How are defects detected in 3D-printed metal parts?

Impulse Excitation Technique (IET) detects defects by measuring resonant frequency and damping. Internal voids from porosity or lack-of-fusion reduce effective stiffness and lower the resonant frequency, while crack faces and pore walls increase damping (Q inverse) by dissipating vibrational energy. In LPBF A205 aluminum lattice structures, frequency differences reliably flagged selectively placed internal defects without the cost of CT scanning.

What quality control methods exist for additive manufacturing?

X-ray CT provides detailed 3D imaging but is slow and expensive. IET offers a complementary approach: a single tap measures the full volume of every part in seconds, delivering GO/NOGO decisions at throughputs exceeding 1,000 parts per hour. Research on PBF-LB metal parts demonstrated that IET not only sorts defective from flawless parts but also classifies parts by their manufacturing process parameters using Z-score statistical analysis.

Can non-destructive testing detect porosity in additively manufactured parts?

Yes. IET detects porosity through two independent indicators: resonant frequency drops as voids reduce bulk stiffness, and damping increases as internal pore surfaces dissipate vibrational energy. Damping is especially sensitive, often flagging porosity before frequency shows a measurable shift. In copper parts produced by 3D micro-extrusion, IET correlated elastic modulus directly with density across the 96-99% range.

What types of defects does IET detect in AM parts?

IET detects lack-of-fusion voids, gas porosity, delamination, layer defects, incomplete densification, and residual stress effects. The technique measures the cumulative mechanical consequence of all defects on the part's vibrational behavior. It does not distinguish between defect types or localize them spatially; for that, CT scanning or ultrasonic inspection is needed as a second-pass method on flagged parts.

Is 100% inspection possible for 3D-printed production parts?

Yes. IET enables 100% inspection at production speed because each part is tapped, measured, and classified in seconds with no operator interpretation required. Automated systems compare resonance signatures against reference populations and reject anomalies in real time. Any repeatable part shape works, including complex lattice structures, since IET compares resonance fingerprints rather than requiring standardized test bar geometry.

How to Detect Defects in Additively Manufactured Parts

Why AM Needs New Inspection

Additive manufacturing creates parts layer by layer, and every layer is an opportunity for something to go wrong. Insufficient laser power leaves unmelted powder between layers. Excessive energy nucleates gas pores. Contaminated feedstock introduces inclusions. The result is a class of defects, including porosity, lack-of-fusion (LOF), delamination, and trapped powder, that are volumetric, invisible from the surface, and devastating to mechanical performance. In laser powder bed fusion (LPBF) of Inconel 718, nitrogen shielding gas alone can elevate porosity and inclusion density enough to shift crack initiation mechanisms and degrade very-high-cycle fatigue life, even when the finer grain structure would otherwise suggest improved properties (Rauf et al. 2025).

The dominant volumetric inspection method today, X-ray computed tomography, delivers detailed 3D imaging but at high cost. CT scanning a single complex part can take minutes to hours, requires skilled operators, and hits physical limits with dense or large components. For production environments printing hundreds or thousands of parts per build, CT is a characterization tool, not an inspection gate. Manufacturers need a method that interrogates the full volume of every part in seconds, at a fraction of the cost. Impulse Excitation Technique (IET) fills that gap.

→ New to IET? Read What is Impulse Excitation Technique? for the underlying physics, vibration modes, and signal processing.

Key takeaway: Additive manufacturing defects are volumetric and process-dependent. IET screens the entire part in seconds, catching porosity and lack-of-fusion that surface inspection misses.

How IET Reveals AM Defects

A single tap excites a part’s natural resonance frequencies. Because those frequencies depend on geometry, mass, and elastic properties, any internal anomaly that changes stiffness or introduces energy-dissipating surfaces shifts the resonance signature. IET captures two independent indicators from every measurement, and together they catch defects that either one alone might miss.

Two Indicators, One Tap

Resonance frequency shift

Internal voids reduce effective stiffness, so a porous or defective part vibrates at a lower frequency than a fully dense reference of identical geometry and mass. IET resolves frequency to better than 0.1 ppm, so even small porosity fractions produce measurable shifts. In LPBF A205 aluminum lattice structures, frequency differences reliably flagged selectively placed internal defects in both lattice and bulk specimens (Celik et al. 2024).

Damping increase (Q⁻¹)

Internal surfaces, including pore walls, unbonded layers, and crack faces, dissipate vibrational energy through friction. Damping rises steeply even when the frequency shift is barely detectable. This makes Q⁻¹ the primary screening parameter for AM defect detection: it responds to microstructural damage that stiffness measurements alone would miss.

The physics is straightforward: defects alter stiffness and introduce dissipative surfaces, so measuring frequency and damping reveals them. IET does not localize a flaw to a specific coordinate; that remains the province of CT or ultrasonic scanning. Instead, it delivers a fast, whole-volume verdict on structural integrity, which is what production screening requires.

What IET Catches

AM processes introduce a characteristic set of defects, and IET detects them all through the same mechanism: their cumulative effect on the part’s vibrational behavior.

Lack-of-Fusion

Irregularly shaped voids between layers caused by insufficient melt pool overlap. LOF defects are among the most mechanically damaging flaws in LPBF parts because they act as stress concentrators and crack initiation sites. Their internal surfaces produce strong damping signatures that IET detects readily.

Gas Porosity

Spherical voids from trapped gas during melting, common at excessive energy density or with contaminated powder. Even when individual pores are small, their cumulative effect on bulk stiffness and damping is measurable. In LPBF IN-718, nitrogen shielding elevated porosity levels compared to argon, and these higher defect populations degraded fatigue life (Rauf et al. 2025).

Delamination & Layer Defects

Unbonded or weakly bonded layers from thermal stress, oxidation, or process interruptions. In FDM polyamide parts, IET vibration sensors detected delamination at defect sizes of 7–10 mm, while acoustic sensors better resolved defect size and position (Jabri et al. 2025). The complementary sensor modes capture different defect characteristics.

Incomplete Densification

Residual porosity in sintered AM parts (binder jetting, material extrusion, SLS) where post-processing must achieve a target density. IET tracks densification non-destructively: the same specimen can be measured after each sintering step. In copper MEX parts, elastic modulus correlated directly with density across the 96–99% range (Kolli et al. 2023).

IET does not distinguish between defect types: a frequency drop is a frequency drop regardless of whether the cause is a gas pore or a LOF void. It detects the mechanical consequence, not the metallurgical cause. Root-cause analysis still requires CT or metallographic sectioning, but for screening purposes the aggregate mechanical signal is sufficient.

Metals, Polymers, Ceramics

IET works across the full spectrum of AM materials because the physics is material-agnostic. Any solid that sustains elastic vibration produces a measurable resonance.

Metal LPBF and DED. The most extensively validated AM application. Research on PBF-LB metal parts demonstrated that IET sorts defective from flawless parts and also segregates parts manufactured with different process parameters (different laser powers, scanning speeds, wall thicknesses, and scanning strategies) using Z-score statistical analysis of resonance spectra (Obaton et al. 2023). For LPBF aluminum alloys (AlSi7Mg, AlSi10Mg), IET tracked elastic property and damping changes during heat treatment, revealing how precipitation reactions and stress relief transform the as-built microstructure (Van Cauwenbergh et al. 2018). In high-strength A205 aluminum lattice structures, geometries where CT scanning faces resolution challenges, frequency-based inspection successfully detected intentionally placed internal defects (Celik et al. 2024).

→ Solution: Process Monitoring for Consistent AM Part Quality: Z-score classification of PBF-LB parts by process parameters using resonant frequency spectra.

→ Solution: Quality Control of 3D-Printed Aerospace Lattice Structures: IET detects selectively placed internal defects in LPBF A205 aluminum lattice and bulk specimens.

Metal extrusion and binder jetting. Sinter-based AM processes produce green parts that must be debound and densified. IET provides the non-destructive feedback loop that sintering optimization demands. In pure copper processed by paste-based 3D micro-extrusion, IET measured elastic modulus across processing iterations, correlating stiffness with density as parts reached 96-99% theoretical density and 90-100% IACS electrical conductivity (Kolli et al. 2023). The same approach optimized infill patterns for copper filament extrusion, where a statistical design of experiments quantified how strand placement strategies affect final part quality (Meng et al. 2024). For Fe-6.5%Si electrical steel manufactured by filament-based MEX, IET verified densification quality as sintered parts achieved 96-99% relative density (Beretta et al. 2025).

→ Solution: Manufacturing High-Conductivity Copper with AM: IET correlated elastic modulus with density as paste-extruded copper parts reached 96–99% density and 90–100% IACS conductivity.

Polymer FDM. Research on FDM polyamide samples with controlled internal defects (0-10 mm at the neutral bending line) showed that IET detects flaws through shifts in peak frequency, damping, and amplitude. A notable finding: 3 mm defects produced an elevated peak frequency, attributed to local hardening at the defect edge, rather than the expected drop, demonstrating that defect signatures can be more complex than a simple frequency decrease (Jabri et al. 2025). Vibration and acoustic sensors responded differently to defect size and type, suggesting that sensor selection matters for polymer AM inspection.

→ Solution: Quality Control of 3D-Printed Polymer Parts: Vibration and acoustic sensors each reveal different defect characteristics in FDM polyamide components.

Process Control Beyond Sorting

Defect detection is the minimum value proposition. The deeper benefit of IET in AM is process control: using resonance data to reject bad parts, diagnose why they are bad, and prevent recurrence.

The PBF-LB classification study by Obaton et al. (2023) illustrates this directly. When eleven groups of parts were printed with systematically varied process parameters, IET classified each group according to its manufacturing settings. The resonance spectrum acts as a fingerprint of the entire build history: laser power, scan speed, scan strategy, wall thickness. A shift in that fingerprint between production runs signals process drift before it produces rejectable defects. This transforms IET from an end-of-line gate into a statistical process control tool.

The same logic applies to post-processing. Heat treatment, hot isostatic pressing (HIP), machining, and surface finishing all modify the stress state and microstructure of AM parts. Measuring resonance before and after each step isolates the effect of that step on structural integrity. If HIP reduces damping, indicating pore closure and improved bonding, the process is doing its job. If grinding unexpectedly increases internal friction, the process parameters need revision. The methodology proven in refractory, cement, and friction material industries for decades transfers directly to AM, because the underlying physics are identical (Bustos & Van den Bossche 2021).

→ Solution: Proven NDT Methods Adapted to Additive Manufacturing: How IET methodology established in refractory and cement industries transfers directly to AM quality control.

Practical Workflow

Build a Reference Population

Print a set of parts under validated process conditions and confirm their quality through destructive testing, CT scanning, or Archimedes density measurement. Measure each part with IET to establish the distribution of resonance frequencies and damping values. This reference set defines what "good" looks like for the specific geometry and material.

Set Acceptance Windows

Define tolerance bands around the reference frequency and damping values. The width of these bands depends on the application: tighter for aerospace brackets and medical implants, wider for non-structural tooling. Parts outside the window receive a NOGO decision. The Z-score method used in PBF-LB research (Obaton et al. 2023) provides a rigorous statistical framework for setting these boundaries.

Test Every Part

Each part is tapped, measured, and classified in seconds. The system compares results against the reference and returns a GO/NOGO decision automatically, with no operator interpretation required. At throughputs exceeding 1,000 parts per hour, 100% inspection becomes the default rather than a luxury reserved for critical applications.

Monitor Trends

Track frequency and damping distributions across builds over time. Gradual shifts indicate process drift (powder degradation, laser wear, gas supply changes) before they produce outright rejects. This is statistical process control applied to AM, and it catches problems that end-of-line GO/NOGO testing alone would miss.

Any repeatable part shape works. IET compares resonance fingerprints, not absolute modulus values, so production parts can be tested as-printed without cutting standardized test bars. Complex geometries, including lattice structures, produce characteristic spectra that shift predictably when defects are present.

Limitations

IET is a global method. It interrogates the entire part volume in a single measurement, which is its greatest strength for screening and its fundamental limitation for diagnosis. A frequency drop indicates that something is wrong, but not where. For defect localization, CT scanning or ultrasonic inspection remains necessary. The most cost-effective strategy uses IET as the first-pass screen, catching the majority of defective parts at near-zero cost per test, and reserves CT capacity for the small fraction that requires detailed 3D visualization.

Surface roughness typical of as-built AM parts generally does not affect IET measurements, since IET depends on bulk vibrational behavior rather than surface condition. However, heavily irregular or non-repeatable geometries, where parts differ significantly from piece to piece, can produce natural frequency variation that complicates comparison against a reference. The method works best when parts within a family share nominally identical geometry.

→ How to Assess Porosity Non-Destructively: deeper treatment of damping-based porosity measurement, including the built-in porosity calculation and comparison with Archimedes and CT methods.

→ NDT Methods Comparison: how IET compares to ultrasonic, X-ray CT, eddy current, and dye penetrant testing across all applications.