Anomalocaris Inspired Undulating Mechanism




Introduction: Anomalocaris Inspired Undulating Mechanism

About: I work in optical design.
This instructable covers the making of an undulating mechanism.
Just a thingy that you can turn and it moves in a wavy way.

Initially it started as a try to make an Anomalocaris:

But with six wings
That proved to be a bit unfeasible with just paperclips.
However the end result is very similar and the movement still looks quite good (especially that spine).
And it turns surprisingly smooth for a paperclip mechanism.

There are some things done that reduced the 'goodness' but I will point them out later.

Just, if somebody tries something similar, just make it bigger than this one, that's the main advice.

Step 1: Materials Used

Materials used:

-Different paper clips:
A lot of smaller, easy to bend ones.
A couple of strong, hard to bend ones.
-Thick plastic (whiteboard) marker

-Some discardable ballpoint pens.

-A flexible, plastic ruler. (often given as advertisement).

-An even more flexible, very flat (almost paper thin), small plastic ruler. (also advertisment).

-rubber band

-some twine

Tools used:

-lots of glue

-Soldering Iron (but can be replaced by whatever can fix two wires together (solder is just very fast and easy)

-Needle nose pliers (or sust fine pliers). One is necessary but a second one is highly recommended.

-Pointy object to make a fine hole in plastic.

-Perforator to cut small plastic circles (out of the small ruler).

-Whatever helps.

Step 2: The Crankshaft

So, in the video it is clear that the wings should move in a sine-wive pattern.  This can be established most easily with a crankshaft, with the cranks positioned in a spiral around the shaft.

Basicly, a rotational movement will be converted into an up-and down kind of movement, which is practicly the definition of a sine.
Look here especially at the animated gif at the bottom. 
Or julst look a bit further in the instructable, it will all be clear.

Since this machine was supposed to look like an anomalocaris with 12 wings, (six at each side). Those wings, in order to be realistic, should tilt when they go up and down. You know, to provide thrust in the water. The simplest solution is to let the front and end of the wings be powered by different cranks that have a small phase difference. The first one, attached to the front of the wing is rotated a bit further than the second one so that the front of the wing goes down first when the entire wing comes down. (look at any slow motion video of a flying/swimming animal, to see what I mean.)
As you can see in the video, the phase difference of the anomalocaris is small. With a dragonfly, there is a huge difference.

To begin, search a strong paperclip (it is an axle that powers all the wings, so it must not be very bendy).
The paper clip is straightened out and cranks are added.
Since it is a very small machine, the cranks must be as small as your pliers allow them to be.
Because 2 cranks power one wing, the cranks are paired up. (look at the pictures)
12 cranks are needed, only 6 of them fit on one paper-clip, so a second one is needed. Fixing those together to one single shaft will be covered later.

As can be seen in the video, the wings of the anomalocaris show only 1 full sine wave. ( sin (x) for x in [0, 2pi] or [0°,360°]) therefore the cranks must spiral only once down the shaft. (or half for each paper-clip)
There are 6 pairs and the first and last wing move exactly the same, therefore the last pair of cranks (or end of the spiral) must have the exact same position as the first pair of cranks. 
In order to achieve this, the crank pairs must be twisted along the shaft about 2pi/5 radians or 72°. 

Then, as explained earlier, the wings must tilt and require a phase difference. So the first and second crank  of each pair are bent a bit in different directions. ( The average angle between each pair must stay 72° so constant adjustment is required.) How much depends on how much space there is between the cranks.  

Finally make sure that all the bases of the cranks are still in a single line. This is done by rotating each paperclip very fast.  You shouldn't see movement in all parts of the shaft that aren't cranks.
The pieces of axle between the cranks should be as constant as possible when turned so that they can fit in a small, fixed hole while turning later on.

Step 3: Connecting the Two Axle Parts

So, the two different axle parts must be secured together.
This must not be done if  the axle is made out of one long piece of iron wire instead of two paper-clips.

A third paperclip is straightened out and a u shape is added in the middle. (a little bit larger than the size of those cranks)
A piece of twine is used to bind the two crankshaft parts onto the third paper-clip so that the two align and form one large spiral.
The excess wire is trimmed so that only a small part of the both parts overlap. (look at the pictures)

Then a large blob of solder is applied to permanently fix the two together.

Now I think of it, this could be simply  done just with some of those "handy hands". No Idea how it is called, but it's a thing with two crocodile clamps and a magnifying glass.

Remove all the twine and straighten out the entire axle again.
Spin the complete crankshaft again to make sure that the entire thing is correctly aligned (in the same way as before).

Step 4: Hooks for the Cranks

Later on I will use a thick marker as a container for the axle. So the movement of the axle must be brought outside its container this will be by means of a rod poking trough holes.

If you look at this gif of a crankshaft on wikipedia you see that the rod is attached to the cilinder by a hinge. Because of the smallness this won't be the case here. The rod will go outside the container via a pointlike hole. Therefore it will not produce an up-down movement outside but rather a mirror image of the movement of the crank (circular but the other way around) and is therefore not directly fit. An adaptation of the cilinder to adjust for this will come later on.

Now, for the rod:

Onto each crank, a rod must be attached loosely so that it can move easily. 
Different straight pieces of paper-clips (preferably out of a slightly more bendable paper-clip than the ones I used) are made. 
Onto each a very fine hook is bent so that it fits very precise (yet loose) around a crank.
(note that in the picture the rods have shifted, but each crank has a rod).

In the second picture is schematically shown how a hook can be formed way smaller than your pliers.

Note that this thing is very small, keep all magnets away because if those small steel rods become slightly magnetic they will constantly stick (you can see that a bit in the photos) and it is just annoying to keep them apart to insert them in the holes later on.

Step 5: Placing the Crank in a Marker

Now, a marker, large enough to contain the crankshaft is used to contain all of it.

The marker is cut in half one part will contain the axle and the other will have holes for the rods to go trough.

But first several circles are cut out of a flexible 30cm ruler. (any kind of flexible yet hard plastic will do)
With a sharp object (in my case a fine screw driver) make a hole right in the middle of each circle. Then a cut is made from the edge to the cender of each cicle. These circles are now via those slits shoved onto the crankshaft in between the different pairs of cranks. Make sure that all the rods are positioned good! (each crank one rod)
These circles are in their turn placed into the largest piece of marker and glued in (with a modest amount of glue because they will be removed several times, (at least in my case, read the next step)).

now the piece of axle that is sticking out is bent a little so that it can be turned to make sure the crankshaft fits loosely and can turn.

The top part of the marker is placed next to the part in which the crankshaft is glued and the rods are laid on top of it as orderly as possible.
Now the places are marked where the holes for the rods should come. (look at the photos) And holes are drilled.

All the rods are carefully pushed into their holes so that the top and bottom part of the marker come together.

Finally, (it is crucial that all the rods are nice in the middle of thier cranks!)  bend the rods sticking all in one direction. 
But be careful!
First push the rod in as deep as possible (this will turn the crankshaft) and thén bend, but a little bit away from the marker. (so there is room to attach something to the hook you have made when the rod is fully retracted)

This last step however, I did with a bit too few precision (after all, they are just paper-clips).  It is still good enough to clearly see the sine wave but sadly from here on, i just forget about the 12 wings and instead think of an undulating plane (like squids have).

Step 6: Reducing the Shifting of the Rods

Now the problem is that those rods shift too much. And when they shift away from the crank they can jam.everything.
To prevent shifting, small plastic discs are glued on the ends of the crank. These limit the sidewards movement of the rod.
I don't have a photo of the final result but I have added a diagram (which will be much clearer than the real thing).

A very thin plastic ruler is used (again, anything made of thin plastic can be used).
With a perforator, 24 disks are cut out of the ruler.

In exactly the same way as before a hole and a slit is cut in each disk which is then shoved onto each end of the crank. But now they are secured to the crank with a small drop of glue. 
(It is important that the marker you use as a container is large enough to let the shaft with those disks turn freely.)

The last picture is a diagram of what the total crankshaft should look like (seen from the side).  But since it is very small it looks a lot messier in real life.

Step 7: Finishing the Container

Now, as mentioned before, the rods sticking out still move in a circular way. The movement must be forced along a line in the same way a piston moves in the cilinder.
Here this will be done by moving tubes on spikes sticking out of the bottom. (see further for more info)

The bottom part of the marker is removed (see was adviced to not use too much glue).
Holes are made in the bottom in the same way as the top part of the marker.

Some strong paperclips are bent in a |_| shape and glued into the marker so that the legs stick out. 

The bottom part is attached back to the total. Now a good amount of glue can be used since it shouldn't be separated again.

Two fairly bendable paper clips are bent in a mickey mouse shape so that the total fits securely around the total marker. This will hold the marker together.
The 'ears' will contain the rod to which all the 'wings' will be attached. 
Look at the pictures and it will be pretty clear what is meant 

Step 8: The Pistons and the 'wings'

First of all, the pistons.

As said before, the rods coming out of the marker still move in a circular movement when the crankshaft is turned.
This movement must be converted to an up-down movement which will be done by sliding thin tubes up and down those spikes which have recently been added.
Each tube will in turn transfer it's movement to two wings via two supports.
So each tube must have two handles in which later a connection to  the wings will be added.

The tubes are just pieces of empty ballpoint ink cartridges. 
Paper clips are bent tightly around the tubes. The ends are then curled in to make two handles.
(see the pictures, a larger sketch is added)
The paper clips are finally glued onto the tubes with most of the tube sticking out the other way than the handles.

Then, 24 fairly bendable paperclips are straightened out. (black ones in the picture). These will be the wings.
Now in the previous step a rod was added which will be the hinge for the wings.
So each black paper clip gets a hook that will later be secured to the rod.
Each one also gets a loop which will be used to move the wings. The moving tubes will be connected to that loop. 

Make twelve in exactly the same way, then make the other twelve as the mirror image of the first set. This will be because they will be added on both sides of the marker. And if they aren't mirrored it will look strange. (which it does already, so no reason to make it worse)
It's just bending that loop the other way around.

Now, that loop is important ! The closer to the hook and the more dramatic the movement of the wing will be. Further away and the movement will be less.
It is simply the lever principle.
If the end of your wing is 3 times as far away from the hinge hook than the loop, the movement will be 3 times as big.
This step must be done as precise as possible (I should have done it with more precision) to keep the movement uniform.

Make the loop closer to the hinge than I did! Later on, piece of plastic bag was added and the movement of those wings was subtle enough to be greatly reduced.

Step 9: Add the Wings

it is time for the wings to be assembled.

But first 24 rods with a hook at both ends must be made. These rods will be the connection from the tubes to the wings. They will push the wings up and pull them down. 
The length depends on where the wings are located. It is a bit puzzling but look at how much the rods move when the crankshaft is turned. Move the tubes in an equal amount on their spike. The wings must be horizontal when the tube is at it's center position.

Among the photos is a diagram of how each pair of wings is added.
Notice that there isn't yet a connection to the moving rods. this connection will come later.

It is a lot of fine tinkering to add each pair of wings(and keep them on) but it works. 

A piece of tube is placed between the hooks of all wings. These tube spacers hold the wings in their proper place. 

The position of the wings can be tuned by making the connection rod a bit smaller by bending it or curling the hooks a bit more.
It is just a lot of tinkering.

It is also easier to hold the wings together temporary with tape.

Step 10: Bringing Over the Movement

In order for the tubes to be able to bring over their movement to the wings, they must be connected to the moving rods.
The movement will be brought over by a paper clip bent around the marker.

Normally two hinges are needed for a nice and fluent up-down movement. But the since the rod can tilt a bit on it's spike (the tube is wide compared to the spike), the paper clip that brings over the motion will be completely secure around the tube. 
The freedom of the tube is presumed to be enough to compensate for the lack of a second hinge.

There is bit a problem with vocabulary since all these paper clips that bring over movement are best described as 'rods'.

Again, 12 fairly bendable paperclips are straightened out and bent in to a rectangular U shape. (the shape doesn't matter)
Then at both ends a hook is made, one vertically (to connect to the rods). One horizontally to connect to the tubes.
Make the horizontal loop by bending the wire around a piece of tubing so that it fits precisely around it.

Make six of those with the horizontal hook bent one way and six others with the loop bent the other way. Again because six will be added on one side and six the other side of the marker.

To add that connecting 'rod' , the vertical hook is shoved around the bent part of the 'rod sticking out of the marker'. Then the tube is pulled a bit out of it's spike, and  the horizontal hook is shoved around the tube.
This is done so that the 'connecting rod' sits at the other side of the wing than the rod that connects the wing with the tube.
This is all a bit complicated in words but it can be seen in the pictures.

Finally, add a small ball of solder to the very end of the bent part of the rod connected to the crankshaft.  This will prevent the rods connected to the tubes from shoving off.
But beware not to solder the the connection between the two!  The solder ball (or whatever is used) should only prevent the rod from sliding off.

Step 11: The Engine

First of all I must say that  in order to stay in the 'office material' theme, I decided to add a rubber band engine. But probably simply a crank would have been a better Idea. Because the engine works, but it is a bit underperforming.
It is probably better to just install a crank.

So, the engine must be very small (it must fit in a marker cap) and produce quite a lot of torque to turn this thing.

The simplest thing I could think of is simply twist a strand of rubber bands so that was what I o-did.

A piece of strong paper clip is bent into a D shape.
Some rubber bands are cut in to manageable pieces with a length just smaller than the length of the cap of (an other) marker.
The D is placed with the straight side among the ends of the rubber bands and glue is appied. Everything is very securely bound together with twine. More glue is added.

The oher side is just the same but with a straight piece of paper clip. This size will be used to wind the engine up and the straight piece can then be secured in some notches.

The end of the crankshaft sticking out of the marker is secured to the D attached to the rubber band.

A slit is given in the marker cap and the rubber band is shoved in. (the cap is resized to make the rubber band fit as precise as possible. 
Everything is glued together.

Also, a piece of twine is used to hold the wings parallel(look at the pictures.

Step 12: Something That Didn't Turn Out Very Well.

The last thing added to the "undulator" is a piece of plastic to mimic some kind of membrane like squids have.

A plastic bag is cut into two pieces of each two times the size(in length) of the area that contains the wings.
(The wings are just covered in plastic).

However, the glue must have stiffened up everything a bit so the movement which was already small is a bit reduced. 
So, if you do this, make sure to make the loops in the wings close to the hinge so the movement is very big. And use some kind of flexible glue.

Step 13: Result

So, the end result is just a little machine that you can turn and it moves in a wavy way.
Here you have more pictures of the final result

Attempt to capture the movement: 
First of all winding the rubber band engine up and then release it just turns the crankshaft very fast. So fast that it is finished in a fraction of a second. (Which is actually a good thing but not so good to make a video from)
Secondly, just turning the engine just winds the rubber band till the torque becomes larger than the static friction, then the rubber band unwinds. So movement is only possible in small steps, it doesn't look good on camera either.
Therefore I ditched the rubber band engine and installed a crank.

To capture the movement adequately I had to position the camera and hold the machine in quite an unconfortable way so the video is shakey and not much is visible.
I added it as a gif and it's the last image (because it is too short and I've never made a gif before and wanted to try that).

Sorry for the bad quality, it was just to show the wavey motion. The pictures are good, look at them an apply some imagination to see the movement in high quality.

I have turned it so much that by now the glue is becoming loose and it now starts to jam.

Office Supplies Contest

First Prize in the
Office Supplies Contest

Be the First to Share


    • Game Design: Student Design Challenge

      Game Design: Student Design Challenge
    • Big and Small Contest

      Big and Small Contest
    • Make It Bridge

      Make It Bridge



    9 years ago on Introduction

    Very cool! Makes me ponder this as an Radio Controlled Blimp motion system. Two motors, one for each side of the blimp, running a set of undulating 'wing'... would make for a very cool indoor RC creature!


    9 years ago on Introduction

    Awesome !…
    Still your too modest.
    You're not "just a student in physics" but an inventor of life movements … a way of saying you're next to God ! ;D

    Lot of fun and thanks for posting !…


    10 years ago on Introduction

    Wow, I'm very excited to see this project! I wanted to build a wind-powered kinetic machine and was trying to figure out the mechanism to move a puppet Manta ray- you have it here. If you have any additional thoughts or adjustments to your design I'd love to hear them, I'm hoping to start building soon. (For the crankshaft I was thinking of using offset discs along the shaft to keep the shaft straighter?). Anyway, thank you, this is fantastic!!


    Reply 10 years ago on Introduction

    That's a great Idea!
    All my thoughts are in the instructable. I've explained some of the problems I had, so they can be avoided. (e.g make sure you connect the moving rod to a good spot on the wings (because of leverage => good movement). And if you do multiple wings and you want to connect them to make one large undulating wing, watch out that it isn't too stiff as was the case with me.

    I had multiple offset discs too, but they didn't all fit. The shaft was small and rigid enough to function without them.

    Good luck!
    Let me know how it turns out!


    Reply 10 years ago on Introduction

    Yes, indeed !
    I've looked to Theo Jansen's strandbeesten to see how it works, but the same mechanism would not have worked here.

    Awesome that you've remembered this. (Cool to know that my instructable stays active in the subconsciousness of other people)


    10 years ago on Introduction

    It might not be ready to swim yet, but I think this looks really cool. This concept would make a pretty sick kinetic sculpture.


    10 years ago on Introduction

    What a good work you have done!

    Maybe the plastic wings added in step 12 are too tight, it would be the reason to its stiffness. Try leaving a wave between a rod and the following. That will allow better movement.


    Reply 10 years ago on Introduction

    I have just removed them and now there is only a very thin glue membrane left. It moves more fluent.
    Later today I will add an other animation of the wing movement.


    Reply 10 years ago on Introduction

    Yes, I've seen that, but that mechanism wouldn't work for two bilateral wings. (with one crankshaft)
    One wing would go up while the other goes down.
    It is more similar to 12 ornithopter mechanisms behind each other.


    10 years ago on Introduction

    I can't tell from the animation - do the wings themselves move very much?


    Reply 10 years ago on Introduction

    Yes, but not good, the glue on the plastic stiffened it up too much I think.
    One moves barely, and the other turns from a wave shape to flat and back.

    Before the plastic was glued on they moved nicely (for paper clips).

    But I will add some more animations tomorrow( hopefully with more light and sharper)