Why run pre-compliance EMC tests in-house?

To find and fix emission problems while the design can still change cheaply, and to close the measure, modify, re-measure loop within a day instead of waiting for a lab slot. A pre-compliance result is not a compliance result: the report the market regulator or the FCC will look at comes from a site and instruments that meet the CISPR 16 series and the arrangement of the product standard. The bench is a filter, not a substitute.

What is the difference between a TEM cell and a GTEM cell?

IEC 61000-4-20 defines a TEM cell as a closed TEM waveguide, often a rectangular coaxial line, whose outer conductor completely encloses an inner conductor (the septum). A classical TEM cell is a two-port waveguide with a symmetric septum. A GTEM cell is a one-port TEM cell with an asymmetric septum; the standard notes that one-port waveguides typically end in a broadband termination combining a discrete 50 ohm load and anechoic material. IEC 61000-4-20 gives no single frequency ceiling for either type: its scope says the range depends on the testing requirements and on the specific waveguide.

Is a spectrum analyser enough, or do I need an EMI receiver?

CISPR 16-1-1:2019 (edition 5.0) applies to both: its term "measuring receiver" covers EMI receivers and spectrum analysers. What matters is whether the instrument meets the requirements, in particular the 6 dB bandwidths (200 Hz, 9 kHz, 120 kHz, 1 MHz by band) and the quasi-peak, peak, average and RMS-average detectors. CISPR 16-2-3 Annex E explains how to check that a spectrum analyser is suitable for compliance tests. A plain analyser in peak mode reads at least as high as quasi-peak, so it errs on the pessimistic side, never the optimistic one.

What does a LISN actually do?

CISPR 16-1-2 clause 4.1 gives three functions: present a defined impedance at radio frequencies at the EUT terminals, isolate the test circuit from unwanted signals on the supply mains, and couple the disturbance voltage to the measuring receiver. For 150 kHz to 30 MHz the network is the 50 ohm/50 μH V-AMN of clause 4.4; the 50 ohm/50 μH + 5 ohm network of clause 4.3 is specified for 9 kHz to 150 kHz. 47 CFR 15.31(l) requires a 50 ohm/50 uH LISN for FCC power-line measurements.

How do near-field probes fit in the workflow?

They localise sources, they do not measure compliance: no product emission limit is written in near-field probe units. For integrated circuits, IEC TS 61967-3:2014 standardises a surface scan of the near electric or magnetic field over the IC, up to 6 GHz, as a diagnostic for floor planning and power distribution. On a product board the same idea is used informally: tune the analyser to a frequency seen in the cell or chamber, move the probe until the reading peaks, change the design, re-measure.

Can a TEM or GTEM measurement replace the chamber test?

Only where the product standard allows it. IEC 61000-4-20 is a basic standard: its scope says product committees choose limits and methods with CISPR. Its Annex A gives correlation algorithms that convert waveguide voltages into OATS, SAC or free-space field strengths. CISPR 32 places GTEM methods in its informative Annex I with "proposed" limits, so for CISPR 32 a GTEM result is not a compliance result.

Which standard covers IC-level emission tests in a TEM cell?

IEC 61967-2:2005 (edition 1.0, stability date 2031). The IC sits on a test board clamped to a wall port of a TEM or GTEM cell, so the board becomes part of the cell wall. The method covers 150 kHz to 1 GHz and was developed with a 1 GHz TEM cell and a GTEM cell, both with a 45 mm septum to floor spacing over the port. It compares ICs; it does not say whether a finished product passes CISPR 32.

What are the most common errors on an in-house bench?

An EUT larger than the validated test volume of the cell; skipping the multi-orientation positions that the correlation algorithm needs; reading dB(μV) as if it were dB(μV/m) without the cell or antenna factor; a LISN of the wrong type for the band; an analyser left at a default bandwidth instead of the CISPR one; and treating a bench pass as proof of compliance when the arrangement of the product standard was not reproduced.

Pre-compliance EMC: TEM cell, near-field probes, LISN

Hugues Orgitello updated 7 Oct 2026 18 min read

A pre-compliance bench lets a hardware team see an emission problem the day it appears, long before the accredited campaign. The instruments are familiar (TEM or GTEM cell, LISN, current probe, near-field probes, spectrum analyser or EMI receiver), but each one has a text behind it, and most bench errors come from ignoring what that text says: IEC 61000-4-20 for TEM waveguides, IEC 61967-2 and IEC TS 61967-3 for integrated circuits, CISPR 16-1-1 for the receiver, CISPR 16-1-2 for the LISN and current probes, CISPR 16-1-4 for antennas. This page goes back to those texts: which edition is current in October 2026, what each clause gives a bench user, what a cell result can and cannot be used for, and the common claims the texts contradict, including several this page used to make.

The texts behind a bench

Text	Current edition	Covers	Bench use
IEC 61000-4-20	Edition 3.0, February 18, 2022 (stability date 2027)	Emission and immunity in TEM waveguides: TEM cells, GTEM, strip lines	Cell measurements, correlation to field strength
IEC 61967-2	Edition 1.0, September 29, 2005 (stability date 2031)	IC radiated emissions, TEM or GTEM cell with a wall port, 150 kHz to 1 GHz	Comparing ICs or IC configurations
IEC TS 61967-3	Edition 2.0, August 25, 2014 (stability date 2028)	IC near-field surface scan, up to 6 GHz	Locating sources on and around an IC
CISPR 16-1-1	Edition 5.0, May 22, 2019	Measuring receivers (EMI receivers and spectrum analysers), 9 kHz to 18 GHz	Detectors, bandwidths, calibration (Annex J)
CISPR 16-1-2	2014 + AMD1:2017, consolidated edition 2.1	Artificial mains networks, current and voltage probes, asymmetric artificial networks	LISN, current clamp, telecom port networks
CISPR 16-1-4	Edition 5.0, October 3, 2025	Antennas and test sites for radiated measurements	Antennas, chamber validation
CISPR 16-2-3	2016 + AMD1:2019 + AMD2:2023, consolidated edition 4.2	Radiated disturbance measurement methods	Annex E: is a spectrum analyser suitable?
CISPR 32	2015 (edition 2.0)	Multimedia equipment emission limits	Limits; Annex I (informative) for GTEM

Editions and dates are from the IEC online store records listed in the sources. IEC 61000-4-20 edition 3.0 cancelled and replaced the 2010 edition; there is no "2010 + A1:2022" version.

Equipment families

Equipment	Role	Reference text
TEM or GTEM cell	Radiated emissions and immunity on EUTs that fit the validated test volume	IEC 61000-4-20
LISN (V-network artificial mains network)	Conducted emissions on the mains port, 150 kHz to 30 MHz	CISPR 16-1-2 clause 4
Current probe (clamp-on)	Common-mode current on cables and harnesses	CISPR 16-1-2 clause 5, calibration in Annex B
AAN (asymmetric artificial network, also called Y-network)	Wired network and signal ports	CISPR 16-1-2 clause 7
Near-field probes (H-field loops, E-field probes)	Locating sources on a board or an IC	IEC TS 61967-3 for ICs; otherwise engineering practice
Spectrum analyser or EMI receiver	The measuring receiver behind every probe, network and antenna	CISPR 16-1-1
Antennas	Radiated measurements in a chamber, from 30 MHz upward	CISPR 16-1-4
Semi-anechoic chamber, FAR or OATS	Final pre-test before the accredited campaign	CISPR 16-1-4; see EMC chamber types

A team can start with a cell, a LISN, a measuring receiver and a probe set, and add layers as the product mix grows. For the chamber types themselves, see EMC chamber types: SAC, FAR, OATS, reverberation.

TEM and GTEM cells

What IEC 61000-4-20 calls them

IEC 61000-4-20 edition 3.0 covers open TEM waveguides (strip lines, electromagnetic pulse simulators) and closed ones (TEM cells), classified as one-port, two-port or multi-port. Its definitions are the useful vocabulary:

TEM cell (3.1.21): a closed TEM waveguide, often a rectangular coaxial transmission line, with an outer conductor completely enclosing an inner conductor.
Septum (3.1.11): the inner conductor, often flat, placed symmetrically or asymmetrically in the outer conductor.
One-port TEM waveguide (3.1.14): a single input/output port, typically with a broadband transmission-line termination at the other end. That termination (3.1.2) combines a low-frequency discrete load matched to the characteristic impedance, typically 50 Ω, and a volume of high-frequency anechoic material.
Characteristic impedance (3.1.3): TEM waveguides are typically designed for 50 Ω; 100 Ω waveguides are often used for transient testing.
Correlation algorithm (3.1.4): the routine that converts TEM waveguide voltages into OATS, SAC or free-space field strengths.
Manipulator (3.1.13): a non-metallic positioner that holds the EUT in the positions the correlation algorithm requires.

Annex D draws the two common closed structures: the two-port TEM cell with a symmetric septum (Figure D.7) and the one-port TEM cell with an asymmetric septum (Figure D.8), which is the geometry usually sold as a GTEM cell.

Frequency range

The scope of IEC 61000-4-20 states that "the frequency range depends on the specific testing requirements and the specific TEM waveguide type". Annex D.7 deals with frequency limitations: in a closed cell, higher-order modes appear above a frequency set by the cross-section, which is why a small two-port cell stops lower than a large tapered one-port cell. The upper frequency of a given cell is therefore a property of that cell, to be read on its datasheet and confirmed by the TEM mode verification of clause 5.2.2. The example uncertainty budgets of Annex F cover 30 MHz to 1 GHz and 1 GHz to 6 GHz.

Test volume and EUT size

Clause 5.2.1 sets the test volume and the maximum EUT size; clause 5.2.2 sets how the usable test volume is validated, with a TEM mode and field uniformity procedure in two variants ("constant forward power" and "constant field strength", Figures 1 and 2). Clause 6 separates small and large EUTs; edition 3.0 added guidance on large EUTs including cables. An EUT that does not fit the validated volume is measured outside the conditions the correlation assumes, which makes the result unreliable. Take the size limit from the standard and from the validation report of your cell, not from a rule of thumb.

Emission measurements (Annex A)

Annex A is normative. A bench user needs four parts of it:

Annex A clause	Content	What it means on the bench
A.3.2	Correlation algorithms	The cell port voltage becomes an equivalent field strength only through the algorithm
A.4	Correction factors, reference emission sources, small EUT correction factor	Characterise the cell with a known source before trusting a number
A.5.2 and Figures A.1 to A.4	EUT arrangement, exit cable routed to a corner of the test volume at a defined angle, positioner, twelve-face orientations	The EUT is measured in several orientations, not once
A.6	Test report	What has to be recorded for the result to be repeatable

Annex F (informative) gives the instrumentation uncertainty of TEM waveguide emission results, and Annex H (informative) correlates emission limits between facilities, with Figure H.5 showing TEM waveguide Class A and Class B limits correlated from CISPR 32.

What a cell result counts for

IEC 61000-4-20 is a basic standard. Its scope says that for radiated emissions "product committees select emission limits and measurement methods in consultation with CISPR standards", and that its methods are separate from IEC 61000-4-3 and "may be used when so specified by product committees". In practice:

CISPR 32 emissions: Annex I is informative and gives "proposed" GTEM limits. It is not a compliance route under CISPR 32, so a GTEM pass is an engineering indication.
Radiated immunity: EN 55035 allows TEM waveguides (IEC 61000-4-20) and reverberation chambers (IEC 61000-4-21) as alternatives to IEC 61000-4-3; see EMC chamber types.

What changed in edition 3.0 (2022)

The foreword lists the technical changes from the 2010 edition: information on testing large EUTs including cables; measurement uncertainty work adapted from CISPR and TC 77; an updated validation procedure for field uniformity and TEM mode verification; information on two-port and four-port waveguides; a new informative Annex I on transient waveguide characterisation; and information on dielectric test stands for EUTs.

IC-level methods: IEC 61967-2 and IEC TS 61967-3

Two parts of the IEC 61967 series (IEC TC 47/SC 47A) bring the TEM cell and the near-field probe down to the integrated circuit.

	IEC 61967-2:2005	IEC TS 61967-3:2014
Method	TEM or wideband GTEM cell with a wall port	Near-field surface scan (electric, magnetic or both)
Frequency range	150 kHz to 1 GHz	Intended for use up to 6 GHz
Set-up	The IC test board is clamped to a port cut in the top or bottom of the cell and becomes part of the cell wall	Probe scanned over the IC on any board the probe can reach
Reference geometry	Developed with a 1 GHz TEM cell and a GTEM cell, both with a 45 mm septum to floor spacing over the port	Resolution set by the probe and the positioning system
Comparison between ICs	Other cells may be used for comparative measurements; a conversion factor may relate cells with different septum spacing	Use the standardised test board of IEC 61967-1
Purpose	Measuring IC radiated emission	Diagnostic for floor planning and power distribution

The IEC record of IEC 61967-2 names the septum to test board spacing as "the primary factor affecting the measured RF voltage". Neither method tells you whether a finished product meets CISPR 32; both tell you which IC, firmware setting or decoupling scheme radiates less.

Reverberation chamber

IEC 61000-4-21 (edition 2.0, 2011) covers immunity, emission and shielding measurements in reverberation chambers, where a mechanical tuner or stirrer makes the field statistically uniform. CISPR 32 Annex I proposes reverberation chamber limits above 1 GHz only, again informatively. For a bench, a reverberation chamber is rarely the first purchase; see EMC chamber types and IEC 61000-4-3 radiated RF immunity.

LISN, conducted emissions on the mains port

CISPR 16-1-2 clause 4.1 gives the three functions of an artificial mains network: provide a defined impedance at radio frequencies at the EUT terminals, isolate the test circuit from unwanted signals on the supply mains, and couple the disturbance voltage to the measuring receiver. The standard notes that "LISN" and "V-AMN" are used interchangeably.

CISPR 16-1-2 clause	Network	Band
4.3	50 Ω/50 μH + 5 Ω V-AMN	9 kHz to 150 kHz
4.4	50 Ω/50 μH V-AMN	150 kHz to 30 MHz
4.5	50 Ω/5 μH + 1 Ω V-AMN	150 kHz to 108 MHz
4.6	150 Ω V-AMN	150 kHz to 30 MHz
4.7	150 Ω delta-network	150 kHz to 30 MHz

For CISPR 32 and FCC Part 15 mains measurements the reference is the 50 Ω/50 μH network of clause 4.4; 47 CFR 15.31(l) requires "a 50 ohm/50 uH line-impedance stabilization network (LISN)". The CISPR 32 mains-port limits (Tables A.9 and A.10) and 47 CFR 15.107 are voltage limits over 150 kHz to 30 MHz. CISPR 22 has been replaced by CISPR 32; in the EU, EN 55032 replaced EN 55022 on March 5, 2017. The full comparison of limits and networks is in Conducted emissions: LISN method.

Current probes and asymmetric artificial networks

Beyond the mains port, disturbance flows as common-mode current on cables. CISPR 16-1-2 clause 5 specifies current and voltage probes, and its informative Annex B covers their construction, frequency range and calibration, including the transfer impedance that converts the receiver reading into a current. Clause 7 specifies the asymmetric artificial networks (AAN, also called Y-network) used on wired network and signal ports; the AAN requirements were copied from CISPR 22 into CISPR 16-1-2.

In CISPR 32, current limits belong to the wired network ports (Tables A.11 and A.12), not to the mains port. On a bench, a clamp around a harness is still the quickest way to see which cable carries the common-mode current that later shows up as a radiated peak.

Antennas for radiated measurement

When the bench grows beyond a cell, the next step is a chamber and a set of antennas. Antennas and test sites are specified in CISPR 16-1-4 (edition 5.0, October 3, 2025). The CISPR 32 radiated limits start at 30 MHz; above 1 GHz the limits use peak and average detectors. A biconical antenna and a log-periodic antenna, or a hybrid of the two, typically cover 30 MHz to 1 GHz, and horn antennas cover the range above; check the band of each model on its calibration certificate.

The calibration certificate gives the free-space antenna factor in dB(1/m) versus frequency. Field strength in dB(μV/m) is the receiver reading in dB(μV) plus the antenna factor plus the cable loss. In a TEM cell the same role is played by the correlation of IEC 61000-4-20 Annex A. For uncertainty budgets, see calibration and measurement uncertainty (GUM).

Near-field probes, locating the source

A near-field probe does not measure compliance; it tells you where the energy comes from. H-field loop probes respond to current (switching loops, return paths, inductors, traces); E-field probes respond to voltage (high dV/dt nodes, heat sinks, open-ended metal parts). Smaller probes resolve finer detail, larger ones collect more signal. Probe sets and their sizes vary by manufacturer; read the probe factor and recommended preamplifier from the datasheet.

At IC level the method is standardised: IEC TS 61967-3 maps the near electric or magnetic field over the IC, records measurement data, post-processing and data exchange (clauses 9.4 to 9.6 in edition 2.0) and defines coordinate systems (Annex D). On a product board there is no equivalent standard, and the result is a map, not a pass or fail.

Orientation

The magnetic field circles the current that produces it. A loop reads most when the field threads it, that is when the loop plane contains the axis of the trace and stands perpendicular to the board. Rotate the probe at each position and keep the maximum; a probe held at a fixed angle can miss a source entirely.

Workflow

Identify the frequency of concern in the cell, the chamber or the LISN measurement.
Tune the analyser to it with a narrow span and peak hold.
Scan the powered board with an H-field loop, rotating the probe, then with an E-field probe.
Mark the position of maximum reading and photograph it.
Repeat for the harmonics and for other suspect frequencies.
Change the layout, filtering or shielding.
Re-measure in the cell or chamber, then re-probe.

The probe reading itself is relative: compare before and after on the same probe, amplifier, cable and analyser settings, not against a limit.

Spectrum analyser or EMI receiver

CISPR 16-1-1:2019 (edition 5.0) specifies measuring apparatus from 9 kHz to 18 GHz, and its IEC record states that its specifications apply to EMI receivers and spectrum analysers alike. The question is not which box you own but whether it meets the requirements; CISPR 16-2-3 Annex E explains how to check that a spectrum analyser is suitable for compliance tests. Calibration requirements for measuring receivers are in CISPR 16-1-1 Annex J.

Bands, bandwidths and quasi-peak time constants

CISPR band	Frequency range	6 dB bandwidth	Quasi-peak charge / discharge
A	9 kHz to 150 kHz	200 Hz	45 ms / 500 ms
B	150 kHz to 30 MHz	9 kHz	1 ms / 160 ms
C	30 MHz to 300 MHz	120 kHz	1 ms / 550 ms
D	300 MHz to 1 GHz	120 kHz	1 ms / 550 ms
E	1 GHz to 18 GHz	1 MHz	No quasi-peak detector specified

Source: CISPR 16-1-1 values as tabulated in the manufacturer white paper listed in the sources (secondary source). The bandwidths are 6 dB bandwidths; an analyser whose resolution bandwidth is specified at 3 dB does not match them by setting the same number.

Detectors

CISPR 16-1-1 specifies quasi-peak (clause 5, up to 1 GHz), peak (clause 6), average (clause 7) and RMS-average (clause 8). For any signal, the peak reading is never lower than the quasi-peak reading, and the quasi-peak reading is never lower than the average reading. A peak scan compared with a quasi-peak or average limit therefore errs on the pessimistic side: it can flag a line that passes in quasi-peak, it cannot hide one that fails. The usual practice is a fast peak pre-scan, then quasi-peak and average final measurements on the lines close to the limit (CISPR 16-2-3 clause 8, automated measurement). On the FCC side, 47 CFR 15.35(a) lets the responsible party use a peak detector at the same bandwidth instead of quasi-peak below 1000 MHz, and 15.35(b) sets a minimum 1 MHz bandwidth above 1000 MHz with the peak capped at 20 dB above the average limit.

Calibration and uncertainty

Every element of the chain contributes to uncertainty: cell correlation, antenna factor, LISN voltage division factor, current probe transfer impedance, preamplifier gain, cable loss, receiver accuracy. The texts give the tools rather than a calendar: CISPR 16-1-1 Annex J for receivers, CISPR 16-1-2 Annex B for current probes, IEC 61000-4-20 Annex E for calibrating E-field probes in TEM waveguides and Annex F for the uncertainty budget of TEM waveguide emissions. Set the recalibration interval from your quality system and the instrument history. The GUM framework is covered in calibration and measurement uncertainty (GUM).

Building the bench in stages

Board level: near-field probe set with preamplifier, measuring receiver with CISPR bandwidths and detectors, a 50 Ω/50 μH LISN for the mains port, a current clamp.
Product level: a TEM or GTEM cell sized for the products you make, characterised per IEC 61000-4-20 Annex A, with the positioner the correlation needs.
System level: a chamber with calibrated antennas, in-house or rented, for a final pre-test before the accredited campaign.

Which stage pays off depends on the product mix; many teams stop at stage two and rent chamber time for stage three.

Workflow during product development

Layout review before fabrication: switching nodes, clocks, connector launches, return paths and decoupling. See PCB design for EMC.
Bare board at bring-up: near-field scan to find the dominant sources before the enclosure exists.
Assembled product in the cell: within the validated test volume, in the orientations Annex A requires, compared with a limit line correlated to the cell.
Cables: current clamp on each harness, then cable routing, shielding and ferrites.
Conducted emissions on the LISN: mains port against CISPR 32 Tables A.9 and A.10 or 47 CFR 15.107, then input filter changes.
Final pre-test in a chamber before booking the accredited campaign. See radiated emissions EMC test.

Pitfalls

Pitfall	Consequence
EUT larger than the validated test volume (IEC 61000-4-20 clause 5.2)	Correlation conditions not met, unreliable field strength
One orientation only in the cell	The correlation algorithm of Annex A needs several positions; one face can hide the main radiator
Cell factor or antenna factor not applied	dB(μV) read as if it were dB(μV/m)
Analyser left at a default resolution bandwidth	Not the CISPR 6 dB bandwidth (for example 120 kHz from 30 MHz to 1 GHz), readings not comparable to the limit
LISN of the wrong type for the band	The 50 Ω/50 μH + 5 Ω characteristic is specified for 9 kHz to 150 kHz; 150 kHz to 30 MHz uses the clause 4.4 network
Exit cables routed at random in the cell	Annex A routes the exit cable to a corner of the test volume at a defined angle; another layout changes the result
Near-field probe at a fixed angle	Source missed or mislocated
Bench arrangement differs from the product standard	Bench and lab disagree; CISPR 32 Annex D sets the arrangement for conducted tests
Treating a GTEM pass as CISPR 32 compliance	CISPR 32 Annex I is informative with proposed limits

Claims the texts contradict

Common claim	What the text says
"IEC 61000-4-20:2010 + A1:2022 is the current version" (this page said so before October 2026)	Edition 3.0 of February 18, 2022 cancelled and replaced the 2010 edition as a technical revision.
"A GTEM cell reaches 18 GHz and beyond" (this page)	IEC 61000-4-20 sets no such figure: the range "depends on the specific testing requirements and the specific TEM waveguide type". Its example uncertainty budgets stop at 6 GHz. Read the limit of your cell from its datasheet and validation.
"A TEM cell works up to about 1 GHz and accepts an EUT of one third to one quarter of the septum height" (this page)	Both depend on the cell. Clause 5.2.1 sets the test volume and maximum EUT size, and 5.2.2 how the volume is validated. IEC 61967-2 was developed with a 1 GHz TEM cell, which is a property of that cell, not of TEM cells in general.
"Without the EMI option, a peak-only analyser overestimates the margin by 10 to 20 dB" (this page)	Peak reads at least as high as quasi-peak and average, so a peak reading underestimates the margin, never overestimates it. The 47 CFR 15.35(a) peak alternative works for the same reason.
"CISPR detectors are quasi-peak, average, RMS-average and peak-hold" (this page)	CISPR 16-1-1 specifies quasi-peak, peak, average and RMS-average detectors. Peak hold is an analyser display mode.
"CISPR 16-1-1 is for EMI receivers; analysers are a compromise"	Its scope applies to both EMI receivers and spectrum analysers. What counts is meeting the requirements, checked with CISPR 16-2-3 Annex E.
"The LISN is 50 Ω in parallel with 50 μH" (this page)	CISPR 16-1-2 specifies networks by impedance characteristic and band; for 150 kHz to 30 MHz the reference is the 50 Ω/50 μH V-AMN of clause 4.4.
"The current probe is paired with an ICN" (this page)	CISPR 16-1-2 pairs current probes with their own calibration (Annex B) and specifies asymmetric artificial networks in clause 7 for signal and network ports.
"CISPR 32 informative annexes give a GTEM correlation accepted for compliance"	Annex I is informative with "proposed" limits. The correlation algorithms are in IEC 61000-4-20 Annex A.

Key facts

Current editions (October 2026): IEC 61000-4-20 edition 3.0 (2022), IEC 61967-2 edition 1.0 (2005), IEC TS 61967-3 edition 2.0 (2014), CISPR 16-1-1 edition 5.0 (2019), CISPR 16-1-2 edition 2.1 (2017), CISPR 16-1-4 edition 5.0 (2025).
TEM waveguide results count where the product standard allows them; for CISPR 32 emissions, GTEM is informative only.
Receivers: one set of requirements for EMI receivers and spectrum analysers, 9 kHz to 18 GHz.
LISN: 50 Ω/50 μH V-AMN for 150 kHz to 30 MHz, the same quantity for CISPR 32 and 47 CFR 15.107.

Going further

EMC chamber types: SAC, FAR, OATS, reverberation: comparative view of the chamber types referenced above
Calibration and measurement uncertainty (GUM): uncertainty budgets for the bench
PCB design for EMC: upstream layout discipline that pre-compliance verifies
IEC 61000-4-3 radiated RF immunity: the immunity context for radiated test chambers
Radiated emissions EMC test: the formal accredited test that pre-compliance prepares
Conducted emissions: LISN method: networks and limits for the mains port
Glossary: definitions of TEM, GTEM, LISN, quasi-peak, antenna factor

Frequently asked questions

Why run pre-compliance EMC tests in-house?: To find and fix emission problems while the design can still change cheaply, and to close the measure, modify, re-measure loop within a day instead of waiting for a lab slot. A pre-compliance result is not a compliance result: the report the market regulator or the FCC will look at comes from a site and instruments that meet the CISPR 16 series and the arrangement of the product standard. The bench is a filter, not a substitute.
What is the difference between a TEM cell and a GTEM cell?: IEC 61000-4-20 defines a TEM cell as a closed TEM waveguide, often a rectangular coaxial line, whose outer conductor completely encloses an inner conductor (the septum). A classical TEM cell is a two-port waveguide with a symmetric septum. A GTEM cell is a one-port TEM cell with an asymmetric septum; the standard notes that one-port waveguides typically end in a broadband termination combining a discrete 50 ohm load and anechoic material. IEC 61000-4-20 gives no single frequency ceiling for either type: its scope says the range depends on the testing requirements and on the specific waveguide.
Is a spectrum analyser enough, or do I need an EMI receiver?: CISPR 16-1-1:2019 (edition 5.0) applies to both: its term "measuring receiver" covers EMI receivers and spectrum analysers. What matters is whether the instrument meets the requirements, in particular the 6 dB bandwidths (200 Hz, 9 kHz, 120 kHz, 1 MHz by band) and the quasi-peak, peak, average and RMS-average detectors. CISPR 16-2-3 Annex E explains how to check that a spectrum analyser is suitable for compliance tests. A plain analyser in peak mode reads at least as high as quasi-peak, so it errs on the pessimistic side, never the optimistic one.
What does a LISN actually do?: CISPR 16-1-2 clause 4.1 gives three functions: present a defined impedance at radio frequencies at the EUT terminals, isolate the test circuit from unwanted signals on the supply mains, and couple the disturbance voltage to the measuring receiver. For 150 kHz to 30 MHz the network is the 50 ohm/50 μH V-AMN of clause 4.4; the 50 ohm/50 μH + 5 ohm network of clause 4.3 is specified for 9 kHz to 150 kHz. 47 CFR 15.31(l) requires a 50 ohm/50 uH LISN for FCC power-line measurements.
How do near-field probes fit in the workflow?: They localise sources, they do not measure compliance: no product emission limit is written in near-field probe units. For integrated circuits, IEC TS 61967-3:2014 standardises a surface scan of the near electric or magnetic field over the IC, up to 6 GHz, as a diagnostic for floor planning and power distribution. On a product board the same idea is used informally: tune the analyser to a frequency seen in the cell or chamber, move the probe until the reading peaks, change the design, re-measure.
Can a TEM or GTEM measurement replace the chamber test?: Only where the product standard allows it. IEC 61000-4-20 is a basic standard: its scope says product committees choose limits and methods with CISPR. Its Annex A gives correlation algorithms that convert waveguide voltages into OATS, SAC or free-space field strengths. CISPR 32 places GTEM methods in its informative Annex I with "proposed" limits, so for CISPR 32 a GTEM result is not a compliance result.
Which standard covers IC-level emission tests in a TEM cell?: IEC 61967-2:2005 (edition 1.0, stability date 2031). The IC sits on a test board clamped to a wall port of a TEM or GTEM cell, so the board becomes part of the cell wall. The method covers 150 kHz to 1 GHz and was developed with a 1 GHz TEM cell and a GTEM cell, both with a 45 mm septum to floor spacing over the port. It compares ICs; it does not say whether a finished product passes CISPR 32.
What are the most common errors on an in-house bench?: An EUT larger than the validated test volume of the cell; skipping the multi-orientation positions that the correlation algorithm needs; reading dB(μV) as if it were dB(μV/m) without the cell or antenna factor; a LISN of the wrong type for the band; an analyser left at a default bandwidth instead of the CISPR one; and treating a bench pass as proof of compliance when the arrangement of the product standard was not reproduced.