Prototyping The Marble Machine Programming Wheel 

Martin, in trying to find a reason to quit the marble machine, keeps accidentally discovering the perfect solution to his problems.

As a musician myself: It's always a bit painful to hear him say the first Marble Machines "failed". They did not. Especially the first one. It did everything it needed to. It played music, it felt authentic because it wasn't perfect. It had a lot of charm, it had a lot of passion and you could feel it. It wasn't perfect by any means and it never had to be. Just one instrument has caught so much attention, so many fans, and to this day, it is one of the most watched and most genuinely liked videos on the internet. Basically the whole internet knows the Marble Machine song. It moved people, made them think, feel. THAT is music. That's what it is all about. Then came MMX, which was started with a vision, and sparked a lot of hope in the fans for more marble machine music. And that hope is still there. In my eyes, the marble machines did not fail at all.

The change from "i believe" to "i verify" is symbolic of martin learning to be a true engineer. Mm3 will be successful

I think you're going to find that your timing mechanism will not hold up under load. You really need to make an oscillating escapement to keep tight time or you will always be fighting against the influence of the mechanism drag effecting your timing. If you want your machine to run like a clock you need to make a clock.

At this point, Martin is going to just end up building a clock with some instruments lmao

I like your idea for the new wheel but I have one suggestion, either use a regular endmill or a slitting saw to cut a small section from the bottom of the circular groove. The ball endmill will have trouble cutting at the bottom of the ball because the radius of the cutter goes to zero with respect to the rotating axis, meaning that point on the ball will essentially drag across the material and cut poorly leaving a bad surface finish.

You will definitely want to test the machining of that long slot it may be a lot more challenging than you think. Chip evacuation will be very important if you intend on using a spherical tip mill. The type of cutter that you showed is not designed to cut plastic at all and will likely clog and break either the cutter or the work piece. The cutter shown is called a rotary bur and is used to contour metal by hand and is notorious for clogging. You would be far better off cutting a groove with a ball nose end mill and then gluing the bar in place. Pressing in a very long and thin metal rod into a long hole will be VERY challenging to get right with a press fit. Most press fits do not go longer than about 10:1 aspect ratio of length to diameter, or you will have problems of bending the "pin" due to every increasing friction of the press fit. Too loose, and it will slide in easy, but you will have to secure it with glue anyway. That would be preferable to the press fit in my book.

MARTIN, don't forget, that varying marble troughput trough instruments will put varying load on marble lifting mechanism, which will deviate marble machine speed since you will be inputing constant force via huygen drive. Less marble troughput, smaller load on the machine, machine speeds up. Loads of marbles to lift, heavy load on the machine, machine slows down

The amount of force needed to push the rods in seems like the hardest (ha) part of this new design.

The design of every piece of this has been stunning, and seeing it assembled is probably going to be one of the most wonderful things I witness.

Keep in mind at 8:46 that the cutter moves along the axis of the cnc but the cutter itself will be rotating and cutting on one edge, causing force to be put in the direction of your rotation axis. Something that will help with that is always starting the cut from the same side. This way the rotation axis isn't pushed one direction during a cut and the other direction on the back cut. If there is an offset due to the cutting edge it will be the same for all slots this way.

Every time I started to worry he was returning to his old mistakes, he throws in a "but this would be a bad idea for of exactly what you were thinking"

Take a drink every time Martin says "tight".

I like this idea! Using a single, known-consistent metal bar for your precision is clever. As long as you can get the rotary CNC to cut channels which are precise distances from one another, sounds awesome. My only hesitancy would be destroying a large/expensive wheel if you get your machining operation wrong, but small scale tests can hopefully help avoid that. :)

CRAZY IDEA: What if instead of having a programing wheel that activates the gates mechanically, with the pins or prophiles, you had one that activated them electrically? One option is that you could print a board with electrodes and a small battery on the wheel, that would make contact with a loose wire instead of a lever. Another option is to use optical sensors with infrared LEDs in the place of the lever, lke those robots that follow a black line on the floor, having the wheel be printed with the notes Both things would be incredibly precise, but maybe the machine would lose part of it's magic, since the mechanical activation is so pleasing to see

Just a technical comment here. You talked about the cutting forces only being in the "rigid direction" on the programming wheel for version 3. There is actually also a force in the "not rigid" direction. As a plunging cutting tool engaged with material on 3 sides. On 2 of the sides the forces cancel out, but the last side creates a sideways force that actually tries to "turn" the programming wheel. Just something to keep in mind, not critical :)

I really hope the marble machine still has some hand crank/foot crank for you to choose the speed. I know its less precise but the appeal of the marble machine for me has always been that it felt like an instrument not an automaton

There is something special about watching someone learn true design engineering and implement it. These videos also demonstrate design thinking very well and is a great tool to show the difference between an idea and a design.

Now this is promising! I hope the tests with the prototypes are successful and you learn a lot from them!

As my previous worries weren't dissolved, I'll just re-post the comment I wrote in a previous video. The programming wheel don't seems rigid enough to avoid deformation between both ends, the lack of diagonals or structural parts may cause a problem, moreover the driving force is applied to one end and there is resistance all the way whit the gates. One other overlooked requirement is the need to play notes in fast succession, cause you can place a pin every 2 holes at minimum, holes that are worth 1 quarter note each in your design. Some will say that the profile could have multiple edges to play multiple notes but you'd have to redesign the gate, as it is right now, it's way to flat and will be sitting on top of the multiple edges, thus merging everything into a single long note.

I mean, I miss the music, but EVERY engineering student should be following this. Just fantastic information, walking through the problems and the process to address them, etc.

I think an important idea to ponder is why a programming wheel. If you want to design form from function maybe consider a different approach than just assuming the programming wheel is the way to go.

Martin, you can change bpm on the go precisely by adding a attack angle rod to the governor. With this rod, you can manipulate drag coefficient of the governor propeller like in helicopters thus changing the rpm that the flywheel rotates at. This method does not effect precision of the system and might be a good addition possibly. With this flexibility on the BPM, you can use the programming wheel much more efficient for varying tempo of different music tracks.

This programming wheel is brilliant. At last, it seems, a wheel that is machineable and foolproof!

imagine martin demanding this kind of precision from his human band-mates.

You've been doing a great job, love how much the marble machine has been improving over the years

9:28 Wait, that seems like a lot of metal-to-metal surface area that needs to slide across each other while being tight enough to "press-fit". After you get a few centimeters in, are you going to attempt hammering at the end of a meter-long, thin rod? I highly recommend prototyping this before you construct a whole wheel with it. Maybe consider making the rods thinner, and adding rings on both sides to hold them in?

Have taken account of thermal expansion of the different materials, brass bar and 3d printed plastic under the lights on stage I would guess it will be warmer than your workshop

Love it! That is such a nice and forgiving design. It WILL play tight music, for sure.

Your videos make me so happy. It is going to be bitter sweet once you nail this design. I have LOVED watching this evolve from the first machine. It has been a pleasure to learn with you, you are a wonderful educator, and an equally wonderful person.

Design number two makes me think that you could look to bike wheel parts and wheel building techniques to save design effort.

I do not know why, but your videos seem to be so cozy and atmospheric. Keep it up, Martin!

I love this channel. I've learned so much about engineering, prototyping, materials, etc., all while being entertained. Thank you!

Awesome video, once again. When this is all over it would be fun to have a video of everything you've done so far. Like a timelapse or something. Or maybe just a 40 hour video consisting of all your Marblemachine video's back to back. I often just have you talking about your machine in the background while i'm working on my on shenanigans

Dang in at only 347 views! I think this is the earliest I've ever noticed a Wintergatan video! Hyped for the programming wheel and hyped for the world tour!

Fiiine! Your perseverance, passion, and love for efficiency made me resubscribe.

Always incredible to watch your videos, Martin! Much love! Stay awesome!

I really like how you're not just using 3D printing to create parts for the machine but also tools to improve the creation process itself, that's really cool!

Martin, can I challenge your assumption that the programming wheel needs to be a wheel? What if it was a chain drive? You could make your music loop of arbitrary length by using larger chains (and adjusting your idler gears).

I always love seeing you problem solve and innovate.

Finding all these solutions while trying to quit is hillarious, when we're on our lowest the mind becomes free of its shackles after all. I'm happy you are getting better and better at the marble machine, no matter the outcome I'm just glad i get to experience this journey in real time. Best of luck!

I am so proud of your new approach and the storytelling is amazing!

Awesome to see new ideas. I still listen to the songs

4 years from now. "Prototyping The Marble Machine Programming Wheel"

Please note that a full plastic wheel will be prone to a lot of thermal expansion on the world tour. This could result in centimeters of variation in programming wheel diameter. Also, plastic is prone to creeping over time. I would recommend a steel basis for the wheel with small, replaceable, plastic parts.

I love the storytelling and video editing in this episode ❤

Wow, what a great project! I can't wait to see it finished. My sincere compliments.😃

So intense! Can't wait for the next video! It almost feel like watching a series and being at the end of a season, waiting for the conclusion to know if there will be a season 3 or not

Fantastic video and I love how we are seeing the evolution of the designs. Those programming wheels will not be easy to move around but I guess that goes for the whole machine too.

This whole astounding ride from 1 to 3 continues to amaze me!!!

You might want to consider adding a thin “backstop” in the programming wheel for when you’re inserting the rods. You could either stop the circular bit just before exiting the far side, or just attach a ring of sheet metal to the outside. It’d at least help with assembly, but if you did it on both sides it could also help with preventing the rods from drifting out over time. Potentially not a problem, but it should be an easy enough fix that it might be worth doing preventatively.

What I got out of this video is how to 3d print a very cool wheel. Got some GREAT ideas from your version 2 wheel. Love your videos!

The video are always so much incredible, SO happy when i see a new one

Buckling up! This will be awesome!

Wonderful, Amazing, Incredible . . . . . . really good stuff !

Man this is one of the greatest youtube cliffhanger ever, will Martin end up continuing marble machine 3 or will he drop the project? can't wait for the next video

Im no engenier, but im a mechanic and i can tell you, depending on the material you use, V3 will be a pain in the ass. 1) the entire ball part of the mill bit is in use. That will most likely mess up the bit pretty fast and you will burn throu them like cracy. You have to use a pretty soft material for the wheel or take it incredebly slow, orherwise the wobble you create by full contact milling will mess up your meassurments. 2) you have a lot of different cuts. If you use a soft material like PE, as soon as your milling bits gets dull your edges start to look like chewed. Its doable, no doubt but if you do, do yourself a favor and outsorce it to a professional. You get your money double back in safed lifetime.

i get so happy when i see an upload from you Martin! I hope it all works so we can finally see your vision come to life! :D

I so love how this project is turning out! From "I believe" to "I verify" - wonderful!

I'm here for that good designing and crafting, absolutely lovely! XD

I really love how 3d printing has allowed the creative process to flourish! Keep up the great work! I'm excited for what the future holds, and am enjoying seeing the progress being made.

I hope it plays the right music because we all need to get the marble machine on stage with you guys when you go on your. Plus I imagine it opens up what music you can make by having the marble machine be a multi instrument leaving band meme era to play other instruments for the songs that otherwise wouldn’t be possible with the number of people on stage.

Great lesson about prototyping!

Martin let be short and simple on Point!!! You're nailed it once again❤ Love it🫠

As soon as I saw design requirements, you got my full attention ❤

these videos always make me grin! martin! you will succeed!

Look into compressionless brake housing for transferring the movement at the programming wheel to the marble gate. Bikes with mechanical brake systems use brake housing to route the movement of the brake lever to the brake calipers. Compressionless housing is specifically made to be stiffer and less squishy, possibly letting it play tighter.

What a cliffhanger!

great video! what a journey. I have absolutely no engineering background, but i'm still here since Day1. Having some cliffhanger vibes rn. martin is not only getting better at designing this monstrousity, even his storytelling skills are blooming! fingers crossed ❤🎉❤

Good vibes. You got this Martin 🎉

Martin, I'm so happy that you seem to be making fewer crazy decisions every video and slowly becoming the engineer that I know you can be deep down!

Trying to stay with Martin is like a roller coaster ride. Computer technology achieves most all off the requirmenrs except on stage presence of an instrument. The truly retro nature is "pain and pleasure. So is the design journey from performance artist to mechanical engineer. I rate Ivan Miranda cause of his very strong mech engineer skills.

The suspense for next episode is insane!!

The dedication of this guy, it's rare to find professional engineers as thorough!

really cool to see the exact methods i'm learning in my mechanical engineering classes being used in this video

THIS HAS TO PLAY TIGHT MUSIC. YOU CAN NOT GIVE UP ON THE MARBLE MACHINE. You've come so far.

I imagine that press fitting those LONG metal bars into LONG slots will be quite difficult. I think the amount of friction building up will quickly overcome your ability to push them into the slots. You might consider splitting the programming wheel into several narrower drums with shorter slots for ease of assembly. You could even align the narrow drums using the metal bars provided you leave some overlap.

Martin. You have brought me back. I'm so happy watching this video.

Chasing "tight" specs is nice and all, but remember that the original captured hearts and was barely holding together. If it's not perfect, that doesn't mean it's not perfect. The *goal* here is to make music, to spread joy, to create wonder. You don't always need millisecond accuracy to do that.

8:48 it WILL add pressure in the rotary direction since the tool is rotating it pushes to the side when doing a slot cut

I'm not sure about "design three", I suspect you might have some addition DFM (design for manufacturability) work to do there. Getting a press of a 3' long metal bar is going to be tough -- meaning even if you find the perfect ball mill that exactly matches the right press fit tolerances of your metal rod, (a) how will that change as the ball mill wears down, and (b) how much physical force will it take to push that rod three feet through tight tolerances? I think you'd be better off tightly constraining the ends of the rod, or maybe a small area on each side of each slot, and then leaving more of the center free in order to avoid an impossible push fit. Similarly, I'm worried that the outer programming skin is responsible for so much of the structural stiffness of the assembly. That outer skin seems to be trying to do too much.

"Will this new wheel be hard to manufacture?" "It will be easy, barely an inconvenience and it will play very tight." "Ohh tight marble music is TIGHT."

I do like the metal bar idea

Building that printed wheel from segments looks like torture.

you did it good this time, it is gonna work!!!

The perfect is the enemy of the good. Love to see the learning experiences and progress, but I would also have loved to see you going on-tour 2 years ago.😘 Go whatever direction the universe demands😙

Martin is literally reinventing the wheel!

awesome work!

Have you considered the tried and true method of programming mechanical instruments... the paper piano roll? I think you're underestimating how much having to manually reprogram in each song is going to affect things - not just for concerts, but while you're developing the machine and practicing, the ability to swap out songs quickly could be night and day.

I liked the loose music. It made it more human than industrial.

If you replace the programming wheel with an electronic controller and replace the marbles with speakers, the whole system would be simplified and the timing/sound would be near perfect.

We want to see you on that wold tour!

Martin, I feel like you've entirely lost sight of the reason the world fell in love with the Marble Machine. It wasn't because it was perfect, played tight music, and used the best engineering to play a song. Those are problems that have been solved for decades. We loved it because there was a crazy, steampunk-esque guy cranking a giant contraption made out of plywood that played one of the catchiest, most beautiful songs ever created. The entire project was an artistic masterpiece, and the world saw what you had created and held it in the highest regard. While I understand your desire for what you see as perfection, please try to remember, at least sometimes, that what made the original so beloved was the fact that it was EXACTLY the opposite of what you're striving for now. I only hope that when this current journey comes to an end, however that may be, we are lucky enough that you'll consider giving the world more of the wonderous art you were capable of so long ago.

dude,... you're taking things to a whole other level.... you haven't ceased to amase me yet!... :o

You could easily get *TIGHT MUSIC* from imprecise _slow moving_ programming wheel, if the programming wheel only register *WHICH* note should be played next, but actual actuation force comes from _fast moving_ flywheel telling *WHEN* the note shall be played.

One of the only projects ive watched from beginning to end, ive been watching since the beginning and loved seeing where you've been and am excited to see where you're going. The first marble machine has a very special place in my heart as it was the first wooden mechanical structure i had ever seen and to make it even better, it played a beautiful song that caught my ear immediately!!!! Super excited to follow along with this project i have faith it will all play tight music!!

as always : wonderful

been watching since you put together that cnc in that shipping container :D I have no doubt you'll succeed Martin

Maybe you need to consider the size of the final wheel. The one from the previous machines seem to be really big compared to your prototype. As the design relies on machining to work properly, If the wheel gets too large it may be really difficult to find a machine that can mount it properly (or too expensive). Also consider that milling the slots will introduce an axial force that will load the spokes in bending if the wheel is clamped at the hubs. This type of loading doesn't happen during normal operation, so it might be overlooked during the design phase (ask me how I know :D). Keep the machining setup in mind and include some features that make mounting and indicating on the mill easier. The metal rods are a good idea, but I'm not sure if you could press fit a rod the entire width of the wheel. Maybe ask a machinist about this.

Finally caught up with the videos. The suspense is killing me!!