I started this project because I wanted to build my own quadcopter, as many people in this hobby has done.
Instead of cuting metal tubes and screwing everything together to make a robust structure, I was thinking about something a bit more modular and with less screws. Something that could be transported anywhere easily and require no tools to prepare on-the-field.
Having a 3D printer on the desktop had opened me the doors to imagine all sorts of new pieces and mechanisms. Because I like to give a try to even the oddest idea, the project grew bigger, and lasted much more than the expected. Although I’ve spent much more hours developing it than flying it, I think the 3DMAV is a great success and I’ve learned a lot from it.
This page is intended to be a visual presentation of the major parts of the project, which can also be used as instructions to reproduce or enhance the pieces I’ve developed (which you can download from the link at the beginning of this page).
The 3DMAV (3D (printed) Micro Aerial Vehicle) is aimed to be a semi-professional quadcopter made from 3D printed pieces (aprox. 75%) and carbon fiber tubes. All its parts (arms, landing skid, gimbal, etc.) are easily detachable using built-in clips and all it fits in a backpack, along with the transmitter, batteries and the FPV system if it isn’t very bulky.
– Flight time (no payload): more than 30 minutes.
– Diameter: 1.1 m. (depends on the carbon tubes length and propellers you choose).
– Weight (no payload & no batteries): 1.1 kg.
– Max. payload: 1.5 kg.
– Most of the pieces can be 3D printed (except carbon tubes): it is easy to upgrade and repair.
– Detachable parts: build-in clips require no tools at all.
Details of my setup:
– Flight controller: APM 2.6 (external compass and GPS).
• RC system: – TX: Hitec Aurora 9 – RX: Hitec Optima 9
– Motors: RCTimer 360kv
– ESC: RCI 30A 2-5s 400~500hz
– Propellers: Gemfan 16×5.5
– Batteries: Zippy 5s 4000mAh 25C
– 12V voltage regulator for LED system and gimbal.
– Brushless gimbal from Quaternium.
– Camera: Gopro Hero3
– Video TX: 5.8GHz 800mW Airwave
– Video RX: 5.8GHz Airwave.
– Antenas: Circular polarised antenas.
The project, step by step
1- Arm connectors.
The first thing I did was the arm connectors. I wanted to make removable arms and thus a connector had to be ambeded into the arms’ carbon tubes. For this purpose I tested two different prototypes.
The first prototype’s the idea was to install the speed controllers beneath the motors, for better heat dissipation. In the design there were two bigger connectors for powering the ESC and tree smaller for signal and lighting.
Then I changed my mind and decided to place the ESC inside the frame to make the design cleaner. The new connectors have three connectors that wire up the ESC and the motors and two more connectors for the lights, all the same size.
2- Fastening system.
Designing the fastening system was very challenging. It needed to be compact, strong and durable. I spent more than a month developing the first working version.
Not working prototypes:
3- Motor mounts.
Next step was to design the motor mounts. I wanted the arms to be very slim profiled, so the objective was to reduce the motor mounts’ height. Each motor mount is divided into two pieces which are fastened to the tubes using four bolts. There are two LED light strips on each —one on the sides and one beneath—, so they can be seen from all angles. Front motor mount lights are white and rear ones are red, to get oriented when flying in 3rd person view (edit: now I would rather follow the aviation rules and make left lights red and right lights green).
This first prototype looked nice, but the PLA plastic could not withstand all that pressure arround the tube and started to crack. After one or two months flying one of them broke in flight and caused THE DISASTER… (Edit: Someone has told me that, by the time the crash took place, there was a bug in the arducopter firmware that in some cases could cause the copter to start spinning out of control. That was exactly what happened so it might not have been all about a design problem. Anyway, the new motor mounts don’t crack any more).
It took me just two days to reprint all broken pieces and assemble all. Luckily, any electronics were harmed. In this rebuild I decided to change the motor mounts design to something more resistant: (the yellow piece distributes the pressure and prevents the clamp from cracking)
4- Central frame.
The main frame consists of two plates and four pieces, which I called inter-plates pieces, that separate the two plates and firmly attach the fastening system to whole structure.
All the power electronics (power distribution, voltage regulators, ESC, etc.) are installed inside the frame. On top of it are the control electronics (flight controller, GPS, RC receiver, radio telemetry, etc.). This slight separation is intended to reduce interferences (which never caused any problem anyway).
My first idea was to place the controller on top the frame and a carbon tube would keep the GPS-Compass module away from the frame to keep it away from magnetic interferences with the electrical system. But that looked really fragile:
After a month or so I came out with an original idea; to integrate the GPS in an X dome that would also protect the electronics:
Photos and videos:
Endurance flight test with 2x LIPO 5s 5000mAh:
That’s it! It took long to develop — about five months—, but I’m sure it can be replicated in half a month with a 3D printer, electronics and a bit of patience (I have to advice that making the connectors is maybe the trickiest part, but also the coolest, in my opinion 😉 ).
The report is: perfectly stable flies in calmed wind, good look, comfortable to take anywhere and easy to repair, enhance, etc.
I was really surprised about how it performed in GPS position hold mode without any adjustment at all. While 3Drobotics states a 2 meter accuracy for their GPS module, it actually didn’t move more than a few centimeters.
Flying during windy conditions it is not so smooth though, provably because of the low RPM motors. I would go for 500 Kv motors or more the next time, 360 Kv is too low.