign in an ISO/IEC 17025 accredited lab recognised by the FCC typically takes 2 to 4 weeks for a multi-band product (cellular LTE plus 5G NR sub-6 GHz, Wi-Fi 2.4 GHz and 5 GHz, BLE). The main multiplier is the number of configurations to evaluate, each band / channel / position combination potentially requiring a full robotic scan of 30 minutes to several hours. Cases with multiple antennas, active MIMO or beamforming and simultaneous transmission scenarios stretch the campaign beyond 4 weeks.

How is aggregate SAR handled under simultaneous transmission?

The FCC documents the approach in KDB 447498 and several complementary publications. The principle is that a product transmitting simultaneously over multiple technologies (cellular plus Wi-Fi plus BLE) must demonstrate that the sum of the SAR contributions measured independently does not exceed the regulatory limit, or alternatively provide a composite SAR measurement under simultaneous transmission. The direct sum is conservative; the composite measurement, more realistic, requires a specific protocol and a lab equipped to drive the radios in parallel. The RED does not explicitly address aggregation but the report must cover it via IEC 62209-3.

What separation distance should be declared for a body-worn product?

Common declared values on phones and wearables range from 0 mm (skin contact) to 15 mm. The choice should reflect the realistic use of the product and be recorded in the user notice. Under-declaring distance to ease test passage leads to non-compliance as soon as marketing copy or packaging suggests closer use. Over-declaring removes the test risk but forces a user warning that can hurt marketing. Reference practice is 5 mm for a smartphone, 0 mm for a watch or medical patch, and the actual use distance for headsets and earbuds.

Is Hearing Aid Compatibility (HAC) linked to SAR?

No. HAC (Hearing Aid Compatibility) addresses a distinct regulatory requirement (47 CFR 20.19 in the United States, equivalent national programs elsewhere) and characterises electromagnetic compatibility with M3/M4 and T3/T4 hearing aids. It is often tested in parallel with SAR because fixtures and instrumentation partly overlap, but the criteria, phantoms and application bands differ. A product can be SAR compliant and HAC non-compliant, and vice versa.

SAR procedures: absorption rate (IEC 62209, EN 50360)

Author Hugues Orgitello Last updated 27 May 2026 ~14 min read

Guide · RF exposure and SAR

Specific Absorption Rate, or SAR, is the reference quantity to characterise the radiofrequency power absorbed by biological tissue when a transmitter is used in contact with the body or in its immediate proximity. All phones, smart watches, wireless earbuds, smart glasses, audio headsets and body-worn medical devices are subject to it. The measurement method is horizontal, defined by the IEC 62209 series, and taken up by the European Union through EN 50360 and EN 50566, and by the FCC through OET Bulletin 65. This page lays out the regulatory limits on both sides of the Atlantic, the measurement standards, the test hardware (SAM phantom, robotic platforms), the switch to power density for mmWave bands, and the most frequent pitfalls in dossier preparation.

What SAR measures

SAR (Specific Absorption Rate) is defined as the electromagnetic power absorbed per unit mass of tissue, expressed in watts per kilogram. The mathematical definition is SAR = sigma * E^2 / rho, where sigma is the electric conductivity of tissue in siemens per metre, E the modulus of the internal electric field in volts per metre, and rho the tissue mass density in kilograms per cubic metre.

The quantity is local: it varies point by point in the biological volume. Regulations therefore impose two types of averaging before comparison to a limit.

Spatial averaging. SAR is integrated over a cubic tissue volume of mass 1 g or 10 g depending on the regime. The FCC uses 1 g, the European Union 10 g.
Time averaging. Integration is over a 6 minute window for controlled exposure and 30 minutes for the general public, per the ICNIRP guidelines.

This double averaging explains why the same product can show very different SAR values across reports: a raw, instantaneous and local SAR value has no regulatory meaning. Only the value averaged per the protocol of the standard cited in the dossier counts.

Why averaging mass changes everything

The big US-EU split is not just in the numeric value (1.6 vs 2.0 W/kg) but in the mass over which the average is taken. Averaging over 1 g (FCC) captures a stronger local concentration than averaging over 10 g (EU). At the same peak SAR, the value averaged over 1 g is structurally higher than the value averaged over 10 g. Practical outcome: a product that comfortably clears 2 W/kg over 10 g in the EU is not guaranteed to pass 1.6 W/kg over 1 g in the United States, and the analysis must be redone for each configuration.

Two distinct regulatory regimes

SAR is framed by two legal architectures that share the same scientific foundations (ICNIRP guidelines and IEEE C95.1) but with different regulatory transpositions.

EU side: RED Article 3.1(a)

For radio equipment placed on the European market, the health protection requirement rests on Article 3.1(a) of the RED Directive 2014/53/EU, which demands protection of human health and safety. The operational translation goes through:

Council Recommendation 1999/519/EC, which fixes the numerical exposure limits for the general public, aligned with ICNIRP 1998 and then ICNIRP 2020.
The harmonised standards EN 50360 for handheld phones used near the head, EN 50566 for other body-worn equipment, and the EN 62209 series (CENELEC transposition of IEC 62209) for the measurement procedure.

See RED pillar for the directive framework and RED tests for the detail of health tests.

US side: 47 CFR plus OET Bulletin 65

The FCC handles SAR in several places:

47 CFR 1.1307 identifies the categories of transmitters subject to RF exposure evaluation.
47 CFR 1.1310 fixes the numerical limits (Maximum Permissible Exposure and SAR), and 47 CFR 2.1093 details the procedure for portable devices.
OET Bulletin 65 (Federal Communications Commission, Office of Engineering and Technology) and its supplements A, B and C provide the technical application guidance.
KDB publications (Knowledge Database) clarify special cases: KDB 248227 (general SAR procedure), KDB 941225 (mmWave), KDB 447498 (modules), KDB 690783 (5G mid-band).

See FCC pillar and FCC tests for the US framework.

Limits compared

Quantity	European Union (1999/519/EC, ICNIRP)	United States (47 CFR 1.1310)
Localised head/trunk SAR	2.0 W/kg averaged over 10 g (6 min)	1.6 W/kg averaged over 1 g (30 min)
Localised limb SAR	4.0 W/kg averaged over 10 g (6 min)	4.0 W/kg averaged over 10 g (30 min)
Whole-body SAR	0.08 W/kg averaged over total mass	0.08 W/kg averaged over total mass
Frequency range	100 kHz to 6 GHz (SAR)	100 kHz to 6 GHz (SAR)
Above 6 GHz	Power density, ICNIRP 2020	Power density, KDB 941225

Reading: while whole-body and limb limits converge, the critical difference lies in localised head/trunk SAR. In the United States, the 1 g averaging mass yields by construction higher values at the same incident field, and the numerical limit is lower. The margin ratio between the two regimes depends on the spatial distribution of the field and cannot be predicted by a simple scalar factor.

The IEC 62209 series

IEC 62209 is the cornerstone of SAR metrology. Published by IEC under TC 106 (electromagnetic fields in the human environment), it is transposed in Europe as EN 62209 by CENELEC. It provides the practical procedure: calibration protocol, phantom geometry, robotic control, internal field post-processing.

The four parts of the series

Part	Scope	Frequency range	Typical use case
IEC 62209-1	Handheld phones used near the head	300 MHz to 6 GHz	Smartphone held to the ear, DECT phone
IEC 62209-2	Body-worn devices	30 MHz to 6 GHz	Watch, medical sensor, hotspot, glasses
IEC 62209-3	Multi-source and multi-band methods	30 MHz to 6 GHz	5G smartphone with simultaneous cellular + Wi-Fi + BLE
IEC 62209-4	Reduced-scope test methods	30 MHz to 6 GHz	Product variants, validated numerical simulation

IEC 62209-1 covers the historical case of the phone held to the ear with the SAM phantom. IEC 62209-2 covers the body-worn case with a flat phantom. Part 3, published in 2019, formalises methods for products transmitting simultaneously over multiple technologies: this is the default situation on any modern smartphone. IEC 62209-4, more recent, proposes methods to reduce the measurement scope when a numerical simulation (typically FDTD) experimentally validated on a reference configuration is available.

Coverage by European standards

The harmonised EN standards reproduce the IEC content, sometimes with a publication delay or RED-specific annexes. The current correspondence is as follows.

EN standard	Equivalent IEC reference	OJEU RED status
EN 50360	Phone-specific methods, older	Listed, undergoing evolution
EN 50566	Body-worn specific methods	Listed
EN 62209-1	IEC 62209-1	Listed
EN 62209-2	IEC 62209-2	Listed
EN 62209-3	IEC 62209-3	Being integrated

See RED standards for the application rules of harmonised standards and their life cycle.

Test hardware: phantom, liquid, robot

A SAR campaign runs in a very specific environment, not easily transferable from a standard EMC lab.

The SAM phantom and the flat phantom

The SAM (Specific Anthropomorphic Mannequin) is defined in IEEE 1528 and referenced by IEC 62209-1. It is a low-loss dielectric plastic shell reproducing the head geometry of the adult male 90th percentile, with two factory-set ear positions. The SAM is filled with a tissue-equivalent liquid whose relative permittivity and conductivity are adjusted at the test frequency to mimic the average electromagnetic properties of the human brain. The liquid composition changes at each band: a typical water / sugar / salt / surfactant blend for 900 MHz is not suitable for 5 GHz.

For body tests, the SAM is replaced by a flat phantom: a flat-bottomed plastic parallelepiped filled with the same kind of liquid. The product is placed against the flat face at the separation distance declared by the manufacturer.

Dosimetric probe and robot

The internal electric field measurement uses a miniature dosimetric probe, typically a three-axis sensor with individually calibrated Schottky diodes. The probe is carried by a six-axis robotic arm that scans a predefined volume below the phantom skin. The de facto market standard is the DASY platform from SPEAG (Schmid & Partner Engineering AG, Zurich), present in most accredited SAR laboratories worldwide. Alternatives exist (cSAR3D also from SPEAG, Schmid systems, IndexSAR), but DASY still dominates the methodological literature.

Post-processing

The robotic scan produces a 3D map of the field modulus. The SAR software (DASY6 or equivalent) numerically integrates sigma * E^2 / rho over the required volume (1 g or 10 g), with the cubic averaging algorithm defined by IEEE C95.3-2002. The retained value is the maximum over the whole scanned volume for the configuration tested. Each configuration generates a map, an averaged value, and an archived calibration trace.

mmWave and the switch to power density

Above roughly 6 GHz, the penetration depth of the electromagnetic field into tissue drops sharply. At 28 GHz typically, the skin depth in human skin falls around 0.5 mm, and the absorbed energy concentrates in a superficial layer where the notion of SAR averaged over 1 g or 10 g loses physical meaning: integration would be over a volume mostly empty of absorption. Standards bodies therefore switched to a replacement quantity, the surface power density expressed in W/m^2.

ICNIRP 2020 and IEEE C95.1-2019

The ICNIRP 2020 guidelines and the IEEE C95.1-2019 update introduce time-averaged power density (TPD) as the reference metric above 6 GHz. Averaging is over a 6 minute window (controlled exposure) or 30 minutes (general public) and over an elementary surface of 4 cm^2 or 1 cm^2 depending on frequency.

Typical ICNIRP 2020 limits for the general public above 6 GHz are around 10 W/m^2 over 4 cm^2, with step relaxations up to 200 W/m^2 by frequency and scenario.

FCC and 5G FR2 mmWave bands

The FCC documents the mmWave approach in KDB 941225, which applies notably to the 24 GHz, 28 GHz, 37 GHz and 39 GHz bands used by 5G FR2 and certain backhaul links. The procedure combines:

A power density measurement with a directional probe.
An antenna pattern characterisation for beam-steering antennas (beamforming).
A statistical exposure evaluation accounting for beam activation duration.

IEC 62232 provides the IEC normative framework for these measurements, complemented by ITU-T K-series guides for base stations.

Summary table: SAR vs power density

Frequency range	Quantity	Main standard	Method	Typical general-public limit
100 kHz to 10 MHz	Induced current density + SAR	ICNIRP 2020	Induced current measurement	Frequency-dependent
10 MHz to 6 GHz	Localised + whole-body SAR	IEC 62209-1/-2/-3	Phantom + robotic probe	2 W/kg / 10 g (EU), 1.6 W/kg / 1 g (US)
6 GHz to 300 GHz	Power density	IEC 62232, FCC KDB 941225	Directional probe, beam scan	10 W/m^2 over 4 cm^2 (ICNIRP 2020)

Test conditions and product configuration

The test procedure imposes a worst-case product configuration: this is the main challenge in dossier preparation.

Full power, multi-channel, multi-position

The product must be evaluated with:

Maximum power on each band, typically driven through a test firmware or chipset-specific AT command.
All radios active in credible use configurations: cellular, Wi-Fi 2.4 GHz, Wi-Fi 5 GHz, Wi-Fi 6 GHz, BLE simultaneously where relevant.
Several channels per band: typically low, mid, high channels.
Several positions against the phantom: head left, head right, cheek contact, ear contact (the IEEE 1528 cheek and tilt), body front, body back, several orientations.

For a multi-band smartphone, the number of configurations to test can exceed 100. Numerical simulation (experimentally validated FDTD) and reduced test scope (IEC 62209-4) techniques are then used to target worst cases and limit the robotic scan.

Liquid temperature

The tissue-equivalent liquid must be temperature-stabilised (typically 20 to 22 degrees Celsius), as its permittivity and conductivity vary appreciably with temperature. A drift of more than 2 degrees during the session invalidates the measurements. The lab calibrates the liquid at the start of each day and verifies it with a reference probe.

Separation distance

The distance between product and phantom face is a parameter declared by the manufacturer and recorded in the user notice. Common values range from 0 mm (direct contact) for a watch or medical patch, 5 mm for a smartphone, 10 to 15 mm for a belt-worn device. Under-declaring this distance to ease test passage is a classic pitfall: the notice must be consistent, and any marketing copy suggesting closer use creates a non-compliance risk.

Simultaneous transmission scenarios

All smartphones and most modern wearables transmit simultaneously over several technologies: cellular + Wi-Fi + BLE, or less frequently cellular + cellular (Dual Connectivity 4G+5G).

Conservative approach: contribution summation

The simplest method measures the SAR of each radio individually and sums the contributions:

SAR_total = SAR_cellular + SAR_WiFi + SAR_BLE

This arithmetic sum is conservative: it assumes that the spatial worst cases superpose exactly, which is not the case in reality. It eases the compliance demonstration but penalises product margin.

Realistic approach: composite measurement

IEC 62209-3 and KDB 447498 allow a composite measurement by simultaneously driving the transmitters at their worst-case configuration and directly recording the resulting SAR map. The protocol requires specific equipment (synchronous radio control), a real-simultaneity validation procedure, and a statistical post-processing. The margin gain can be significant (typically 1 to 3 dB) but the cost and campaign duration grow.

Simultaneity coefficient

An intermediate approach, described in several FCC KDBs, applies a simultaneity coefficient on the arithmetic sum, based on the real duty cycle of the protocols. The coefficient is justified by duty cycle and coexistence measurements (Wi-Fi in CSMA/CA mode, BLE at 1 % duty, cellular under scheduler allocation). This approach requires the coefficient to be demonstrated by experimental characterisation and by protocol specification.

Classic pitfalls in SAR dossier preparation

First-pass rejections in SAR labs come overwhelmingly from five recurring causes.

Inconsistent or undeclared separation distance. The manufacturer delivers the product without having formalised the separation distance to test. The lab applies a conservative default (typically 0 mm for a wearable, 5 mm for a smartphone), and the SAR value exceeds the limit. The distance must be fixed at specification, justified by realistic use, and printed in the user notice. See RED pitfalls for pitfalls tied to user documentation.
Wrong phantom. Testing a body-worn wearable with the SAM (head) phantom instead of the flat (body) phantom is an obvious error, but the reverse is more subtle: a hybrid product (smart glasses touching the head and temple at the same time as being worn in front) may require a double head + body evaluation with two successive phantoms.
Sub-6 GHz / mmWave confusion. For a 5G smartphone combining FR1 (sub-6 band) and FR2 (mmWave), SAR is needed on FR1 and power density on FR2, in the same campaign but with two distinct protocols. Confusing the two or testing only one is grounds for immediate FCC TCB rejection. KDB 941225 must be read before the test phase, not discovered at the report stage.
Multi-radio configuration under-evaluation. The dossier covers only standalone cellular and omits the cellular + Wi-Fi + BLE combination. The TCB asks for a complementary evaluation or an aggregate SAR calculation under IEC 62209-3. Result: 1 to 3 weeks of additional delay and a second lab session.
Expired liquid calibration. The tissue-equivalent liquid must be re-titrated periodically (typically every 1 to 3 months depending on lab). A report produced with an out-of-validity liquid is rejected by the TCB. The check is administrative but often missed at dossier audit. See FCC pitfalls for dossier audit pitfalls.

Relationship with HAC compatibility

Hearing Aid Compatibility (HAC), governed in the United States by 47 CFR 20.19, requires that mobile phones be tested for their near-field emission in the audio band (around 1 kHz typically) and their coupling with hearing aids. Although distinct from SAR (phantoms, bands and criteria differ), the HAC campaign is often organised in parallel with SAR because:

The product is already on the bench in worst-case configuration.
SAR-equipped labs are generally also HAC-equipped.
The overall campaign schedule is thus optimised.

HAC criteria are M3 or M4 for audio magnetic coupling and T3 or T4 for T-Coil coupling, with M4/T4 stricter than M3/T3. See the glossary for full term definitions: SAR, HAC, MPE.

Campaign timeline and cost

To plan budget and lead time to market, the orders of magnitude observed for a campaign in a European lab accredited ISO/IEC 17025 and FCC-recognised are as follows.

Product type	Typical configurations	Campaign duration	Indicative cost
Single-band BLE wearable	4 to 8	3 to 5 days	EUR 4 000 to 8 000
4G LTE multi-band smartphone	20 to 50	1 to 2 weeks	EUR 12 000 to 25 000
5G smartphone FR1 + FR2 (mmWave)	50 to 100+	3 to 4 weeks	EUR 25 000 to 50 000
Body-worn medical device	5 to 15	1 week	EUR 6 000 to 12 000
Smart glasses (head + body)	10 to 25	1 to 2 weeks	EUR 10 000 to 18 000

Cost depends mainly on the number of configurations to evaluate, and secondarily on robot mobilisation for the longest scans (mmWave under active beamforming can multiply durations). A dual EU + US campaign in an FCC-MRA lab can produce a single report usable on both sides, provided that the dual coverage was specified at quotation. See Certification costs for the overall budget framework.

Key takeaways

SAR measures RF power absorbed per unit tissue mass. EU limits: 2 W/kg over 10 g; US limits: 1.6 W/kg over 1 g. Averaging mass matters as much as the numerical value.
Metrology rests on IEC 62209. Four parts cover phones (1), body-worn (2), multi-source (3) and reduced-scope derived methods (4).
Phantoms are specific. SAM for the head, flat for the body, filled with tissue-equivalent liquid calibrated per band.
Above 6 GHz, SAR is replaced by power density (TPD), governed by IEC 62232 and FCC KDB 941225 for 5G FR2 mmWave.
The declared separation distance is a critical dossier parameter: set it upfront and verify it across all user documentation.
A multi-band campaign takes 2 to 4 weeks for a complete product, more with active beamforming or simultaneous multi-radio transmission.

For practical implementation on the EU side, see RED tests. For the US side, see FCC tests. For term definitions, see the Glossary.