Short feeds are powder bed anomalies that occur during laser powder bed fusion (LPBF) when the recoater distributes insufficient powder across part of the build area. This creates a localized region where the powder layer is thinner than intended.
Depending on its size, depth, location, and duration, a short feed can affect the subsequent melting process and introduce risk to part quality. The challenge is not simply identifying that a short feed has occurred, but determining when it started, where it occurred, how severe it was, and how it developed during the build.
This case study evaluates the use of Phase3D’s Fringe Inspection™ to quantitatively detect and characterize short feeds during LPBF. By measuring powder bed height throughout the build, Fringe Inspection provides objective, unit-based data that can be used to identify short feeds and track their progression across consecutive layers.
Key result: Fringe Inspection identified the initiation of a short feed at layer 652. Over the following 20 layers, its measured area increased from 57.5 mm² to 135.9 mm², while its maximum depth increased from 80 µm to 260 µm.
Fig 1. An example of a localized short feed within an LPBF powder bed. The affected region is outlined in green.
The Challenge: Detecting Localized Short Feeds
During LPBF, each new layer depends on the recoater distributing a consistent layer of powder across the build area. When insufficient powder reaches a portion of the bed, the resulting short feed creates a region with reduced powder height or layer thickness.
These events can present several risks to the build:
- Localized areas may receive insufficient powder for the intended layer.
- Powder distribution can become inconsistent between layers or builds.
- Parts intersecting the affected region may be exposed to process variation that could affect part quality.
- Repeated short feeds may indicate an underlying issue with the recoating process.
Short feeds are particularly challenging because they are not always large or persistent. An event may affect only a small region of the powder bed and may only be visible for a limited number of layers.
Traditional visual inspection also introduces subjectivity. An operator may be able to identify an obvious short feed from an image, but the initiation of the event can be subtle or impossible to distinguish visually. Even when an anomaly is visible, a conventional image does not directly quantify its area, depth, or severity.
For production environments, identifying that “something looks different” is not enough. Manufacturers need objective measurements that can distinguish normal powder bed variation from a short feed that requires investigation.
The Methodology: Quantifying Short Feeds with Fringe Inspection™
To evaluate quantitative short-feed detection, Phase3D analyzed data collected from an EOS M290 using Fringe Inspection.
Build details
- Machine: EOS M290
- Material: IN718 nickel-based superalloy
- Nominal layer thickness: 30 μm
Fringe Inspection uses structured light to measure the surface of the powder bed and generate quantitative heightmap data. Rather than relying on image intensity or visual appearance, the analysis uses measured surface height to characterize the short feed.
For this study, each short-feed event was evaluated according to three primary characteristics.
Short-Feed Size
The lateral extent of the affected region was measured to determine how much of the powder bed was impacted.
This provides a quantitative measurement of the footprint of the short feed rather than relying on a visual estimate of the affected area.
Short-Feed Depth
The depth of the short feed was determined from the measured reduction in powder height relative to the expected powder bed surface.
Depth provides an indication of severity. A shallow deviation may represent normal powder bed variation, while a deeper region may indicate a more significant loss of powder.
Short-Feed Location
Each event was spatially located within the build area, allowing the short feed to be associated with specific regions of the powder bed and, where applicable, individual parts.
This spatial information is important when determining whether a short feed occurred over a critical part geometry or in an area of the build with no components.
Results: Identifying Short Feeds Layer by Layer
Identifying Short-Feed Initiation
One of the challenges with visual monitoring is determining exactly when a short feed begins. Early stages may produce only small changes in the powder bed and can be difficult to distinguish from normal surface variation. For this study, short feeds were defined as powder bed depressions greater than one layer thickness (30 μm) in depth and greater than 10 mm² in area. Additionally, these powder bed depressions were spatially constrained by only considering anomalies at the end of recoater travel, as this is where short feeds are most likely to initiate.
Using quantitative height measurements, the initiation of the short feed was identified at layer 652, when a powder bed depression exceeded the defined depth and area thresholds.
Fig 2. Heightmap of layer 652. The short-feed initiation point is shown in the red box.
Identifying the initiation layer provides engineers with a more precise point for investigating the source of the anomaly.
Tracking Short Feeds Across Consecutive Layers
Short feeds do not necessarily appear on a single layer and disappear. Depending on the recoating behavior, the affected region can persist or change over multiple layers.
For the event shown above, the short feed was detected across 1,184 consecutive layers.
Tracking the event layer by layer makes it possible to determine:
- When the short feed began
- How long it persists
- Whether the affected region grows or shrinks
- How its depth changes over time
- Which parts or regions of the build are exposed to the anomaly
This provides additional context that is difficult to obtain from individual powder bed images.
Measuring Short-Feed Growth
In addition to detecting the presence of a short feed, the measurement data can quantify how the anomaly develops. To highlight how quickly a short feed can develop, the first 20 layers following its initiation are explored in more detail below.
Fig 3. Fringe Inspection powder bed heightmaps following short-feed initiation at layer 652.
For the analyzed event, the affected area changed from 57.5 mm² at initial detection to 135.9 mm² after 20 layers, corresponding to an average growth rate of 3.92 mm² per layer. The maximum measured depth changed from 80 µm to 260 µm over the same period.
Fig 4. Maximum measured short-feed depth from layers 652 to 672.
Quantifying this progression helps distinguish isolated powder bed variation from an event that is becoming increasingly significant as the build continues. The short feed in this build continues to grow, as the affected area increased from 57.5 mm² to 600.1 mm² by layer 802. The short feed then reached a maximum area of 840.6 mm² by layer 1187.
Fig 5. Growth in the measured short-feed area from layers 652 to 802.
Impact: Moving from Visual Detection to Quantitative Short-Feed Inspection
Detecting a short feed is useful. Quantifying it makes the information actionable.
By measuring short feeds according to their size, depth, location, and persistence, Fringe Inspection provides manufacturers with objective data that can be used to determine when an event requires attention.
This creates several potential applications within the production workflow.
Automated Warnings
Quantitative thresholds can be used to automatically flag layers where measured powder bed conditions exceed established limits.
Rather than reviewing every layer manually, engineers can focus on layers containing anomalies that meet their defined short-feed criteria.
Operator Notifications
Detected short feeds can provide operators with information about where an anomaly occurred and how severe it was.
Instead of receiving a generic process alert, the operator can review measurable information including:
- Short feed location
- Affected area
- Maximum depth
- Initiation layer
- Number of consecutive affected layers
- Growth or progression of the event
Fig 6. Progression of a short feed from initial detection through subsequent layers. The affected region is highlighted in yellow to help operators quickly identify its location.
Identifying Recurring Recoater Issues
Because short feeds are measured spatially and over time, repeated events can also be analyzed for patterns.
If short feeds repeatedly occur in the same area of the build plate or at similar points in the recoating cycle, the data can help engineers investigate potential recoater, powder delivery, or process-related causes.
Investigating Affected Parts
Spatially locating the short feed also allows engineers to determine which parts intersected the affected region.
This creates a direct path from process anomaly to part-level investigation. Instead of treating every component in a build identically after a powder bed event, engineers can identify which components were exposed and focus subsequent inspection accordingly.
Fig 7. A short feed intersecting part geometry after a large melted region.
From “Something Looks Wrong” to a Measurable Process Anomaly
Short feeds illustrate a broader challenge in additive manufacturing quality control. Many process anomalies can be visible in conventional images, but visibility alone does not establish an objective definition of the event or its severity.
Quantitative surface measurements provide a different approach. In this study, Fringe Inspection was used to detect short feeds based on measurable characteristics including size, depth, location, initiation point, duration, and growth. The resulting data provides a repeatable framework for identifying short feeds and understanding how they develop throughout a build.
For manufacturers moving toward production-scale additive manufacturing, this changes the question from “Does this layer look acceptable?” to “Does this layer meet our defined, measurable process criteria?”
That distinction enables faster review, more consistent decision-making, targeted part investigation, and a clearer path toward automated in-process quality control.
Frequently Asked Questions
What is a short feed in LPBF?
A short feed occurs when the recoater distributes insufficient powder across part of the build area, creating a localized region where the powder layer is thinner than intended. Its potential significance depends on its area, depth, location, and persistence across subsequent layers.
How can short feeds be detected during an LPBF build?
Fringe Inspection uses structured light to measure powder bed height throughout the build. In this study, a short feed was identified when a localized powder bed depression exceeded the defined depth and area thresholds.
What measurements can be used to characterize a short feed?
A short feed can be characterized by its affected area, maximum depth, location within the build area, initiation layer, duration, and progression across consecutive layers. These measurements provide more objective information than visual identification alone.
Does detecting a short feed mean that a part is defective?
Not necessarily. Detecting a short feed confirms that a measurable powder bed anomaly occurred. Its effect on a component depends on factors including the anomaly’s location, severity, duration, part geometry, material, and process parameters. Additional analysis or post-build inspection may still be required.
Share: