Using an RF Shielded Test Enclosure at Low RF Frequencies

Using the chamber from 10 kHz to 30 MHz: what changes, what doesn’t, and how to get results you can trust

Overview

Most of the traffic through an RF shielded test enclosure involves radios operating at hundreds of megahertz and up — Wi-Fi, Bluetooth, cellular, GPS. But a surprising number of products live much further down the dial: NFC and HF RFID readers at 13.56 MHz, access-control fobs and animal ID tags at 125 and 134.2 kHz, wireless power systems operating between roughly 85 and 360 kHz, AM broadcast receivers, hearing-aid telecoils, and utility metering gear that talks over the powerline. If you test any of these, you have probably wondered whether a shielded enclosure specified at “−100 dB, DC to 1 GHz” really delivers that number at 50 kHz.

The short answer is good news for nearly every real test scenario, but it comes with one piece of physics worth understanding: at low frequencies, electric fields and magnetic fields interact with a shielded wall in very different ways. This note explains what actually happens, shows calculated performance for a standard JRE aluminum enclosure, and offers practical guidance for making clean, repeatable low-frequency measurements. One formula is involved; nothing worse.

Two Kinds of Field, One Aluminum Wall

Above a hundred megahertz or so, the signals reaching your enclosure are true radiated waves, with electric (E) and magnetic (H) fields locked together in a fixed ratio. A shielded enclosure handles these with ease: the conductive aluminum skin reflects the wave the way a mirror reflects light, and the small amount of energy that enters the metal is absorbed on the way through. This is the classic Faraday cage behavior, and it is why the isolation numbers at Wi-Fi and cellular frequencies are so comfortable.

At low frequencies the picture splits in two. Sources near your bench — and at 100 kHz, “near” extends hundreds of meters — present themselves not as balanced waves but as predominantly electric or predominantly magnetic near fields. Electric near fields remain easy: a grounded conductive box terminates them almost perfectly, at any frequency, all the way to DC. If your low-frequency interference is electric-field coupling from fluorescent fixtures or open wiring, the standard enclosure stops it essentially completely, and nothing in the rest of this paper needs to worry you.

Magnetic near fields are the challenging case. A low-frequency H-field barely reflects off a nonmagnetic conductor, so the wall must earn its isolation a different way: by absorbing the field as eddy currents in the metal. How well that works depends on one number.

Skin Depth: The Number That Runs the Show

When an alternating magnetic field penetrates a conductor, the eddy currents it induces oppose it, and the field decays exponentially with depth. The distance over which it falls to about 37% of its surface value is the skin depth, δ:

δ = 1 / √(π f μ σ)

where f is frequency, μ is the permeability of the metal, and σ its conductivity. Each skin depth of wall thickness buys roughly 9 dB of absorption. The whole low-frequency story is contained in one comparison: how thick is the wall, measured in skin depths?

At 13.56 MHz the wall is nearly 100 skin depths thick — effectively infinite, which is why NFC-band isolation is never in question. At 10 kHz the same wall is not quite three skin depths, and absorption alone yields only about 24 dB. (Fortunately, eddy-current circulation around a small closed box contributes substantial additional cancellation, which is why the real numbers below are considerably better than 20 dB.) Notice the square root in the formula: each decade down in frequency makes the skin depth only about 3× larger, so the isolation degrades gradually, not off a cliff.

What a Standard Aluminum Enclosure Does at Low Frequency

Figure 1 shows calculated magnetic-field isolation for a standard welded-aluminum JRE enclosure of JRE1812 size (16.5″ × 10.5″ × 8″ interior) with a 0.090″ wall, modeled as a closed conductive shell in a uniform low-frequency magnetic field. The frequencies of common low-band applications are marked.

Figure 1. Calculated H-field isolation of a standard aluminum enclosure. Above roughly 60 kHz the metal is no longer the limiting factor; door gasket, seams and I/O penetrations set the practical ≈100 dB limit, just as they do at microwave frequencies.

Three things stand out. First, from about 60 kHz upward, the aluminum wall meets or exceeds the full product specification — the door gasket and I/O penetrations, not the metal, define the enclosure’s performance there, exactly as they do at 2.4 GHz. Second, below 60 kHz the wall-limited isolation tapers smoothly: about 93 dB at 50 kHz, 75 dB at 25 kHz, 57 dB at 10 kHz. Third — and easy to forget — these reduced numbers apply only to magnetic near-field coupling. Electric-field and radiated isolation remain at full specification across the entire range.

What This Means at Your Bench

NFC / HF RFID (13.56 MHz), AM broadcast, and anything above ~360 kHz. Full specified isolation, no special considerations. The enclosure behaves exactly as it does for Wi-Fi work.

Wireless power (85–360 kHz) and LF RFID (125/134 kHz). At and near full specification. A 125 kHz access-fob reader inside the enclosure is isolated by roughly 100 dB from the world outside — more than enough to keep a production-floor reader bank from triggering your unit under test, or vice versa.

Below about 50 kHz. Here the magnetic isolation is a real number you should design your test around: call it 55–95 dB depending on frequency. In practice this is still ample for most measurements, for a reason worth stating explicitly: at these frequencies the ambient interference in a lab is dominated by magnetic near fields from nearby sources — switch-mode power supplies, LED drivers, monitor deflection circuits, motor drives — and near fields collapse with the cube of distance. Moving a source from 30 cm to 1 m away reduces its field by roughly 30 dB before the enclosure contributes anything. Fifty-plus dB of enclosure isolation stacked on that distance falloff puts most interferers below your noise floor.

A few habits make low-frequency work in any shielded enclosure noticeably cleaner:

• Keep switch-mode supplies outside the enclosure. Power the device under test through the filtered DC feedthroughs or a PEM-series power entry module rather than placing a wall adapter inside the box — an adapter inside is a strong LF magnetic source with no wall between it and your device.
• Mind the cables. Below 100 kHz, interference rides into the enclosure as common-mode current on cable shields far more often than it leaks through the wall. Use the filtered I/O connectors, keep leads short, and dress cables against the enclosure floor.
• Know your ambient. A small shielded-loop probe and a spectrum analyzer will tell you in one minute what your lab looks like at 20–200 kHz, and whether your problem is even one the enclosure is being asked to solve.
• Verify with loops, not whips. To check enclosure isolation at low frequency, use a small transmitting loop outside and a matched receiving loop inside, and compare coupled level with the door open and closed. This is the same open/closed method described in our verification note, moved down in frequency. Keep the loops at fixed positions — near-field coupling is strongly distance-dependent, and the repeatability of the geometry matters more than its absolute calibration.

When Aluminum Truly Isn’t Enough: The Steel Option

A small class of tests genuinely needs more magnetic isolation below 50 kHz than any practical aluminum wall can provide: susceptibility testing that places an intentionally strong low-frequency field source near sensitive equipment, military and aerospace low-frequency emissions work in the spirit of MIL-STD-461 RE101/RS101, or measuring microvolt-level LF signals a bench away from a kilowatt induction heater. For these cases the answer is not simply thicker aluminum — doubling the wall gains just over 20 dB at 10 kHz but still falls well short of full specification, while the weight doubles with it. The better answer is a different metal.

Mild steel has lower conductivity than aluminum, but its magnetic permeability (μ in the skin-depth formula, several hundred times that of free space) shrinks the skin depth dramatically. A 0.075″ steel wall — slightly thinner than our standard 0.090″ aluminum — is about sixteen skin depths at 10 kHz, where the aluminum wall is not quite three: nearly 150 dB of wall-limited isolation at a frequency where aluminum provides 57 dB. Figure 2 compares the two.

Figure 2. Calculated H-field isolation, same enclosure geometry, standard 0.090″ aluminum vs. 0.075″ mild-steel wall. The steel wall holds full-specification magnetic isolation down to roughly 5 kHz.

Steel brings its own bookkeeping: it is heavier, its permeability varies with manufacturing history and drops when the material saturates in very strong fields, and at extremely low frequencies — hundreds of hertz and below — even steel tapers off, with only modest attenuation remaining at DC for an enclosure of benchtop size. No sheet-metal box of any material is a substitute for a mu-metal instrument shield when the requirement is shielding a static magnetic field. But across the 1–50 kHz band where aluminum runs out of skin depths, steel construction transforms the picture.

JRE Test builds steel-wall versions of our standard enclosures on special order. If your requirement lives below 50 kHz, contact us with the frequency, the field levels involved, and the isolation you need, and we will tell you plainly whether the standard aluminum enclosure, a steel build, or neither is the right tool.

Summary

A standard aluminum JRE enclosure delivers full specified isolation for electric fields and radiated signals across the entire low band, and full specified magnetic isolation down to roughly 60 kHz — which covers NFC, LF RFID, wireless power fundamentals, and AM broadcast testing without qualification. Below 50 kHz, magnetic isolation tapers gradually (not abruptly) to a still-useful 57 dB at 10 kHz, which combined with sensible bench practice — supplies outside, filtered I/O, cables dressed — is sufficient for the large majority of low-frequency test work. For the exceptional cases that need full isolation below 50 kHz, steel-wall construction is available on special order.

Related reading on our resources page:

• Measuring and Verifying Shielding Isolation
• Understanding Data Signals Through Filtered I/O Connectors
• Mounting Devices Inside an RF Shielded Test Enclosure