Why Your Enclosure Rings & How to Predict It

Every Metal Box Is a Resonant Cavity

Ever had a product sail through radiated emissions testing on your bench, then absolutely crater once you get it into the chamber? There's a decent chance chassis resonance is biting you. Here's the thing most people forget: every closed (or nearly closed) metallic enclosure behaves exactly like a resonant cavity. Same physics that heats your lunch in a microwave oven. At specific frequencies, the internal dimensions of your box line up with half-wavelength multiples of the electromagnetic field, standing waves form, and suddenly energy at those frequencies gets amplified instead of shielded. Any slot, seam, or cable penetration turns into a surprisingly efficient antenna.

Figuring out where those resonances sit should be one of the first things you do when you're laying out a new product enclosure. The open the Chassis Resonant Frequency calculator makes this literally a 10-second exercise, and you'd be surprised how many headaches it can prevent.

The Governing Equation

A rectangular metallic cavity supports transverse-electric (TE) and transverse-magnetic (TM) modes. The resonant frequency for the $\text{TE}_{mnp}$ (or $\text{TM}_{mnp}$) mode is:

where $c$ is the speed of light ($\approx 3 \times 10^8$ m/s), and $L$, $W$, $H$ are the interior length, width, and height of the enclosure in metres. The integers $m$, $n$, and $p$ tell you how many half-wavelength variations you've got along each axis.

For TE modes, at least two of those three indices have to be non-zero. In a typical enclosure where $L > W > H$, you'll usually see the lowest-order modes being $\text{TE}_{101}$ and $\text{TE}_{110}$. The calculator reports both and tells you which one gives you $f_{\text{min}}$ — the frequency where your trouble starts.

Why It Matters for EMC

At resonance, the shielding effectiveness of your enclosure can drop like a rock — we're talking 20 to 40 dB compared to off-resonance performance. If a digital clock harmonic or a switching-converter spur happens to land right on one of these cavity modes, you'll see emissions spikes that no amount of ferrite beads or input filtering will fix. The box itself is the problem.

I've seen this play out in a few different ways. You get unexpected radiated emissions peaks at frequencies that have no obvious connection to anything on your PCB. Or you've got coupling between boards in a multi-board enclosure, where one board's noise excites a cavity mode that couples straight into another board's sensitive analog front end. Test results become maddeningly inconsistent — move a cable slightly or reposition a PCB and suddenly the measured amplitude changes by 10 dB.

Most engineers skip thinking about this until they're already in the test lab, and they regret it.

Worked Example: A Typical Industrial Controller Enclosure

Let's walk through a real-world example using a standard extruded-aluminum enclosure. Interior dimensions are:

• $L = 250\text{ mm}$ (0.25 m)
• $W = 150\text{ mm}$ (0.15 m)
• $H = 50\text{ mm}$ (0.05 m)

These are pretty common dimensions for an industrial controller or a small instrumentation box. Let's calculate the first two resonant modes.

TE₁₀₁ Mode

TE₁₁₀ Mode

So the lowest resonant frequency is about 1.17 GHz, and it's set by the $\text{TE}_{110}$ mode. The corresponding free-space wavelength is:

Now here's why this matters: 1.17 GHz sits squarely in the range scanned during CISPR 32 / FCC Part 15 radiated emissions testing, which typically runs up to 6 GHz for many product classes. If your design has any digital clock harmonics, high-speed serial links like USB 3.x or PCIe or HDMI, or switching converters with spectral content anywhere near 1.17 GHz, this enclosure will amplify those signals instead of attenuating them. You'll see a big fat spike right at the resonance, and you'll be scratching your head wondering where it came from.

Plug these same numbers into the open the Chassis Resonant Frequency calculator and you'll get the results instantly, along with the wavelength at $f_{\text{min}}$. Saves you from doing the arithmetic by hand every time.

Practical Design Strategies

Once you know where the resonances are, you've got several options for dealing with them. Some are easier than others depending on where you are in the design cycle.

Change the enclosure dimensions. This is the cheapest fix if you catch it early. Even a 10 to 15 percent change in one dimension can shift the resonance away from a problematic frequency. If you're still in the CAD stage, this costs you nothing. If you've already cut metal, well, it's expensive. Add absorber material. Placing RF-absorbing foam or loaded elastomer on an interior wall damps the Q of the cavity, which reduces the resonance peak. You see this a lot in high-frequency enclosures above 1 GHz. The absorber doesn't eliminate the resonance, but it takes the edge off and can buy you 10–15 dB of margin. Just make sure the absorber material is rated for your operating temperature range. Partition the enclosure. Internal walls or shields break one large cavity into smaller ones, which pushes the lowest resonance higher in frequency. This can be as simple as a grounded metal divider between two sections of your PCB. The trick is making sure the partition is well-bonded electrically to the enclosure walls — a few screws aren't always enough at GHz frequencies. Manage apertures deliberately. A resonant cavity radiates most efficiently through slots whose length approaches $\lambda/2$. Keeping seam lengths and ventilation slots well below $\lambda_{\text{min}}/2$ is critical. If your lowest resonance is at 1.17 GHz, you want slots shorter than about 128 mm. Longer than that and you're asking for trouble. Relocate noise sources. Standing-wave patterns have nulls and maxima at predictable locations inside the cavity. If you can't move the frequency (because it's tied to your clock tree or your power supply topology), you can sometimes move the physical source to a field null. This requires some EM simulation or a lot of trial and error, but it can work when you're out of other options.

Quick Sanity Check Rule of Thumb

For a fast mental estimate, some people use this approximation for the lowest resonance:

where $L_{cm}$ and $W_{cm}$ are the two largest interior dimensions in centimetres, assuming $H$ is much smaller. For our example: $\sqrt{25^2 + 15^2} = \sqrt{850} \approx 29.2$, giving $f \approx 150/29.2 \approx 5.14$ GHz. Wait, that doesn't match. That's because this approximation is actually estimating the half-wave resonance along the diagonal, not the proper cavity mode. The real cavity calculation (as shown above) gives 1.17 GHz, which is way different.

The lesson here: use the real formula, not shortcuts, especially when compliance is on the line. Rules of thumb are great for cocktail-party estimates, but they'll steer you wrong when you're trying to debug a failing test.

Try It

Before you finalize your next enclosure design — or if you're currently debugging a mysterious emissions peak that doesn't line up with anything obvious on your schematic — open the Chassis Resonant Frequency calculator and plug in your box dimensions. Takes about ten seconds and could save you a very expensive re-spin or a week of frustrated troubleshooting in the test lab. Pair it with a shielding-effectiveness calculation or an aperture-leakage estimate for the full picture of how your enclosure will actually behave once you get it into the EMC chamber.