I completely relate to that feeling of “oh no it’s broken again, time to rebuild everything”. This tends to almost always be the case with DIY flying projects! That was really flying though! Can’t wait for part 4.
well, birds have had a good couple billion years of development time. I bet anyone could build a decent ornithopter if they had that long to work on it. :P
I've found out that using just a sligthly more expensive ESC instead of the yellow one that comes with those motors can make a big difference in motor torque
Yeah, those cheap ESCs aren't up to the task. I found out that with one particular receiver, the motor stuttered like it was skipping. I would not recommend these cheap ESCs.
Same experience here, I tried using 4 of them in a 3D printed quadcopter and they were spinning at different speeds, and had a very obvious step/notch to their speed curve. Try a better ESC and maybe another motor as I wasn't too impressed with those no-brand silver and orange ones. I love your perseverance though James, keep going!
There are plenty of examples of ornithopters that fly by utilising this far more simple approach - counterintuitive as it seems. The flapping creates mostly thrust and the position of the tail gives it an overall angle of attack. So you don't need all the complicated imitating bird biology.
If the tail is a bit more adjustable, up and down, and a bit side to side, you can compensate for the stalling and turning out in the field. This is how smaller DIY ornithopters work, the tail piece is attached to a thin piece of metal so you can slightly bend it in different directions to trim/tweak the flight. Either way, keep it up!! This is a great series!
i think a good idea would be to remove the rod you already cut up, and extend the shorter angled ones up to the main CF rod. i think this would achieve this "air pocket effect" without the wing deforming and letting the air escape. another design i frequently see is that parts of the wing have their own joint and move freely to remove some air resistance when the wing moves upwards, this would probably help a lot too. great video!
I love that you are showing how much work engineering and design is. Some people dont appreciate how many iterations it takes to make something that looks like the obvious answer.
I don't know a whole lot about these things, but the toy version created some artifical lift by angling the tail up a little, if i remember the demo of the toy right. May be worth looking into. i cant wait to see the next version. Thanks you for keeping making interesting and challenging videos.
This project gets more cleanly at the interplay of deliberate thought and action, and tinkering that I think is at the heart of good science and engineering. There are other projects I am more excited about by virtue of what they aim to accomplish, but I continue to be excited about this one because of the nature of the development process.
Feathers on the trailing edge of a bird's wing separate when wings are moving up to let air pass more easily, and then come together to create a larger surface area to push more air when flapping down. (At least on some birds.) If you need more lift, you could try somehow designing this into the trailing edge of your ornithopter wings.
An angled hinge would allow differential movement, but the added weight may be an issue...unless, maybe semi-rigid (celluloid? Plastics? ) feathers and a cloth hinge?
What if you ditched the belt drive and went with a compact planetary gear? A planetary gear would eliminate any of the issues you have with the belts slipping and may even end up a hair lighter then your current setup.
I'm not sure if you could do this with a planetary but a non-circular gear might be a good idea. Unlike a propellor, the torque needed to flap the wings changes a lot through the cycle and a changing gear ratio might help to counteract that.
Try dropping it from a high building just to get it flying first. Then you can work on the launching but knowing it can fly properly! Very inspiring work indeed!
Found it! There's a man named Sean Kinkade who spent many years refining these. (Sean Kinkade Slow Hawk Remote Control Ornithopter) He was finally starting to manufacture and sell them when he was killed in a car accident. I found one for sale and with a bit of research you should be able to find his design. I do know CG is a bit further back than you think, they flap with high alpha and glide level. I've seen one fly and it's brilliant! Loving this series James...
Instead of gluing the rods to the wing using fabric strips, you could use a sewing machine to create piping in the wing fabric, with removable rods slid inside them.
Really liking this series James, definitely nice to see you mixing things up. Now strap an Arduino on that back of that birdy and make her fly waypoints!!
I love how battery,motors and 3d printers have developed in leaps and bounds,so there's a way to reliably design,build and run something awesome., ThankQ. TkEZ
If you look at the way a birds wing works, it is hinged on the leading edge half way so on the up stroke the wing bends slightly lessening the up-force but on the down stroke it locks into place giving it full down-force resulting in more lift. You are getting there tho. Great video ✌✌
You may want to consider some sort of curved skis to attach to the Ornithopter during the testing phase so that you can minimize the amount of damage to your structure and get a softer landing each flight.
I would try using 6mm rods for the wings. Then in the center of each wing rod, putting a hinge on it so that when the wings are flat they hold in place, but when the wings lift up they bend. upwards at the hinge. I would probably put some light elastic over the top of the hinge so the wings want to naturally remain flat when not being forced up by the motor.
I first started watching your videos many years ago with your homemade vaccuum forming machine, you are very inspirational and im glad I subscribed God bless you and keep you safe.
I would try making the body heavier so that it moves less vertically when the wings flap thus enabling the wings to exert more force on the bird. I would also put the braces on the top side of the wings in sections which are tight so that on the down movement they are pushed together and act like a solid brace but on the up motion they disconnect and let air flow around the wing better. Love your videos by the way !!
Just a suggestion from an old RC flier: Start with controlled glide tests. I suggest setting the wings slightly up and locking them in place to do glide testing. The goal is to adjust your center of gravity and tail for a smooth glide at what you would consider "Approach Speed". That will take a lot of the crashing out of your testing. It would also allow you to design some kind of control mechanism. Many years ago I had a tiny little ornithopter, barely a 6" wingspan, that changed the tension of the wings alternately for steering. It did this with a moving arm at the trailing edge of the wing sails, and this functioned as the mounting point for the aft edge of the sails as well. It didn't take much movement of this arm to induce a turn. You could also add elevator control to the tail, just move the whole thing up and down. Remember that these flat, sail wings produce very little lift of their own, needing forward speed and angle of attack to maintain flight (Think "Hang glider"), so control of the pitch will make this *much* easier to fly. Hope this helps.
This is a really cool project, but I think you’re overlooking some important aspects of designing something that flies. The first is longitudinal static stability. I’ve noticed that on some throws, the ornithopter pitches up immediately, slows down, and then drops like a rock, because it’s going too slow for the wings to produce enough lift. Does the ornithopter glide alright without power or control input? If not, it won’t matter how good of a mechanism you have, it will crash every time. The second thing is drag. Having the drive mechanism exposed to the free stream is going to produce a lot of drag, and that may prevent the ornithopter from gaining enough speed to remain airborne. Tom Stanton has a good video on how to design and print a single-layer shell that could act as a fairing and reduce drag. It could also serve as a sacrificial energy absorber during crashes and save the drive mechanism.
I feel like you need to build a rig to hang it from to really test it and dial in the movements... almost like the equivalent of a wind tunnel (not a wind tunnel, but something that serves the same purpose but for an Ornithopter) -- something where you can maybe just hang it and flap the wings and see if it goes up or tilts at all and just get it to where you can dial in the movements.
I had a large ornithopter that came from China and from what I can remember, it had a lot of power and thrashed those wings really quite fast. It would climb really well to quite a height. The three things that were apparent to me were to be as light as possible, have a really powerful motor and to really thrash those wings. Holding it while testing it was quite a scary thing to do and bits would break all the time.
Might be interesting to look into where you can figure how to make it so on the up stroke it flexes in a way that provides more forward motion and in the down stroke it flexes in a different way that provides more lift.
Food for thought - motor speed reduction through an optimised / light weight version of the cycloidal drive. This would present it's own design challenges for the symmetrical drive to the wings, but final drive ratio is easy to achieve and the compact design would remove all the issues with belt tensions and stress loadings, as well as centralise the weight which could be kept close to the centre of gravity to aid stability. With careful design, perhaps using fibreglass or aluminium for the cycloidal disc to help minimise weight, I think the total weight would not be too different from that of all the "beefed up" printed frames, bearings and pulleys in the current design. As a bonus, it would certainly remove most of the failure points in the event of a crash.
I think the simple flapping shoulder works by creating thrust and then you just need to give it an overall angle of attack. Anything else gets ridiculously complicated mechanically. I have a large Kinkade ornithopter that works on this principle.
Also, I think your wings don't flap like a birds, they're more mechanical and less flexible. If you add more points of movement to add power to the strokes downwards and less power to the strokes upwards to maximize the effectiveness of the wings. You want more surface area when stoking down than when stroking up. This would make the machine a bit bigger but I think it'd still be doable with the tech we have right now. Hell even easy if you know what you're doing which I have faith in you in ^^
Hey James Bruton, well done on your vehicle! It is coming along and taking shape! Fantastic!!!! IMO, You should design your next one with a launching mechanism so that way the launch is the same each time, and maybe put a tether above, like a "dog run," to keep the Machine from hitting the ground breaking everything. Also, 😅 make a "slide of some sort to distribute weights easier and faster with marks to show how far each move is.
Back in the early days of electric airplanes everything was brushed with gear boxes. If I remember right the gears were to increase the rpm's. A brushed motor might be what you want
There are a couple manufactures that make a lightweight PLA that can potentially reduce your weight by half, at the cost of some structural integrity: ESun PLA-LW or ColorFabb LW-PLA
dragonfly backpack? I see that as the end result of you exploring the next step in this design which is more than simple linear up and down flapping. you gotta scoop it
I know its way easier said than done, but if you had the wings have individual "feathers" of wing structure, that grouped up to make the wing, and then had them tilt on the upstroke, you could reduce a lot of downward force on your wings. good job so far mate, looking forward to part 4 :)
Perhaps consider motorizing the tail a bit, then using accellerometer and gravimetric info, in addition to wing stroke timing, to decide when to point the tail up or down it might be good to syncronize the pulses mechanically with the main drive, then have a seperate motor operate some gimbal range for the tail's movement
I have been really enjoying this series but I have to say that it hurts every time I see the chassis glued to the carbon fiber rods. especially when prototyping as that means you cant replace / retention the belts on the fly and just makes repairs that much harder. I really hope you come up with a better method for the next version. glue is quick easy and strong but you know as well as I, it is a bit of a cop out / short cut.
Yeah, definitely a better/less permanent solution than glue needed. My ideas include: 1) Set screws. No idea how these would fair with carbon fiber. I'm guessing poorly, as the hollow rod might just crush. Or, it could damage the print instead, because the force is decently concentrated. 2) Screw clamp. Probably the most materially reliable option I can think of, but might lack grip strength. The wide contact area means that both the print and rod are unlikely to be damaged, save for around the screw on the print. A set screw might have an advantage in terms of grip, because it would dig into the material somewhat, unlike a clamp. 3) Low temperature hot glue (Replacing glue with glue... Sometimes, I even impress myself). If the temperatures match up, this could be perfect. Or, it could be nightmarish. The main advantage here is high strength, as well as being a drop-in replacement for the current glue. Whether this is practical is a matter of whether or not the glue can be made to melt without also melting/weakening the print. A precursory Google search tells me that ABS's glass transition temperature is up above 100 C, and that some hot glues melt as low as 65 C. If the structure can be brought to between that range, the glue should be workable, and the print should remain intact. The placement could be readjusted, and then the glue allowed to cool. I think the idea has its merits, but either of the two listed above are a safer bet.
I think the best luck you could have is to add a little extra to the back to control the pitch of the tail, if you want to mimic a bird allow the back before the tail to have a decent amount of flex
Using a V-tail or maybe crescent shaped, could give you both rudder and elevator if you articulate it in two axes; a third axis and you can even get roll as well with some clever linkage (or even just sequence of individually powered articulations.
for the wing struts/reinforcement rather than using a thinner carbon tube that is more flexible in all directions and could crumple and buckle/kink if over driven. Instead keep using the thicker rod but only cut on the top and bottom surface 90% of the way through, you could make a rod that would be flexible only in the desired direction/orientation. You ever think it might be time to start laser cutting thin sheet metal and getting a simple sheet metal bender for components? you could still bolt and glue things together, and you could unlock much smaller/stronger design forms.
Love your videos, this is off topic from the current video but a project that would be cool for you to build / walk us through would be to build a mini robot like Bolt. The one used by pros to do programmed camera movements. But you could scale this down for a go pro or smartphone.
One idea I had was to make the wing able to bend upwards but not downwards. When flapping up, the wings will bend resulting in less surface area, but when downwards, the bend will straighten out like a knee joint, increasing downwards surface area.
I think you need to automate motor speed in a feedback loop with actual lift. Maybe you can even setup a rig holding the ornithopter with strings and letting it learn by itself.
So you're looking to make the gearing lighter and smaller (and more rigged). How about something like that then? Probably not one of your big cycloidal drives (though, that'd be a lot of power), but how about a harmonic drive, or, better yet, a planetary gearbox? Those are pretty simple to design and work out the ratio (once you know the right equation) and would be more compact and more rigid overall. Just a thought, love to see flying machines of all description :)
Cycloidal drive is too heavy. Harmonic drive too lossy. Planetary would be ideal, but 3D printed gears would hardly be up to the task. No, I think belts _are_ the way to go here, only the front one would need to be made more robust (perhaps wider, and idlers could certainly help). An interesting alternative could be to avoid the crank mechanism entirely, and instead move the wing with _rope winches_.
Ditch the gears and make a rotating oblong angled peg that slots in an a slot affixed to the wing. The rotation of the peg (which should describe an hourglass in profile/3d) will lift and lower the wing via the slot affixed. Also, feather the beast, rather than single wing, with each feather being supported by a sliver of fiberglass rod (thinnest you can find) or if you want to spend the coin, carbon fiber rods (smaller than one mm) that are affixed to a main rod at the wing's leading edge.
Couple of suggestions (as if you need more) Why not use a geared dc motor? Could you design wings a bit like a rowing boats ore. So it gets updown motion but also frontback motion. Easier said than done of course
I don't have experience with ornithopter, but having experience with regular RC planes it seems to me your CoG is off. My advice would be to calculate your Center of Gravity, and adjusting nose weight accordingly. Then doing unpowered gliding flights, to check your CoG is spot on.
Cool design! But I still thing it is gonna be hard to make it fly properly. In my opinion part of the problem is when the wings flap up, they create force pushing the bird down which is useless force. I would suggest some mechanism that folds the wings while going up and unfold them while going down.
I'd wager you could add a pulley behind the motor and put the belt over the motor onto the second pulley, that is currently the first pulley which would give you a shallower angle for the belt.
James can you at some point do a video on options/feasibility of recycling old 3d prints? I think it's a modern problem that requires a bit more attention as PLA in a dump just means more microplastic in the environment so maybe there is a way to prevent this, options for other materials etc. Would be interesting what an avid 3D printer guy such as yourself thinks about it.
absolutely love this series - inspired me to look into this for my robotics dissertation! I'm wondering if, once a robust design was reached, it would be possible to have it learn to fly?
Very interesting project... now imagine if you could install thousands of servos to control individual little sections of the wings... pivoting and micro-adjusting small areas, opening up vents to let air pass through on the up-stroke for a more efficient lifting stroke or keep them closed to generate more thrust. Then an incredibly powerful CPU, able to make those micro-adjustments on the fly in real-time to respond to every imaginable scenario that might happen in the air or on launch and landing. Add in an advanced chemical powerplant able to convert just about any old hydrocarbon chain into useful fuel, and then a self-maintenance system that can repair it if it gets damaged, and you'll have _some_ idea of how complex birds are. xD Actually... the wing gaps idea is something I've often thought might be interesting to include in an ornithopter. Obviously you can't control it individually like a bird can its flight feathers, but bulk control might be possible. Alternatively, overlapping wing sections kind of like those "wind buster" umbrellas that let them get much bigger without collapsing in the wind, by kind of being two overlapping umbrellas in one: a small one overlapping a bigger ring-shaped one. I'm not sure if it'd be worth it overall, as it would be a method for increasing lift, and thrust seems to be more the problem here, but it's an interesting thought. Of course, a small part of me can't help but wonder at the futility of it. I mean, I get the challenge aspect, and the aesthetic triumph of recreating something that at least vaguely resembles how birds fly, but ultimately, if nature could make wheels or propellers, it totally would. They're just _better_ in most cases.
Although torque doesn't seem to be a problem, I've been using an A50S V2.2 (very lightweight) VESC, also in an ornithopter with a direct drive, 6008 KV180 motor.
![เปิด อาณาจักร หมื่นล้าน "สาระตั้ม" l [Nickynachat]](http://i.ytimg.com/vi/3XE0nPF4YkQ/mqdefault.jpg)
![[LIVE] : ONE ลุมพินี 87 วันนี้!! คู่เอก "ก้องชัย vs โชคปรีชา"](http://i.ytimg.com/vi/YN6e7HVnekw/mqdefault.jpg)
I completely relate to that feeling of “oh no it’s broken again, time to rebuild everything”. This tends to almost always be the case with DIY flying projects! That was really flying though! Can’t wait for part 4.
The beauty of any flying project is: they always come back down :D
@@inventiveowl395 And depending on how they come down they can even disassemble themselves
From kit to kit. :)
@@atvheads Unless your doing model rocketry and than its usually from Kit to Crater
Not everyone is rich though
Love following the journey on this.
My takeaway point in this is that it seems a miracle that birds can fly at all.
well, birds have had a good couple billion years of development time. I bet anyone could build a decent ornithopter if they had that long to work on it. :P
Same goes for walking in two legs. You don't realize how much goes into it unless you try to build a bipedal robot.
It helps to be able to feel what you're doing, as opposed to watching and guessing
Yup, reverse engineering nature is a science in and of itself.
Millions of years of evolution and natural selection can do some amazing things.
I've found out that using just a sligthly more expensive ESC instead of the yellow one that comes with those motors can make a big difference in motor torque
Yeah, those cheap ESCs aren't up to the task. I found out that with one particular receiver, the motor stuttered like it was skipping. I would not recommend these cheap ESCs.
Same experience here, I tried using 4 of them in a 3D printed quadcopter and they were spinning at different speeds, and had a very obvious step/notch to their speed curve. Try a better ESC and maybe another motor as I wasn't too impressed with those no-brand silver and orange ones.
I love your perseverance though James, keep going!
could flash it with blheli can then change settings on esc, probably easier to buy a better one though
Have you looked at the paths of wings during flight for various animals? They tend to figure 8 about their pivot, not just up and down.
There are plenty of examples of ornithopters that fly by utilising this far more simple approach - counterintuitive as it seems. The flapping creates mostly thrust and the position of the tail gives it an overall angle of attack. So you don't need all the complicated imitating bird biology.
This is more like butterfly wings, not a bird. I think
@@fast1nakus great point I think it's definitely more insect than bird
“Smashed everything to pieces”. Yet you continue onward … unbothered. That is extraordinary.
If the tail is a bit more adjustable, up and down, and a bit side to side, you can compensate for the stalling and turning out in the field. This is how smaller DIY ornithopters work, the tail piece is attached to a thin piece of metal so you can slightly bend it in different directions to trim/tweak the flight. Either way, keep it up!! This is a great series!
Very based you are the one making the birds for the government
Pardon?
Talk like yoda you do !
Lol
You should catch a bird and smash it with a hammer to prove its a robot
i think a good idea would be to remove the rod you already cut up, and extend the shorter angled ones up to the main CF rod. i think this would achieve this "air pocket effect" without the wing deforming and letting the air escape. another design i frequently see is that parts of the wing have their own joint and move freely to remove some air resistance when the wing moves upwards, this would probably help a lot too. great video!
I love that you are showing how much work engineering and design is. Some people dont appreciate how many iterations it takes to make something that looks like the obvious answer.
I don't know a whole lot about these things, but the toy version created some artifical lift by angling the tail up a little, if i remember the demo of the toy right. May be worth looking into. i cant wait to see the next version. Thanks you for keeping making interesting and challenging videos.
Yes tail up, less weight and more power… belt drive 🙅🏼♂️
I wonder if it was because of the center of mass on that toy, to keep it from nose diving.
This project gets more cleanly at the interplay of deliberate thought and action, and tinkering that I think is at the heart of good science and engineering. There are other projects I am more excited about by virtue of what they aim to accomplish, but I continue to be excited about this one because of the nature of the development process.
Feathers on the trailing edge of a bird's wing separate when wings are moving up to let air pass more easily, and then come together to create a larger surface area to push more air when flapping down. (At least on some birds.) If you need more lift, you could try somehow designing this into the trailing edge of your ornithopter wings.
An angled hinge would allow differential movement, but the added weight may be an issue...unless, maybe semi-rigid (celluloid? Plastics? ) feathers and a cloth hinge?
What if you ditched the belt drive and went with a compact planetary gear?
A planetary gear would eliminate any of the issues you have with the belts slipping and may even end up a hair lighter then your current setup.
I'm not sure if you could do this with a planetary but a non-circular gear might be a good idea. Unlike a propellor, the torque needed to flap the wings changes a lot through the cycle and a changing gear ratio might help to counteract that.
Try dropping it from a high building just to get it flying first. Then you can work on the launching but knowing it can fly properly! Very inspiring work indeed!
Found it! There's a man named Sean Kinkade who spent many years refining these. (Sean Kinkade Slow Hawk Remote Control Ornithopter) He was finally starting to manufacture and sell them when he was killed in a car accident. I found one for sale and with a bit of research you should be able to find his design. I do know CG is a bit further back than you think, they flap with high alpha and glide level. I've seen one fly and it's brilliant! Loving this series James...
Instead of gluing the rods to the wing using fabric strips, you could use a sewing machine to create piping in the wing fabric, with removable rods slid inside them.
somebody will love this series in 100 years from now looking back
6:39 it's FALLING...with style.
Awesome!!! 🔥🔥 We need part 4, 5, 6!!! Let's go 💪
Seeing the progress on this one is exciting, ready for the next one!
This is my favorite project you’ve done in a long time. Please see it through can’t wait to see the next video
You are so close! Don't stop!!!! I can't wait to see the new redesign
Getting closer. keep at it, and you'll get it!
Hell yeah thank you once again for bringing us along as you design, build, and test.
Really liking this series James, definitely nice to see you mixing things up. Now strap an Arduino on that back of that birdy and make her fly waypoints!!
I love how battery,motors and 3d printers have developed in leaps and bounds,so there's a way to reliably design,build and run something awesome., ThankQ. TkEZ
If you look at the way a birds wing works, it is hinged on the leading edge half way so on the up stroke the wing bends slightly lessening the up-force but on the down stroke it locks into place giving it full down-force resulting in more lift. You are getting there tho. Great video ✌✌
You may want to consider some sort of curved skis to attach to the Ornithopter during the testing phase so that you can minimize the amount of damage to your structure and get a softer landing each flight.
James, Now that you figured out the main wing - its really all about an articulating tail for bigger "birds". Cheers
Good work as always James! Don't be discouraged by failure - any engineer can relate to the pain in it all
Imagine how little time has passed since the Wright Bros? This is thrilling to watch.
I would try using 6mm rods for the wings. Then in the center of each wing rod, putting a hinge on it so that when the wings are flat they hold in place, but when the wings lift up they bend. upwards at the hinge. I would probably put some light elastic over the top of the hinge so the wings want to naturally remain flat when not being forced up by the motor.
I've built toy planes + drones for years, you're very ambitious doing an Ornithopter first!
I first started watching your videos many years ago with your homemade vaccuum forming machine, you are very inspirational and im glad I subscribed God bless you and keep you safe.
I'm really enjoying this ornithopter series. thanks
Amazing technical video and production quality as per usual! Keep it up - tuned in for part #4
I would try making the body heavier so that it moves less vertically when the wings flap thus enabling the wings to exert more force on the bird. I would also put the braces on the top side of the wings in sections which are tight so that on the down movement they are pushed together and act like a solid brace but on the up motion they disconnect and let air flow around the wing better. Love your videos by the way !!
Keep at it James. This thing will fly to the moon someday.
It's getting real close now! Great series. Mel looks really interesting as well, checking that out.
Just a suggestion from an old RC flier: Start with controlled glide tests. I suggest setting the wings slightly up and locking them in place to do glide testing. The goal is to adjust your center of gravity and tail for a smooth glide at what you would consider "Approach Speed". That will take a lot of the crashing out of your testing. It would also allow you to design some kind of control mechanism. Many years ago I had a tiny little ornithopter, barely a 6" wingspan, that changed the tension of the wings alternately for steering. It did this with a moving arm at the trailing edge of the wing sails, and this functioned as the mounting point for the aft edge of the sails as well. It didn't take much movement of this arm to induce a turn. You could also add elevator control to the tail, just move the whole thing up and down. Remember that these flat, sail wings produce very little lift of their own, needing forward speed and angle of attack to maintain flight (Think "Hang glider"), so control of the pitch will make this *much* easier to fly. Hope this helps.
Nifty ! Was hoping you didn't give up on it, its been a while. You'll get it, just stick with it.
This is a really cool project, but I think you’re overlooking some important aspects of designing something that flies.
The first is longitudinal static stability. I’ve noticed that on some throws, the ornithopter pitches up immediately, slows down, and then drops like a rock, because it’s going too slow for the wings to produce enough lift. Does the ornithopter glide alright without power or control input? If not, it won’t matter how good of a mechanism you have, it will crash every time.
The second thing is drag. Having the drive mechanism exposed to the free stream is going to produce a lot of drag, and that may prevent the ornithopter from gaining enough speed to remain airborne. Tom Stanton has a good video on how to design and print a single-layer shell that could act as a fairing and reduce drag. It could also serve as a sacrificial energy absorber during crashes and save the drive mechanism.
My favorite project so far, keep on going you'll get it flying eventually.
I feel like you need to build a rig to hang it from to really test it and dial in the movements... almost like the equivalent of a wind tunnel (not a wind tunnel, but something that serves the same purpose but for an Ornithopter) -- something where you can maybe just hang it and flap the wings and see if it goes up or tilts at all and just get it to where you can dial in the movements.
I love your project. One strong recommendation is to use a much larger motor and a larger ESC.
Very much looking forward to seeing this project finally get to where you want it. :)
I had a large ornithopter that came from China and from what I can remember, it had a lot of power and thrashed those wings really quite fast. It would climb really well to quite a height. The three things that were apparent to me were to be as light as possible, have a really powerful motor and to really thrash those wings. Holding it while testing it was quite a scary thing to do and bits would break all the time.
Might be interesting to look into where you can figure how to make it so on the up stroke it flexes in a way that provides more forward motion and in the down stroke it flexes in a different way that provides more lift.
This is insane, I was literally just starting to build one and then boom ur video came out
THIS CHANNEL IS SO AWESOME!
Food for thought - motor speed reduction through an optimised / light weight version of the cycloidal drive. This would present it's own design challenges for the symmetrical drive to the wings, but final drive ratio is easy to achieve and the compact design would remove all the issues with belt tensions and stress loadings, as well as centralise the weight which could be kept close to the centre of gravity to aid stability.
With careful design, perhaps using fibreglass or aluminium for the cycloidal disc to help minimise weight, I think the total weight would not be too different from that of all the "beefed up" printed frames, bearings and pulleys in the current design. As a bonus, it would certainly remove most of the failure points in the event of a crash.
James,
Pivot it at the elbows not the shoulders, so the inboard goes up when the outboard goes down
I think the simple flapping shoulder works by creating thrust and then you just need to give it an overall angle of attack. Anything else gets ridiculously complicated mechanically. I have a large Kinkade ornithopter that works on this principle.
Nice!
As you found tail heavy doesn't do well.
There's a saying in the rc plane hobby; nose heavy flys difficult, tail heavy flys once.
Well this is totally new to me! Very interesting stuff and lovely to see you enjoying it! 👍
Thank you for sharing your experience and stepwise success.
In my opinion a compact transmission for lightweight and less air resistance could help.
Also, I think your wings don't flap like a birds, they're more mechanical and less flexible. If you add more points of movement to add power to the strokes downwards and less power to the strokes upwards to maximize the effectiveness of the wings. You want more surface area when stoking down than when stroking up. This would make the machine a bit bigger but I think it'd still be doable with the tech we have right now. Hell even easy if you know what you're doing which I have faith in you in ^^
Hey James Bruton, well done on your vehicle! It is coming along and taking shape! Fantastic!!!!
IMO, You should design your next one with a launching mechanism so that way the launch is the same each time, and maybe put a tether above, like a "dog run," to keep the Machine from hitting the ground breaking everything. Also, 😅 make a "slide of some sort to distribute weights easier and faster with marks to show how far each move is.
Back in the early days of electric airplanes everything was brushed with gear boxes. If I remember right the gears were to increase the rpm's. A brushed motor might be what you want
There are a couple manufactures that make a lightweight PLA that can potentially reduce your weight by half, at the cost of some structural integrity: ESun PLA-LW or ColorFabb LW-PLA
dragonfly backpack? I see that as the end result of you exploring the next step in this design which is more than simple linear up and down flapping. you gotta scoop it
Do not underestimate the difference a bigger space for flying and a proper throw can make. Especially at this point where it's almost flying!
I love your projects. Thank you for sharing.
oh, i love these! went through a phase for a week after i watched dune where everything i said was ornithopter-related
That ornithopter looks really cool 💚
I know its way easier said than done, but if you had the wings have individual "feathers" of wing structure, that grouped up to make the wing, and then had them tilt on the upstroke, you could reduce a lot of downward force on your wings.
good job so far mate, looking forward to part 4 :)
Perhaps consider motorizing the tail a bit, then using accellerometer and gravimetric info, in addition to wing stroke timing, to decide when to point the tail up or down
it might be good to syncronize the pulses mechanically with the main drive, then have a seperate motor operate some gimbal range for the tail's movement
Nice work! I like the project choice!
I have been really enjoying this series but I have to say that it hurts every time I see the chassis glued to the carbon fiber rods. especially when prototyping as that means you cant replace / retention the belts on the fly and just makes repairs that much harder. I really hope you come up with a better method for the next version. glue is quick easy and strong but you know as well as I, it is a bit of a cop out / short cut.
Yeah, definitely a better/less permanent solution than glue needed. My ideas include:
1) Set screws. No idea how these would fair with carbon fiber. I'm guessing poorly, as the hollow rod might just crush. Or, it could damage the print instead, because the force is decently concentrated.
2) Screw clamp. Probably the most materially reliable option I can think of, but might lack grip strength. The wide contact area means that both the print and rod are unlikely to be damaged, save for around the screw on the print. A set screw might have an advantage in terms of grip, because it would dig into the material somewhat, unlike a clamp.
3) Low temperature hot glue (Replacing glue with glue... Sometimes, I even impress myself). If the temperatures match up, this could be perfect. Or, it could be nightmarish. The main advantage here is high strength, as well as being a drop-in replacement for the current glue. Whether this is practical is a matter of whether or not the glue can be made to melt without also melting/weakening the print. A precursory Google search tells me that ABS's glass transition temperature is up above 100 C, and that some hot glues melt as low as 65 C. If the structure can be brought to between that range, the glue should be workable, and the print should remain intact. The placement could be readjusted, and then the glue allowed to cool. I think the idea has its merits, but either of the two listed above are a safer bet.
I think the best luck you could have is to add a little extra to the back to control the pitch of the tail, if you want to mimic a bird allow the back before the tail to have a decent amount of flex
A 3d printer sponsor, that's the life right there
Using a V-tail or maybe crescent shaped, could give you both rudder and elevator if you articulate it in two axes; a third axis and you can even get roll as well with some clever linkage (or even just sequence of individually powered articulations.
8:30 that looks so cool!
for the wing struts/reinforcement rather than using a thinner carbon tube that is more flexible in all directions and could crumple and buckle/kink if over driven. Instead keep using the thicker rod but only cut on the top and bottom surface 90% of the way through, you could make a rod that would be flexible only in the desired direction/orientation.
You ever think it might be time to start laser cutting thin sheet metal and getting a simple sheet metal bender for components? you could still bolt and glue things together, and you could unlock much smaller/stronger design forms.
Keep plugging away. Great stuff!
I'm excited to see the next iteration!
Maybe you could put a soft bumper on the front - foam plastic or rubber so it protects it. :D
Love your videos, this is off topic from the current video but a project that would be cool for you to build / walk us through would be to build a mini robot like Bolt. The one used by pros to do programmed camera movements. But you could scale this down for a go pro or smartphone.
th-cam.com/video/IN8tjTk8ExI/w-d-xo.html
I love the passion of this kind of people they are really good inspiration for me :)
Looking good so far. Looks like you can still reduce a little bit of weight by shaving away unnecessary corners
One idea I had was to make the wing able to bend upwards but not downwards. When flapping up, the wings will bend resulting in less surface area, but when downwards, the bend will straighten out like a knee joint, increasing downwards surface area.
I think you need to automate motor speed in a feedback loop with actual lift. Maybe you can even setup a rig holding the ornithopter with strings and letting it learn by itself.
Check the bird ornithopter to see what the body weight to wing ratio is. It is possible you need a heavier or lighter body.
You are awesome James
So you're looking to make the gearing lighter and smaller (and more rigged). How about something like that then? Probably not one of your big cycloidal drives (though, that'd be a lot of power), but how about a harmonic drive, or, better yet, a planetary gearbox? Those are pretty simple to design and work out the ratio (once you know the right equation) and would be more compact and more rigid overall.
Just a thought, love to see flying machines of all description :)
Cycloidal drive is too heavy. Harmonic drive too lossy. Planetary would be ideal, but 3D printed gears would hardly be up to the task.
No, I think belts _are_ the way to go here, only the front one would need to be made more robust (perhaps wider, and idlers could certainly help).
An interesting alternative could be to avoid the crank mechanism entirely, and instead move the wing with _rope winches_.
Ditch the gears and make a rotating oblong angled peg that slots in an a slot affixed to the wing. The rotation of the peg (which should describe an hourglass in profile/3d) will lift and lower the wing via the slot affixed. Also, feather the beast, rather than single wing, with each feather being supported by a sliver of fiberglass rod (thinnest you can find) or if you want to spend the coin, carbon fiber rods (smaller than one mm) that are affixed to a main rod at the wing's leading edge.
Great !!! Thanks for share all !!!
Very cool, always liked 'thopters.
Couple of suggestions (as if you need more)
Why not use a geared dc motor?
Could you design wings a bit like a rowing boats ore. So it gets updown motion but also frontback motion.
Easier said than done of course
I don't have experience with ornithopter, but having experience with regular RC planes it seems to me your CoG is off.
My advice would be to calculate your Center of Gravity, and adjusting nose weight accordingly. Then doing unpowered gliding flights, to check your CoG is spot on.
Cool design! But I still thing it is gonna be hard to make it fly properly. In my opinion part of the problem is when the wings flap up, they create force pushing the bird down which is useless force. I would suggest some mechanism that folds the wings while going up and unfold them while going down.
I'd wager you could add a pulley behind the motor and put the belt over the motor onto the second pulley, that is currently the first pulley which would give you a shallower angle for the belt.
Getting closer!!
Good bird creation from your side
congrats for 1.04M subs
James can you at some point do a video on options/feasibility of recycling old 3d prints? I think it's a modern problem that requires a bit more attention as PLA in a dump just means more microplastic in the environment so maybe there is a way to prevent this, options for other materials etc. Would be interesting what an avid 3D printer guy such as yourself thinks about it.
absolutely love this series - inspired me to look into this for my robotics dissertation! I'm wondering if, once a robust design was reached, it would be possible to have it learn to fly?
"Learn to fly" what the hell are you on about?
Very interesting project... now imagine if you could install thousands of servos to control individual little sections of the wings... pivoting and micro-adjusting small areas, opening up vents to let air pass through on the up-stroke for a more efficient lifting stroke or keep them closed to generate more thrust. Then an incredibly powerful CPU, able to make those micro-adjustments on the fly in real-time to respond to every imaginable scenario that might happen in the air or on launch and landing. Add in an advanced chemical powerplant able to convert just about any old hydrocarbon chain into useful fuel, and then a self-maintenance system that can repair it if it gets damaged, and you'll have _some_ idea of how complex birds are. xD
Actually... the wing gaps idea is something I've often thought might be interesting to include in an ornithopter. Obviously you can't control it individually like a bird can its flight feathers, but bulk control might be possible. Alternatively, overlapping wing sections kind of like those "wind buster" umbrellas that let them get much bigger without collapsing in the wind, by kind of being two overlapping umbrellas in one: a small one overlapping a bigger ring-shaped one. I'm not sure if it'd be worth it overall, as it would be a method for increasing lift, and thrust seems to be more the problem here, but it's an interesting thought.
Of course, a small part of me can't help but wonder at the futility of it. I mean, I get the challenge aspect, and the aesthetic triumph of recreating something that at least vaguely resembles how birds fly, but ultimately, if nature could make wheels or propellers, it totally would. They're just _better_ in most cases.
Hey James! While I really like your solution with the belt reduction, wouldn't a worm drive be better suited here?
Hoping to see part 4 soon!
Although torque doesn't seem to be a problem, I've been using an A50S V2.2 (very lightweight) VESC, also in an ornithopter with a direct drive, 6008 KV180 motor.