The RF Absorbing Foam Inside Your Test Chamber: What It Does, and What It Doesn't

Open the lid of any JRE Test shielded enclosure — take the JRE 1812 as an example — and the first thing you’ll notice is the charcoal-gray foam lining the interior walls. Ask ten engineers what that foam is for, and eight of them will tell you it’s part of the shielding. It’s a reasonable guess. It’s also wrong.

The shielding — the −100 dB of isolation that keeps outside signals out and your test signals in — comes entirely from the welded aluminum shell, the door gasketing, and the filtered I/O connections. The foam contributes essentially nothing to it. What the foam does is a different job, and an equally important one: it soaks up reflections inside the chamber so your device under test sits in a quiet, well-damped RF environment instead of an echo chamber. In a JRE enclosure that lining is 3/4″ thick LS-30, a carbon-loaded polyurethane absorber, and it knocks down wall reflections by roughly 40 dB — about 20 dB on the way into the foam, and another 20 dB on the way back out after the wave bounces off the aluminum behind it.

This paper explains that phenomenon in detail: what a metal box does to radio waves when there’s no absorber (it becomes a reverberation chamber), what the foam changes, why the numbers come out the way they do, and what the foam can and cannot do for your testing. We’ll keep the math light in the main body; the full treatment lives in the appendix for those who want it.

It helps to be crisp about the two jobs, because they’re done by different parts of the enclosure:

Shielding is about isolation — keeping external RF (cellular towers, Wi-Fi access points, radar, FM broadcast) away from your measurement, and keeping your transmitter from interfering with the world outside. Shielding is the business of the continuous conductive shell: the welded aluminum body, the door gaskets, the filtered I/O plate, and the waveguide-below-cutoff penetrations. A standard JRE enclosure provides −100 dB of isolation from DC to 1 GHz, −95 dB to 3 GHz, and −85 dB to 6 GHz, and every bit of that comes from the mechanical sealing. How that works — and what it takes to hold the sealing together at millimeter-wave frequencies — is covered in our paper on using an RF test chamber at 5G and microwave frequencies.

Damping is about the environment inside the shield. A perfect shield is, by definition, a perfect mirror on all six sides. Whatever your device radiates has nowhere to go: it bounces, and bounces, and bounces again, building up standing waves that make the field strength at any given spot inside the chamber almost meaningless. The foam’s whole purpose is to break up that echo. It absorbs the energy that strikes the walls, so each reflection comes back a tiny fraction of what it was, and the chamber behaves much more like open space.

A good way to remember the division of labor: the metal keeps the outside world out; the foam keeps the inside world honest.

Consider what happens without the foam. The interior of a JRE 1812 measures 16.5″ × 10.5″ × 8″ (420 × 270 × 205 mm). At 2.45 GHz the wavelength is about 4.8″ (122 mm), so the chamber spans several wavelengths in every direction — plenty of room for wave interference to do its thing.

Your device’s antenna launches a wave. It travels outward, hits an aluminum wall, and reflects with almost no loss — bare aluminum reflects better than 99.9% of the incident power. The reflected wave crosses the chamber and reflects again. And again. Each of these traveling waves is still coherent with the original, so at every point in the chamber the direct wave and dozens of reflected copies add together — interfere — according to their relative phases.

Where the waves arrive in phase, they reinforce, and the field can build far above what the antenna alone would produce. Where they arrive out of phase, they cancel, and the field collapses into a deep null. The result is a fixed three-dimensional pattern of hot spots and dead zones: a standing-wave pattern. Move your device an inch, or change frequency by a megahertz, and the pattern reshuffles completely.

To put numbers on this, we modeled the JRE 1812 cross-section with a full-wave electromagnetic solver, using a small antenna driven at 2.458 GHz — right at one of the chamber’s natural resonances. Figure 1 (left panel) shows the result. The bare-walled chamber rings up like a bell: averaged over the interior, the field energy sits about 40 dB — ten thousand times — above what the same antenna would produce in free space, with knife-edge nulls slicing through the pattern where the reflections cancel. Sample the field along a line across the chamber (Figure 2) and you’ll see it swing more than 50 dB between a hot spot and a null just a couple of inches apart.

There’s a name for a metal box used deliberately in this mode: a reverberation chamber. Reverb chambers are legitimate test tools — with a mechanical mode-stirrer and heavy statistical averaging they’re used for EMC immunity testing. But for the everyday work a benchtop enclosure does — checking a transmitter’s power, verifying receiver sensitivity, running a throughput test on a Wi-Fi or 5G device — that environment is poison. Your measured signal level would depend violently on exactly where the device sits, which frequency channel it’s on, and whether someone’s torque wrench is lying in the corner of the chamber. Repeatability goes out the window.

The bare box also rings in time. Kill the transmitter, and the stored energy keeps sloshing between the walls for hundreds of nanoseconds — thousands of round trips (Figure 5). In radio terms the chamber has a quality factor, Q, of several thousand. For wideband modulated signals like Wi-Fi (OFDM) or 5G NR, that long decay smears symbols into one another, degrading EVM and throughput numbers for reasons that have nothing to do with the device being tested.

LS-30 is the heaviest grade of the Eccosorb® LS family: an open-cell polyurethane foam impregnated with a precisely controlled dispersion of carbon black. The carbon makes the foam slightly conductive. When an RF wave propagates through it, the wave’s electric field drives currents through billions of high-resistance carbon paths, and the wave’s energy is converted to heat — the same principle as a resistor terminating a transmission line, spread throughout a volume. (At the power levels of device testing, “heat” is figurative: microwatts. The foam is rated to 90°C and the chamber stays stone cold.)

The numbers for 3/4″ (19.1 mm) of LS-30 are dramatic. Computed from the measured material data (Figure 4):

Here’s the mechanism, one step at a time (Figure 3). A wave crossing the chamber reaches the foam-covered wall and passes into the foam. Crossing that 3/4″ layer costs it roughly 20 dB or more — 99% of its power, gone into the carbon. What’s left reaches the aluminum behind the foam and reflects, essentially perfectly, just as it would from a bare wall. But now the survivor has to cross the same 3/4″ of foam again to get back out — another 20 dB or more. What re-emerges into the chamber is down a total of about 40 dB: one ten-thousandth of the power that went in. For practical purposes, the wall has stopped reflecting.

Two refinements to that simple picture are worth knowing, and both work in your favor:

The absorption grows with frequency. As the table shows, the 20-in/20-out figure is the conservative, low-band number. By Wi-Fi frequencies the round trip through the foam is far more than 40 dB, and at 5 GHz and above the aluminum behind the foam might as well not exist — nothing that enters the foam ever comes back. This is why, as we noted in the 5G and microwave paper, the foam becomes more effective as you go up in frequency — a pleasant contrast with most things in RF, which get harder.

Not every ray takes just one trip. LS-30 is a dense, heavily-loaded absorber, and a portion of an incoming wave — particularly one arriving square-on — bounces off the foam’s front surface without ever entering it (the foam’s impedance is lower than free space, so the surface itself reflects; the appendix covers this honestly). That surviving slice isn’t absorbed on its first encounter — but it’s now just another wave crossing the chamber, and its next wall encounter runs the same gauntlet, and the next, and the next. At roughly 30% energy loss per encounter and a wall every couple of nanoseconds, even the “lucky” rays are ground down almost instantly. The one thing the chamber can no longer do is the one thing that made the bare box so nasty: build up a coherent, high-Q resonance. Which brings us to the measurable payoff.

Figure 1 shows the modeled JRE 1812 cross-section both ways — same antenna, same drive level, same 2.458 GHz resonant frequency. The two panels are on the same color scale, and the difference is night and day. With bare walls, the average field energy sits ~40 dB above free space, with peaks more than 60 dB above and deep nulls scattered through the volume — a total spread between hot spot and null of about 80 dB. With 3/4″ of LS-30 on the walls, the average energy lands within a few dB of free space — a 41 dB drop in the chamber’s stored energy — and the violent peaks and nulls collapse into gentle ripple.

That 41 dB, computed by full-wave simulation from the foam’s measured material properties, is the model’s version of the “typically 40 dB damping” figure we quote from chamber measurements. Two very different routes to the same answer.

Figure 2 makes the practical consequence plain. Slide a device along a line across the bare chamber and its received signal lurches up and down by 50+ dB. In the foam-lined chamber the same traverse shows a few dB of ripple around the free-space level. That’s the difference between a measurement you can repeat tomorrow and a number that changes when you breathe on the setup. (At microwave frequencies the residual ripple shrinks further still, since the foam absorbs even the first-surface bounce more completely.)

And Figure 5 shows the time-domain view: the ring-down. Shut the transmitter off and the bare chamber’s stored energy decays with a time constant of roughly 600 ns — a Q of about 9,400 — while the foam-lined chamber is 60 dB down in about 24 ns, a Q of a couple dozen. If you like acoustic analogies: the bare box is clapping your hands in a tiled stairwell; the foam-lined box is clapping in a recording studio. Same clap, very different room.

See it run

The same model, live in your browser. Two JRE 1812 cross-sections, one antenna at 2.45 GHz in each, identical drive and one shared color scale — the only difference is 3/4″ of LS-30 on the right-hand walls. Red and blue are the crests and troughs of the wave. It opens on the ring-down of Figure 5: a single pulse is fired into each chamber, and the readout under each panel is the energy still circulating, relative to that chamber's own peak.

Now back to the misconception we opened with. The foam does not shield, and it’s worth understanding why.

First, geometry: the foam is inside the Faraday cage. Any signal trying to enter or leave the chamber must pass through the aluminum shell, and −100 dB of metal doesn’t care whether there’s foam on the other side of it. Second, magnitude: even if you propped the lid open, 40 dB of lossy foam stuffed in the gap would be no substitute for 100 dB of continuous metal-and-gasket sealing — that’s a factor of a million difference in leakage power.

The division of labor runs the other way, too: no amount of extra gasketing will quiet the inside of the chamber, and doubling the foam thickness won’t improve the isolation. If your problem is outside interference getting in (or test signals getting out), look to the seams, the gaskets, and above all the I/O filtering — see measuring and verifying shielding isolation and the effect of adding multiple I/O filters. If your problem is ragged, position-sensitive, unrepeatable readings inside a chamber that isolates just fine — that’s the foam’s department.

A few field-tested points about living with the foam lining:

• Leave it in place. Every so often a customer removes foam to gain a little interior room. The chamber still shields perfectly — and the first sweep shows 10–20 dB ripple across the band where there was ripple of a dB or two before. The room costs more than it pays.
• Don’t cover it. Metal tape, foil labels, or a sheet of aluminum in front of the foam creates exactly the bare-wall reflection the foam was installed to prevent. Mount fixtures through the foam to the wall, or on the floor, and keep the wall areas clear — see mounting devices inside an RF shielded test enclosure.
• Respect the near field. Foam damps traveling waves that cross the chamber; it doesn’t repeal near-field physics. Keep your device and measurement antenna sensibly spaced from the walls (an inch or two of standoff helps) and from each other — our paper on antenna performance inside an enclosure covers positioning in detail.
• It’s tougher than it looks. LS-30 is rated to 90°C, and JRE enclosures use it with an anti-dust surface treatment. It doesn’t shed under normal use. If a chamber takes mechanical damage, replacement die-cut foam sets for any JRE model are available from JRE Test.

The mechanical body of a JRE enclosure — welded aluminum, continuous gasketing, filtered I/O — is the shield: it provides on the order of 100 dB of isolation between the inside and the outside world. The 3/4″ LS-30 foam lining is the damper: it turns what would otherwise be a high-Q reverberation chamber, with 40+ dB standing-wave swings and hundreds of nanoseconds of ring, into a quiet, nearly echo-free volume where the field your device sees is within a few dB of free space. It does that by absorbing roughly 20 dB of a wave’s power on the way into the foam and another 20 dB on the way back out — about 40 dB per wall encounter at the low end, and far more as frequency climbs — so reflections die in nanoseconds instead of circulating for microseconds.

Metal for isolation. Foam for quiet. Two different jobs, and a properly built chamber needs both done well.

For the full catalog of enclosures see the JRE Test product line, and for more application notes like this one, browse our resources and notes page.

This appendix uses the measured electromagnetic properties of Eccosorb LS-30 (complex permittivity ε* = ε′ − jε″ versus frequency, from the manufacturer’s material data) and standard transmission-line theory. Nothing here changes the story in the main body — it just shows where the numbers come from, including one subtlety we glossed over.

A.1 Path attenuation through the foam

A wave traveling through a lossy dielectric decays exponentially. The attenuation rate in dB/cm follows from the imaginary part of the propagation constant, γ = jk₀√ε*, where k₀ = 2πf/c is the free-space wavenumber. Selected values for LS-30:

(The 3 GHz value of 23–27 dB/cm agrees with the manufacturer’s published 24 dB/cm.) Note also the wavelength compression: at 2.45 GHz, √|ε*| ≈ 9.3, so the wavelength inside the foam is about 13 mm versus 122 mm in air — the 19 mm layer is well over a wavelength thick inside the material.

A.2 The subtlety: front-face reflection

The same high ε* that makes LS-30 so lossy also gives it a characteristic impedance well below free space: Z = Z₀/√ε*, with |Z|/Z₀ ≈ 0.13 at 3 GHz. A wave striking the flat foam face head-on therefore sees an impedance step, and the single-bounce reflection off a metal-backed slab, computed the standard way —

Zₐₔ = Zᶜ · tanh(γd), Γ = (Zₐₔ − Z₀)/(Zₐₔ + Z₀)

— comes out at only about −1.5 to −4 dB across 1–28 GHz. In other words, for one normal-incidence encounter, most of the energy bounces off the front of the foam and never samples the absorber at all. If a single specular bounce were the whole story, the foam would be a poor absorber indeed, and the simple “20 dB in, 20 dB out” picture would only describe the fraction that enters.

A.3 Why the chamber gets quiet anyway: multiple encounters and Q

Inside a closed chamber, no wave stops after one bounce. The front-face reflection loses ~30% of its energy per encounter (|Γ|² ≈ 0.70 at 2.45 GHz), and encounters come fast: the mean free path between wall hits in a JRE 1812 is 4V/S ≈ 18 cm, or about 0.6 ns of flight time. The chamber’s quality factor follows from the energy decay per bounce:

τ = (mean free path)/(c · fractional loss per bounce), Q = 2πf · τ

The stored energy in a driven cavity at resonance is proportional to Q. Collapsing Q by a factor of ~360 wipes out the resonant buildup — and on top of that, the few surviving reflections are too weak and too short-lived to organize into the deep, stable interference nulls that make the bare box treacherous.

A.4 Full-wave check

To check the whole picture at once, we solved the 2-D Helmholtz equation over the JRE 1812 cross-section (420 × 270 mm) on a 1 mm grid: aluminum walls as perfect conductors, the foam lining as a 19 mm layer with the measured ε* = 33.8 − j80.1 at 2.458 GHz, and a point source driven at a bare-cavity resonance. The results (Figures 1 and 2): with bare walls, the mean interior energy sits +46 dB relative to the same source in free space, with peaks above +60 dB, and a Q consistent with the estimate above. Foam-lined, the mean energy drops to +5 dB relative to free space, with residual ripple of roughly ±10–15 dB near the first-bounce zones at this (2.45 GHz, worst-case) frequency, shrinking as frequency and absorption climb. The difference in average stored energy is 41 dB — the full-wave model’s version of the “typically 40 dB” damping observed in chamber measurements.

So the honest accounting arrives at the same place as the simple story: at low microwave frequencies the foam’s ~40 dB of round-trip absorption (for what enters) plus rapid multi-bounce grinding (for what doesn’t) removes about 99.99% of the reflected energy circulating in the chamber; by mid-microwave frequencies the absorption term alone is overwhelming. The chamber keeps its shielding from the metal, and gets its manners from the foam.

References

• Laird Technologies, Eccosorb® LS — Lossy, Flexible Foam Microwave Absorber, datasheet RFP-DS-LS.
• Laird Technologies, Theory and Application of RF/Microwave Absorbers, Tech Note ABS-CS-RF.
• JRE Test measured material data for the Eccosorb LS series (complex permittivity vs. frequency).