We often encounter questions about the accuracy of 3D printing. Since the topic is far more complex than it might seem, I decided to describe and explain this issue in some detail and answer the question of how accurate 3D printers really are. Let us start by explaining what dimensional accuracy actually is and break this concept down into several components.
Accuracy vs. precision – a brief overview of the concepts
- Dimensional accuracy (accuracy) indicates how close the average dimension from a series of 3D prints is to the nominal dimension (e.g., specified on a technical drawing or defined in a 3D model).
- Precision (also referred to as repeatability) is a measure that indicates the differences between the dimensions of successive parts in a series.
- Dimensional tolerance is the permissible range of deviation for a specific dimension (e.g., ±0.2 mm).
Distinguishing between these concepts is crucial to understanding the topic of this article. A given printing technology may be precise (i.e., repeatable) but not accurate. In such a case, all measured parts will have very similar dimensions, but these dimensions will significantly deviate from the CAD model. The opposite situation is, of course, also possible.
The main factors affecting 3D print tolerances
- Printing technology (FDM, SLA, SLS, PolyJet) – each technology has its own limitations, related for example to motion accuracy or the precision of laser beam control, as well as extrusion accuracy, material quality, and material type.
- Material – in this case, the main culprit is thermal (processing) shrinkage. Within the same printing technology, different accuracy levels can be achieved. Some materials, such as PLA or PET-G, exhibit relatively low shrinkage, while engineering materials like Nylon (polyamide) or PPS (polyphenylene sulfide) show linear shrinkage reaching up to 1% of the dimension.
- Part orientation – dimensions can have different accuracy along the Z axis compared to the XY axes, which results from the nature of the process and the layered characteristics of most additive manufacturing technologies.
- Geometry (wall thickness, holes, thin features) – shrinkage does not have to be linear; thin-walled parts behave differently than solid components with a large cross-section.
- Use of support material – removing support material or performing additional mechanical post-processing (e.g., sanding) most often affects the final dimensions. This change is not always predictable, especially in the case of manual processing.
- Machine calibration, printing parameters, and environmental conditions – even ambient temperature and humidity can affect the achieved dimensional accuracy. These deviations are usually small, but they cannot be ignored, as they represent another variable in the equation that determines the final part accuracy.
With technological progress, some of the factors listed above are increasingly being compensated for and are becoming more predictable. However, even shrinkage compensation cannot eliminate it 100 percent. An iterative approach to producing parts using 3D technologies is often necessary, along with manual compensation through changes in geometry and process parameters.
Resolution vs. accuracy – the most common mistake
Throughout my nearly ten-year career in the 3D printing industry, I have repeatedly encountered claims of accuracies on the order of a few micrometers, both from machine manufacturers and service providers (often my industry colleagues). One of the biggest myths of 3D printing technology has grown out of this. Print resolution determines only how small a feature will not be ignored by the 3D printer’s software. It is like saying that if we build a house wall using very small bricks, its dimensions will be closer to the design than if we used large concrete blocks. This is, of course, an untrue statement. Dimensional accuracy depends on the care taken during execution; in this case, the role of the contractor is assumed by the physics of the process and the factors described in the previous paragraph.
Typical (realistic) tolerance ranges in different technologies, i.e., 3D printing accuracy
Let us start with the most popular technology, FDM 3D printing. The lower limit is usually assumed to be an error of about ±0.3 mm, which is true for dimensions not exceeding 100 mm. For larger dimensions, a percentage error is most often specified, typically around ±0.2% of the nominal dimension.
Resin technologies such as SLA and DLP generally exhibit errors similar to FDM. A commonly repeated myth is their very high dimensional accuracy. Once again, we must return to the concept of resolution. These technologies offer excellent reproduction of fine details, such as surface patterns, but shrinkage during photopolymerization plays a major role. To such an extent, in fact, that achieving perpendicularity between two surfaces can be difficult with these technologies.
In the case of powder-based technologies, we are dealing with slightly higher dimensional accuracy than in FDM, but significantly worse repeatability. Accuracy is usually specified as ±0.2 mm or 0.15% of the dimension. Unfortunately, even parts printed in the same process can differ from each other by as much as 0.1–0.15 mm. The mere placement of parts within the build chamber can affect the final dimensions, and much also depends on the operator’s experience and part orientation.
Finally, I have left the accuracy of parts produced using metal 3D printing technology. This technology is often compared to CNC milling, which is a fundamental mistake. In terms of dimensional tolerances and surface quality, it is closer to metal casting. I have personally measured parts with errors on the order of 0.5–0.6 mm. Machine manufacturers in DMLS technology often specify a parameter called “best achievable tolerance,” with a value of around ±0.3 mm for parts smaller than 100 mm.
Metrology and dimensional validation
To measure and validate the above statements throughout my career, I have used calibrated calipers, micrometers, and 3D scanners. Each of these tools also carries its own measurement error. Nevertheless, we are talking about errors one or two orders of magnitude smaller than those encountered in 3D printing. It should be remembered, however, that in many cases an accuracy of ±0.2 mm is more than sufficient. This certainly does not disqualify 3D printing technology from many applications. It is simply that when accuracy measured in hundredths of a millimeter is required, conventional technologies such as electrical discharge machining or CNC machining are a much better choice.
Is it possible to achieve higher accuracy?
At this point, I would like to cite an example of a project carried out by my company. The client needed a part made of PA12 with a length of 630 mm, and the overall dimension had to fall within a tolerance of ±0.3 mm. I personally convinced the client to use a material reinforced with glass fiber, which significantly reduced shrinkage. A part produced without corrections was nearly 1.5 mm shorter after cooling. After applying corrections along two axes, it was possible to reduce the error to 0.13 mm. Repeatability across a series of several dozen parts was within ±0.04 mm. As an interesting note, the same part produced with the same corrections on a different industrial 3D printer deviated dimensionally by as much as 0.2 mm.
The dimensions were verified using a 3D scanner calibrated against a micrometer setting standard. Such standards have an accuracy on the order of ±1 to 2 micrometers. Of course, at this level of precision, ambient temperature can no longer be ignored. The 3D scanner repeatedly measured the standard with an error of approximately 0.04 mm.
Summary and conclusions
Dimensional accuracy in 3D printing technology is a more complex topic than it might seem. There is no single universal value that can be assigned to all technologies, nor even one value that can be assigned to all parts within a single technology and specific material. There are many variables that influence the final result and the achievable accuracy class.
3D printing offers a very high degree of design freedom but limited dimensional accuracy. Awareness of the limitations of additive technologies has been increasing year by year and, in my opinion, grows in proportion to hands-on experience with 3D printing in a broad sense. At our company, we carefully analyze the submitted geometry before every project. We help select the appropriate technology and material and often also assist with design optimization to achieve a compromise between accuracy, durability, and cost. For highly demanding applications, we offer inspection of the manufactured parts and the delivery of sample prints. More information about our 3D printing services can be found on the following page – custom 3D printing.
