So why not choose SLA for the whole print if it's cheaper than SLS and has a better surface finish?
Larger sections of the model, such as the main body, would be perfect for SLA. The pieces are big enough that there is ample strength, and the high-resolution capabilities of the printer can capture the finer details and produce a smoother surface finish, requiring minimal post-processing before painting. However, the strength of the front and rear wings poses an issue when looking long-term. As mentioned above, thinner features have already been thickened; however, this can only be done so much before the model looks out of proportion. Additionally, these thin features could break when the model is being transported. As a result, a test was done to see if we printed the front and rear wings in SLA and needed additional strength. Would it be better to print a hybrid model using SLA for the main body and SLS for the two wings, taking advantage of the greater strength properties while minimizing costs, compared to printing the entire model out of SLS?

Overall, the print of the rear wing was successful, as the print's overall quality met our expectations. However, there are remaining concerns regarding its structural strength due to the fragile shape of the wing. In the last image, there is a clear deformation in the wing, which is a result of multiple different factors but they all point to the same conclusion that the wing is not strong enough using the SLA printing method. A judgement call was made to follow the hybrid model approach and print the wings using SLS. 

Problems, Problems and More Problems

Any engineering project will always have hurdles and barriers that must be overcome to accomplish the end goal. Some areas of a project's timeline are smooth, and some are full of complications. As shown above, the first section of the main body had been printed successfully, with a few defects that could be resolved in post-processing. However, upon moving ahead to the middle section, the manufacturing process started to go wrong. We faced two problems throughout this project.

1. Problems you can predict

2. Problems you can not predict

Anticipating possible issues in advance is a good practice, but predicting every situation is impractical. For instance, as mentioned earlier, it was hypothesised that the wing sections would be too weak, so the printing type was changed from SLA to SLS to enhance their strength. This is an example of data-driven engineering since we made a prediction, tested our theory and implemented the necessary changes based on the results. Despite this one example, which was predictable, we faced many problems due to our lack of experience and the two stand-out ones are explained below. This shows how, through these complications, we learned something new about the additive manufacturing process, building upon our existing knowledge.

Temperature

During the printing process, a significant issue arose due to colder ambient temperatures at night as winter approached. The finished parts showed substantial surface defects caused by these lower ambient temperatures. As each body section took an average of 31 hours to print, the printer had to run continuously, day and night, to finish one part. Due to this prolonged printing time, the resin was exposed to cooler temperatures at night, which increased its viscosity. This reduced the flow rate of the resin, which ultimately resulted in issues in final print quality.

It was a simple fix to resolve the issue. We placed a temperature-activated heater near the printer to maintain the resin at an optimal temperature range. This issue, although foreseen, provided a valuable learning experience that will help us identify and prevent potential problems in the future.

Print Weight

The weight of each subsection was the second problem. Even though the main body was divided into sub-sections and had only an internal hollow mesh structure, reducing total weight, it was still an issue. In SLA printing, the build plate is flipped upside down and dipped in and out of the bottom resin tank. This means the model must be periodically lifted from the tank against gravity, becoming heavier at each added layer. In addition to this, due to the high viscosity of the resin, any uncured resin would grip the model, creating a resistive force when pulled from the bottom resin tank. These combined points led to the model detaching from the build plate halfway through the print. This required not only restarting the print, but its weight and sharp edges pierced the FEP film in the printer, rendering it unusable until the FEP film was replaced.

This issue could have been solved by dividing the main body into smaller sections. However, this solution may not be as effective as it would require extra modifications to split the geometry further and more pre-processing for each new part. Moreover, assembling multiple separate components could lead to misalignment issues and produce an unesthetic final product, an essential requirement for this model.

Decision Time

As we encountered delays and had to reprint several times to resolve initial issues, we decided to use FDM printing to produce a backup model of the body in case the problems persisted and we couldn't meet our revised deadline. Eventually, we fully transitioned to FDM printing, which had a significantly higher success rate during its first print. However, it required more post-processing work as a downside. The images of the entire body printed using FDM are provided below.

After printing the model's body using FDM, it is necessary to perform an additional post-processing step to conceal all the visible layer lines. This step is critical because it helps achieve a smooth surface finish resembling a professionally manufactured model. We experimented with several options on test pieces before making our final decision.

In comparing XTC-3D, Humbrol, and clear lacquer spray, each has its advantages and disadvantages. In this case, the clear lacquer is the best option. This is because it can be sprayed on in a very thin layer at a time, unlike the other two options, which require a brush to be painted on, resulting in a much thicker layer. With clear lacquer, it is much easier to control the thickness of the coating, ensuring it is just thick enough to smooth all the layer lines while still being thin enough to preserve the finer details of the model.

After considering all the options, we sought advice from an experienced model painter to exchange ideas and receive valuable suggestions on the next steps. Following our conversation, we reconsidered our approach based on the painter's recommendation.

Surface Repair and Bonding

As explained earlier, the model's body and two wings were 3D-printed separately and would be assembled later. The main body was joined and glued at this point, but the front and rear wings were kept apart. They would be fully processed and painted separately, then attached to the model at the final assembly stage.

The main sections of the body were designed with dovetail joints, where each section was joined together using a transition fit. This means that the friction between the mating surfaces held together each part. After the glue had dried, filler was added to cover up any visible join lines or surface defects caused by the printing process. Finally, the filled surfaces were sanded smooth, producing a single body that appeared as if it had been printed in one go.

Priming

After repairing any surface defects, the next step is to apply primer to achieve a smooth finish while preserving the surface details. Since FDM printing can result in visible layer lines, it is necessary to use a few coats of primer. We used two different primers, as recommended by our painter. The first primer is cheaper and is used to smooth out the bulk of the layer lines, while the final coat is a higher-quality primer that creates the final surface finish.

Base Layer

To achieve the desired aesthetic outcome and streamline the painting process, a decision was made to apply a gloss black base coat to the entire body. Subsequently, a second coat was added to the windows and top surface in anthracite grey and gloss white respectively. This method offered the advantage of reducing the time required for masking off the non-black areas and adding an extra layer of paint before applying the anthracite and white, leading to a smoother surface finish.

Masking and Top Layer

Once the base coat of black paint was applied to the model, the next step involved masking off the areas not meant to be painted black. This was the most tedious part of the process, as it called for a steady hand and close attention to detail. A 0.5 mm step was 3D printed in the model to make this process easier, which helped outline the boundary between the top gloss white and body gloss black coat.

Auxialley Components

Once the main body of the model was finished and painted, attention was turned towards the auxiliary components. These included the front windshield, door window, and wheel assemblies. With the outlines of the windows masked off and the pre-processing steps completed, the final step was to apply the last layer of grey paint. 

Moving on to the wheel rims, they required two different painting processes. The first process involved using specialised rubber-like paint to give the tires a realistic look. This process was done by hand. The second process involved painting the wheel spokes and nut as a single component. The same painting process as the main body was applied to the wheel rims. The surface was repaired if necessary, and primers were used before painting the entire wheel rim with the correct grey colour.

To paint the wheel nut black, a clever approach was used to save time. Instead of masking all surfaces, a solid mask was printed and slotted into the wheel spokes, exposing only the wheel nut. The wheel nut was then painted black, the mask was removed, and this process was repeated for all four-wheel spokes.

Final Assembly and Decal Application

The last stage involved assembling and bonding all the different parts to the main body. The front and rear wings were aligned using the pockets built into the model during printing, which helped minimise any errors in the positioning and alignment.

After painting the model, it was time to add the decals to the model using transfer paper. The first step was to align the stickers with the model. This was done by placing a small piece of tape on the sticker and positioning it onto the model. Once aligned, the backing was removed, and the sticker was applied to the model. A card was used to remove air bubbles to ensure a smooth and bubble-free finish.

What we Learned

After reviewing the entire model-creation process, from the design stage to the final assembly and finishing, we discovered numerous specific issues that, if addressed, may improve the final product. However, if we treated each item as a separate problem, it would be incredibly difficult to resolve all of them, and resolving one issue may have a negative impact on another resulting in a never ending cycle. Sometimes, it's best to reduce the list to a few high-level items to make it simpler to see where most changes can be made. This will aid in optimising the strategy the next time we create a similar model.

Problem 1: Better Design Control

What do we mean by better design control? As the designer, we can control as many variables as possible, reducing guesswork in the manufacturing process. In our model, we have implemented this approach by adding a 1mm step to guide the masking between the black and white livery design and an alignment pocket for both wings. These two features have worked well, but two areas of the model could have benefited from the same treatment.

Problem 2: Better Time Allocation

As this was our first attempt at building a scale model, we anticipated some timing issues. We learned that there are two areas where we could improve in the future. Firstly, we used SLA instead of FDM, resulting in delays. Secondly, we underestimated the time required for the painting and finishing phase. 

As mentioned earlier, using the SLA method could have provided a better surface finish than FDM, reducing the time required for post-processing. We encountered too many issues during the printing process that could have been resolved if we had more time. Eventually, even with considerable post-processing, we realised that FDM worked much better than expected. We could consider using the SLA method in the future, but we may need to reduce the size of the prints and divide the main body into even smaller sections to avoid the issues we faced earlier. 

The second area that needed more attention was allocating sufficient time to finish and paint the model. Unfortunately, this task was underestimated, and the original deadline had to be pushed back due to a delay in the delivery of the front and rear wings from an external supplier. As a result, we had less time to complete the work, which rushed the process and affected the final product's quality. Our painter advised that the model needed more time to be worked on slowly and accurately, and each coat of paint required sufficient time to dry fully. In the future, we will ensure that we give ourselves enough time to complete each task thoroughly.

Summary

When completing any project, it's always valuable to take a step back and reflect on what went well and what could have been done better. While we had high expectations, there were some areas where we faced challenges. For instance, we encountered continuous print failures and delays in delivering both SLS wings, which led to the project deadline being pushed back and a sense of urgency to finish the model.

However, there were also some positive aspects that we can build upon and use again, including the decision to print the wings and wheels separately from the main body and use stickers to apply company logos instead of directly painting them onto the model surface. These approaches allowed us to make changes as we went along and saved us time in small iterations in the design and the final manufacturing.

Despite our challenges, we're happy with the final model and have learned a lot from the process. Moving forward, we're confident that we can apply what we've learned to produce even better results in the future.