The Wacky Wyvern

A few weeks ago, Joe Coles of the (amazing) Hush-Kit online magazine (@Hush_Kit, https://hushkit.net/) posted a challenge to create "the craziest, most imaginative model based on a Westland Wyvern." Having always loved the crude, badass Wyvern, how could I resist that challenge?! I decided to really push the boundaries and modeled a Twin Wyvern. If you love a Wyvern, this is 2x more lovely! Two of those amazing engines and props. And over 2X the payload.

I was interested on two additional levels. I have 3D-printed airplane models over the years, but I have never "finished" one with paint and decals. Secondly, I had been wanting to get an airbrush and learn how to use it. I used this project as an excuse to experiment in both areas. I am quite happy with the results, and I am amazed at what you can do with a consumer 3D printer combined with some hand finishing. I decided to document the process in case it might help others who are thinking about this medium. Before we get into the build details, here is a shot of the finished project. More at the bottom!

This was a from-scratch effort that started with dreaming up what I wanted to build and then 3D modeling it in a CAD program. This is also referred to as "lofting" in the aerospace world. My idea is a crazy "what-if." What if you built a Twin Wyvern with two of the same engines/propellers, increased gross weight, and a much higher load carrying capacity, while maintaining similar performance? I spent a total of 5 minutes estimating the numbers. Therefore, I am in no way claiming this would work as envisioned. But it is plausible, I think.

With 2X the thrust, I think you could double the Max Takeoff Weight of the plane and maintain similar performance, assuming you enlarge the wing to keep a similar wing loading. The wing weight would go up, but the fuselage would stay roughly the same. The wing-mounted engines would give some inertia relief to the wing, so the wing weight increase would be manageable. The landing gear would need strengthening as well. All in, I think the payload capacity could more than double. With my really crude estimation, I think the plane could carry a Tallboy earthquake bomb, weighing in at 12,000 lbs. Don't quote me on this. (Recall that this is a crazy fantasy project.) Scroll to the end of this post to see the other crazy things this plane might carry.

I'll start with a little background on how to create an aircraft 3D CAD model. Aircraft lofting is one of the things I do professionally and I don't want to imply this is a simple task for a novice to pick up quickly. But don't let that stop you! It is not as hard as you would think. And like anything, your skill can develop quickly with practice. I currently use Rhino3D as my lofting program. It has tremendous capability for the price. Higher-end programs with only slightly better lofting capability (CATIA, NX) cost 15x more.

I start with a good 3-view of the aircraft which I digitize (trace) into the CAD program. These drawings are easy to find online at this point, or in books. The better the quality/accuracy of this first drawing, the better the results. If you are making a model like this for fun (as opposed to building it to fly), perfect accuracy doesn't matter much. It just needs to look right. Here is a view of the 3-view I found, and the tracing process:

This was a particularly good 3-view with a lot of detail, including fuselage cross sections and even a wing airfoil. Those help tremendously in creating an accurate loft. After the views are created, the next step is to scale and align them in 3D space. I then generate surfaces through the curves, starting with the primary surfaces (wing and fuselage skins) and then adding details like fillets, inlets, and landing gear. These smaller details usually take more time than the primary surfaces. At this point, you decide how much detail you want to model. I chose to print this at 1/48 scale, and resolution of my printer converts to about 1" at full scale. Any detail smaller than that is not worth modeling. If the purpose of the model is to generate a rendering, particularly with close-ups, you might need to model in more detail. Here is a picture of the beginning of the lofting process:

Very thin parts do not 3D print well, particularly wing and tail trailing edges. A full-scale airplane might have a trailing edge that is 1/8" thick (or less). At 1/48 scale, this becomes .0026" thick—far too thin to print. I modify the loft to have thicker trailing edges. I try to keep everything a minimum of 0.020" (.5mm) thick when printed, which results in a scaled-up dimension of one inch. Here is an example of that. To an aerodynamicist, this looks awful. But it is hardly noticeable in the finished model. Plastic model kits have the same adjustment made.

I decided to finish the "normal" Wyvern model first before moving on to the modifications for the twin. Here is that model.

I was quite happy with this. As you can see, there are a lot of little details missing. And a few big details are missing as well, like the tailpipes. I'll add these if I ever use this model in the future. But I was anxious to move onto the "Wacky" version. Here is that!

I started by scaling the wing larger. And adjusting the wing-body fairing. I lofted a new nose shape. I decided to move the four 20mm cannons from the wings to the nose. (I love the look of this!) I scaled the landing gear up to give slightly more ground clearance. I copied the original engine installation and propellers onto the wings and then blended the nacelle into an aerodynamic shape. I decided to switch the exhaust into a straight-aft exit since the engines are no longer buried in the fuselage with the need to bifurcate around the pilot. I think this new layout might increase thrust a bit as well. With the modeling complete, it was time to break it up into pieces that were printable, with added joint details. The following shows the pieces to be printed.

Time to print! I like to hold the pieces together with 1/8" brass pins. This provides accurate alignment and adds strength to the joint. Because small printed details are never dimensionally perfect, I model the holes for these pins at 0.130" diameter and the pins fit well.

I have a Prusa i3 MkIIIS printer with PLA filament. The Prusa "slicing" software which converts the CAD model to a printable file is very easy to use. I am impressed by how easy 3D printing has become. It is probably the easiest step of everything I share here.

Here are a few notes on printing. The modern FDM 3D printers are remarkably accurate, but the printing resolution is the limiting factor. The print "direction" makes a big difference on this, with the printer Z direction (the build direction) having the biggest limitation. I prefer to print wings vertically, so the layer marks are only noticeable on the wing tip. On a tip, the curvature is changing fast enough that you barely notice the print lines. On a the gradually sloping surface you get by printing normal to the wing planform, the layer marking is much more pronounced and harder to finish. The next pictures show examples of this. The first picture is my preferred printing direction. The second shows the difference in print direction. The black surface was printed in the planform direction. The yellow was printed in the wingspan direction. You can fill and fair either of course, but the vertical printing direction is much easier.

Another topic is printing resolution. There is a tradeoff between resolution and print time. The following shows the difference between the 0.2mm (left) and 0.1mm (right) settings. The printing time difference is roughly 2x. For models like this, I prefer the 0.1mm setting as it requires less finishing work, especially on small parts like propellers. This complete model took over 40 hours to print with this finer setting. I usually print overnight, and I rarely get blocked by printing time.

You might ask, why don't I print the propellers in my preferred vertical direction? I do for larger propellers. But these small blades are not stiff enough, and do not print well in the vertical direction.

A final topic is printing supports. FDM 3D printers cannot print new material in unsupported space. They need to build from a surface below. For overhung details, the software prints "support" material that is removed later. The follow picture shows this for the propeller. I generally dislike supports. But they are a necessary evil. They often don't print properly, and they can be a pain to remove, especially on small, delicate parts. They also produce a worse surface finish where they connect. Whenever possible I design parts that do not require supports. But for some things it is unavoidable. Sometimes it is also a good trade for part and joint simplicity. This propeller is a good example, as shown in this picture:

After printing, the parts need to be filled and faired to be ready for painting. There are many ways to do this, and it is an area I am still exploring. I don't sand the original parts from the printer, but instead I apply a relatively heavy coat (or several) of sandable primer. This fills the printing contours and results in a smooth surface after sanding. I usually do this in two steps, initially using 220-grit sandpaper. I then spray a lighter primer coat (if needed) and finish with a 600-grit wet sanding. This gives a very good surface for a matte finish. For metallic or glossy, I would probably go with a finer grit as the final sanding step. The following picture shows these three stages of raw, primed, and the first sanding with 220-grit. I do the initial priming and sanding before any assembly, as the parts are easy to handle. I then assemble, fill the joints, and do the final sanding at the end.

After the model was fully filled, faired, and sanded (which I didn't get a picture of), it was time to paint! I recently bought an Iwata airbrush and air compressor. There was a bit of a learning curve with this new tool, including having a bottle of bad paint with little particles in it. Being a newby, I didn't realize what the problem was and thought for a while that I was just terrible at airbrushing. Once I sorted that out, it went great! The following are some of the painting and masking pictures. I did the stripes by masking and used decals for the roundels and tail marking. I eventually want to learn to use water-slide decals. For this build I used printable vinyl stickers for the roundels, and cut vinyl for the tail. These stickers look OK for photos but are kind for thick compared to decals. My finishing effort is FAR from perfect, and I have a lot more to learn. But I was happy with the results as a first try!

And here is the final product! The first pictures show it with a Tallboy earthquake bomb hung from the belly.

I didn't have time to fabricate any other ordinance. But here are some renderings showing a few other possibilities. The Tallboy, a Little Boy atomic bomb, and a pod holding a GAU-8 Avenger 30mm Gatling gun of A-10 fame! Brrrrrrrrrrt.

It's been a fun project!

Questions or comments? Connect with me on Twitter: @joe_wilding

Hmm. What to build next?!