EMI Shielding Fails When the Material Number Is Treated Like an Enclosure Rating

EMI shielding often looks like a material selection problem.

Pick copper.
Pick aluminum.
Pick steel.
Make the wall thicker.
Assume the interference is solved.

That is where many shielding mistakes begin.

A metal sheet can calculate to an impressive shielding effectiveness value in dB. But that number usually describes a continuous material barrier, not a finished enclosure with seams, fasteners, ventilation slots, cable penetrations, gaskets, paint, bonding issues, and real installation geometry.

For a first-pass material check, the shielding model is built around three terms:

SE = A + R + B

Where:

SE = total shielding effectiveness in dB
A = absorption loss in dB
R = reflection loss in dB
B = multiple-reflection correction in dB

That decomposition is useful because it shows where the attenuation is actually coming from. A design may depend mostly on reflection. Another may depend mostly on absorption. A thin barrier may lose some performance because of multiple internal reflections.

The mistake is looking only at the final dB number and assuming the whole product will perform that way.

The first thing to check: skin depth

Skin depth is the distance into a conductor where the electromagnetic field falls to about 37% of its surface value.

The screening formula is:

δ = 1 / √(π × f × μ × σ)

Where:

δ = skin depth
f = frequency
μ = material permeability
σ = material conductivity

Skin depth matters because absorption loss depends on how many skin depths the shield thickness represents.

A simple absorption relationship is:

A = 8.686 × t / δ

Where t is barrier thickness.

That means one skin depth gives about 8.7 dB of absorption loss. Two skin depths give about 17.4 dB. Three skin depths give about 26.1 dB.

The relationship is simple, but the consequences are important.

At higher frequencies, skin depth becomes smaller, so even a thin metal layer can absorb strongly. At lower frequencies, skin depth can be much larger, so a thin shield may provide less absorption than expected.

This is especially important for low-frequency magnetic fields from transformers, motors, loops, and high-current conductors.

The common mistake: confusing material SE with enclosure SE

A continuous sheet of metal can look excellent in calculation.

But a real enclosure is not a continuous sheet.

It has openings.

It has seams.

It has cable entries.

It may have painted surfaces that interrupt conductive contact.

It may have poorly bonded panels.

It may have a shielded cable terminated incorrectly.

Any one of those paths can dominate the actual EMI leakage.

That is why a material shielding result should be treated as a first-pass material estimate, not a final enclosure rating.

The calculator can help answer:

Is this material and thickness theoretically strong enough for the frequency?

It cannot prove:

Will the assembled enclosure pass EMI testing?

Those are different questions.

A practical example

Suppose an engineer is checking a thin aluminum barrier for plane-wave shielding.

Inputs:

Material = aluminum
Thickness = 25 µm
Frequency = 1 MHz
Field type = plane wave
Target shielding effectiveness = 100 dB

At this frequency and material condition, the skin depth is about:

δ ≈ 84.6 µm

The barrier thickness is:

t = 25 µm

So the shield is only:

t / δ = 25 / 84.6 ≈ 0.30 skin depths

Now calculate absorption loss:

A = 8.686 × 0.30

A ≈ 2.57 dB

That is not much absorption.

The barrier is thin relative to skin depth, so most of the theoretical shielding result must come from reflection, not absorption.

In this example, reflection loss is very large:

R ≈ 105.99 dB

But because the shield is thin, the multiple-reflection correction matters:

B ≈ −4.12 dB

Now total shielding effectiveness is:

SE = A + R + B

SE = 2.57 + 105.99 − 4.12

SE ≈ 104.4 dB

Against a 100 dB target, the result appears to pass with:

Margin = 104.4 − 100

Margin = 4.4 dB

That is a moderate first-pass material margin.

But this is exactly where engineering judgment matters.

A calculated 104.4 dB material value does not mean the finished enclosure will deliver 104.4 dB. A slot, seam, cable penetration, or poor ground bond can reduce real-world performance dramatically.

So the correct interpretation is not:

“This enclosure is good for 104 dB.”

The correct interpretation is:

“This continuous aluminum barrier has enough theoretical material SE for this plane-wave case, but the enclosure details still decide the real result.”

Why low-frequency magnetic shielding is harder

Low-frequency magnetic fields are often the most difficult shielding case.

For plane waves and electric near fields, reflection can be very helpful because the impedance mismatch between the wave and the metal is large.

For magnetic near fields, the wave impedance is low. Reflection may be weak or even floored at zero in a simplified model. That means the design depends more heavily on absorption, permeability, thickness, and geometry.

This is why a thin copper or aluminum shield may work well for high-frequency electric-field noise but perform poorly against a nearby transformer or motor magnetic field.

A common mistake is using the same shielding intuition for every EMI problem.

That does not work.

The source type matters.

The frequency matters.

The distance from the source to the shield matters.

The material permeability matters.

The geometry matters.

The near-field distance mistake

Another common mistake is entering the wrong distance for near-field shielding.

For near-field calculations, the relevant distance is the distance from the interference source to the shield surface.

It is not the cable length.

It is not the enclosure size.

It is not the distance from the observer.

If the wrong distance is used, the wave impedance estimate can be wrong, and the reflection-loss result can become misleading.

The far-field boundary is:

r_t = λ / 2π

or:

r_t = c / (2π × f)

Beyond that distance, the plane-wave model becomes more appropriate.

This matters because a source that is “near field” at one frequency may be effectively far field at another frequency and distance.

The thickness mistake

Required thickness should not be rounded down casually.

If a calculation says the minimum thickness is 0.208 mils, choosing a slightly thinner material because it is convenient may push the result below the target.

That is why a reverse thickness calculation should be rounded up, not down.

The same logic applies to coatings, foil laminates, plating, and thin conductive films. Bulk-metal formulas may not represent those materials accurately, especially when the film is discontinuous, oxidized, cracked, laminated, or poorly bonded.

The engineering takeaway

EMI shielding should not be judged from one impressive dB number.

A good first-pass review should ask:

What is the interference frequency?

Is the source plane wave, electric near field, or magnetic near field?

What material is being used?

What are the relative conductivity and permeability?

What is the skin depth?

How many skin depths thick is the barrier?

How much of the result comes from absorption?

How much comes from reflection?

Is the calculated value material SE or finished enclosure SE?

Are seams, apertures, cable penetrations, gaskets, and bonding being handled separately?

The dangerous part is that the calculation can look excellent while the real enclosure still leaks.

That does not make the calculation useless. It makes the interpretation important.

A shielding effectiveness calculation is a screening tool. It helps compare materials, estimate required thickness, and understand whether the physics is working in your favor. But final EMI performance still depends on enclosure construction and measurement.

For a fast first-pass review, use the EMI Shielding Effectiveness Calculator — Material SE in dB to estimate shielding effectiveness, skin depth, absorption loss, reflection loss, multiple-reflection correction, and required thickness before moving into enclosure design or compliance testing.