Dream 2700 | A Tailless Tale
Part III: Lets Build It!
Those who have not yet done so may want to read the first two parts of this series, then continue with this article — Ed.
Every time you build something you have designed, you get scared! Is it going to be a single prototype? Do I want to build more than one or even a small series? Is it worth to go all-in with molds? And even more tricky: is it really going to fly or will it be a disaster? What I want to demonstrate with this build? Do I want to just make a sanity check of my design, and what is the confidence level I have? Things can get quite expensive depending on the manufacturing process you choose and the quality of the result is affected as well.
Agile Methodology Applied to RC Building
For professional reasons in the last two years I’ve been exposed to the Agile for Hardware methodology (see link in Resources, below). This is a product development process that started in software development and, lately, has been succesfully applied to hardware industrial development. In a nutshell, Agile focuses on faster deliveries of products by applying a step-wise release process which minimises the risk of failure and maximises the value delivered at each step. Each incremental delivery is called an MVP (minimum viable product). An MVP includes all the main features a product should have but with a level of details and industrial readiness that grows from MVP1 to MVPx. After each iteration, leading to the release of an MVP, you can make a sanity check on your product to make sure you are still in line with the final objectives. If you see a deviation — no worries — this is the right time to stop, analyze the issue, and maybe pivot towards another solution, limiting the efforts you put into the project.
Personal note: I was introduced to the Agile methodology by my former manager and great friend, Norbert Neumann. He challenged me in embracing this new challenge in my role as an innovation manager. I fell in love with Agile. Once I tried it, I discovered there’s nothing special or magic — it is simply a very human-centered and natural way to approach issues and define priorities. I believe everyone is Agile in the way we naturally takle everyday problems in our personal life. Thanks, Norbert!
To test my comprehension of Agile, and to see how that methodology works on an ‘out of the office’ project, I applied to use it as much as I could with my building process, and some decisions has been taken according to this methodology:
- 3D-Printing for the Fuselage: Molds would be too expensive, as I may need to heavily modify the design after the first flight tests.
- Foam Core and Vacuum Bagging for the Wings: If you just need to build one prototype, the quality you can get is similar to full composites molds but much much cheaper.
These decisions would make the building process faster but I wanted as well to test some full-scale features found on full-scale sailplanes. As reported in the previous parts in this series, there is a long-term goal to build a full-scale sailplane. Therefore I decided to design and build a spar joiner system that can be representative of what needs to be done on a full-scale equivalent: there is a specific paragraph on this below.
Fuselage Build
The fuselage was 3D-printed in one single piece, and the selected material was nylon. It’s 0.6m long: not something you can print easily at home. Home 3D-printers are based on an FDM (filament deposition) process that is good enough for certain applications. In my specific case I went for SLS (selective laser sintering): this technology produces very strong and durable parts. The cost is not negligible, but way cheaper than molds. I used Shapeways, and there’s a link in the Resources section below with more information about the process.
The surface quality is very good, and only required some finish sanding and nothing else. The material thickness is just 1mm, to minimize weight. Structural stresses on the fuselage are quite low, and this is again connected to the special wing joiner design.
The internal volume is huge and therefore I designed a structure that facilitates the layout of the electronics and increases the overall stiffness. All these ribs have been printed at home in PLA material, with a thickness of 2mm.
In the next two pictures, you can see how this structure simplifies the internal layout:
Wing Build
The wing manufacturing is based on the method developed by Professor Mark Drela for the Supra TD/F3J. The wing structure is based on a solid foam core, carbon spar caps, and fiberglass skins. Hot-wire cutting all the foam cores was already a challenge. The wing is characterized by a large amount of non-linear twist from the root to the tip. With hot-wire, I can just go for linear twist variation on each foam block: therefore I decided to split the wing in six blocks, according to the twist distribution shown immediately below:
If you are interested in the cutting process, see my Dream 2700 Scale Sailplane Wings Foam Cutting video link in Resources.
The wing spar is made by cutting the wing foam core, gluing the prefabbed carbon spar caps (20mm width, 0.5mm thick) on top and bottom, and wrapping everything with 55g/m² glass at 45°. For further details, see Professor Drela’s Supra 3.4m TD/F3J Sailplane link in References.
At this point, I have the wing spars ready, but this was not the right time to glue them in place with the rest of the wing. Before doing that, I needed to finalize the wing joiner construction. The wing geometry is very complex and requires lots of alignment checks before gluing anything: I had to think twice, and after, think again before gluing!
The Challenging Wing Joiner System
As mentioned previously, I wanted to build a wing joiner system which is very similar to the one found in full-scale gliders. The design is well illustrated in the picture below:
The wing joiner core is made of lightweight balsa with vertical grain, followed by top and bottom layers of birch plywood, and finally covered with carbon rowings for the full length. It is likely structurally oversized, but don’t forget that all the bending and torsional loads are concentrated here.
I started cutting the birch plywood spar caps, and prepared a mounting frame to bend them with the right dihedral angle: hot steam was used for bending.
The next step was to glue all the balsa cores:
The balsa cores are kept together with the two wing joiners to maintain a perfect alignment. The next step was to prepare the steel rod boxes:
After gluing the steel rod boxes, I can fill the remaining space with balsa, and only now I can separate the two wing joiners:
As a last step, I glued the top spar cap in position:
It was a tedious process, but finally things were getting into shape!
The next challenge was to glue the wing joiners to the spars, making sure everything was correctly aligned. To do that, I used the female foam templates to build a jig and check the alignments: gluing was done in place with slow curing epoxy.
Before gluing the spars to the wing, I added some carbon fiber reinforcement where the spar connects to the wing joiners. The last step was to vacuum bag the wing with the spars.
Then it was time to reinforce the wing joiners with some carbon rowings, and vacuum bag for the second time:
I’m getting closer, I promise! The last step was to level the wing spar recess with balsa and foam, and to sand everything flush:
Wing Skins Vacuum Bagging Can (Finally) Start
From this point on, I return to a more conventional process, applying some reinforcements on critical areas and for the integrated flaps and elevon hinges, applying the glass fiber skins (80g/m², diagonal), and vacuum bagging everything. There’s nothing special with that process, I just applied a well known tecnique used for F3K gliders. If you want to go more in depth, I suggest Scratch Built DLG glider: Vacuum Bagging Wing in Resources.
Being the first time for me with this process, I must say the result was stunning! If you want to see the full process, see the Molding the Main Wing video linked below.
This is the first picture of the full sailplane. Note that all parts are still kept together with tape:
Final Assembly
With such a complex shape it is important to stress absolutely correct alignment of all components. The first step was to glue the wingtips (6° anhedral) to the main wing sections (6° dihedral). Wing dihedral alignment on a swept wing is really a headhache. Luckily I still had the foam beds available, and I used them as a template, adding as well a 3D-printed reinforcement jig where the dihedral change is located:
The joint was reinforced with carbon fiber cloth at 45° and everything was clamped with a soft foam tool: this provides an almost uniform pressure and a perfect finish, thanks to thin polyethylene sheet.
The most difficult part was related to the mid-span fins: I wanted them to be removable, for easy transportation, but with a very precise assembly. To do that, I glued two carbon rods inside the fins and 3D-printed the fin fairing. Two cylindrical carbon tubes where glued in position with the fairing, but kept separated from the fin. Now I have a good reference guide to drill two holes in the wing. Final operation was to glue the fairing to the wing, with the right alignment. All those steps took a lot of time, but the result paid off.
And finally, the time came for finishing and painting. The wing came out on the heavy side; I didn’t want to take any risk on structural loads. On the next prototype I can probably go lighter on both spar and skins. For that reason I decided to go for a very light finish, with just one layer of primer before the final paint. The wing surface is not shiny perfect — I’m not a perfectionist when it comes to painting.
Servo and Electronics Installation
The wing section is very thin (10% thickness), therefore I used thin servos, and I had to prepare a specific support for them since I wanted to make them to be easily dismountable. Guess what — 3D-printing helps (again).
If you want to see the details of the servo box preparation, see Servo Mounts video available linked in Resources.
For servo linkages, in that case I took inspiration from F5J standards (see see Flight Comp videos for more detail). The control rod is 1.5mm steel, and the control horns were manufactured with a sandwich of 3D-printed core and two carbon layers on both sides.
The radio equipment installation was flawless, due to the huge amount of space available in the fuselage pod. If you want to see more details, there’s a specific video — Assembly Process, below in Resources — where I show as well how the wing joiner concept works.
A Final Aerodynamic Touch: Turbulators
The Reynolds number on the wing, at the design speed of circa 11m/s varies from 50,000 to 150,000. This brings a potential risk of flow separation. For that reason I decided to implement turbulators on the upper whole surface of the wing, and on the lower surface of the wingtips: again, the easiest way to get those turbulators was to print them!
Final Considerations
It has been a really long journey. It took me way too much time to build this first prototype. I’m sure there was an easier way to do it but the pleasure not only comes from flying, at least for me. Designing and building from scratch makes you feel in close touch with your project. It’s a continous challenge in solving issues as long as you progress. Sometimes you feel very frustrated, being aware that your creature might not even fly, or crash in few seconds.
I must say I was very lucky, since the first flights are already done and the Dream 2700 flies great, with no particular issues.
In the next, and final chapter, I will provide you with a detailed report of the flight characteristics, and some ideas I want to implement on a second prototype (MVP2, according to the Agile methodology) to make it even better!
©2023 Domenico Bosco
Resources
- Le parkour by Norbert Neumann and Sarah Koch — Our guide will show how to overcome long standing rules and chart a clear path
toward a more agile and competitive company, drawing inspiration from the trend sport »Le Parkour«. - Agile for Hardware Development by Dorian Simpson and Gary Hinkle. — For those who are interested, an ebook on the Agile for Hardware process.
- Shapeways Selective Laser Sintering — “begins as thin layers of polymer powder are dispersed over the build platform. A computer-controlled CO2 laser traces the cross-section of the 3D design on the powder. It then scans each layer, fusing them all together…”
- Dream 2700 Scale Sailplane Wings Foam Cutting by the author on YouTube. — “Cutting of the first two sections of my tailless sailplane design. Results are not bad.”
- Supra 3.4m TD/F3J Sailplane by Professor Mark Drela as posted on the Charles River Radio Controllers website. — “The Supra wing is a slight modification of the Aegea wing. The sweep has been eliminated, mainly to reduce the flaps-down launch torsional loads by a factor of…”
- Scratch Built DLG Glider: Vacuum Bagging Wing by Wind & Wings on YouTube.
- Molding the Main Wing by the author on YouTube. — “A timelapse of the process I followed to mold my wings. I have to improve the leading edge finish, as i did already for the second half wing.…”
- Servo Mounts by the author on YouTube. — “In this video I will show my method for installing servos. I’ve used tailor-made servo mounts, 3D-printed in PLA…”
- Flight Comp on YouTube. — “Flight Comp is all about RC sailplanes. We are passionate about what we do and want to share it with you.”
- Assembly Process by the author on YouTube. — “This is the final assembly procedure for my sailplane rc scale model. Everything fits tight and strong, sign of a precise construction…”
All images by the author. Read the next article in this issue, return to the previous article in this issue or go to the table of contents. A PDF version of this article, or the entire issue, is available upon request.
