Introduction: 3D Printed Ornithopter - Micro UAV Drone

Picture of 3D Printed Ornithopter - Micro UAV Drone
In the past couple years there has been quite a bit of activity into r&d of tiny insect/drone/ornithopter type vehicles. Two intriguing designs I've been following are Harvard's RoboBee and Cornell's Ornithopter. While some designs, like Harvard's, are out of reach of the average DIYer (check out their white paper, it's awesome) the Cornell 3D printed ornithopter is attainable and I took it as a challenge to make my own. Cornell's use of 3D printing techniques makes it easy to try out and make modifications. Unfortunately they don't give out their files and they built theirs on a 3D printer that cost 100k. My work has been to model my own design off of their work, but make it using commonly available 3D printers. You can see the results below.

Quickly I would like to discuss how amazing desktop 3D printers are. So far I have been able to make two versions: the first is blue and printed by my schools 3D printer (6.07 grams), the second is clear plastic and made by a Makerbot Replicator 2 (4.729 grams). For a reference point Cornell's weighed 3.89 grams. Between my first and second designs the weight of the wings has gone from 4 grams to 2.6 grams because the wing thickness went from 0.02 inches to 0.008 inches (200 microns). After receiving comments about more flexible wings, I have made a new version with a total wing weight of 1.968 grams making my ornithopter a total of 4.10 grams(the total weight of 4.729 grams minus the difference in the wings (2.6 - 1.968) grams) making it only 0.21 grams off of Cornell's weight! I am not done with this project yet and would really love to get a design that flies, so please give me your comments and ideas!

Step 1: Parts/Tools

Picture of Parts/Tools

The majority of the parts for this design are 3d printed. The rest are as follows.

Motor - The exact one Cornell uses I coincidentally already owned and is ideal. It is a 1g pager motor with a planetary gearbox that can be found HERE.
Paperclips - Try and find very thin (diameter wise) ones.
Batteries - Can use anything around 3V to test and 7.4V for test flights. The arduino's 3.3V pin or a power supply can also work. Power supply is definitely the best but not needed.
The rest of the parts are 3D printed!

3d Printer - Pretty vital.
Drill - With some smaaaalllllll drill bits.

Step 2: 3d Printing

Picture of 3d Printing
You can get the files for my parts on this page or possibly HERE. Let me know if other file types are needed, I made them on Inventor 2013 and I'm not sure what file types other programs/3d printers need. I tried uploading the IPT and STEP files but couldn't.

There are five parts to the ornithopter that need to be printed: the body, two driving arms, and two wings. I would like to note that there are two versions of these parts below. The 'Ver3' ones are much more refined and precise with the wing thicknesses of 200 microns and 100 microns and holes drawn into the sketches, so no drilling is needed. If you don't think your printer can print to that quality level then just download the parts that don't say 'Ver3' and they are the exact same ones used on the version 1 (blue model) I made. Apparently some programs have had trouble loading the files to the correct dimensions so I have added pictures to this step that show dimensions (in inches) for all the first version parts. This should allow anyone to scale them properly.

Cornell, as far as I know, doesn't offer the files for the parts they 3d print (this was a little irksome as it seemed like the point of Cornell's project was that they could be easily manufactured with 3d printing). Theirs has a resolution of 16 microns and the Makerbot Replicator 2 has a resolution of 100 microns, but that's plenty close for this project. I wasn't sure of the precision so the first version parts are a little beefier than they probably needed to be. Anyways they both work and are great places to start out before modifying the design!


Step 3: Motors

Picture of Motors

The motor for this project is quite cool. Weighing in at only 1.2 grams it packs a lot of power. Recently there have been several more planetary gear pager motors that are available at robotshop and other sites so feel free to experiment with those. They all run on 3v (although Cornell ran them at 7.4V for the flight tests) and will need the motor nib to be drilled out.

The one thing to note and pay attention to is how fragile the wires are. They pull out EXTREMELY easily. My advice is to use a zip tie or tape to keep them against the body and take the stress of the wires off of the base connection with the motor because that is where they lose connection and stop working (see my picture if you are confused what this looks like). Also soldering very thin wire to the leads of the motor helps from pulling on them too much. It is also necessary when flight testing so the ornithopter only has to lift its weight off the ground and not the wires.

I took one apart to see if I could get it working again and found out how cool the insides are. If the top and bottom pieces of the motor separate it's no big deal because the bottom is the coil of wire and drive shaft while the top part is the gear box. They can be placed back together with no trouble. Unfortunately I found fixing the motor by soldering new wires into it impossible. My one semi-solution is to drill two holes where the wires came out of the motor and if you slip two wires into it and jimmy them around a bit you can get the motor to work, albeit with an intermittent connection. Just things to keep in mind!

Step 4: Drill

Picture of Drill

NOTE: If you used the 'Ver3' parts with built in holes then you can skip this step.

I didn't print the holes for the paperclips into my design for the ornithopter parts because of the issue of precision. To ensure a tight fit I pulled out my tiniest drill bit and went to town (I feel there's a few jokes here). There are many holes that need to be made, one through both wings, ones on each end of the forks, one directly above the motor mount on the body, and two through the chamber for the wings. Check out my photos of the assembled and drilled parts because it clears up a lot of questions. The last thing you need to drill is the top piece of the motor (the part that spins). This is quite tricky and leaves little room for error so I first suggest making a little groove that the drill bit can sit in nicely by stabbing it with a needle or something similar. Make sure when doing any drilling to use a vice to hold the fragile parts steady.

What I've found to be the best way to do this is to spin the bit at the highest speed possible and use very little force to push it through the plastic. This prevents stress fractures in the thin layers of plastic. For the fork that drives the wings you will probably have to drill a few holes at different positions and test to see which gives your wings a full range of motion.

Step 5: Assembly

Picture of Assembly

Now you are going to have to put it all together. Pull out your paperclips and straighten them. For the offset crank you should put a hook at the bottom for the motor then bend it out slightly making two offsets for the driving arms, they should be approximately 30 degrees apart. Leave a short straight segment at the top of it after the offsets to go through the top of the body. Check my pictures, it's much easier to see than explain. Bend away. Expect lots of trial and error. The size of the offsets depends on where you drilled your hole in the wings. It should be half the distance between the hole when the wing is at its mid position and the hole when it is at its fully extended position. 

For pinning the wings to the body and the driving arms to the wings use little L shaped segments of paperclips (tiny, straight sewing pins also work well). Check my pictures for a clear example of this. 

I would like to note in my first design for the body I didn't have a way to get the wings into the section of the body where they would be mounted so I cut a slit through one wall as seen in the pictures. This allowed me to slide the thin center beam of the wings through it. You may want to trim the body anyways to reduce weight. 

For mounting the motor just get some sand paper and slowly increase the size of the hole at the bottom of the body until there is a tight push fit. I didn't have to use anything other than friction to keep the motor in place.

Step 6: Wings

Picture of Wings

Alright this is the part that I would like lots of suggestions on! Obviously the wings have to be super light and Cornell did a special 3D printing technique to create theirs so I'm not sure how to replicate it other than by printing very thin layers. The Makerbot Rep 2 is capable of 100 micron thickness and Cornell used wings 40 microns thick so it's impossible to get the same resolution but we can get close!

I have also included files for my open wing title 'EmptyWing' so you can try using cling wrap or whatever you would like instead of the 3D printed filling.

Step 7: So Fly

Picture of So Fly

You are ready to flap. Make sure when you are holding the ornithopter that your fingers don't come in contact with the wings, driving arms, or offset crank. I just tried to hold it but a clamp would be ideal. For power I used an arduino uno and hooked the two motor wires up to 3.3v and ground (doesn't matter which goes to which, it just changes the flapping direction) but this is optional, three AA batteries or anything around 3v would also work.

For flight tests you will need to suspend your ornithopter. I did this similar to how cornell did, by using fishing line through two holes in the frame that is nearest to the ornithopters center of mass. Finding the center of mass is all just guess and check. Try and make the fishing line as tight as possible, in my video you can see it vibrating, allowing the ornithopter to move around too much. Tie a small knot in the fishing line for the ornithopter to sit on so it won't be in contact with the table. To try and actually get it to fly I trimmed the frame down and put very thin wires (individual strands from stranded wire) connected to the motor so it wouldn't have to haul up the weight of the wires up. I also used a 7.4V battery like Cornell did. Cornell recorded a rate of flapping around 30 Hz and while I couldn't give you an exact number for mine it was definitely MUCH higher than with 3V.

Step 8: Final Thoughts

Picture of Final Thoughts

I sincerely hope you enjoyed this, if it got you thinking then it has done its job. There are a few difficulties I had that I wanted to discuss. I think the precision of the 3d components can definitely be improved with iterations, but unfortunately my limited access to 3D printers forced me to make beefier parts to ensure they worked.

I have a few ideas for variances, the most interesting one would be using a 2g brushless motor I saw on some hobby site (capable of much higher torque offering around 28g of lift) in a slightly larger over all build. This would eliminate a great deal of the challenge of such flimsy, tiny parts and would perhaps even be able to carry a payload. The other thing I was hoping to see was the use of a small camera and/or use of motor control to perform wireless controlled flight. The last thing I thought would be just plain cool was if rather than carrying a payload of batteries a laser could be used to give energy wirelessly to drive the motor. Really how awesome would that be?

I will post when I am able to make my next iteration, but if you have any ideas for existing modifications please let me know! This means you all need to take it up and show me what amazing places you take it. Let me know. Thanks!

Step 9: Updates

Picture of Updates

Since posting this there have been many great ideas from the comments! Here I will post my trials of these ideas:

So far I have drawn up two revised wing designs, both with tapered thicknesses to allow for more flex when flapping and thus more upward thrust. I will plan to have test videos of them soon. I have put the stl file for the new wings down below, they work with all the 'Ver3' files. The wings weigh under 1 gram each bringing the total weight of the ornithopter down to slightly over 3 grams!

I have also done a few experiments trying to print straight onto plastic wrap so I wouldn't have to print out the filled in part of the wings. This would reduce weight, but so far the plastic has not bonded well with the plastic being printed onto it. I will try using some light adhesives and hope to have an update soon.


Ben Finio (author)2013-06-16

Cool! I worked on the RoboBee as a graduate student at Harvard (that's my "Steerable RoboBee" video you linked to at the beginning) and coincidentally am in the Creative Machines Lab at Cornell now, although to the best of my knowledge no one here is working on the ornithopter project anymore.

FYI I just posted directions to make a simple paper RoboBee model - this isn't an actual flying version, and really better suited as a craft project for kids, but lets you see how the mechanism works: In case you haven't seen it they recently got the real RoboBee to hover.

Also, not sure if you're interested in walking robots at all - but the Biomimetic Millisystems Lab at U.C. Berkeley has a bunch of really cool hexapod-inspired robots made from laser-cut cardstock. So if you're into DIY versions of robots that started in academia, that might be worth checking out - I think they have some of the drawing files on their website. For the record the RoboBee project actually has its roots in the "Berkeley Micromechanical Flying Insect" project from the early 2000s.

Tomdf (author)Ben Finio2013-07-07

0.o I need to gets me some piezo actuators, those tiny robots are boss!

Patrick S (author)Ben Finio2013-06-16

Thank you for commenting! I had actually read your paper RoboBee when it came out and saw you said you worked on the real one, that is TOO cool! I'm going into undergrad as an ME next year and hope to be able to do something like that for grad work.

That is awesome, I had seen the RoboBee hover but I haven't heard anything about that lab at U.C. Berkeley I will definitely read into it! I've actually been wanting to look into bipedal or hexapods recently so that is too neat. Keep me posted on your future work!

Ben Finio (author)Patrick S2013-06-17

Wow, that's great that you're getting started with this stuff so early. I didn't get started with robotics until grad school. I'm not working on the RoboBee anymore, the best place to check for updates will always be the Microrobotics Lab's publications page. They also tend to put the latest/best YouTube video on their homepage. I'm happy to answer any questions you have in the meantime, or can try to point you to the right publications.

ruma29 (author)2016-01-29

Patrick did you post step files? did you make it fly? Thank you.

Tomdf (author)2013-07-07

Very Awesome! That is a really good idea trying to print onto the wing material, I would be very interested in hearing how that goes.
I've also attempted to design a 3D printed ornithopter but ran into the same snag that you have, the wings. I'm fairly certain that getting flex on the wings is crucial. I did have some luck with mylar before shelving the project. 

Here is some nice technical info on wing design if yer interested:

Patrick S (author)Tomdf2013-07-07

Thanks! Yeah I actually read yours while I was doing this 'ible and although we take different approaches it seems some stuff is in common. The printing onto the wing material unfortunately hasn't gone too well, I'm going to try increasing the extruder temp a bit more to see if that'll melt the wing material and the plastic together. Usually the plastic pops right off. I checked out that site you linked, good stuff.

JamesRPatrick (author)2013-06-30

Don't the bottoms of each wing need to be flexible?

TREX ZoaR0K (author)2013-06-28

is any lift generated

kitemancake (author)2013-06-17

Similar to a lot of other suggestions, I say that you should need to remove the lower part of the wing frame completely, and stick on Mylar wing surfaces. I believe wings gets a lift by making a sweeping action, which pushes the air down each time with a backward and forward action. It has nothing to do with an aerofoil section whatsoever.
There is a French company that makes a flying bird called Tim. It operates with a wind up rubber band, and flies very successfully. This has a stiff leading edge and the free-floating Mylar wing. A downward curve on the leading edge should prevent the wing surface from flapping horizontally, which would obviously produce no lift.
the ultimate problem with this little miniature flyer is the power supply. I doubt it will ever be able to lift a battery.

Patrick S (author)kitemancake2013-06-17

Interesting, I had never heard of Tim! But yeah I will definitely be redesigning the wing.

zenTaurus (author)Patrick S2013-06-21

found out that it's called the Timmy Bird. please see: and


Patrick S (author)zenTaurus2013-06-22

Haha I really enjoyed the music in those vids.

zenTaurus (author)Patrick S2013-06-22

and a TEDtalk -- A robot that flies like a bird:

Ben Finio (author)2013-06-16

Also - some random advice after reading some of the comments below and looking at the videos: fine-tuning the wing flexibility is actually a big challenge for any flapping-wing MAV, and can definitely have a huge impact on lift generation. Long story short there are generally two approaches: you have a rigid wing with a hinge at the top that lets the angle of attack change throughout the stroke (RoboBee approach), or you have a flexible wing that bends along the chord as it flaps (Cornell ornithopter approach). It's hard to tell without high-speed video but judging by your video, you'd get more lift if the trailing edge of your wing was more flexible, which you can probably accomplish by tapering the outline to be thinner at the trailing edge (depending on the resolution of your printer).

If that doesn't make sense in writing I can try to find an appropriate figure from my thesis or one of the RoboBee papers.

Patrick S (author)Ben Finio2013-06-16

Interesting stuff, yeah the tapering of the wing structure is a great idea. I'll draw up a new one with a thinner bottom and side edge but maintain the stronger axle to mount it. I looked back at Cornell's videos and can definitely see a lot more bend when theirs flaps versus mine.

Ben Finio (author)Patrick S2013-06-17

You also might want to try wings with veins made out of carbon fiber rods with a thin plastic film (search for Mylar or Kapton) as a membrane. 3D printing the whole thing is kind of a cool novelty, but carbon fiber is much stronger and lighter than the plastic from consumer 3D printers. At a minimum, I would just print the veins and then stretch a membrane across them. You might be able to try Saran wrap if you don't want to order Mylar or Kapton - might require a couple layers so it doesn't tear.

Patrick S (author)Ben Finio2013-06-18

Yeah a film would make printing much easier for sure as well. I'll look into it. So just print the veins onto the plastic film? Sounds do-able, I'll draw some stuff up and look into it.

Ben Finio (author)Patrick S2013-06-18

Not sure if you'll be able to print directly onto the plastic film - that will depend on the printer (e.g. could you stretch film over the print platform then expect the extruded plastic to stick? not sure). If not, you can always print the membrane first then stretch the film over it, using a SMALL amount of glue or adhesive to bond around the perimeter (don't want to add too much weight to the wing).

Patrick S (author)Ben Finio2013-06-19

I'll test both. I think I'm also going to try printing the frame onto the stretched out saran wrap, perhaps the hot plastic would bond more easily than with glue. Should be interesting.

nazdreg2007 (author)2013-06-17

You could try use some light tissue to infill the wing and to increase it's strength, soak the tissue in ABS juice so when it dissolves, it'll reinforce the tissue nicely and bond it to the wing. That could cut the weight even more for you.

Kiteman (author)2013-06-16

To test the lift capability, attach the bottom of the thread toa weight sitting on a scale - the lift will show as a drop in weight.

I would also be tempted to remove the bottom part of the frame around the wing. This would let the skin of the wing flex more, changing the angle of attack, and (hopefully) generating more lift.

Patrick S (author)Kiteman2013-06-16

Good ideas for testing thanks! Yeah I'm drawing up some more wings right now, hopefully I'll be able to make them soon and let you guys know how it goes.

Any other weight reduction ideas?

Kiteman (author)Patrick S2013-06-17

I would try tapering the thickness of thd frameas well, make it thinner near the bottom. That might need careful sanding if you are near the limit of printing thinly.

Not only would that be lighter, but it would allow thd lower part of the wing to flex more as well.

Otherwise, look to nature for lighter wings - bird bones have an air-filled structure. Try printing with gaps in the structure, or CNC milling from a foamed material.

tschnizlein (author)2013-06-17

While I haven't had a chance to attempt this build yet, from watching your video my first suggestion would be to shorten the stroke of your wings. They seemed to be colliding with each other, in which case your losing a lot of energy and stability that could be directed towards lift instead. Secondly, you might try tapering your wings, thickest at the top, thin at the bottom. This should allow for a little more flexibility as they flap and generate more upward force rather than side to side. Just my thoughts from my experiences. Good luck with the project, hope to see it flying soon.

Patrick S (author)tschnizlein2013-06-17

Yeah the clicking noise made me think they touched too but I tried putting paper between the wings and there wasn't any contact with it. I think it might be rubbing the frame a bit though... But thanks, good ideas!

he3r0 (author)2013-06-16

that is what i was wondering too... IMHO pointless instructable if it aint working (good material for blog), though interesting design.

Patrick S (author)he3r02013-06-17

It get's people thinking and they can give me new ideas.

eng_Andy (author)2013-06-16

If you want to upload STEP files of your designs here, I think you can try putting them in a .Zip or .RAR archive, as I've seen compressed archive files on some other instructables.

Patrick S (author)eng_Andy2013-06-17

Cool I'll give it a try.

scotttaggart (author)2013-06-16

I only see flopping and no flying. Am I missing something?

dgateley (author)scotttaggart2013-06-16

That's what I saw too. Is there an explanation?

Patrick S (author)dgateley2013-06-16

Yep, this project so far is my first two test trail versions. I have designs for a third version that will be much closer to the weight of Cornell's robot but I'm not sure when I will be able to print it so I thought I would get some ideas while I waited.

dgateley (author)Patrick S2013-06-16

Got it. Following along to see how it goes. I don't think passive control has a prayer of working but I'd love to be wrong.

farabide (author)2013-06-16

Great idea:
CanI suggest a hanging balance for testing (like a wide baby's mobile) and removing most of the bottom of the wing section, leaving the outer curve and a point for more of an analogue sweep (made up term but i like it) and greater downdraft.
I love the vertical crank - very elegant :)

Patrick S (author)farabide2013-06-16

Cool stuff, the baby's mobile might work! I'll see if it has enough tension but thanks for the ideas.

nomasfilms (author)2013-06-16

Your work is wonderful! If I could offer a suggestion, I would ditch the 3D printing of the wing surface, and use a thin plastic film which I would glue to your frame. The lightest RC models in the world do just that. See here:

Also, varying the wings frame thickness might get you a little more flex in your trailing edge, as mentioned before. Great work though... Any chance this device would fly for a few seconds on a supercapacitor? :-)

Jared_Reabow (author)2013-06-16

after observation i am presuming that the way this model acheives air flow from the flapping is from the wings flexing as they flap to create the air flow, the issue with your design is there is far to much rigidity in the wing, you should make a wing where the root is thick and it constantly gets thinner as it gets to the other end, this would allow rotational flexing but stop warping from the movement.

sitearm (author)2013-06-16

@Mizchief100; Fascinating! I had no idea this kind of work was being done. I've watched your video and looked up your Cornell link. Sharing with my friends now. Cheers! Site

Patrick S (author)sitearm2013-06-16

Thanks! Yeah the stuff Cornell and Harvard have been doing were just so cool I thought I'd give it a go! Fun stuff.

rimar2000 (author)2013-06-15

Great work!

Seeing the video I think you could make thinner the lower part of the wings's frame. So it would be a little more flexible, and will create greater air flow down.

Patrick S (author)rimar20002013-06-16

Thanks! And yeah I was pondering that, they do seem a little stiff.

About This Instructable




Bio: I hope to help people with the things I make.
More by Patrick S:Guitar BackpackDIY Prosthetic Leg Foam Cover (Cosmesis)Tact: Low-cost, Advanced Prosthetic Hand
Add instructable to: