Magnetic Marble Picker 

Martin, I am studying in aircraft design and one of the thing we learned was that designing something has an iterative process to it. The reason behind that is that you can always refine the design, make it 10%, 5%, 1% better by marginal gains. Although i thought the design of planes was perfect all the way through, my studies showed me that the planes flying nowadays are far from being designed to the perfection. The chief engineers sometimes have to tell their engineers "okay your design might be able to be 2% more efficient if you work for 10 more months but we need to deliver the design next week". I feel like I am seeing a version of that in the MMX project of you trying to make every component of it 100% perfect. It is definitely not a bad thing as you are converging towards the optimal solution but maybe it is sane sometimes to not scrap parts which are 98% perfect to make a 100% version of it. In aviation the rules are written in blood and all the experience gathered by engineers before is put towards the next generation aircraft. For the MMX you are trying to anticipate all the risks and issues but maybe you should try seeing them happen? Finally, i wanted to say that you are assessing the reliability of the MMX only in the funtion "play the marbles" but maybe you should also think about assessing its reliability in the other type of functions such as "being able to compose", "changing the song" and so on. It would be a shame that during the world tour you can reliably play 1000000 marbles but have something that brakes or malfunctions once every 10 song changing. Anyway maybe that could help you! Your work is really inspirational and good luck to you and Wilson ! :)

Martin whacking the moles in the beginning sounded like the mmx snare.

Martin's doctor: "Have you been getting enough exercise?" Martin: "Well, I smack a table with a mallet. A lot."

Martin: I'm not against prototyping a redesign for the divide by 4, but I DO request that you wait until you have most of the channels operational before you actually consider doing it. Because the more channels you have running, the less activity the divide by 4 sees, and I suspect that a fully operational MMX will use that function very little. So the problems it's causing now might be an overrepresentation of how "fragile" it is

Divide by 4 did work really well on previous tests even when you played the machine for hours. It'd be sad to see it go.

Another quick trick for using setscrews without causing damage on a soft part: Make your hole a little larger than the PMMA pipes, and insert a metal "shim" between the plastic and the set screw. The shim can be made from rolled strip metal, or by slicing and slotting a metal tube. If the shim is thick enough, it will clamp down evenly in all directions when pushed by the set screw.

NGL, I love the divide by 4, but your arguments makes sens. I would let it on though. Test it more thoroughly as you continue working on the machine, because, as for now, it worked great. I think that you have much bigger ploblem on the machine to overcome. Focusing on this little easy (maybe) problem may be satisfying now, but could be a mistake in the long run because the divide by 4 is cool. Stick a note on it saying "Big Martin is watching you" and when time comes, decide of its fate. Mabe he will be scared and work like a charm.

Who is this Wintergatan everyone's always talking about? I only know Wintergatan 2...

He should do a live stream of one of his marble tests.

Martin:put a microphone to take the slightest sound off the taping to make it asmr for headphone user. also martin : smash a hammer all along the video.

Be aware that surfaces will wear over time, changing the characteristics of things like gears and balances and counterweights (especially with wood and plastic parts). This is why machines are designed to be a bit "stiff" out of the factory and need to be "broken in" before everything will move smoothly.

I wise man once said “pain is temporary but glory is forever“

Martin : Any setting that I have to tweek to an exact value for it to function over a hundred thousand marbles is a velnurability. Also Martin : spends days tweeking marble path length to 0,2 mm, yet has lots of interfering factors. Keep it up dear genius. You'll get there!

The stress of school was getting to me, then I saw this notif. My day has now been made

It occurs to me that you may want to build a “fast demagnetizer” for your concert tours, so you can demagnetize all the marbles before the concert each time.

You COULD just attach a giant bucket filled with thousands of marbles...and use those to play the songs. But that's not what you want. The MMX is not just a music machine. It is a beautiful and interesting music instrument. Please keep it that way. You have solved so many 'moles'. The fish stair and divide by 4 will be fine. Dont take them off yet. :)

Nothing better than watching a good wintergatan video in the evening

MARTIN the springsteel at the fishstair started catching because you angled the tracks down! What used to be flat is now pointed down so the springsteel at the top protrudes more!

I’d like to take the time to not only thank you for making the MMX, but taking us with you on the journey. It can’t be easy to show us the failures as much as the triumphs. But, I’d also like to thank you for the wisdom you’ve shared with us along the way, and about how you overcome different aspects in your life. I’ve been a procrastinator all my life, but through this channel and your words of wisdom, I’ve begun to change my outlook on how I do things, and that is 100% your doing. Seeing how you tackle a problem and work through it, helped me to realize that I can too, I just have to put in the work, so thank you more than you can imagine for helping me discover that I can change, I just have to want too. Thank you Martin.

It’s so beautiful. I don’t get how you ever can feel discouraged when just looking at what you have already achieved never stops being so awe-inspiring.

It is Magnetic Marble Monday my dudes...

6:42 Girlfriend: "hey, do you have a condom with you?" Martin: "no, but I have a stack of washers"

This has been such a journey. I really feel like we’ve been there with Martin, watched him grow as a person, as an engineer and a designer

BTW, any huge change like that should really go to the main channel.

As long as the fish stair moves in time with the music, the marbles will look like they move in time with the music. I hope you'll at least keep that element! Making sure every step is neatly filled with marbles isn't that important, but the in-sync idea had been there since the very very beginning. It's almost like I too have an emotional attachment to it hahaha

with the 1 to 4 divider relying on swinging freely until full, any other vibrations through the machine could effect it! if its going to with with 3 marbles waiting for the 4th to trigger it? maybe this could be fixed by having the 5th marble in the cue activate the divider? instead of a counterweight you want a trigger!

Yay another awesome episode. Imagine one day Martin calling it DONE.

Way to go Martin. Remember you can always come back to it later. Every engineer has parts on every project that they don't like but sometimes you need to move on and save the "ideal" solution for the next version.

"I'd rather have a musical instrument than a Rube Goldberg machine." Well said Martin!

The mmx is fastly becoming an ongoing prototype

Martin, I am so delighted to see the evolution of you as a troubleshooter. Developing that part of the brain and not getting frustrated is the hard part!!! I cannot wait for the finish line!!!!

Your machine is beautiful, no one except perhaps you is going to miss dem marble pipes 😂! Remember, this machine could end up playing zero songs at this point and this project, at least in my mind would still be a raging success!!! You have built a wonderful machine regardless of its functionality, that for me brings out this feeling of pure joy and curiosity. You remind me to keep pushing myself to achieve my goals and to keep persuing the things that I find interesting regardless of what others may think. I hope you realize how amazing this project is and what it has already accomplished, though there will be much more amazing things to come as well!!! Thanks for everything, can't wait to hear and see more!

2:31 engeneering ASMR...only if you like it. Okay, it's Wintergatan. this is something else :)

if you have a doubt on speed, you should stop guessing solutions, you should try it. Maybe totally other things break. Why not do a 200bpm test instead? You know you're going to have to...

Martin's gonna cure cancer with a stack a washers.

I'm REALLY enjoying watching the troubleshooting, because your look on things is quite abstract, and the solutions you come up with work, and when they don't, you don't hide it. You are open and willing to share flaws with your project, and that's hard to find. Keep up the great work.

Don't remove the current mechanism! It's one of the most wonderful parts of the machine and hasn't really caused problems as far as I can tell. The spring steel was only catching at the end because you adjusted the angle of the tracks downwards without readjusting the spring steel

Martin's Marble Machine Madness! You are spoiling us Martin...and we LIKE IT!

I think the best mentality to help you sacrifice aesthetics for functionality is that the most aesthetic aspect of the MMX is its ability to successfully perform its function!

Your ending assessment was so on point. I completely agree, get rid of anything that looks pretty but is vulnerable. I really liked the divide by 4 but I like your music production even more.

It's almost terrifying seeing those marbles jumping with each other there above the channel divider, under the Zero Marble Fails perspective

imma comment something just because I love the content and want too boost the interactions in the algorithm lol, this channel definitely has the same quality as the main one and it's ridiculous!

Martin is providing learning material for generations of design students to come ^^

Don't worry about the aesthetics of the parts--the way this project delves down into the gritty details of the design and engineering process is cool enough as is, and, by extension, any part you make will still be just as unique. Also, don't forget that you're making a *music playing marble machine!* No single part can get in the way of that, as long as it works.

Martin please don't go crazy trying to get it perfect. The hard truth is it'll never be perfect, since you'll need to disassemble and reassemble it many times and many external variables could affect it. I'd suggest you try and get the other channels done and whack any apparent moles that affect it greatly. From then on try and "debug" it and squash any other problems that pop up. But I feel like it's totally fine for it to drop some marbles here and there. I do believe in this project and I do believe in you being able to deliver it, but I'd rather see it playing in a year rather than not drop any marbles in who knows how many years.

I know I'm not alone here but might still possibly be in the minority when I say that I really don't think you should get rid of that part. Its okay to focus on function over form, and maybe its just sunken cost fallacy speaking, but it seems to me that you're just sacrificing years of effort to make the MMX become the MMY. Calm down, you just had good progress on the divide by 4. The moment it started to fail again was exactly when you introduced a previous problem back in it with a temporary piece that didn't belong there. The divide by 4 and fish stairs work well enough as they are.

You're breaking my heart Martin :( I think it's all those cool, functional, but interesting and beautiful mechanisms like the divide by 4 that make the marble machine what it is. From its initial design, it's not made as efficiently and functionally as it could be. It's a piece of art as much as it is a machine. With this and the recent funnel designs, I fear you are going to strip away it's character. Who am I to tell you how to build your masterpiece though. You're the artist, you do you.

I can't help but notice how good these table whacks sound. They're like a perfect snare drum hit every time.

Would be sad to see the divide by 4 taken off. Hoping there's a better compromise there

Yippee-yi-ay, yippee-yi-o... Ghost marbles on the floor!

Your engineering is continuing to improve. Taking a problem one at a time is the only way to really know what is going on. Well done Martian !

Look man, you've gone this far, you have 8,500 patron supporters, and so many viewers watching, waiting for the end. Giving up is not an option, we all are supporting you and not a single person agrees that you should give up (that I've seen). Please, just keep going, I'm a patron supporter, and I support you and your work because it's a thing that should be in this world. It's unique, and nobody has the time and the mindset to make something this complicated in today's society, but you. You have the brains, the financial support, the emotional support of thousands of people, and a world tour to look forward to. Don't give this up. ♥

The moles have gotten to the point that he is forced to provoke the machine before they emerge😱 Ever onward on the march of mines!!

Oohh, that cliffhanger ending!

I love how the whole think works and looks, but I understand what you mean when you say it’s fragile, I’ll leave it up to you Martin :) your decisions have always been enough for me!

I'm so glad you ended the video the way you did. I spent the whole time watching thinking "there are so many ways that could go wrong". Hoping you end up happy with both the aesthetics and functionality of lifter version 2!

No !! Please it will work do a fast test before changing the divide by 4 please !

All I can hear is "engineering guacamole".

"I'm going to add a washer stack" had me laughing, washers have to be one of Martin's favourite fixes 😂

The idea of this synchronized pickup system is really neat, but I think you don't have to worry about the looks of that part to much. I mean, what makes a mechanism look awesome and satisfying? The fact that you can see it working reliably and doing what it's supposed to do. So if you come up with a simpler way that works even better, I'm sure it will still be great fun to watch it work :)

No-oooo! Leave it as it is. You could spend forever going back and redesigning everything to make it marginally better. If it works then leave it be. Sometimes good enough has to be good enough.

Martin - You should set up a ton of cameras to record every working element in the MMX. This will help you see any flaws as you run your tests It will also make for some killer B-Roll

Balancing the four lane marble pickup was truly inspiring.

Wow that intro. I could watch it work all day. I grew up going to the Boston museum of science and their art installation of a giant ball machine. Golf balls, bowling, bearing balls. Your machine without the music is fantastic. Thank you for this work of art.

5:09 this is when wintergatan knew , he discoverd a new MOLE

I may be alone on this one but if you want a perfect, tight note playing machine you could just open a DAW on your computer and program in the exact notes. This machine you are building is not perfect, but it has character. It being a bit whacky and having some "silly" marble paths is the reason so many people decided to follow this journey. Those individual parts are the reason why this musical instrument is so interesting in the first place.

This continues to be my favorite journey. Thank you, as always, for continuing to work on and share your progress!

With the bass drum and snare in the background, the opening slow-mo looks and sounds like a victorian-era steam contraption.

Maybe that sounds easier than it is, but "why don't you just" make the pipes coming from the divider a little steeper? Raising the divider by a few millimeters could suffice to load the stairs with slightly more force, which it looked like (to me) could fix your problems

Piano tuners - "This will take a few hours, I need some spanners" MMX tuner - "Give me six months and an angle grinder"

I say this every time, my favorite Wintergatan videos are the ones where Martin fixes like 5-20 things in a single week, and... Oh he's completely eliminating the divide by 4... okay.

Martin, this channel is definitly my favourite ! I'm glad you made this change !

I feel like the removal of the divide by four is a bit premature. It seems like it was working fine. The focus should really be on critical marble fails and not noncritical ones.

Yes, redesign of this divider seems needed. This look vulnerable, as whenever one marble gets stuck at the stairs it will eventually lead to the divide-by-four to stop working on all channels at once, instead of just that one blocked channel. Or did I miss something? Maybe the divider works different than I can see

This might not be what you want to hear.. but I could watch you fix this machine for years! (In fact I have) but don't lose heart. You're videos are so good and so entertaining, the journey is a joy. At least for us watching!

11:30 It's so nice to see the marbles flush are the marble devider entrance knowing the efforts it took ! Love the daily uploads, we belive !

The current system is quite close to working, is beautiful and may take less time to get working than starting anew. Using a stiff rubber, silicone or felt instead of sheet metal strips to guide the marbles into the stair would do a lot for reducing noise. Adding curvature to that strip, conforming to the shape of the marble instead of a flat strip may also help in preventing the marbles from moving to the side as they enter they fish stair. Also instead of the tape on the bottom fish stair attaching a larger guide to the pipes covering both the side of the pipes and the bottom fish stair might work better. I'm curious, how many marbles are in the system in total? Is there a way to calculate the likely hood that the staircase occasionally going up with only 2 or 3 marbles would cause a channel to be empty and unable to play a note? Would adding more marbles to the system make it more resilient? I've been following this epic since the very beginning and am starting to feel the melancholy one feels towards the end of a great book. The machine really feels like it is getting close. While I know that you need to finish this machine for the sake of your sanity, and I'm looking forward to seeing you in Ottawa Canada on the world tour, I'm gonna miss this mix of engineering and art that speaks right to my soul. Thank you to you and Hannes these daily videos. Best thing to happen this decade!

IDEA: have different modules you can swap in when you need. A simple marble tray for the stair lift for consistency and performance and the divide by the 4 for music videos/final stage song. So why not have spare modules that function the same but have different pros and cons. Like even a drum or visual fingers can be swapped in when needed. The MMX is Alive and will never be finished, only awaiting its next upgrade. MM(infinity sign) #savethedivideby4 #savetheART

MMX running like clockwork! Your dedication is inspiring

Martin, I totally agree with the change on the fishstair entrance. The music has to be the priority. There is enough other visual candy going on that the marble divider is not worth the headache.

In this rate, Martin will be doing wild circus tricks just to whack some moles

Hi Martin, honestly, running 100,000 marbles is a fruitless effort. Remember, a machine's "life" is built to run for only a finite number of cycles. I'm not sure what you expect to gain, other than you may have to start replacing parts that have already met their quota of cycles. From an Agile POV - it's time to start giving the marble machine it's voice. Fixing and enhancement will come. Plus, you need to test how the machine performs when actually making music.

Oh Martin you've made me so happy that those are even now. Thank you so much. My OCD needs a nap lol.

As far as I was concerned, this was a lifelong project. I didn't expect that someday it would come to an end. I need to rethink my life.

As a software engineer, I know that developing something is a constant refactoring process. However I’m begging you to stop fixing issues that are not really an issue. Yet... when you add all of the vibraphone tracks you will have a great number of new issues. Marbles not filling up certain channels due to the speed of marbles through certain channels. The drums could stop playing due to a intense melodic part. Yes this is theoretical, and I know you have a solution for evening out marbles in the marble divider. Please please finish this machine and make more music. All the best /D

I honestly wouldn't mind if it took him another 5 years to finish the machine.

These videos have been the best thing to happen for a while. You have officially become part of the daily routine!

You'll get there, Martin! Let's goooo.

4:13 is so elegant

One day the table would break from too many hammer smashes

i can't tell you how much i am appreciating these daily videos in lockdown.

I‘m a web developer and see so many similarities in code refactorings. It is hugely satisfying to see you whack so many moles. I‘m gonna use that metapher in my team aswell to fight for the time to whack all our moles aswell. Thank you! But.. perfection aside, isn‘t it fascinating how much of the system still works even with this many flaws and deviations from the initial idea and concept? So crazy how hard reality hits you and your ideas in the face just to prove you wrong and show how naive you are. Even after all those years in the job :D

OMG THE COINCIDENCE OF ME LOGGING INTO RU-vid TO SEE YOUR UPLOAD...... HI MARTIN 🥺

Never clicked on something so fast in my life! :D

Martin have to say - some really great photography in this episode~! Love the pragmatic and relentless problem solving, but also -- beautifully shot! Thanks again for so many hours of entertainment and insight.

You can do it Martin! I can’t imagine working on one project for so long, has to be strenuous at times.

Anyone else watching this while waiting for SN9 to launch?

Can't wait for the next big test! Hopefully we'll get to see it this week. I'm not really sure why the "divide-by-4" is nessaccary? Why couldn't it just be a basin like where the conveyor picks up the marbles? Sure, the farthest tube would fill up quicker than the rest and it wouldn't be even, but that would remove the need for balancing and it would remove the noise. And, the rest of the set up wouldn't need to be changed, so it would be a relatively quick fix I think.

The ending of this makes me think someone’s been listening to Elon musk too much. Paraphrasing “the best part is no part”

Even though Martin is getting down when he has to sacrifice design, I think that’s part of being an engineer. You can either make something that’s functionally beautiful, or you can make a statue that’s exquisitely detailed. I’m here for the music that the MMX will play, so if something give you trouble Martin, axe it and get to the end of this build. Would love to see you down here in Australia, so you’ve got until the vaccine is freely available and the borders are open!! 😂