Join my DISCORD! discord.gg/fc6MedG7eW INSTAGRAM: instagram.com/xrobotsuk TWITTER: twitter.com/xrobotsuk MASTODON: fosstodon.org/@XRobotsUK

@loveblowsbad

Жыл бұрын

if you make the pins buoyant and put it in a boat will waves drive the vehicle like Veritasiums wind car? wind car seen here kzread.info/dash/bejne/nK2F2cl7ccSokco.html

@munchobell2890

Жыл бұрын

I think the best way for that to move if Added a flexable bottom like point, so that can bend and cause a force forward by leaning it in one way, but this only walks one way.

@DmitrySinyukov

Жыл бұрын

What if you attach short "static" legs perpendicular to the rubber strip you have on the bottom? In other words, if you cover the bottom side of the rubber strip with needles pointing down. Maybe they'll act similar to centipedes legs.

@gshaindrich

Жыл бұрын

the error is, why would you think that?, is that "the wave" generates a "force" ... after all those years of experimental mechanics, haven´t you still not understood it? The cam is only pushing up and down, so where should the forward/backward force come from?

@WetDoggo

Жыл бұрын

12:27 you need a pivot point in the toast center... otherwise the default pivot point is on the ground. the initial experiment was kinda pointless, but I guess you expected that anyways. i was curious about the rubber strip too and I thought it could work to be honest 😅 but the centipede leg version is very promising, just needs a dedicated pivot point, so it won't simply pivot on the ground.

OK, so it doesn't walk brilliantly... But how well does it play piano?

@jasonree

Жыл бұрын

Was going to ask the same..my guess was a piano playing robot…

@nkronert

Жыл бұрын

It will play piano better than it walks, depending on your music taste of course 😉

@bornach

Жыл бұрын

Or the pipe organ! Send it to This Museum Is Not Obsolete

@pawesliwinski7234

Жыл бұрын

Hahahaha 😂

@RealAndySkibba

Жыл бұрын

The world needs to know!!!

Really really like the slightly updated format where you talk over the pieces as they’re printing. It provides more context for your build compared to the old bit where it was just a montage with music.

@ddegn

Жыл бұрын

Agreed! The toast made me hungry.

The previous wave mechanism gives you the solution if you look at it carfully: I believe the rubber ribbon you used here IS part of the the answer, as long as you fill it with 2-3cm long studs like a brush. The studs will follow the orientation of the wave so their tips will move side to side as they foollow the sine, transforming the sine into horisontal micro-movement. Oh and you also need good grip between these studs and the surface the machine is walking on.

@AV1461

Жыл бұрын

Yes. I like the idea. I just thought of tilting it forward so each leg would hit at an angle (or shave a bit of each leg, so it sits tilted forward). But your idea I think fits better with the original concept.

@claytonfs

Жыл бұрын

Yep, I'm surprised James didn't see it. The studs on the original belt stay perpendicular to the belt as it flexes over the sine shape. Like this: /|\|/|\|/|\|/|\ The result is each stud traces an inverted arc while in contact with the ground. Oddly, making the mechanism walk in the opposite direction as the wave progression.

@ldiajvclxizo

Жыл бұрын

I think this will work

@chrismofer

Жыл бұрын

I also noticed that the only one that walked well had extended rubber feets that act like little levers and actually cause the forward motion to happen. simply lifting up and setting down your feet won't get you anywhere.

@janilepisto5084

Жыл бұрын

this is correct solution

I think you could get this to work with a different cam shape. Your final slotted design has a leg only touching the ground briefly. The forward back motion during this time is too small, and vibration is overriding any potential wins. You want to move the leg back (to move robot forward) only when the leg is touching the ground. So the cam should have 4 distinct phases: Down and back, lift, up and forward, lower. Where up/down are controlled by the cam radius, and forward/back by the slot position. Your final cam will have two quarter circles and two shapes connecting them. You might also need to put a thinner section in the middle of the toaster slots, so that the forward/back motion has something to lever off of.

@keksdieb4020

Жыл бұрын

Thank you, thats exactly what I wanted to suggest!

@James02876

Жыл бұрын

It's the "lobe" of the cam that needs to change. Rather than being egg-shaped, the peak should be wider so the foot stays extended while the groove moves the foot backward.

@codybuzzard1

Жыл бұрын

@@James02876 This is exactly what I was thinking. If you made the top of the "toast" pieces come to a point, you could have the bottom of the cam, when it is pushing the furtherest down, nearly flat but with the groove for the pointed cam follower angled towards the front. Then the toast would dwell at the bottom (and be touching the ground) while simultaneously being pushed backwards.

@matt7403

Жыл бұрын

Exactly! I might add that you could have a bit of a bulge or a ridge at the halfway/middle of the new thinner toasts, which would provide a precise fulcrum location and allow a larger and more controlled motion at the bottom. That is, controlled front back motion at the top and a controlled pivot at the middle gives controlled front back motion at the bottom. Raising or lowering that fulcrum modifies the magnitude of this front back stroke.

@gordonhutchison8098

Жыл бұрын

Yep! Exactly what I wanted to say. Have the cam lobes changed so the leg (toast) touches the ground for, say, 180 degrees of cam rotation, while the cam groove moves the leg slowly back. Then say 60 degrees each for moving up, forward and down. So a slow push back on the ground, followed by a quicker up, forward, down motion. Combine the last three for a smoother leg motion.

If the primary force from each piece of toast is vertical, then the body will only move vertically which is what we observe. If you were to tilt the toast bed (reduce the height of one side) in the desired direction of travel, each piece of toast would then be pushing up and forward resulting in forward motion. That said, the cam wobble concept is brilliant. I dont believe the reason for lack of horizontal motion with this design is due to the concept, but more the execution. The toast needs to make contact with the ground at the most forward part of its ellipse, maintain contact with the ground during wobble travel (as the cam groove exerts horizontal force), and then retract after the maximum horizontal length of travel. This would require that the part of the cam with the 'bottom travel wobble' to be nearly flat. Likely a more resilient contact groove between the toast and the cam would be required, but I dig this concept. As it is, the wobble would likely work in water if the buoyancy level allowed the toast pieces to be in water during its wobble travel on the bottom half of the cam rotation. Either way, looking forward to the next stage!

@AV1461

Жыл бұрын

So many cool ideas on the comments to solve it! xD. Fiddling with the cam profile also sound like a quite nice idea to me.

@SnowyYukiThe2nd

Жыл бұрын

I would love to see if the cam w/ grooves design would work if it was driven a lot slower. I think the bouncing and lack of friction on the plastic feet might have played a big role in its subpar performance

@SmallSpoonBrigade

Жыл бұрын

It looks like the only reason it's moving forward is that the design isn't balanced. If the motor were in the middle it probably wouldn't move at all and if it were on the backside then it would move backwards. Adding the tracks would allow the normal force of the track to be something other than straight up and down and give it a bit of forward or backward push depending upon the direction of the cam shaft rotation. Or at least that's what it looks like to me.

@MrIrondog55

Жыл бұрын

I agree with Brians post above. Personally I would keep the current cams and "split" the toast vertically. One hemisphere will be lower or higher depending on the cams rotation. This will give it more of a walking movement, much like the screw tank, where each "foot" is exerting opposite horizontal forces to propel it. The other smoother option is to overengineer and use 2 cams at alternating rotations, but you could get away with the above section with one. Also, if you add angled rubber feet, it will add a more gravitational aspect where it will "lean" providing applicable force to each side and step of the robot. Kinda like going downhill. Hope this helps.

You know someone is a good engineer when they make something three times as complicated but one third as effective. 😉

Always fun to see you test your hypotheses and optimize. You always have a better idea of the concept you are engineering by the end of the video. Excited to see your next installment in this wave drive series.

I always tell my engineers that failure will teach you way more and honestly I think I enjoyed learning through this video and seeing James' struggle so much more than usual.

The tracked version with the spiral works because the track is hinged and the feet are fairly wide and grippy. The connections to the other feet cause the contact point/"feet" to pivot while remaining in contact with the ground. The motion is actually pretty similar to biological feet when you think about it; the back part of the foot makes contact with the ground, then the entire foot pivots, moving the contact patch to the front of the foot before it lifts off of the ground.

I love it when you publish your failures. We all learn a lot and I think it sends the right message to young folks that even really smart people sometimes have setbacks and to keep trying.

James Bruton, the man whose failures are far more interesting than other KZreadrs successes! Great video, Jimbob!

I lost it at "eventually I want to make a giant land worm that I can ride on" at 14:07

The original track worked because it had more effective "legs." Think of each tread moving up and down along the spiral guide. Each point along the guide - relative to the side - is going up and down without any forward or backward bias. However, since the angle of the guide - again from a side perspective - is changing like a sin wave. This means if you have an offset from that point on the guide, like a rubber foot on a track segment, the foot swings like a lever or a "leg." This is what provided the forward motion in the original design. If you want to improve on the design, you might consider increasing the length of the track, and the spaces between the legs (or simply staggering them on the inside and outside of the track), and lengthening the legs themselves to increase the travel of the end of leg.

@TimmmmCam

Жыл бұрын

Yeah I did point that out on the original video but maybe my explanation wasn't clear enough: kzread.info/dash/bejne/goWlyriznaW0ibg.html&lc=UgySaBczRsi15sOXjqR4AaABAg

@hitnovak

Жыл бұрын

No, the original track worked because the legs actually moved back and forth, as clearly illustrated in this clip: kzread.info/dash/bejne/f6SkpNGBZ8TVdbg.html (blue line represents the rotating spiral as viewed from the side, and the orange dots represent the legs, equally spaced along the spiral). The clip also illustrates where the jerky movement in the original track design comes from, because in the clip the rightmost dot has a lot of sideways motion, while in the original track the last leg is attached to the frame which doesn't allow for that kind of motion (this wouldn't be a problem if the spiral was exactly one full rotation, see the dot around 6.2 [= 2 * pi] which has almost no sideways motion).

@TimmmmCam

Жыл бұрын

@@hitnovak Yeah you are right. I thought that's what @billeethesciencegeek was saying but I actually read his comment this time and he also seems confused. Honestly this is all really obvious. I don't really understand why people - even ones as smart as James - have difficulty with it. In this video the legs are clearly just going up and down. Of course that isn't going to work.

@0osk

Жыл бұрын

7:36 he identifies why his new system doesn't work compared to the old system in the video.

Props for you posting this video. Some makers would not even showed it due to it not working as you wanted it to. Learning how to make something work is figuring out what doesn't work.

I Respect your integrity for sharing your failures as well as your triumphs

2:06 That is where a laser cutter would come in handy, flat parts could be made in minutes as a pose to hours 3d printing. once again you came up with simple yet brilliant solutions to little problems, I really like the rubber band toaster raisers,

@Shocker99

Жыл бұрын

I agree on the laser cutter.

@SmallSpoonBrigade

Жыл бұрын

Yes, is there anything that lasers don't make better?

It's so much fun to watch an engineer (or anyone, really) just take an idea and test it to see what happens.

7:43 On the original screw drive, we can see that the little black feet change angle quite significantly so that they are actually taking steps like a centipede. If a strip like the black one from this video was attached to each vertical paddle and allowed to change angle with axles, with feet attached to the strip, they could also change angle and achieve walking. The result would be a finely detailed version of the screw drive's walking.

I love it that someone is on the ball setting up Sponsorblock portions.

Thanks for being willing to show us some of your failures among your many, many successes and triumphs.

This is one of my favorite things ive ever seen a youtuber build.

I love that you show simi failures too Thanks for sharing your expirence with all of us 👍 😀

"Eventually I want to build a giant land-worm that I can ride on" that sentence really cracked me up xD

Your persistence in the face of failure is amazing, although your "failure rate" is amazingly low. This also reminds me of the strandbeest videos which are incredible.

I find it surprising how much I enjoy seeing you fail upwards (testing and developing). I enjoy it way more than videos where you just see the final project, presented as if it needed no development and when perfectly on the first try

Thank you for this kinematic gem! I've been toying with wave driving for some time... Here is the deal. The (ideal) waves by themselves do not carry the mass. You need some motion biasing to actually move. The roll of tape moved because the gravity was pushing it off the rolling hill so to speak. It was the right size for it, comparable to the period of the wave. Your initial version with the toast pieces should walk as well, if you run it over a bunch of eight-sided pipes for the same reason. The previous version (with the screw) worked because the track was imperfect, as in wobbling back and forth, and acting on the ground at varying angles. The newer version got a lot closer to a good wave generator but still was able to move a bit because the toast pieces could wobble back and forth a bit. Millipedes and worms have a different kinematic, it is closer to one-dimensional wave. They use clever tricks to bias their wavy motion forward. Bias is the key here. The millipede achieves that by angling its legs that are not straight. The worms have small bristles and irregularly shaped segments to drive them forward. Snake uses terrain in normal motion (they have a weird other mode of commotion too, but use it rarely) to push itself off. I hope this little setback will not discourage you from attempting v3!

@Benco97

Жыл бұрын

THANK YOU for saying this, My god I was going mad watching this video.

I like how you put in the effort to make a flawwed design just to show that it needed to be the screw. I went into this knowing that just up and down was not going to be enough but it's interesting to see the exploration of the idea

The joy you feel at having to rebuild 30 odd pieces that are exactly the same over and over again really comes through in this video.

"Eventually i want to build a giant land-worm" this is why I love this channel. especially these experimental robot designs, they're such interesting concepts to see take shape and be experimented with!

Brilliant as usual. Absolutely incredible the amount of talent here.

I thought the carpet would have done the trick. Wonderful visually stunning bot, can't wait to see the next iteration!

It can't go forward unless it's providing a forward force. It should stand in one spot unless you can transfer some vertical force to horizontal. It would be great on a slope! Try using a secondary cam like the coupling rod on a locomotive and almost like the spider legs you have as an example at the end of the video.

@nanorex_designs

Жыл бұрын

I feel like getting to this point without realizing the problem is a failure of basic engineering design methods. Even the most basic of force analysis done before building a whole-ass prototype and wasting all this plastic and time would have immediately revealed that this thing wouldn't work. Heck, even just intuition would suffice in this case. This channel has so much creativity but it pains me to see so much of it being squandered by poor design methodology.

Never been so entertained by a project that simply did not work at all ❤️

This is lesson in learning from the failures. Really good to see.

Don't give up James! You are awesome!

On the upside... you made a really fancy tap-dancing machine! 🕺💃

I can't believe how much work you've put into that thing! Bravo for doing it.

This was really interesting and I found it a lot more compelling to see the designs failures and what you changed to make them work compared to just jumping to the final product that is functional. Love your channel and I aspire to be half as productive as you are!

You might want to check out the biomechanics of snail and earthworm locomotion. Worms have a travelling wave, where alternating sections of anchoring and extension travel down the body. Snails use a traveling wave to create shear which solidifies mucus to create a travelling anchor.

If the toaster slots were at a bit of an angle, the leg would push the ground at an angle, so the motion will have an X component. As is it moves only the Y direction. This fix would make it go, but only in one way/forwards.

@ddegn

Жыл бұрын

Wouldn't it pull back in the x direction as the toast raised?

@SmallSpoonBrigade

Жыл бұрын

@@ddegn Yes, but if it's no longer in contact with the ground, then it should just counteract the momentum that the leg had, not whatever friction was doing to push the vehicle forward. And if there were a sliding mechanism to change the angle of the foot when it touches the surface, it should be able to speed up and slow down a bit, as well as change directions in dependent of just messing with the rotation of the camshaft.

...doesn't walk..but looks so dang cool! I could watch that thing sit and spin in place for hours! Love it!

I like that you stick to mentioning the sponsors on every video where the product placement is used. as one should.

I love the way your mind works! Some of the things you design may not be practical, but it's so cool seeing the process an results!

It seems maybe this would work okay like a paristaltic pump design. I'm really glad you put out the vid. Makers do a lot of things and some just don't work out. Its good to see others out there like me that put time and effort into something and it doesn't produce some amazing new thing and instead just goes toward more learning.

Looking at the proposed legs at the end seems like inventing your way towards the Strandbeest machines

I think the offset weight and the vibration of the motor was the only thing causing the motion. Its like a gigantic bristlebot

You could reshape the cams on this design to have the toasts be fully extended for longer. Essentially the inverse of what you had in this video, where you have a positive impulse wave with extension downtime at the fully retracted position. So instead shape the cams to make a negative impulse wave, with extension downtime at the fully extended position. If you combine that with some ridges sticking out in the toaster slots to constrain the toast to more defined fulcrums, then that should at least get it walking a little bit.

I've seen an invention similar to this which was a drive for a watercraft. This had the advantage of no cavitation and drove the boat along at a very high speed.

I believe your hypothesis about their time contacting the ground is correct. I think that, instead of egg shaped cams, it may contact longer with simple offset circles. The membrane is a good idea. One really expensive but fun little test is to replace the whole cam thing with the strandbeest leg system moving the slices of toast up and down with the leg movement.

Even though it didn’t work out as you hoped I found this to be one of your best and most interesting / instructive videos

Your slots for the toast need a pivot point in the middle of them that is narrower than the openings at the top and bottom. So you get a mechanical lever motion between the cam and the toast slot that forces the toast to pivot properly.

Most satisfying robot motion ever

Am at 04:30 and this is such a satisfying device! Love the rubber bands doing what they do. The repetitive nature of the wave is so mesmerizing! Edit: at 07:30 No! It isn’t just vibrations pushing it downhill… it’s pushed ‘towards’ the location of the motor irrespective of the direction of the wave because the virbratuobs are pushing it towards its tilted position. The device itself is not perfectly parallel to the table. It it ‘were’ just going down hill you’d never be able to go in the other direction… which you WERE able to :-) Anyway… that is me being picky! Carry on! Edit 2: at 09:30 it no longer works because what I described earlier no longer happens. The tilt of the entire device cause by the weight of the motor and caused by the feet retracting earlier is now being countered by the fact the rubber matt is attached at the end and keeps the device parallel to the table surface. It now ‘does’ infact only move as a result of the table being somewhat off level. Edit 3: I bet you can make it work by going back to the straight cams and give each ‘leg’ a hinged knee at the end that (just a little bit) hinges towards the drive direction as it lands on the ground. This could even be angled TPU flaps that all point backwards from the intended drive direction so that that weight of the device causes them to land, grip the ground and and then bend backwards as the weight is above it and the foot is being pressed further down. The only downside is that this allows only one direction of travel.

@ddegn

Жыл бұрын

"this is such a satisfying device!" Agreed. It was really a wonderfully awful contraption.

I've seen a concept similar to one at the end of the video in one book I've read. But the legs there were solid rotating details formed in such a shape they were forming a wave while moving together.

Just line up all the toast pieces then cut a diagonal line across the bottom of them all so they are all at different lengths and drive it with everything at an angle. When the cam pushes the legs down it will also push them back a little so you will get reliable forward movement with minimal change to the design.

I'm actually using the legs shown in the end for my engineering project, it basically works like a tank threads, it's simple and easy to build.

Thanks for putting this video out even though it was a very rare fail. It's great to see the thought process in action, and also that you are human :)

What previously worked to provide locomotion was the rotation of each of the links in the tread caused by the connection to a link's neighbor at a different height. Its not the up-down motion, but the rotation given to the links. Try adding a chain linked tread to your cam drive. You might see an improvement.

You learn more from failing than from not. Thanks!

James is next level.

"I want to make a giant land worm that I can ride on." This is the clarity of vision that gets funding kids!

I kind of love the one that doesn't walk in any direction with the wobble on the cam. It's beautiful. I don't know what that would be or is for, but it's definitely beautiful.

@zaibartdp

Жыл бұрын

I think you just described art.

for the next you should use that mechanism but do a grid of them so it could go sideways also though they would probably need a more complex mechanism for that or be motor controlled

It's really pretty though if nothing else!

Thanks!

It moved towards the motor end as it tips forward slightly when it picks up the first few bits of toast. The other end is lighter so it doesn’t tip backwards when the toast gets picked up

Your commitment to failure is admirable.

I think this design is promising! You should try adjusting the cam profile so the feet contact the ground for longer, and maybe work on tightening the tolerances and precisely mating the cam profile for the forward motion. I think the cam on cam to drive 2 axes is brilliantly simple.

it looks extremely cool though!

😅 Just having such a déjà vu: this must be the most complicated Moonwalk 🤣

The original track works because the contact point on the ground is lower than the hinge point on the track, and this means that as the segment is tilting while touching the ground it pushes forward on the segment contact point.

Very cool video, Even though i could see the outcome before watching, still loved to see the process its almost there, but the key is the angular motion of the flat parts of the previous design's track causing a forward leveraging . having a connection point between 2 toasts with a tall-ish rubber foot will sort it out 1- front leg comes down pivoting the flat toward to rear 2- second leg comes down shortly after maintaining the flats angle 3- flat then contacts ground through the foot while pivoted to rear 4- the second leg continues coming down pivoting the flat toward the front - this causes reverse motion through the foot's contact point being a distance away from the pivot center line repeat

The track at 6:00 worked, seemingly because the "toast" are not angled straight up and down, but instead slightly diagonally; this results in consistent movement in the opposite direction.

it moonwalks beautifully 🙂

Missed you saying, "that's all for now"! Great vid, just proves that some things you just have to build, to see that it doesnt work...

I think the last mechanism have a chance to work. Just needs a thicker cam so that it will have a wide movement (left to right). Then the center of the feet should have joints to the slot so the movement will have more motion and constant than the feet is just loosed to the slot.

That sucks that your iterations didnt pan out but the fact that you were thinking ahead and already had certain things printed into the parts was awesome. And even though it didn't work well it still looks really cool and I think you could use that same mechanism for something else so that it's not a total waste 😭 but we all learned something so it's not a bad thing

I think it moving despite physics is because of where the machines center of gravity is combined with the play you gave the slabs. It ever so slightly gives the legs a greater tendency to bend away from the motor than to it which moves the machine towards the motor. The guided slabs only make contact with the ground at their apex for a very short time in which they hardly change angle. A simple solution is to extend the time the slabs spend being fully extended.

You could also try a cam with a flat point. On the tip of the cam you will let the slot go diagonal while the height is constant. The cam would reverse the legs back to start when the legs are up, can be over the whole small diagonal, and during the ~20 degrees that it is fully extended it will have the slot go from front to back "quickly". Cam sideview would be aprox. The same as it is now but the tip looks "smashed flat" (or more approximately like a square with rounded corners). Basically increase the radius on the cam tip and add flat. Don't know how to put it in words but I hope you understand what I'm trying to say. Great video!

There is a 1950's sci-fi movie called "KRONOS" which featured a giant robot that looked like a modern fridge on 4 pillars for legs. The legs moved up and down only, but of course, this is a movie, so it worked and the robot was able to walk about the country side and trampled on much of California! Your machine reminded me of that movie with the way the T legs were going up and down and doing basically nothing. Also, check out that movie, it's lots of fun!

Hey James, I think one way to make your machine move is to attach a short leg to each bottom piece of toast, so it can rotate forward/aft slightly. The rotation should be linked to the position of the toast directly in front of it (or behind it). The linkage to the neighbouring piece of toast therefore must allow a sliding motion. This should give you the circular motion you need for walking. One problem could be the loose mounting of the toast pieces making the legs jam.

I have two suggestions for movement. Wider cams would let the legs be in contact with the ground longer, and adding some material to the bottom of the legs that can grip better than plastic

You would have to print more cams, but maybe it would help to make the wave gradually increase in depth towards the rear. That way, your robot will be leaning forward, and the toast would push down and back.

I’d love to see the toaster design with the rubber belt tested on sand. I think with a surface that isn’t completely smooth and hard it can push against the contour of the surface.

For the wave motion to drive forward, it not only have to have a height difference changing in one axis. If you look at one point travelling along the edge of a screw or a spiral in three dimension, it not only travels up and down, but at the same time, it is also travelling diagonally left and right from the center line. So just cut those "toasts" vertically into at least two pieces (three or more would be better), with each piece having their own retention rubber. Although, just cutting the "toasts" would make the setup wobbly, it would still go forward. To mitigate that, you would have to redesign your cam pieces, where the forward and backward facing plates are offset horizontally, or just make a screw with thick a edge. But if you don't want to destroy your original setup, putting a rolling track around the whole mechanism could also be a solution, but then each piece of track should have a bearing with an axis in line with the "toasts", which they can push against and force itself to move forward. (Just how the ball was pushed in one direction in the video.) It would turn basically the thing into a flattened out cycloidal drive.

I think it can be salvaged somewhat. Have a track of spheres or cylinders that the toast pushes along. The ground will likely keep the cylinders in place through friction enough for the bot to move forward. It's like making uneven ground that it carries around with it. Not quite the same type of drive, probably, but still interesting. Getting more complex, though.

Ingenious!

You should make a "Brutonized" version of a walking strandbeest. Also see if you can implement that brilliant print through a piece of fabric technique people are using to make flexible 3D printed clothes. I bet that technique could be used to make some kind of locomotion.

the real reason it's only moving in one direction is because the wedges are tilted slightly (you can see if you look closely) and it sort of creates the same mechanism that hex-bots (does anyone still remember those?) so it can only move in one direction regardless of wave direction which is why it still was able to move uphill for a short while, if you increased the angle then (whilst is would increase the stress and complexity) it could work just fine. The reason that it doesn't work with the webbing is because it is altering the angle of the legs and creating a neutral motion thus creating 0 thrust

your main issue is that you need a diagonal surface pushing against the ground. That is what propels a screw forwards. It's friction combined with the angle of attack. Just going up and down won't impart any forwards or backwards force. The main reason the track was able to move better was the angle of attack as well as there being a much larger surface area.

Very Interesting, Thank as always. Half way down the video, you demonstrated that vibration was moving the object. Or, the fact that it has "near air time" (a step section with no friction), cause gravity to move it down the slope. Making me feel you should build a "skipping robot". - use a gyro to generate forward vector - legs that drop to the ground, collect ground friction energy into a spring, and release.

I'd suggest adding flexible toes to the feet, basically nubs sticking down from the ends, with angles tips. Thanks to flexible material, the machine should push in the opposite direction of the angle on the toes.

If you draw a line at the zero of your sine wave and then draw a line at maybe 10 degrees from that line, it will make a triangle. Print and attach the triangle on both sides and it will turn your mechanism into a tilted free body diagram which will travel forward!

The mechanism you're describing at the end of the video is exactly how the Strandbeest by Theo Jansen locomotes.

Great video James! I'm thinking you might be able to get the rubber bottom to work if you added in some brush bristles to the rubber to get that circular motion. The mechanics are similar to how some piezo motors work.

I think you need a foot with rubber grip between each pair of legs. This would mimic sections of your track from the previous version. As the foot pivots back and forth this would give the forward force at the ground you need

I'm convinced The v1 screw drive worked so well not because of the wave like nature of the screw, but because each section of the track while at the bottom of the wave is turning around the bottom of the wave just like a wheel while it follows the edge of the curve. Having those rubber standoffs are like sticking tiny little legs on the end to amplify the turning radius. I don't believe any wave drive will work until the bottom of the wave has the ability to move significantly. The individual legs like in the centipede are moving significantly and so I think will definitely work much better. I would be interested to see what a flexible track like on v1 but attached to the bottom of v2 would do since it would combine the simplicity of the cam with the circular motion of the individual track pieces of v1. I would also be interested to see what it would look like if you attached significantly longer "legs" instead of rubber standoffs to the bottom of the v1 track as I suspect it would end up looking much more like the centipedes motion where the legs bunch up on the inside of the curve and spread out on the outside of the curve creating a torque against the ground while they are spreading out.

Sorry if this has been said before and I don't know how much more development you want to put into this type of mechanism, however I think that if you attached tracks to all of the pieces of toast with a pin on the toast and a free floating slot in the track and than every link had a thinner link that was not connected to anything besides it's adjacent links that it the up and down motion would cause the middle link to perform that sudo rolling motion that you are looking for. This may require widening the distance between the toast slices to accommodate Sliding pin slots and an extra free floating track link.

Close. Have an adjustable body for the cam shaft to angle those legs forward at some angle. That will provide forward thrust. Youll want to develop something that enables it to tilt back for backwards thrust. Something that tilts to the left or right for that thrust. Might benefit from 4 independent cams and leg segments to drive those changes