Surface Roughness

The surface roughness of a part is critical to its function and long-term performance. Metal Additive Manufacturing (AM) processes alone cannot usually meet surface roughness requirements. This necessitates slow, expensive post-processing such as machining or polishing. To choose the optimal manufacturing workflow, one must understand the surface roughness capabilities of metal AM, as well as post-processing techniques and their associated time and cost.

Surface roughness is a measure of the variance in a part’s surface topology. The engineering requirements for most parts include surface roughness specifications. Roughness affects part aesthetics (e.g. shiny or matte) and mechanical behavior like crack initiation, wear resistance, fatigue life, mating, sealing, bearing, and fluid dynamics. Because metal AM processes produce relatively rough surfaces, secondary post-processing is required.  This requirement has a large impact on total production time and cost.

This post provides an overview of surface roughness including: the ranges achievable by various metal AM and secondary processes; key process parameters that influence surface roughness; and the effects of processes and parameters on total production time and cost.

Measuring Surface Roughness

Surface roughness is usually specified in Ra, which measures the surface’s average absolute deviation from its mean height. Ra can be measured physically (below, left image) or optically. Either method can provide a precise profile of surface height which is then used to calculate Ra (below, right image).

https://bit.ly/2kk5Hxa

The typical definition of Ra is less useful with large melt pool, thick layer 3D printing technologies like wire DED. This is because DED-printed surfaces can be relatively smooth in a small region but have large overall variance due to the undulating surface contour. Nevertheless, Ra is used across all technologies in this post because it serves as a useful and consistent comparison.

Causes of Surface Roughness

In 3D printing metal, there are three major contributors to surface roughness:

• Surface artifacts due to low process resolution and layering effects,
• Granular micro surface textures from melting and binding powder feedstocks, and
• Support structures, and the remnants and surface marks left by their removal.

The following 5 categories of factors are most important in determining the surface roughness that can be achieved by a metal AM process:

1. Core process resolution and precision

The resolution and precision of a 3D printing process is the most important determinant of surface roughness. Because 3D printing builds parts by layers, the process resolution can be broken down by XY and Z axis resolution.

In the XY axes, resolution is dependent on the specific mechanism of the process. In Powder Bed Fusion (PBF) processes, resolution is determined by the diameter of the laser or electron-beam.  In Binder Jetting, the resolution of the jetting process is measured in dots per inch (DPI). The resolution of wire-based process (like DED and Joule Printing™) is defined determined by the width of the deposited bead.  In the Z axis, resolution for most processes is defined by the layer thickness.  Typical XY and Z resolutions for the most popular 3D printing processes are shown in the table below.

LPBF “as-printed” surface roughness example https://bit.ly/2k4n19d

2. Material feedstock – type, size and quality

For powder-based processes (PBF and Binder Jetting), the morphology (size and shape) and quality of the material feedstock affects the surface roughness.  In these processes, the size and shape of powder grains stuck to the outside of the part impact the surface roughness. Powder feedstocks can vary from highly spherical particles as small as 5 μm, up to irregularly shaped 120 μm particles. The quality of the powder is also an important factor because low quality powders can clump, preventing proper flow and distribution in the process and further exacerbating surface issues.

3. Surface orientation with respect to process

The orientation of the surface with respect to the printing process also plays an important role in surface roughness. Below is an example of PBF surface roughness as a function of surface orientation (measured in degrees to horizontal for both upward facing (“upskin”) and downward facing (“downskin”) surfaces).

PBF surface roughness as a function of surface’s orientation in print process https://bit.ly/2lE4X60

4. Support Interface

Some metal AM processes require support structures in order to build overhangs and to anchor the part or certain features to the build plate. Where support structures connect to the part, the standard physical removal process (manually with pliers) leaves remnants that create roughness.

5. Other key processing parameters

There are many other processing parameters that can influence the surface roughness of metal AM parts. These include the power input, print speed, location in build, and cooling rates. Many parameters must be controlled and refined in order to optimize and ensure surface quality.

Joule Printing™ surface roughness example

Post-Processing to Improve Surface Roughness

Near-net-shape manufacturing processes create parts that are close to the final design but require secondary material removal to reach the desired final dimensions and smoothness. Most engineered metal parts have surface roughness requirements that exceed the capabilities of metal AM processes (and of conventional processes like casting and forging). As a result, for most applications metal AM technologies – regardless of the particular technology chosen – are near-net-shape processes. (Accuracy requirements also drive the need for secondary operations.  We will cover accuracy in a future post).

The below chart compares the range of typical surface roughness achieved in by different additive and conventional manufacturing processes. Metal AM processes produce higher surface roughness than almost all conventional processes. For this reason, post-processing techniques are generally included in metal AM workflows. The post-processing methods most widely used include CNC milling and turning, grinding, and polishing. Many conventional processes can achieve less than 1 μm Ra surface roughness whereas the smoothest surface possible from any metal AM process (Binder Jetting) is in the range of 3-13 μm Ra.

For additional context, these examples of typical engineering surface roughness requirements and capabilities can be useful:

Most engineering drawings specify an Ra surface roughness for any sealing surface (meaning preventing gases or fluids from passing) in the range of 0.05 - 0.5 μm. This roughness level is only achievable with polishing, grinding, and or the highest precision tuning and milling operations.

The standard “as-machined” (CNC milling or turning) surface roughness is ~3 μm. An additional finish machining pass can be performed to reduce the surface roughness to ~1 μm. For many engineering applications this roughness is acceptable.

There are other post-processing options, not referenced in the above chart, that are worth mentioning. For interior surfaces that cannot be reached via CNC operations, such as conformal cooling channels, there are secondary operations such as abrasive fluid machining which can smooth surfaces relatively quickly and cost effectively. There are also low cost, automated post-processing techniques like tumbling, shot peening, and centrifugal finishing that are frequently used on exterior part surfaces. However, these techniques only make subtle improvements to surface roughness, and may damage fine features and geometries like corners and edges.

If there are specific surface roughness requirements for critical surfaces where geometry and tight accuracy must be maintained, then precision CNC operations are needed. These operations are usually expensive and time consuming.

Cost and Time Implications of Post-Processing to Meet Surface Roughness Requirements

For complex parts with demanding surface roughness requirements, the production time and cost of post-processing operations can exceed that of the actual printing. Increasing the amount of printed material that needs to be removed increases post-processing time and cost, and contributes to high-cost material and energy waste. Excess printed material is most pronounced in Wire DED processes where sometimes more than half of the printed material is removed in post-processing. An example of wire DED waste is shown in the image below.

Norsk Titanium printed part through stages of machining https://bit.ly/2k1gD2l

The direct post-processing steps and associated material waste are not the only time and cost driver. Secondary CNC processes require complex programming and fixtures, which can add significant overhead to a production program. You can read more about AM and CNC lead times, economics, and workflows in our Comparison of AM and CNC blog post.

Surface Roughness as a Lever for Production Improvements

“Of the 5 major categories of factors affecting metal AM surface roughness listed above, the best candidates for improving overall results are: design, process resolution, and process speed. These three can all be leveraged to lower production time and cost.

Design optimization, including part orientation, geometry, and supports, is the most effective way to improve surface roughness without making significant changes to the core manufacturing process. An optimized design can reduce support interfaces and mitigate other surface defects, yielding time and cost savings.

Process resolution and print speed can be manipulated for time and cost savings.  However, this usually increases surface roughness. In most metal AM processes, reducing the resolution increases both surface roughness and print speed. The time and cost savings from the increased print speed can often outweigh the small increases in post-processing needed to address the added surface roughness. 

Peter Rogers, Additive Manufacturing Specialist at Autodesk, reinforces this point about PBF:

“We find that the surface quality improvement compared to build speed decrease doesn’t make for a strong commercial argument, so most companies are looking at how thick they can produce layers without compromising density. If they can achieve the density and near-net-shape, that’s perfect, as they would have been milling and polishing anyway.”

The trade-off of surface roughness for print speed applies across processes. One of the reasons the industry is excited about DED and Joule Printing™ technologies, despite their relatively high surface roughness, is their fast deposition rates. The time and cost savings during the printing process can outweigh the additional post-processing time needed to achieve final accuracy and surface roughness specs.

There are many levers for optimizing the entire metal AM process with respect to surface roughness.  The key is understanding all the dependencies and trade-offs so that the whole system can be optimized to achieve certain goals.

EPBF “as-printed” surface roughness example https://bit.ly/2lE6SYg

Conclusion

The relatively high surface roughness of metal AM processes is unlikely to improve, which means they will continue to be near-net-shape processes for the majority of applications. A successful metal AM program requires an understanding of the tradeoffs between the printing and post-processing steps.  Surface roughness is a key consideration.

When developing a metal AM workflow for an application, one must consider both the metal AM technology and its required post-processing steps.  Faster printing technologies lower printing cost, but at the penalty of higher surface roughness - increasing the amount of post-processing required.  When efficient secondary processes can meet the surface finish and resolution requirements of an application, a workflow that leverages a fast, lower-resolution printing process may yield the best combination of production cost, speed, and quality.  As highlighted in our recent post Economics of Metal AM, certain metal AM technologies are well positioned to provide this type of solution.

This post only scratches the surface (pun intended!) of surface roughness, but we hope it provides a good foundation to build upon. Future posts will provide continued perspective and education around other important technical and business considerations in metal AM. If you found this post useful, please fill out the form below to join our mailing list and receive updates on future posts in Digital Alloys’ Guide to Metal Additive Manufacturing:

We would like to acknowledge and thank the following individuals that contributed to this post:

Chelsea Cummings - Barnes Group Advisors

Joris Peels, Editor in Chief – 3DPrint.com

Peter Rogers - Autodesk

Please check out other posts in our blog series:Digital Alloys’ Guide to Metal Additive Manufacturing

Learn about our technology:Joule Printing - The Fastest Way to Make the Hardest Parts