Insights · Technical note

Below 30 MHz, the chamber now has to prove itself too

CISPR 16-1-4 added a site validation requirement for 9 kHz to 30 MHz: normalized site insertion loss, measured with a pair of loop antennas. A short introduction to the method and its geometry, what it means for laboratories that measure magnetic fields at 3 m or 10 m, and a measured result from a semi-anechoic chamber.

19 September 20269 minute readSite validation · Standards

For as long as semi-anechoic chambers have been validated, the validation started at 30 MHz. Normalized site attenuation covered 30 MHz to 1 GHz and site VSWR covered the range above it. Below 30 MHz a laboratory measured magnetic field with a loop antenna at 3 m or 10 m in the same chamber, and nobody asked whether the chamber was any good there. Amendment 2 (2023) to the fourth edition of CISPR 16-1-4 closed that gap with a new Clause 5, and the fifth edition of 2025 carries it unchanged. The method is called normalized site insertion loss, NSIL, and it is the low-frequency counterpart of NSA.

In short
  • An OATS or SAC used for radiated disturbance measurements from 9 kHz to 30 MHz shall be validated by the NSIL method of CISPR 16-1-4 clause 5.5, at the distance used for testing: 3 m, 5 m or 10 m.
  • Two loop antennas at 1.3 m height, three orientations, five transmit positions, 15 site insertion loss sweeps. The measured insertion loss, corrected for both antenna factors, is compared with a value computed for the same geometry by numerical simulation.
  • The acceptance criterion is ±4 dB at every frequency, orientation and position. A 10 m chamber that misses it may still be used with an increased measurement uncertainty.

The method in one equation

NSIL borrows the idea behind NSA. The site's contribution is separated from the antennas' contribution, so that any pair of calibrated antennas measures the same property of the site. The transmit loop is driven at a fixed level and the receive loop's output is read twice: once with the two cables joined through an adaptor, which gives VDIRECT, and once with the cables connected to the antennas at their positions in the site, which gives VSITE. The site insertion loss deviation is then

ΔAi = VDIRECT − VSITE − FaH,T − FaH,R − ANi

where FaH,T and FaH,R are the magnetic field antenna factors of the transmit and receive loops, calibrated to CISPR 16-1-6, and ANi is the normalized site insertion loss: the insertion loss an ideal site would show for that geometry, with the antenna factors taken out. The deviation ΔAi is what the site adds, and it is the number judged against the criterion.

The difference from NSA is where ANi comes from. For NSA the standard prints tables, because Hertzian dipoles are a good enough model of the antennas. For loops below 30 MHz they are not. Annex H shows that the NSIL value depends on the diameter of each loop and on where its feed point sits during the measurement, so no universal table is possible. Instead, ANi shall be calculated for the actual loops by a numerical electromagnetics code, NEC, using the decks the annex prints, and Annex J sets out how accurate that simulation is. A calibration laboratory offering NSIL therefore has to bring its own simulated reference for its own loops, not a table copied from the standard.

The geometry

The arrangement is fixed by clause 5.5.1 and Figures 5 to 7.

  • Site. An OATS, an OATS with a weather-protection enclosure, or a SAC. A semi-free-space environment with a ground plane, the same as for NSA.
  • Distance. 3 m, 5 m or 10 m, whichever is used for the product measurements, measured between the reference points of the two loops and kept constant for every position.
  • Height. Both loops at 1.3 m above the ground plane. No height scan for either antenna.
  • Orientations. Three: Hx with the loops coaxial, facing each other along the measurement axis; Hy with the loops coplanar in the vertical plane; Hz with the loops coplanar and horizontal.
  • Positions. The transmit loop is placed at the centre of the test volume and at four points on its perimeter: front, rear, left and right. The receive loop moves so that the distance stays constant. Five positions by three orientations is 15 insertion loss sweeps.
  • Frequency steps. No coarser than 1 kHz to 20 kHz, 5 kHz to 150 kHz, 50 kHz to 1 MHz and 100 kHz to 30 MHz, which is 345 points over the band.
  • Housekeeping. A signal-to-noise ratio of at least 20 dB is recommended and the achieved value goes into the uncertainty budget. The transmit and receive cables must not form a ground loop through the chamber's bulkhead connectors, and ferrites every 20 cm along the cables inside the test volume are recommended.
Test volume 1.5 m diameter (NSIL Hx) both loops at h = 1.3 m Front Centre Back Left Right transmit loop positions on the turntable Front Centre Back Left / right: receive loop stays here, both loops turned to face each other receive loop positions d = 3 m, receive loop to transmit loop receive loop relocated over 1.5 m The three orientations, side view of one loop pair Hx, coaxial loops face each other along the axis Hy, coplanar vertical both loops in one vertical plane Hz, coplanar horizontal both loops horizontal, same height
Top view for the Hx orientation, after Figure 6 of CISPR 16-1-4, drawn for a 3 m distance and a 1.5 m test volume. The transmit loop takes the centre, front and back positions on the axis and the receive loop is relocated by the same amount to keep d = 3 m. For the left and right positions the receive loop stays at its centre position and both loops are turned to face each other. The same five positions are repeated for Hy and Hz.

The criterion, and the 10 m exception

Table 2 sets a single limit: the deviation ΔAi shall be within ±4 dB at all frequencies, for all three orientations and at all five positions, whether the distance is 3 m, 5 m or 10 m. The table carries a footnote for 10 m semi-anechoic chambers. Measurements have shown that some cannot meet ±4 dB across the whole band, and the standard accepts their use provided the increased uncertainty is carried into the compliance decision. Annex K shows the arithmetic: the CISPR uncertainty for magnetic field measurements assumes a ±4 dB site contribution, and where the site exceeds it, the laboratory's own uncertainty grows, and the excess is added to the measured field strength before comparison with the limit. That is a real penalty at the limit line, which is why a laboratory would rather know its chamber's deviation than assume it.

Who this applies to

The requirement follows the measurement method, not the type of product. Any OATS or SAC in which radiated disturbance is measured with a loop antenna at 3 m, 5 m or 10 m under CISPR 16-2-3 is a site that clause 5 expects to be validated. In practice that means:

  • CISPR 11 industrial, scientific and medical equipment where magnetic field limits below 30 MHz apply at 3 m or 10 m, including induction heating and wireless power transfer.
  • CISPR 14-1 and CISPR 15 appliances and lighting products where the laboratory chooses the loop-at-distance method instead of a large-loop antenna system. The large-loop antenna system has its own validation in Annex B and does not use NSIL.
  • CISPR 36 electric and hybrid vehicles, whose below-30 MHz radiated limits are measured with a loop at 3 m or 10 m.
  • Any laboratory whose accreditation cites CISPR 16-1-4 for its site, since the assessor reads the current edition and its Clause 5.

Three consequences follow for a laboratory. The site file gains a third validation alongside NSA and site VSWR, on the same periodic cycle. The validated test volume below 30 MHz is defined by where the transmit loop was placed, so a larger EUT needs the perimeter positions moved out and the measurement repeated. And the calibration certificate has to carry the simulated NSIL values for the loops that were used, because without them the deviation cannot be reproduced or checked.

What a validated chamber looks like

The plots below are from an NSIL validation we performed in a 10 m semi-anechoic chamber at a test laboratory, at a 3 m measurement distance, a 1.5 m test volume and a loop height of 1.3 m. The transmit antenna was a 60 cm passive loop and the receive antenna a 60 cm active loop, with the reference NSIL for that pair computed as Annex H requires. Each plot shows the deviation ΔAi for the five transmit positions against the ±4 dB criterion, over the 345 frequencies of Table 1.

-6 -4 -2 0 +2 +4 +6 10 kHz 100 kHz 1 MHz 10 MHz 30 MHz +4 dB criterion −4 dB criterion Hx, coaxial largest deviation -1.2 dB at 26.2 MHz -6 -4 -2 0 +2 +4 +6 10 kHz 100 kHz 1 MHz 10 MHz 30 MHz +4 dB criterion −4 dB criterion Hy, coplanar vertical largest deviation +2.1 dB at 19.1 MHz -6 -4 -2 0 +2 +4 +6 10 kHz 100 kHz 1 MHz 10 MHz 30 MHz +4 dB criterion −4 dB criterion Hz, coplanar horizontal largest deviation +1.6 dB at 30 MHz Centre Front Rear Left Right ΔAi in dB against frequency, log scale
Measured NSIL deviation in a 10 m semi-anechoic chamber at a 3 m distance, 1.5 m test volume, loops at 1.3 m. Five transmit positions per orientation, 345 frequencies from 9 kHz to 30 MHz, judged against the ±4 dB criterion of Table 2.

Two things stand out. The five positions lie almost on top of one another in every orientation, so the ground plane and the surroundings treat the whole 1.5 m test volume alike. And the deviation stays within about 2 dB across the band, with the largest value of 2.1 dB in the Hy orientation near 19 MHz, half of what the criterion allows. A chamber built for measurements from 30 MHz upward, with a proper ground plane, passes at 3 m with room to spare. The same chamber at 10 m, with the loops closer to the walls and the site insertion loss far higher, is the case the Table 2 footnote was written for.

Before you book a validation

One decision is the laboratory's. The rest is the calibration laboratory's job, and a good one will not ask you to do it.

Your part

Fix the distance and the test volumeThe validation is done at the distance you test at, and the five positions define the volume you may then use. Decide both from the products you measure, not from the chamber's nominal size, and tell us the largest EUT you expect below 30 MHz.

The calibration laboratory's part

The simulated referenceThe NSIL values on the certificate are derived per Annex H for the loop pair actually used, from their diameters and feed positions. We bring the simulation for our own loops; a generic table is not what the standard asks for.
The antenna factorsThe magnetic field antenna factors of both loops are calibrated to CISPR 16-1-6 at the Table 1 frequency steps, or the sum of the two per Annex L, before the visit, and they are stated on the certificate.
Cables, ground loops and noiseThe dynamic range check of 5.5.1, the routing that avoids a ground loop through the bulkhead connectors, the ferrites on the cables and the signal-to-noise ratio are the calibration laboratory's set-up, and the achieved values go into its uncertainty budget.
Where we stand

NSIL from 9 kHz to 30 MHz is included, with NSA and site VSWR, in the site validation work we perform on site under LAB Support Ltd.'s accredited scope, with a 60 cm loop pair and NSIL values simulated for that pair as Annex H requires. The site validation guide lists the parameters. If your chamber is used for loop measurements below 30 MHz and has never been validated there, send us the measurement distance and the largest product you test, and we will plan the positions with you.

Ask about an NSIL validation More insights

Tell us what you need to test

We reply within one working day with a quote or a plan.

Contact us