Good luck on the next flight. Beautiful job!
I've been a bit busy and unable to work on the modular plane as much as I would have liked these past couple of weeks, but the third version of the jet is coming together, this time in a peppermint flavor.
View attachment 237808
The different colors show the different sections of the plane that bolt together using the modular connection. This most recent version includes further improvements to reduce weight and increase structural rigidity in a couple of key locations. I'm hesitant to give a timeframe for when it will fly, but I hope to test it sometime soon.
This looks amazing! I recently found out that my grandfather worked on this plane as part of Lockheed's SkunkworksI've been toying with the idea of a modular plane for a while. The two big advantages are that it would give a similar versatility to Flite Test Swappables for experimenting with different configurations, and, if a part was broken, you could just replace that part instead of rebuilding large parts of the aircraft. There are also a number of challenges to overcome, with the largest two being weight and the durability of the connections. That said, I think I've built up enough experience with 3D printed planes to make it work. If this system works as I hope it will, I plan to use it for all of my future designs. I've been working on this for quite a while in the background, but didn't want to post a thread about it until I was pretty confident I could make it work.
I have a few goals with this project.
1. Make a modular system able to be used for many different designs
2. Make it easy to print (limited use of supports, normal PLA, fits on an Ender 3 or equivalent)
3. Make it able to be used for prop or 64mm EDF or 70mm EDF
4. Make the propulsion system removable and compatible with foam board power pods (via an adapter)
5. Design it for 3s or 4s to make the batteries more affordable and safer
6. Design a number of planes for the system ranging from a trainer to warbirds to 6 channel + EDFs
7. Design connection hardware that can be printed independent of the rest of the system to allow for experimentation with foam board parts
8. Make flight-tested STLs available on this forum
9. Continue to refine the design based on experience to make it as simple as possible to use
10. Design/build data logging hardware/software that can be installed in multiple configurations
11. Actually take video of each design flying (probably onboard RunCam on the tail)
The design I am currently working with uses circular connection hardware that twists together and locks in place with two screws per joint for the fuselage. For the wings and tail I decided not to reinvent the wheel and am using connections very similar to what you'll see on commercially available RC planes. All of the fuselage sections for the initial tests are cylindrical and fit tightly around the internal components, but I plan on designing additional fuselage shapes in the future. I took some inspiration from the appearance of @Power_Broker 's modular design ( https://forum.flitetest.com/index.php?threads/large-modular-uav-design.69987/ )without really knowing the geometry of the internal structure.
I am now on my fourth set of test hardware and the first that should be flight-capable. My first design was mostly to experiment with fits and connection system geometries. It never even got printed. My initial thought was also to make the wing sections modular to allow for as much versatility as possible, but I couldn't make the aerodynamics work.
View attachment 236638
My second iteration was meant to experiment with fits for the connection hardware and got partially printed. It took me several hours to get everything to go together right for just two pieces, so it definitely needed some optimization. It was also too heavy.
View attachment 236639 View attachment 236640
Version three was the first that would probably have been flight-capable had I finished the design. However, I still wasn't happy with how everything fit together, and I did some back of the envelope calculations that left me concerned with the strength of the connections (the margin of safety was about 0.02 for a 10g maneuver which was too close for comfort). The connection hardware fit better and only took about 30 minutes to get to fit, but still required post processing with a hand saw.
View attachment 236641 View attachment 236642
Which brings me to version 4 - the current version. I changed the orientation of the print for the connection hardware which almost doubles the strength and improves the tolerances I can achieve but requires it to be printed in 6 parts instead of 2. However, it's an enormous improvement and the time to assemble a connection is about 10 minutes plus the time for the CA to harden. The only post processing required is breaking all the edges with a knife. I decided to build the first model in the shape of an F-104 (I have a reputation to keep up after all). The real advantage is the enormous amount of time I spent characterizing the design of the F-104 for my 50mm EDF and 80mm EDF versions which saves a lot of time. However, I expect a trainer jet with different wings and a tail optimized as a jet trainer to be ready to fly about a week after I first test the F-104 version. Then some prop designs will follow and then more and more varying jets.
View attachment 236643
View attachment 236645
Another big improvement over the previous versions is that the cheater inlet is now 130% FSA for a typical 70mm EDF (I went and measured all the 70mm EDFs in my possession) which means that additional conventional inlets are strictly optional which will make experimenting with widely varying designs simpler. This F-104 in particular has inlets worth 215% FSA and I didn't measure any noticeable change in thrust between having the fan installed and in free air which is a first for me. The final AUW should be about 1550 grams and the EDF (a Powerfun 70mm 4s unit running on a 3600 mAh 4s pack) outputs about 1600 grams so the final TWR should be about 1, even if the weight grows a bit (as it always does). Built light with fixed gear and 4 channels, this plane should weigh in at about 1 kg so it should also be possible to fly it on a strong 64mm EDF which I will eventually test.
Thanks! I've still just not had time to get the last few touches in place it feels like the next test is still "a week away." Hopefully soon.Good luck on the next flight. Beautiful job!
That's really neat that your grandfather worked on it. I had the chance to see one of the Starfighters Aerospace ones taking off a few months ago, and the full-scale looks even crazier in person.This looks amazing! I recently found out that my grandfather worked on this plane as part of Lockheed's Skunkworks
That's a good idea with the tufts. It would be useful to see where the wing is seeing flow separation, especially around the flaps. The cheater holes do already wrap around most of the fuselage, including around the bottom. The idea was to have the cheaters able supply the EDF even if I didn't have any additional inlet area so as to make modular construction as simple as possible.Like the cheater holes for the EDF. When the plane is nose high, the wing will cause some blockage flow, suggest next time you think about putting them so the bottom is exposed as the nose goes higher alpha. Try some tufts on the wing to show flow pattern, had trouble with a purchased F-16 and that solved it for power.
That's a good idea with the tufts. It would be useful to see where the wing is seeing flow separation, especially around the flaps. The cheater holes do already wrap around most of the fuselage, including around the bottom. The idea was to have the cheaters able supply the EDF even if I didn't have any additional inlet area so as to make modular construction as simple as possible.
View attachment 238523
That I do. But actually for this project, I have used a surprisingly small amount. Only About 3 kg between all four planes and quite a few test parts. It's a bit hard to track since each plane uses small quantities of specialty filaments like TPA and PETG.telnar , boy you must have a lot of pla
Flight video, as promised
I use Fusion 360. They have a free hobbyist license and it's pretty easy to use as far as CAD software goes. Most modern CAD programs will let you generate usable STLs though. There's a bit of a learning curve, but nothing too bad. As far as actually modeling the plane, there are a ton of different approaches and definitely there are better tutorials out there than anything I can offer. I started by buying a 3D lab print design and trying to replicate their techniques before eventually settling on methods I liked better bit by bit. My preferred method is to model the plane using solids (so the walls have thickness) since this lets me play around weight distributions more easily and it just clicks better in my brain, but a lot of people use thin walls instead since it gives you a bit better control of where the print head will go. The only thing that is particularly difficult is keeping how everything will fit together in mind. When I scratch build foam planes, I tend to just go with the flow, but that approach just doesn't work with 3D printing (see my MiG 21 project and how it didn't work at all) so you need to plan everything out meticulously pretty much from the beginning right down to the wire routing, or go through multiple stages of redesign when stuff doesn't fit or is too heavy (for example with the modular F-104 the main gear didn't quite fit when retracted in the first few versions of the fuselage and I needed to redesign the EDF mount a couple times to get the ESC to fit above it, both of which required multiple reprints). Finally, it's really easy to design parts that are a huge pain to print or are too heavy. That's why pretty much all the structure of this F-104 is at 45 degree angles and only 0.4 mm thick. Apart from that, if you're more technically minded, I use airfoil tools (http://airfoiltools.com/) to try and find the ideal airfoil for the plane (although I frequently fall back on the good old Clark Y and NACA 0012 and 0009) and Simflow (which also has a free trial version) as well as XFLR5 and XFOIL to do CFD, and AVL and XFLR5 to help evaluate flight dynamics, but none of those is even remotely necessary. If you have specific questions about parts of the plane or process feel free to ask, though, as I've said, I'm definitely no expert.I understand about the 3D printer and files needed to print it. My question to you is If I was going to design a non existent plane from start to finish, could you explain what would I need to generate the plane files so it could be printed in 3 D parts?
Here are the STLs for the F-104. Each file is labeled with material, infill, and wall thickness (e.g. aileron_hinge_L_PLA_10_infill_2_wall must be printed in PLA with a 10% infill and 2 walls). If a part does not have infill and wall thickness called out, all features are explicitly modeled and it should print fine with a wall thickness of 2 or higher and any infill setting. Everything is designed for a 0.4mm nozzle. The common parts folder contains some parts that will be common across many or all modular planes I design in the future like the mating ring inserts.
I have also included the STLs and fusion 360 files for the mating hardware in a separate zip folder called mating hardware in case anyone wants to play with making their own parts using this. The mating hardware has certainly been the biggest success of this project so far with it surviving every crash even when the rest of the plane was destroyed.
I've seen some of your work. It's really great stuff. That's a big printer you're talking about to do it all at once in 15" sections. Probably about $300 at the cheapest. But you could also print in smaller sections (up to 10") which means that what you're describing could be done on the lowest end printers. Materials wise, one roll of PLA should be plenty and goes for about $15-20Thanks for all that info, that gave me an excellent overall view of what's needed to 3D print a plane in components. I am into designing EDF ducting(inlets, y-ing it, exhaust nozzles, etc.) for the NGAD stuff. What would be the price range of a good printer that can handle 2 round(64-70 EDF's ) ducts and merge those for a single rectangular exhaust(done in segments obviously) say over 15 inches? Also, cost of material for something like that.