A feature published in 2023, initially in the May/June issue of 3D ADEPT Mag. For this discussion, “large format” refers to machines with build volumes approaching or exceeding 1 m³. This is the article’s scope, not a standardized threshold. Job titles, products and outlook refer to that period.
Producing large parts and manufacturing at high volume are distinct questions. Large-format additive manufacturing (LFAM) can serve tooling, prototypes and some repeat production. This feature examines the conditions that make these applications appropriate, focusing on software, materials and costs.
Large format depends on the process and application. Here we consider yacht-hull moulds, autoclave tooling, furniture, boats and building elements. A 600 × 300 × 600 mm machine may be large in its own segment while remaining outside this article’s scope. Producing a large component in one piece is an option; splitting and assembling it can also be appropriate.
Industrialization requires speed, precision, strength, customization and repeatability to be considered together. High deposition capacity alone does not qualify a process for an application. Three areas structure this feature: software, materials and costs.
The dossier below aims to help industrials:
- Understand the major differences between designing for Large-Format AM and designing for a “standard AM process”
- The key material considerations that could open up (new) large-scale AM applications
- Understand the common pitfalls of large-scale additive manufacturing and where exactly the cost consideration stands in the midst of all of this.
To discuss this topic, we have relied on the expertise of Justin Ferguson, Autodesk Senior Solutions Engineer for the software part, Kyle Calvert, Composite Applications Engineer at Ingersoll Machine Tools for the materials and costs standpoints, as well as Andy R. Bridge, Director of Business Development at Additive Engineering Solutions, LLC for the user perspective.
Designing for LFAM
Several process families can serve large-format applications. The following grouping distinguishes manufacturing mechanisms, in line with general AM categories, without claiming to list every installation available in 2023.
- Material extrusion: pellets (FGF), filaments (FFF), and some cementitious or clay mixtures.
- Powder bed fusion: localized powder fusion, distinct from binder jetting, which initially binds particles.
- Directed energy deposition (DED): wire or powder feedstock combined with an energy source, including lasers, electron beams or an arc in WAAM.
Material jetting and cold spray involve other mechanisms. They should not be conflated with powder bed fusion merely because some formulations use particles.
Large powder-bed machines and DED systems address different constraints. Wire-based processes, including WAAM, offer opportunities for high-volume deposition. Selection still depends on geometry, required precision, material, finishing and qualification conditions.
As far as design is concerned, knowing the different design tools in the DfAM toolbox– is one thing, making wise use of them is another one, especially when you know that each AM process and each machine comes with its share of challenges.
While he acknowledges the peculiarities of each machine and each process, Justin Ferguson from Autodesk outlines that “one task that is universally difficult, independent of process, is recognizing areas (during the design process) where there may be a collision of the printing hardware with sections that have already been printed, or non-printed items.”
A single-piece build can reduce assembly operations. It does not automatically ensure greater strength or shorter lead time: defects, residual stresses, post-processing and inspection also need consideration.
How do you develop a strategy to safely increase the scale of your 3D print?
Ferguson identifies five focal points that may help the designer to differentiate designing for LFAM from designing for a “standard AM process”: the machine itself, the shrinking of part size during or after printing, the dependence of strength upon the print direction, the division into separate planar sections as well as the need for inserted, non-printed materials.
The explanations below reflect the configurations Ferguson studied. Coordinate frames, available axes and usable size remain machine-specific. Anisotropy also occurs at smaller scales and is not exclusive to LFAM.
He details the reasoning behind each of these considerations below and shares a couple of tips to address issues like shrinking during and/or after the printing process:
“Depending on the LF-AM process, metals (DED, Directed Energy Deposition), plastics (FFF, Fused Filament Fabrication/FGF – Fused Granule Fabrication), aggregates (clay and concrete), and others, different strategies for design become critical to adopt.
- In small-scale additive, we normally have a fixed machine with a single coordinate system and a fixed build envelope. With large-scale, we might be using a robot system, or printing on top of another part, but usually, we don’t have a fixed coordinate system or a fixed build volume. While desktop printers often only move in three axes, large-scale printers often have more axes of motion, whether that’s to move on a track or tilt the tool head. These large format processes performed with multi-axis-capable machines add the ability to create features that are not dependent on overhang angle. This enables us to think outside the ‘regular AM process’ of design for additive manufacturing (DFAM). For example, building features on features, and designing without support structures becomes possible, but can introduce new complexities such as how to slice, surface prep (when building features on features), etc.
- Most LF-AM processes will see a shrinking of part size during or after printing due to conditions such as the cooling of metals and thermoplastics or drying/curing for aggregates and thermosets. Also, as LF-AM prints increase in size, upper layers ‘squish’ lower layers if they aren’t cooled or cured adequately.
If, by design, there is a stop during printing and then restart, such as printing a new feature, or allowing for adequate cooling, the part size and shape are likely to change before we continue.
In Autodesk’s Birmingham Technology Centre we use a thermoplastic pellet extruder fitted to a Kuka 6-axis spherical wrist robot. During one project, we determined that the best way to accurately continue was to over-print and then cut away to a known height using a milling spindle, then restart printing. This is a major difference for LF-AM, because separating the model into print sections, determining overbuild amounts, then the machining amount, all need to be accounted for in the design process. If hybrid (additive and subtractive in the same setup) is not an option, then the shrink amount of the print sections is very important, as is determining where to stop and for how long, so the shrink is predictable and compensatable.
- Another area that LF-AM is quite different from ‘regular AM’ relates to strength being dependent upon the print direction. In a laser powder bed-style machine, for example, the pattern that the laser sinters can change the grain structure and resulting strength. One of our customers, using a powder laser machine, found they could achieve a better grain structure if the whole part was created in one continuous path. The desired grain structure can propagate along continually with consistent speed and constant heat.
- When creating complex geometry such as a bent tube or other complex shapes, a decision could be made to divide into separate planar sections, each with a different build direction – like building a tube with wedge sections. Or an engineer might decide to change the build direction on every layer, and modulate the speed, feed, or other process parameters to achieve a non-planar layer shape.
- When considering some of the largest prints, such as homes, there is a need for inserted, non-printed materials. Windows, doors, HVAC, electrical, posts, beams, and other items may need to be accounted for in the DFAM process. If the non-printed materials are integral for the structure, such as posts or beams, determining when to stop a print to then insert these, and make sure they won’t cause collisions with the deposition equipment later, are required steps that are unique to this LF-DFAM use case.”
What happens when your part is just too large for even the largest machine?
When a part exceeds the chosen system’s capacity, splitting the model into components for assembly is one option. A different machine or strategy can also be considered. For an assembly, CAD allows alignment features to be planned and interfaces positioned with loading and finishing operations in mind.
For Ferguson, “physics really starts to play a role as parts get bigger. For instance, a wall structure will need to be designed differently if your bead is 1cm wide vs 15cm wide. Think about the weight of the part as layers increase. For DED and other processes, bottom layers can squish or wall structures can even buckle with the tremendous weight of large parts.
In the case of FFF/FGF, if printing in a non-vertical or non-planar orientation, as weight extends out there is the possibility of both compression and tension in layers. This could cause failure in prints and should be accounted for in the design process, which is much more difficult as part size increases.
Lastly, determining if support structures are needed and how they should be applied really becomes quite a difficult design question with large final part sizes, like printing buildings or rockets. If there are supports, they have usually been designed very specifically.”
Key material considerations for LFAM applications
Materials and costs have always been mentioned as the biggest challenge in AM – no matter what technology is used. These challenges are the same if not exacerbated for LFAM as applications span different technologies.
As far as materials are concerned, surface roughness, damage tolerance, inferior fatigue & tensile strength, etc. are a number of problems that may affect a 3D print. Those problems are often linked to an inconsistency in material properties. Whatever the process used, these problems can still occur when fabricating LFAM applications.
“There can always be a defect. Sometimes, when you’re printing, it’s hard to see where the defect is. For a lot of parts we make – most of our parts are for the aerospace industry -, we make a mold or a hot forming tool; those parts that will allow the customer to create the end-use part. And when those problems occur, you add a factor of safety. So, even if there is a defect that alters material properties, you still have a strong enough part at the end; and we combine that with non-destructive testing,” Calvert, Composite Applications Engineer at Ingersoll Machine Tools states. According to Ingersoll’s expert, there is a certain amount of deflection that is allowable in a mold during the fiber laying process that is allowable to still produce a high quality part. Since their printing process aims to replace metal tooling with large format additively manufactured polymer tooling, the manufacturing process should withstand the forces of the layup process without deflecting (temporarily changing shape) over a certain amount. This deflection can be measured in order to quantify if the tool meets the requirements needed for this manufacturing method.
A safety factor does not replace material and process qualification. In Calvert’s tooling example, allowable deflection, inspections and acceptance criteria need to be defined for the intended use; a detected defect is not necessarily acceptable.
On another note, it’s also possible to identify these defects as they occur by tracking the process parameters using thermal and vision systems. However, because it’s such a new technology, most machine manufacturers in this area are still in the data collection phase. As we get more data on what happens during the print process, we could use machine learning to identify the parameters that needed to be adapted and address the problem to predict future errors,” Kyle Calvert adds.
Keep in mind that Calvert shares his expertise here with polymers in mind – especially pellets used for Ingersoll’s polymer extrusion technology whose build volume is greater than one cubic meter.
Needless to say that solutions to address material inconsistency vary from one manufacturer to another, and from one technology to another.
For the production described at Additive Engineering Solutions in 2023, Andy R. Bridge reports using statistical process control during printing, followed by dimensional checks and, where relevant, vacuum-integrity tests. These practices complement material and geometry control.
Moving forward, materials that will bring a substantial advantage in terms of performance and resistance (to UV and water in particular) are likely to open new applications in the LFAM segment. To these items, Calvert adds materials with lower cost and materials with better isotropic properties. The Composite Applications Engineer lays emphasis on the fact that some of these requirements can be more stringent for certain users – like aerospace part manufacturers.
“For aerospace part manufacturers, the Coefficient of Thermal Expansion (CTE) is not quite as good as metals because the material is a polymer. When you’re trying to put the mold into an autoclave process, it works more than metal would. The better those engineering polymers get, the closer we get to replacing metals in mold making,” Calver notes.
This comparison concerns the polymer tooling discussed in the interview. Thermal expansion depends on polymer, reinforcement, orientation and temperature, rather than a universal contrast between every polymer and every metal.
Apart from these engineering requirements, it should be noted that recycled/repurposed/upcycled material formats are the next areas of interest that manufacturers would like to see expand.
A hard time for plastics, in general
“Plastics in general have a hard time being outside,” Ingersoll points out. “Material experts have been working on adding additives into the plastic so that it can better withstand UV and water, but until we get a plastic that can last many years outside, under the sun or rain, it’s going to be harder to adopt LFAM at a wider scale.”
This concern relates to the LFAM formulations considered by the contributor at that time. It does not mean that no plastic can provide durable outdoor service. UV, moisture, temperature and ageing performance need assessment for the chosen formulation and use.
In addition, cost is another key item that prevents a larger adoption of LFAM. For our expert, machine manufacturers in this niche segment currently compete with conventional manufacturing processes, wood or other cheap prototyping methods of making parts.
This means that:
- If production ramps up and the price of materials drop,
- If newly developed engineered parts with better material properties enable the manufacture of end-use parts and not just molds or tooling,
- There is a great chance that more industries adopt pellet-based LFAM technologies.
Materials used in other processes
Other process families also present scale-dependent challenges. Thermal control, stresses, material delivery, monitoring and finishing need reassessment for the intended geometry and volume, even when the alloy or polymer is already familiar.
To that, Autodesk’s Ferguson warns that “there is a trend toward larger-format powder bed machines, which requires greater care during the build preparation, toolpathing, and simulation stages because as printed parts get larger, the cost of failure dramatically increases. Predicting potential issues before starting the print, a specialization of Autodesk Netfabb, is of tremendous value”. “Netfabb also allows for the simulation of DED toolpaths for thermal and mechanical analysis. This enables users to look at the internal temperature of the whole model as it is being printed, and aid in deciding when to add breaks in the toolpath for cooling. It even can predict what the final distorted shape might look like,” he adds.
The software capabilities cited are those discussed in the 2023 interview. Distortion and temperature predictions depend on models and input data; they need comparison with measurements and do not alone guarantee part conformity.
That said, Fusion 360 has other tools that can aid in LF-AM as well, for plastics and other materials that can have complex infill patterns. The software provider recently introduced Volumetric Latticing, a tool that enables users to create infill structures that can provide strength while using less material, meeting this way cost, weight, and/or sustainability requirements.
The common pitfalls of large-scale additive manufacturing and other cost considerations
Just like AM in general has led many industrials to have a misleading appreciation of the technology, LFAM processes may often induce disguised expectations for those who are not familiar with the different technologies.
Based on insights shared by Ferguson, Calvert and Andy R. Bridge, we can already point out the pitfalls below:
- LFAM complements other production methods and may replace some for a defined application. Comparison should include manual, subtractive and additive approaches suited to the need.
- Robotics do not make an installation ready for production without preparation. Training, programming, safety, qualification and maintenance remain necessary even with greater automation.
- Bigger prints and more material mean it’s expensive to have a print failure. Process test prints to fully understand your process, try to know your problems before you have them. Determine if any bad sections can be cut out and/or repaired.
- In the projects discussed in 2023, some systems exceeding about 1 m³ are customized or still in development. Maturity, software, support and operating conditions need assessment system by system.
- Not knowing a process well enough could be seen as a pitfall, costing more (waste), introducing stress concentration points, under-printing or over-printing, etc. Material flow rate out of a nozzle, for example, adds its own difficulty and isn’t specific to one process. In a lot of cases, it is simply not an option to pause or change the extrusion speed on the fly, or it’s very wasteful to do so.
- It’s worth keeping this in mind: in addition to the capital costs for the Large Format printing equipment, there are many other areas of cost to consider. Some of these are software, polymer expertise, design expertise, process control, engineering investment, and training. Certain end-use products like aerospace tooling also require large-scale precision CNC mills.
- Last but not least, sometimes, it’s not about costs, it’s about lead times.
In 2023, the contributors described applications ranging from prototypes to small series in aerospace, marine and construction, at different maturity levels. Progress toward regular production depends on application, process control and organization of the complete chain.
What we find the most interesting is that there is momentum in the construction industry –where there is a need to reduce man-hours – and applications in this field reveal that concrete 3D printing is not always the only production candidate that can be used to achieve that.
This dossier has first been published in the May/June edition of 3D ADEPT Mag.
A few words on the contributing companies
Additive Engineering Solutions (AES) is an advanced manufacturing company that provides large format polymer additive manufacturing (LFAM) solutions. Whether it’s 3D printing large tools and molds for the composite aerospace & defence industry or printing large end-use parts for the AUV markets; AES ambitions to demonstrate the ability to manufacture a variety of products with more efficiency and customization than traditional methods. The company is looking to explore vitrimer materials in the near future for LFAM applications.
Autodesk is a software company that provides several solutions that can assist in LFAM, ranging from simple 2.5D slicing to multi-axis, non-planar slicing, the ability to drive desktop machines, hybrid machines, and robots, and the ability to simulate metal prints in order to detect possible issues before getting on the machine. Autodesk offers software that covers most of the additive space, including LFAM. One issue seen across industries is the idea that data is separate, usually programming in a separate software from the design. For additive, especially LFAM, having to switch back and forth from separate design and programming packages for design changes means a lot of exporting and importing. The company’s software solutions are designed to eliminate this issue, enabling a more seamless workflow.
Ingersoll described gantry and robotic large-format polymer printers and automated fibre-placement solutions in 2023. Depending on configuration, printing, milling, automated fibre placement (AFP) or tape laying (ATL) may be offered. The company also described contract manufacturing for aerospace tooling, marine applications and furniture; these functions should not all be assumed to be present on every machine.
THE 3D ADEPT WEEKLY NEWSLETTER
The industry moves fast. Stay ahead.
Every week, the latest additive manufacturing news, in-depth features and real-world applications, straight to your inbox. Plus, be the first to know when a new issue of 3D ADEPT Mag is out.
Free · One email a week · Unsubscribe anytime
Clicking loads our email provider’s sign-up form, including its analytics and Google’s anti-bot protection. You’re only subscribed if you complete the sign-up. Read our privacy policy.