3D Robotics


From Matterhackers, a profile of the making of Freebird One, a Pixhawk-powered commercial drone. 

Technology and the latest gadgets have always been an interest to me, but they were confined to hobbies during my limited free time as a financial analyst for nearly 20 years. I decided to take some time away from that world in early 2014 after  the 2008 financial crisis effects sapped the industry of a lot of the bright minds that had attracted me to it in the mid-‘90s. I was hoping to unleash my creativity in developing something new where I could make a difference, and I was excited to finally have some time to do that.

Naturally, I became interested in 3D printing, and bought a Cubex printer. I began teaching myself how to do basic prints with the out-of-the-box software, but I soon discovered Fusion 360 from Autodesk and was quickly making more complicated designs. First on the list were custom child-safety items like protected wall plates, knobs for stoves, and doors to make the house safer for our twin boys who were 1-year old at the time. I found I could make these items in just the right dimensions to fit the hardware in our house in contrast to generic store bought items that typically fell short of being truly useful.

A couple months later, I got a small drone, another new technology I was interested in, and which were just becoming more widely available. A few weeks after that purchase, like most new drone owners, I crashed it in our back yard and broke one of the arms. As I was trying to locate replacement parts, I glanced over at the printer and it occurred to me I could probably make replacements. Having successfully done that, and learning more about the drone technology in the market at the time, a couple of months later, I thought I’d try to develop something better than what was available – in particular something larger that could be used for all sorts of applications – not just video – and that was more weather tolerant. That was summer 2014.


By early fall, I had a working prototype – made mostly from ABS and PLA. I had also upgraded to a Lulzbot TAZ 4 printer, which opened up a whole new universe of materials that I couldn’t use with the proprietary Cubex platform. That’s when I discovered MatterHackers, which quickly became my single source for filament. In October or November 2014, they introduced Proto Pasta’s carbon fiber filament, which I quickly tried. That was perhaps the biggest enabler for me to develop the current design. I could now print much thinner, much faster, and with absolutely no warping. The design at the time, and which I stuck with for a couple more months after the carbon fiber was introduced, more closely resembled a tradition quadcopter with four arms. Even with the carbon fiber, I still had problems with arms breaking off over time because of the heavy load (up to 25 pounds) being carried by the arms in conjunction with vibration from the motors.

As I became better at printing the carbon fiber and improved my CAD design skills, I decided to overhaul the design completely in order to print even thinner parts while at the same time evenly distributing the weight of the UAV and stress from the motors’ torque all around the frame. The current enclosed frame design came to me one day – I still remember – on a Friday in early January, just about a year ago. The whole picture was in my head. It was to be an airframe that enclosed the blades (for safety), but in a way that actually increased the strength of the overall frame by distributing the vehicle’s weight and motor torque stress evenly to avoid concentration of stress at just four points – something I now call a SurroundFrame. Even until now, I have not seen an enclosed frame design that is anything more than a body cover that sits on top of a traditional frame with four arms – or a variation that has partial enclosures on the outer corners. In any case, while helpful from a safety perspective, they represent added weight that reduces battery and flight time. It took me exactly a week to design and print the whole frame and transfer the motors and other electronics over from the prior design. That following Friday, the 3-foot diameter UAV (now called Freebird One), was in the air.

I knew this was the design I was after, but much work was still needed. It was lighter than my old design, but still too heavy. Flight times were 8-10 minutes with large batteries. It weighed 20 pounds or so. The wiring couldn’t stand up to the high current pull. A couple of times it even caught fire in the air! As it was winter at the time, electronics and motors were repeatedly destroyed by flying in the snow. Even so, I discovered the first alternative use for a large UAV with a lot of prop wash: an airborne snow blower. It was good for up to four inches or so of snow and could clear our driveway in minutes operated from inside the warm house! But it could do even more than a regular snow blower. It cleared snow off the cars and then off the roof of our house.


My mission was clear at that point: I had something powerful that could fly very precisely and on its own. But it needed to fly a lot longer, be a lot lighter, handle any kind of weather, carry a lot of equipment and attachments that I might want to develop for various tasks -- and do all this without sacrificing strength. I stuck with carbon fiber for the most part, though successfully produced one unit entirely with wood fiber filament from MatterHackers last summer. Additional materials in the current version include Ninjaflex for weather covers and vibration dampening and transparent Bluprint for the cockpit cover to make the lights from the 3D Robotics Pixhawk flight controller visible.

The bottom line is that numerous iterations and a year later, I have created a 100% printed large UAV from an idea in my head, a fantastic CAD package from Autodesk, a solid workhorse printer from Lulzbot, and high quality, reliable (always 2-day free deliver across the country) filament from MatterHackers, in particular Proto Pasta’s carbon fiber. The only purchased components are the electronics and propellers, though I have successfully printed propellers that work, and may eventually switch to printed props after testing is completed. The UAV is even assembled using printed carbon fiber pins, which replaced traditional stainless steel screws I had been using earlier.


Not only is Freebird One 100% 3D-printed, but I would argue its performance specs beat anything on the market today. It now weighs just 8 pounds without a battery (half of what it weighed the first time I made it). It can fly for up to 35 minutes, has maximum speeds of 70MPH horizontally and 3,000 feet/minute vertically, can carry an additional 15-20 pounds of payload thanks to more 6,000 watts (9+ horsepower) from the four brushless motors, and is completely weatherproof (snow, rain and winds up to 50 MPH or so). It’s also the safest UAV.

My hope for the future is to add artificial intelligence features to assist in obstacle avoidance and expanded failsafe features such as a parachute and ability to find safe places to land in the event of a problem that precludes the UAV from returning home. Equally important, I hope to develop a number of attachments to expand functionality of Freebird One as a useful “tool”. Snowblowing is just one example, but even that could be improved with attachments I have in mind. I also successfully used it this fall as a leaf blower. And because the blades are enclosed, I was able to remove impacted leaves along fences by bumping the UAV along the fence and angling the airflow in a way to get under the leaves. In a similar capacity, the UAV works well as a gutter cleaner with a 3D Camera (in a 3D-printed weatherproof case) sending a live video feed  to a pair of goggles, allowing for precise maneuvering in tight spots. The possibilities are endless – both for consumers and businesses.

Freebird One will be commercially available in the near future.  After having issued RFQs for injection molded parts, it is not clear that I can get parts made to the tight specs I have been achieved through 3D printing.  Given the nature of this project, every gram of weight matters a lot, and in particular the strength-weight dynamic. Every 100 grams represents roughly one minute of flight time, and every gram of material needs to be exceedingly strong. As a result, the current plan is to actually build up a 3D printer “farm” to produce parts quickly. I have improved print settings to the point where the parts coming off the printer require little to no post-processing. At 20 um resolution using carbon fiber, the lines are not very visible and look more like a pattern than what traditional 3D-printed parts look like. Using 3D printing as a manufacturing solution also enables me to revise and improve parts quickly, introduce new attachments, and provide custom modifications without using a typical update cycle.

To date I have been working on this venture alone, but will be expanding in the near future. I really treasure the time I’ve spent on this so far. Not only have I been able to spend more time with my family, but I have been able to just dream up new ideas and then actually produce them right before my eyes. Even more than the UAV I’ve created, I remain amazed at the possibilities opened up by 3D printing and high quality consumer-accessible CAD software. It’s a great age we live in where someone like me with no engineering background can design and produce a high quality aircraft in the confines of my home with nothing more than an idea, a computer, a 3D printer, a few spools of filament… and some patience.

If you are interested in learning more about Freebird Flight, you can visit Roger's site or follow them on FacebookInstagram, and Twitter.

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones


  • Hi Jerry,

    I totally agree with Rob on that one although the PC - Apple battle continues.

    And I am sure a lot of you out there will disagree with me.

    I had a job for several years that required me to use an Apple for text editing, worked great for that.

    But although Apples one source hardware platform was great for a few things, in spite of the difficulties the versatility of the PC always won out.

    And in serious CAD that battle has - I think - already been fought and pretty much entirely won by the PC.

    The Mac is at best a distant secondary platform for CAD, Apple's hardware, the interfaces and the add on's never measure up to what is available for a PC.

    I have a screaming huge multi GPU 970 NVidia video board and 4K monitor and it is a superb platform for Fusion 360 both CAD and CAM.

    Also at this point many of commercial CNC machine manufacturers are actually switching to a direct PC based system, casting aside the specialized (and horribly over priced) built in dedicated systems.

    Commercial CNC routers are now by far mostly PC based.

    And as for Inventor, Fusion 360 is just a compilation of Inventor and AutoCAD along with an enterprise supporting interface and from what I can tell AutoDesk is actually moving away from Inventor to Fusion.

    Fusion definitely has better and easier to use 3D construction, modeling and modification tools.

    @ Rob, happy to hear that the printer interface works so well.



  • I think my problem is OS X is a server client graphic environment, could never beat a micro kernel system. On PC I don't have any advantage to use fusion. If you want autodesk way, use inventor, want old way, cero. want fast way, solidworks. Want sketch up way, try Ansys Spaceclaim.

    The problem is, I don't want to use PC when ever possible.
  • I think your problem is not Fusion, your problem is you're using a Mac.

  • I am sorry I didn't say fusion is a toy cad like other package from autodesk for children. I have 10 years experience in solidworks, tried as many 3D cad as I could, proficient surface in 3ds max and rhino.
    Here I am not debating how fusion will change the market nor ultimatemaker prints looks. (I have 3 um2 clones on my desk made by myself) but you guys could use some info of my current work.
    Fusion on Mac is the only serious cad on Mac I could find. A British cad company who made a genetic algorithm structure generation have a Mac cad may worth a try. 3d print supports in fusion is designed for their SLA machine, a very small and tedious peeling process it use. (TI DLP and rotational exposure process).
    Fusion on retina for working is totally no go, frame rate is insufficient even on my latest Mac Pro with graphic card. I only use fusion for desktop360, demo the models on my iPad pro.

    UM2 has serious problem on the extruder, it wear out in just few hundred hours. I am building a gear/belt delivery system, right now using a 100% fill mod from youmagine.

    Printing 100% fill, um has no advantage comparing to my form1+. SLA parts are much easier to sand compare to PLA. Cura is toy after you have fully confidence in simplify3D.

    10 grams for a non strength critical mounting is a bit luxury... You can use simplify3D to slice it to single wall weight in 4 grams, or use SLA to print a exotic genetic algorithm generated organic bone form in half weight I think.
  • Gary, Fusion's 3D Printing hook-up is stunning.  You pull down the menu, select "3D Print", click on the part you want to print, tell it to send it to your slicer of choice (in my case Cura), click enter.  It automatically fires up Cura, and the model pops in.  That's it.  From there I make whatever changes needed in Cura, and I'm done.  The "it just works" factor is 10/10.

    I've worked with many CAD programs, from basic AutoCAD, ProE, SolidWorks, all the way up to CATIA (which requires the user to be slightly insane if you want to become proficient, from what I've seen).  I'm not sure it has all of the features of all of the systems (ie: FEA, etc.).  But it will do anything I ever needed to do in the auto industry. Even better, it's so easy to learn and use, an engineer can just do it himself instead of needing to interface with an (insane) CAD operator.

  • Whatever else is true, Fusion 360 is not "Toy CAD".

    It is probably the best CAD system available anywhere right now, AutoDesk has just taken a really bold move and made it freely available to people for non-commercial work.

    However, even for commercial stuff, their annual fee is miniscule compared to everyone else.

    And it comes with totally full featured 3D CAM as well.

    It is in every aspect "Pro" and it's price basically makes all the other stuff obsolete even on the big CNC machines.

    Only companies already commited to other packages have any reason at all to continue paying the outrageous prices for the pro CAD/CAM software they are already paying (and I am sure they are not happy about it.

    By all measures AutoDesks Fusion 360 is a Enterprise level CAD/CAM system.

    A long time ago I was an authorized AutoCAD developer and I have great respect for the company, but this is the best thing any CAD company has ever done for the public period.

    I'd suggest you look into the free "full" edition of Fusion 360 and get one for yourself.

    Rob, great to see it is working so well for you, wasn't sure about 360's "additive" machining functionality, but from what you have said, it looks good.

    Also, really happy to see you like it too.



  • Toy CAD?  Have you used it?

    And terrible accuracy additive machine, in reference to what?  Are you aware of the capabilities of these machines?  The point of 3D Printing is not that it is the be-all end-all manufacturing technology.  It fills a couple niches, however.  It does allow fabrication of parts that are not possible any other way.  (ie: semi-hollow structures) and it also allows low-volume companies to produce parts they could not have any other way. 

    Here's a GPS boom mount for a helicopter, which features ziptie tunnels.  The printed part weights I think 10g.  It's not produce-able any other way. 


    Or a motor cover:


     The printed part gives the prototype a degree of professionalism not possible otherwise.

    I do see way too many people using their printers to make flat slabs however, which is disappointing.

  • Today anyone could make a 'drone', thanks to Chris. But good design doesn't have to come out from a cool looking Mac running toy CAD nor a terrible accuracy additive machine, one have to master multi disciplined engineering skills and experiences.

    Very nice photos and website though.
  • On the very first image, the motor size appears to be on a scale with the diagonal dimension of the GPS.  There no way that thing is lifting 20lbs.  It's just not clear if that is a small-size prototype they are showing, or what.  I'm also quite unimpressed with the design.  All flat sided slabs, not befitting of what a 3D printer is capable of.

  • Yeah, I didn't to be debbie downer again, but the claimed specifications make no sense to me at all either. The AUW, payload, top speed and flight times are quite suspicious. The photos are low quality and hard to see what we're looking at.

This reply was deleted.