Lego Technic Huygen Chain Drive

Support Wintergatan:
- Patreon ► / wintergatan
- KZbin membership ► bit.ly/4cQVM7C
Marble Machine Engineering Discord Server:
/ discord
Video edited By Martin and Hannes from the Trainerds KZbin Channel:
/ trainerds
-
PATREON ► / wintergatan
KZbin MEMBERSHIP ► bit.ly/4cQVM7C
WINTERGATAN RECORDS ► www.wintergatan.net/#/shop
SPOTIFY ► bit.ly/2oKxXWd
ITUNES ► apple.co/2ntWNsZ
ENGINEERING DISCORD ► / discord
COMMUNITY DISCORD ► / discord
-
- The Huygens Chain drive allows us to play continuos music!
I´m not posting a feedback link this week, will only post when i have specific questions for the feedback, currently I´m still digesting the ins and outs of a weight driven system.
Good luck with everything you are doing :)
Marble Machine 3 - Episode 7
#4k #wintergatan
-
MUSIC DOWNLOADS ► wintergatan.bandcamp.com
WINTERGATAN RECORDS ► www.wintergatan.net/#/shop
SPOTIFY ► bit.ly/2oKxXWd
ITUNES ► apple.co/2ntWNsZ
MERCH ► teespring.com/stores/wintergatan
DISCORD ► / discord
SECOND OFFICIAL CHANNEL:► / wintergatan2021
-

[@maxedbricks589]

To spare you the R&D Martin, the next mechanism to look into is the clock winding differential as it won't cause backlash on the output like Huygen's ratchet does.

[@Iangamebr]

Can someone explain how the mechanics of that contraption works?

[@thomasbecker9676]

But Martin has to recreate everything in his own way!

[@maxedbricks589]

@@Iangamebr A differential operates on the simple rotational relationship out = (in1 + in2)/2. If the output is set as the power storage and in1 is ratcheted (meaning its rotation is 0 unless acted upon by a manual force), that will continuously force in2 to wind down. Please see segment 2 of "Making a Billion-year Lego Clock" for visual reference

[@kadehaviland3920]

Martin, I always loved watching your marble machine x videos in high school as I aspired to be a mechanical engineer. Now that I am at college to become one, I have loved watching your videos of the process you are using and the solutions you are finding to become a better engineer. You inspire me!

[@almerindaromeira8352]

That would be a hefty spring

[@BartJBols]

What you need is an expanding flywheel, its a flywheel built like a governor. It balances the speed of the flywheel to a constant by moving its weight arms inwards or outwards as it loses or gains kinetic potential (instead of slowing down or speeding up like a flywheel does). You will lose some potential due to friction of the mechanism, but you gain a LOT of consistency.

[@InventorZahran]

Great suggestion! I hope Martin sees this...

[@ADRIAAN1007]

One of his design goals is to keep components as simple as possible. A flywheel with an integrated governor is a lot more work to build and adjust quickly compared to a flywheel with a separate governor. Plus, with the governor separate, it can be placed anywhere on the machine and easily adjusted.

[@SlaxX]

I see fellow UEF, i comment

[@TheStaffmaster]

And a Planetary Gear torque converter (or two) might help with different BPMs. Need something that runs in 6/8 time? Shift one gear box to 3/4 and then halve the second one, for instance, etc,.)

[@MysteryTako]

He would need to also add a brake mechanism directly onto the flywheel and dive into the most finger-eatey module of his machine to adjust it between every song😬Not to mention winding it all the way back up from a standstill afterwards.

[@sephalon1]

Yes, he's definitely inventing "the clock"

[@thomasbecker9676]

Yeah, and he's doing it before the Greeks!

[@almerindaromeira8352]

Cool name is that how we are calling it from now on: "the clock"? I guess what other applications it might have in life😁

[@lucasvignolireis8181]

😂😂

[@lucasvignolireis8181]

next step: i discovered what i need is an analog mechanical computer!

[@pauljs75]

He's making a "mechanical synthesizer", so this is true. One has to establish a reference BPM source to drive the sequencer part.

[@grunions9648]

I'm learning to appreciate this channel as an infinite R&D process, now that I understand that it isn't about creating a music box any more.

[@aronseptianto8142]

that's fun on its own right, but yeah i do kind of feel a bit sad knowing how much the project has stalled over perfectionism i couldn't tell if the perfectionism is warranted or not because I'm not a trained musician but as someone that enjoy watching the marble machine grow in both the mechanical whimsy and visual art aspect of it, it's a bit sad to let it go, but i also welcome this new chapter as it's still interesting nonetheless

[@leo-paulbonvalet3686]

​@@aronseptianto8142 you said it all... I'm also very sad to leave the MMX behind, as it certainly was full of little failures, but it also was beautiful of complexity, of things going in every direction but still making sense... All of that, I'm afraid of not finding this in the MM3...

[@grunions9648]

@@aronseptianto8142 Me too, totally. I guess the alternative is some demos of the MMX and then nothing else, so at least we get to see someone figuring out problems again. I just hope that the result (or at least the end-goal) will be as beautiful as the MMX was.

[@armr6937]

@@grunions9648 This machine has to tour, I mean... Setting it up and playing a full concert anywhere it goes can't be a massive headache or it just won't be worth it. Martin has had to become an engineer in order to make this happen, first he needs to get the basics right first, then we'll see him add the complexity and visual pizzazz we all long for.

[@aronseptianto8142]

@@grunions9648 i think it would be beautiful, just in different ways. It's a different vibe and it's going to feel different. It's just going to cater to the more industrial aesthetics

[@PiercingSight]

Also, side note, I LOVE that the music's tightness and rhythm would be controlled almost directly by you with a hand-powered crank. That just makes it so much more charming in so many ways, and I think that automating the whole thing, and making it inhumanly tight kinda detracts from the magic of the machine. But you want tight music, and I think the machine is going to be cool regardless, so I'm looking forward to whatever you choose to do!

[@conorstewart2214]

Did you watch the video? The tightness and rhythm is controlled by the falling weight. All the cranking does is keep the weight raised, it won’t affect the tempo or rhythm in any meaningful way.

[@aronseptianto8142]

@@conorstewart2214 i think that was his point, the MMX was only crank and flywheel which was more messy and as such charming.

[@conorstewart2214]

@@aronseptianto8142 that may be what was meant but it wasn't what they said.

[@hanvyj2]

@@conorstewart2214 That's the point. A laptop would also be perfectly tight... The tightness of the previous machines wasn't really an issue. The ability to change tempo was a pretty big part of the first marble machine song, at the end. I find the flywheel was a pretty big part of the character of the machine.

[@offbeat4772]

@@conorstewart2214 I think that is what he said. It's exactly as I understood it given the context of previous videos

[@Bbeaucha88]

Here is a set of hypotheticals: A. Would you go to a concert where a guy wheeled out an automatic piano he built that plays PERFECTLY tight music and he feeds in new programming sheets between songs? B. Would you go see a concert where a guy built a complex hurdy-gurdy type instrument where you watch him play it and make music you wouldn't expect possible? Personally, every single thing Martin has been saying lately is inching him closer and closer to hypothetical A. That's fine, the engineering is cool and fun to watch but at some point this is going to stop being a musician-instrument relationship and instead will be a machine-operator relationship. I would gladly go watch a musician play his instrument, I would not pay to watch an operator operate his machine. This is just my subjective personal opinion so anyone who disagrees with me is totally cool.

[@ivandagr857]

Maybe he could go for hypothetical A, while implementing AI to make programming sheets

[@IvoTichelaar]

This crosses my mind, too. But I love his music box as well. I think when he mentions stadium tours, the marble machine will maybe play one or two songs by itself and have a more supportive role during the rest of the concert.

[@conorstewart2214]

@@ivandagr857 why would using AI to make the music be any better? It would make it much worse, you are completely removing the human factor then other than for powering the machine.

[@conorstewart2214]

Exactly, it is becoming too automated. If the weight was high enough up then it could play a whole song by itself with no human input, him just cranking the weight up doesn’t add much to it.

[@johnnyjohnson6643]

I’m afraid I’m going to have to agree with you completely

[@kiqrocha]

I will never understand his obsession with music being so tight

[@Darkblitz9]

I think he wants a machine that after sitting in a museum for 100 years will still play music exactly as expected on day 1. Getting music extremely tight now allows for age to have little impact in the future. I think he's upset about the fact that the MM1 worked for one song, poorly, and even though it's in a museum now, it's unplayable, and no one can hear it play live again.

[@paulramsey2000]

Small errors introduced by a human will be organic/random and often pleasant. Errors introduced by machinery could easily have patterns which we will actually hear. So an easy target with machinery is to just get rid of all error even though organic/random error would be fine.

[@kiqrocha]

@@grunions9648 After this one is complete, he will make a MM4 that will feature human-like imprecisions

[@DaftFader]

Don't forget he plays alongside midi controlled arps. etc. Aside from the drift which absolutely cannot happen for this machine to work in his sets, there's a huge difference between gradual swing timing errors from a human, and random and possibly even repeating patterns of timing errors which could become quite noticeable and extremely off putting, even on the + a few ms scale. Especially if they are both before and after the beat randomly one after another. It really messes with the feeling of flow, especially when alongside actually highly timed music (like a midi controlled arp for example). 20ms is all it takes for an instrument to start to begin to really sound out of time for a trained ear. If you already have tightness, you can then manually play with timings to get a non random, more natural and less robotic like performance if you really want, but when you're playing alongside midi arps anyway, that robotic tightness of midi might be the sound he's going for (it certainly sound like it at this point).

[@bongi6811]

He gives his explantion at the end of this and the last video. He says, that he woudn't like to use a machine that isn't precise enough to replicate the music he wants to compose for it and fulfill his vision of a marble machine is general. I my opinon thats completely understandable, since as the composer of the music for the machine, he is the only person who knows whats good enough. The question is just how much time he should spend on this, before moving on and working on something with better chances on quick improvement. Spending a few weeks with lego prototypes to potentially find a revolutionary solution isn't doing any harm and at some point martin will either find a great solution that's up to his high expectations, like with the 3d printed marble gate or he will have to accept a compromise between development time and quality and move on.

[@jonassandstedt]

Isn't a variety in tempo and tightness part of the joy of listening to live music instead of a computer? It makes it more human.

[@tommaniacal]

Inb4 martin removes himself from the machine because humans are too inaccurate

[@BaddouOurs]

THANK YOU! This is my thought since the MMX. Why chasing perfect tightness of the music since the small variations give life and colour to the music. Isn't it why composers in electronic music give small random alterations of few milliseconds to their notes?

[@kuzakiv3095]

also the clunkiness of the first design made it look really cool in my opinion, like howl's moving castle type

[@mickys8065]

MM4 will be midi machine 4, because marbles and gravity aren't precise enough

[@jonathanschubert9052]

With MMX, maybe. MM3 has way more channels with more stacking tolerances. That being said, MM3 is almost certainly too big, but I'm just a guy on the internet.

[@xxportalxx.]

It's funny to me how this is basically building up a mechanical analog of a linear electrical power supply, you have an alternating power source coming in (your foot peddle), some form of rectification (converting the peddling to turning of the shaft), a capacitor (in the form of the weight), an inductor (in the form of the flywheel), and now a linear regulator dumping excess energy using resistance (in the form of a governor).

[@jdmjesus6103]

I suggested using something similar a few weeks ago. But I also suggested using an escapement and pendulum. You probably wouldn't need the flywheel then, or at least a smaller one. Bpm can be adjusted with the length of the pendulum. A governor is going to be less consistent and less predictable. You could literally mark the pendulum with bpm increments and it would be the same every time, in any conditions, unless you take it to the moon!

[@dedoslav]

I was going to make the same comment, but just checked if anyone made this observation already and indeed you have. Instead of trying to fine-tune a continuum mechanism, pendulum-based (or maybe some other) escapement mechanism brings the tightness that only a discrete system can achieve.

[@GordieGii]

Going to need a really short pendulum to get 120 BPM...

[@Nitro8996]

​@@GordieGiiif I'm not wrong someone invented gears a while ago...

[@GordieGii]

@@Nitro8996 You're not wrong. Someone did invent gears a long time ago. What's your point?

[@GordieGii]

@@quirinr96 Maybe a drum machine...

[@Djellowman]

The hand crank makes this an instrument. Using a constant weight makes this a music-playing robot.

[@justinnaramor6050]

You obviously have not been around the marble machine project long enough, then. Turning a crank does not automatically make something an instrument. That is actually a pretty stupid comment. The marble machine is not a human-played instrument where each individual sound is manually triggered by the musician. It is playing a pre-programmed piece of music which is encoded onto a wheel that rotates. It's still a "music-playing robot" either way. the hand crank is nothing more than a power source for the rotating wheel, and it's a logical fallacy to argue otherwise. If I turn a hand crank for a barrel organ which is playing pre-programmed music, I am most certainly not "playing an instrument". The thing that really makes the marble machine more of an instrument than just a musical robot, and the thing that got me fascinated with it in the first place, has nothing at all to do with whether or not it's hand cranked. There are hundreds of hand-cranked music boxes, so it's foolish to act like that's something special. Instead, the special part, is the fact that instruments can be muted or un-muted, and the strings of the bass guitar (which are triggered by the programmed music on the rotating wheel) can have their pitches altered, using the fretboard. And there's a way to manually drop marbles onto the instruments. All of that is controlled by the human operator. Pretty much all other mechanical musical "instruments" are just playback devices, where once the pre-programmed music starts there's not much else you can do musically. The closest thing to the marble machine as an instrument would be something like a player piano, which is just like every other piano in that it can also be played entirely manually with no pre-programmed music. And I've also heard of something called the Gloggomobil, which is a xylophone-like instrument played by mallets. Mallets can be operated by the pins on the rotating wheel for the pre-programmed music, but the instrument can also be played manually with hand-held mallets. That is what really makes something like this an instrument. It is not the hand crank. Hand cranks are nothing special... they only serve as a power source. Nothing more.

[@Vastafari34]

So pretty much what is happening here is that Martin has a goal for an instrument, primarily to be able to play crazy tight music, and he has come to the realization that just making a giant instrument with modified and enlarged clock-work mechanisms is the best way to keep time (It makes so much sense since clocks are pretty good at keeping time...).THAT IS EPIC and beautiful, and organic. I have LOVED every second of this journey all the way from MM1

[@Corn_Curls]

Perfect. Every concern I had with the weight drive from last episode evaporated with this one. This is already a much improved power delivery system than you had three weeks ago and improvements can only dial it in. Keep it up Martin!

[@mackado]

If you want to compare your flywheel and weight drives accurately, you have to make sure the angular momentum of the flywheel is comparable to the potential energy of the weight. The more mass (in either system) the more reliable the rotation will be.

[@philippsaure7545]

Going back to the first principles approach: While it isn’t technically the easiest or most rudimentary approach a inverter/motordriver with closed loop control is more or less a plug and „play“ solution. Trying to engineer a beautiful mechanical system is more hassle than it is worth if you are going (as I understand) the form follows function approach.

[@GordieGii]

@@philippsaure7545 If Martin wanted plug-and-play I'm sure he would buy a synthesizer.

[@philippsaure7545]

@@GordieGii You are totally right. But it is known (at least since the invention of the mechanical clock) that there are ways to let a shaft rotate with fixed rpm even mechanically powered. I think that is a somewhat solved problem. And if you can’t solve this you have a reliable solution (the motor) in your backpocket. But there are a lot more problems on the table. And once you got the whole system and it’s requirements figured out you can design a more fitting „power source“.

[@GordieGii]

@@philippsaure7545 A weight or spring or a steam engine that is controlled by a governor is technically a servo system, whether it modulates a brake, a valve, or a variable ratio transmission. The marble machine has always been a human powered instrument. Using an electric motor, servo or not, would be cheating. I think the governor is going to be a problem, but it is an interesting problem to solve. I don't think a speed governor alone will give the required performance (or 'tightness') that Martin wants. Too much lag. But I think in combination with a torque governor to compensate for the dynamic changes in load could be made to work. I also think the ball elevator needs to have a separate drive with it's own governor, but that doesn't need to be nearly as precise.

[@philippsaure7545]

@@GordieGii you are right. Once Again. Still: Don’t you think it would be more reasonable to design the powersupply once the requirements of the machine are better known? Especially because you can only guess the sizes of the shafts needed and the required torques. I am not against a human powered machine but I think you can build a better (as small and as simple as possible) once you get to know your machine.

[@vidiot5533]

I think there's been an oversight. Lifting marbles takes energy. If a song (or section of such) uses more marbles, the machine will require more power. Thus the goal of having steady power delivery should take into account unsteady power consumption. The governor is what helps with this, so 100% this device NEEDS to be spec'ed to have the minimum and maximum resistances at the desired speeds capable of providing resistance equal to the minimum and maximum operating resistances of the marble machine to effectively balance the power

[@blacklabel6223]

Oh shoot this is an overall oversight. Moving so many marbles so fast so high is already going to be a wildly exhausting feat, much less the actual energy requirements vs the weight size

[@petecomps7260]

The power input must not only overcome the friction, but the energy required to lift the marbles. If at any moment in time, the machine is on average lifting 100 marbles, the energy required to lift the marbles is also a component of "power usage." If the number of marbles being lifted varies from moment to moment (which seems likely), the power usage will correspondingly vary.

[@TheUncleRuckus]

Excellent point! 👍👍

[@Temperans]

That may not actually be true if there are enough marbel at the bottom to act as a buffer than the amount being lifted at any given time would remain constant; Even if the buffer pre and post left are changing in value.

[@Luddevige]

It's crazy that you can get continuous tight music without an electrical motor. This is so interesting.

It's like clockwork!

[@angellestat2730]

it is crazy that he can get continuos tight music just using an electric motor with the option to choose any music speed that he wants and forget about all this complexity :) It is like he still wants to find the hardest path as he did in the first two machines.

[@kentonroush]

@@angellestat2730 Well yeah. Anyone could make a 'marble machine' using electric motors. Or, frankly, a synthesizer. If the goal was just to create music, the whole project is largely 'pointless'. The clockwork aesthetics, the limitations, the machine itself, is all part of the goal too - The trick's just figuring out the right balance between practicality and complexity to get a machine that's beautiful to watch in operation, but not excessively difficult to build and maintain with sufficient reliability.

[@damicapra94]

@@angellestat2730 just play an mp3 at that point

[@angellestat2730]

​@@damicapra94 so according to your definition all modern air church organs and even old pipe organs are mp3?

[@bradleydalton2170]

I'm not sure if you're looking for theme or aesthetics, but I'm struck by the imagery of the Sword of Damocles - a sword hanging over someone's head. And so I have this silly image in my head of having to constantly crank the Marble Machine to keep something heavy or dangerous-looking from dropping onto you; literally playing for your life. Either way, this is a neat and fascinating improvement!

[@Palerider1942]

That would be an awesome aesthetic to the marble machine XD a slowly falling sword as the show goes on until it sinks behind the MM3 and the set is over

[@EdbbieRosado]

I've learned so much about engineering and physics from this channel in the past few years that i can't even assign a proper value to it. Thank you, Marteen!

[@IanZainea1990]

Martin. You cannot use the Lego models for fine measurement. You can only use them as a guide, rapid prototyping to get an idea of what might work. If you try to test using Legos you're going to burn yourself when you move to real materials. Only use them as a guide!

[@thomasbecker9676]

But the tight tolerances on LEGO blocks! /sarcasm

[@Sembazuru]

I hope that I'm right that he is just using the Lego models to test out different proofs of concepts. I could be wrong, but I think I hear that in the subtext of his explanations.

[@conorstewart2214]

I saw a video of a PhD project of a 3D printer for printing biological tissue, it was made of Lego apparently so it would be easier for researchers to copy and it would be cheap, everyone in the comments defended it saying that lego was absolutely suitable for a “complex” machine that is apparently meant for use in research labs. It was using lego Mindstorms motors. It was complete BS and the machine wasn’t even that great, they would have had much better results just modifying a cheap 3D printer. A cheap 3D printer would have been cheaper anyway. The only new or innovative part was the tool head for extruding the tissue, the rest of it was just a crappy, tiny, imprecise 3D printer made of lego. Just modifying a cheap 3D printer would have been, easier, cheaper and much more precise with many more features. But no according to them and everyone in the comments a crappy 3D printer made of lego was the best option for something meant to be used in research labs. It was pretty insulting to researchers saying that they made it out of lego so that researchers could copy it. I think if you are doing that sort of research you are more than capable of modifying a 3D printer.

[@jetison333]

@@conorstewart2214 its not even like the lego mindstorm sets are cheaper than a cheap 3d printer anyway lol.

[@yerwol]

I'd love to see where a human comes on that chart of "how tight a music can X play?"

[@yerwol]

@@Josh_728 .... what a helpful answer. Can you elaborate? I'm just interested to see how well a human can keep time. And I'm no drummer myself, but have an engineering background and played piano and guitar in my youth. So don't just jump down with the "musically illiterate condescending" comments thanks

[@JamesScholesUK]

@@yerwol the third bar in the bar chart is Martin attempting to wind to a metronome, so that's probably a fair approximation of "how tight can a human get with this particular mechanism?"

[@gitsurfer27]

@@JamesScholesUK No he was winding a small lego crank, very unreliable method of keeping time especially with no practice. A decent drummer with some sticks could play 80% tighter.

[@icycooldrink6085]

@@gitsurfer27 80% exactly? Might it not be interesting to measure and compare that directly instead of using a proxy? If only Martin had access to experienced musicians to record. Sounded more snide than I intended, but bottom line is, I think the OP has a good point. It'd be nice to have a real guide to how tight a person plays, rather than '80% better than a proxy movement'.

[@yerwol]

@@icycooldrink6085 exactly. I'd love to see a comparison of the wintergatan members to see how accurately they can play on a drum or piano or typewriter or online bpm calculator whatever else they can find!

[@TnOO]

Somehow, I feel like this is the precise instant where the marble machine slips off the "instrument" category. A "human" power input is what makes an instrument an extension of the body, and I think, is the reason why music makes sense, because you feel the rythm of one's body. Then again, there's electronic music, and probably many other exemple where this doesn't apply... Anyway, good luck

[@charlesgalant8271]

I think this just makes the marble machine a little more like a pipe organ, which is getting consistent power input from a set of bellows. That concept is kinda combined with a synth or loop station where there's a 'recorded' set of tracks and the artist is choosing which combination play at any given time. If it's anything like the previous machines, he'll probably have a way to play manually as well.

[@SukacitaYeremia]

Giant church Or- oh someone beat me to it hehe

[@Phriedah]

I swear someone comments this every week. It isn't called "The marble instrument", it's "The marble machine". It contains instruments, but is a machine. Martin will manually interface with the machine to make music. It isn't worth debating, ultimately it's a matter of opinion.

[@MrGonzonator]

I disagree. It has always been a mechanical machine running on human power. Is a bagpipe not an instrument because you don't directly blow in the pipes, but instead store air in the bag?

[@johnnyjohnson6643]

Feel free to direct me to the timestamps that make it seem like Martin will be directly interfacing with this machine in order for it to play music, fellas. Is it gonna be when he hits the “on” switch? That’s the clear trajectory his project has taken.

[@DavidGuyton]

I have a feeling this is going to be one of those parts added to the machine that will get deleted in the future for a far simpler idea.

[@randolflp8626]

Why exactly? 1: I think the flywheel of the MMX Was a good Idea; 2: Then you can install a Motor to hold Speeds and Timing. But overall, I think exactly this Clicking sounds, Imperfections in Music Timings are, what makes you, your machine and the Music you produce so unique and beautiful.

[@toddcamnitz6164]

I think the thing I’ve come to appreciate most about this channel over the years is the evolution of what makes this struggle so inspiring. Regardless of my personal opinions about design/engineering/approach/mmx/ and so on, I have immense gratitude and respect for the fact that you’re continuing to carry on in spite of everything. Incredible.

[@techdaemn]

Centifugal Govenors actually were meant to control the throttle of a steam engine (as the balls go up, it pulls on a central rod) , but in your case you could probably just use air friction with flat paddles instead of balls, such that as they fly outward during to centrifugal acceleration, they quickly cause greater air resistance since it is trying to both spin faster and on a longer arms. Side note.. This is where the phrase "balls out" comes from. Is what happens when you're at top speed (the ball govenors balls fly outward).

[@H34...]

Wait till martin discovers how tight a midi keyboard can play! On a more serious note, a sprag or roller clutch will fix your backlash issues. The bigger issue is the mmx does not have a constant power usage, it depends on how many and how often notes play, and adding a governor will sap significant energy. A simpler governor as opposed to the centrifugal one, is a fan or using air resistance. Little music boxes often use this.

[@NojRel]

Part of the beauty of the first marble machine was the variability of the speed.

[@vincenemeth6522]

I really appreciate how much time effort and work he puts in and the r&d that goes into this. However I have a feeling that Martin went to another extremity. He went from barely thinking about the practial design of the machine to overthinking and over engineering. At this point he wants to implement something to make his machine play tighter music which is useless/unnecessary. The perfectionism will be the archnemesis of Martin and the machine. I don't think he will ever finish the mm3 if he continues his probably unrealistic perfectionism.

[@WeArePharmers]

to be quite honest, it's kinda why I lost interest in following the construction of the next Marble Machine - whether it's MMX or MM3. as the machine starts progressing, the perfectionism takes over and becomes this rabbit hole that he can't climb out of. at this rate, the machine will never be finished.

[@cutlery]

I don't think so, he's already made a fair amount of progress and isn't showing any signs of slowing down

[@O-.-O]

So far the only perfectionism he follows is generating new "tight" buzz words

[@STAG162]

something that may have been overlooked in this joyous discovery, is the excess noise transmitted from the ratcheting system.

[@matthiasengh7935]

So you can use weight drive and still stand and crank the machine continuously!! This is the number one most awesome video in marble machine history!!

[@Rickkepik]

Hearing Martin pronounce Huygens' name hurts my dutch soul (Love you Martin)

[@pieface1726]

If you want to avoid backlash, you can add a differential. Also side question, how do you intend to play music with different BPMs?

[@GordieGii]

He said the governor would be adjustable.

[@m00str]

By exchanging the 10t weight with a 6t one mid-song, obviously ;) He really hasn't thought about it yet. There is just no way. He takes principles that work for fixed loads and tries to apply them to this very dynamic load. Let's just hope he doesn't waste too much time on those ideas

[@CameronPenner]

It is very possible that the wieght drive is only more tight given consistent friction. Once you have variation in the forces applied against the drive, it may lose precision. For example, when you start playing many notes, you may lose tempo. Edit: I made this comment at the beginning of the video before you introduced the governer concept. This looks very promising!

[@thomasbecker9676]

There's no "may" about it.

[@GordieGii]

Yes, you want to have enough weight to drive the max tempo at the max friction/load and use the governor to hold back any excess power when the tempo and/or friction/load is lower.

[@Beakerbite]

@@thomasbecker9676 Theoretically, if the governor has variable resistance, then it should smooth out the variable demand. That's entirely dependent on how variable the governor is and how responsive it is. I'm not sure a wind driven governor could add enough resistance for the amount of weight Martin is going to need.

[@smit5983]

The explainer for why you're obsessed with timing at the end was great!

[@DeafAlone32]

I love the videos!!! They are super informative and super consistent It's so good and I'm following this every step of the way, i love it!!

[@roboman2444]

I think a clock escapement for adding power to the flywheel is a good idea. You would be able to change the temp by adjusting your pendulum length.

[@Sejiko]

If you have already a flywheel and a govenor why do you need a gravity based system to give you a smooth power input? You do double the work but get only the same output for it.

[@huub8]

Correct me if I'm wrong but: If you slow down a flywheel (to slow down the music), energy is just lost. It won't be able to speed up again unless you manually put energy back into it to rev it up. If you slow down a falling weight, you partially save energy (/you reduce the rate of potential energy consumption ). So you can speed it back up without putting in additional energy. So a governor with a flywheel will be able to limit the maximum speed, not so much be able to keep it constant unless you constantly try to rev it up yourself.

[@a11y93]

I love your videos. Im an engineering student and love doing little electronics projects in my free time and im trying to get motivated to tackle some of my more complex project ideas. I always feel like im getting a liitle bit closer to feeling confident starting a bigger project the more i watch your videos:) Im also an artist and can really relate to being stuck in an artistic idea, but watching your stuff has really taught me to drop some of them when making stuff for the real world.

[@gamingscientist7445]

I'm learning a lot from videos like this. Thank you for sharing the step by step process

[@plewelly]

This doesn't look like over-engineering to me. I had thought with your last video that you needed to find a way to be able to input power to the weight drive system without speeding up the machine, and then you found just that. This seems awesome.

[@mikeperry4262]

Its almost like a set of bagpipes where you have to keep "blowing" in order to keep the music playing. And you can stop when ever you want. This is a massive game changer and should be implemented.

[@HueMann]

A set of bagpipes was exactly my first thought too! I think this is a good compromise with the analog feel of the machine

[@tinwhiskerSC]

This is probably the best idea you've had yet. It's very functional, well understood, and potentially looks interesting/entertaining.

[@bradley3549]

I absolutely love the mechanical contrivances. Makes me tingle in very peculiar ways.

[@alimahmoud6966]

martin i hope you havent forgotton the main principal upon which mm3 was built: "remove dumb design requirments". do you relly need a whole complex mechainism for powering the machine?

[@benjaminbarratt-peacock2633]

Martin, have you considered using a bellows drive like a player piano? It would be much easier to regulate the BPM of the music and you would still be cranking but you could try pneumatic devices to run the machine

[@roycsinclair]

Well done in prototyping things using Legos so your build time for each iteration is reduced. You're getting MORE of the basic research done without taking as much time as it would without Legos.

[@dudefromkeene]

You interacting with the machine is what makes it worth watching.I like the idea of a foot or hand crank with a governor

[@struanpeat5116]

it might be worth looking into how the balance on a watch works, a carfully balances wheel continuously powered by a weight drive can keep time in a matter of seconds over decades

[@maxedbricks589]

I've experimented with this idea. The locking of the escapement causes a reaction impulse which will cause the system to achieve static equilibrium. To get continuous power, a governor is the best option.

[@alessandroscognamiglio9699]

Hey Martin, in order to play different tempos you could try to implement somthing like a CVT system: a Continuosly Variable Trasmission that will gave you infinite gear ratios, it will be a bit overkill, but worth the effort in my opinion

[@PKamargo]

I am happy that you are including more horology in your project. A modern alternative to weight drive is a constant tension mainspring.

[@NiksSofa]

Ah, and there I thought the next mmx would be boring because it's designed sensibly. Good to see the madness is still alive and kicking :D

[@bigsarge2085]

😄 Find someone who looks at you the way Martin looks at a Huygen drive!

[@Nanarchy_2k7]

Find someone that looks at you the way I look at a 1970 Dodge Demon with a hellcat engine installed.

[@br52685]

Why would you want to find someone you’ll just abandon in a few weeks, after moving on to something new/fresh/exciting?

[@christiankuziara8355]

Yes this is amazing!!! I can’t wait to see how you fit these pieces into one 😂

[@Toastmaster_5000]

The Huygens chain drive is a fantastic idea to implement in this - minimizes weight, complexity, electronics, and timing inconsistencies.

[@lii1Il]

Nice! I see some great uses of these principles! Ty

[@CristiNeagu]

The tightness doesn't come from the weight drive itself, it comes from the fact that you no longer used your hands as the input.

[@funkenjoyer]

which is pretty much the point, handcranking -> inconsistent power input, weight drive -> consistent power input

[@CristiNeagu]

@@funkenjoyer The point of that there's way better ways of doing this than using complicated weights and governors.

[@JoeHannouch]

a motorized motor has always been the best solution, he even used one on the previous MMX, this is becoming weird to watch

[@funkenjoyer]

He used a motor on MMX to be able to run long consistent tests, he never intended it to be the main drive for the machine

[@Yorque]

How tight does the music need to be? And what is the determining factor for this?

[@GaiusCaligula234]

Martin's obsession

[@Timberwolf69]

This marble machine will play even tighter than a synthesizer when Martin's done. And then the band has an issue, as no human is able to play this tight.

[@JmsNmnn]

MMX videos are what got me interested in this channel. I followed the project from around the middle of the build process to the end, and I can say that even though that project ultimately didn't match the set expectations, those are still some of the very best videos on KZbin in my opinion. Martin's marriage of ingenuity with artistry is so inspiring. MMX was a beautiful disaster and I have NO DOUBT that MM3 will live!

[@roippi3985]

While everything is on the table, consider a pump/tank/waterwheel system as an alternative to provide continuous gravity power. Major kudos for keeping an open mind after you did all that work on the flywheel assembly!

[@tobyCornish]

I now want Martin to have a Lego channel

[@dangerrangerlstc]

The governor and flywheel idea is a good one, and an old one. They've been used for centuries. Old farm equipment relied on governors to keep equipment running at a constant even speed. Threshing machines for grain processing were powered by steam engines, whos power output could vary depending on the available steam pressure in the boiler. You could automate the recharge of the weight using some limit switches and an electric motor. When the weight reaches the bottom, it could turn on a motor attached to your ratchet which lifts the weight to another switch that turns the motor off.

[@svenenglund]

I laughed so hard at the end, it's such a pleasure to see you back on track that brings you so much joy Martin!

[@megavali1001]

At this point, why not just use an electric motor that drives the machine and an encoder for the crank? This way, it would be the most flexible in terms of BPM and input filtering (letting more or less of the crank speed variations through). Parameters could be changed even while playing a piece.

[@Defaultsound]

Don't know if this is any use. But you can also use a differential to allow continuous power. Look up 'Making A Billion-Year Lego Clock' it uses this idea.

[@evenbrattestad8364]

you truly are a marvel of the human existence. I enjoy your videos and the enthusiasm you put into all of you decisions. They all go way above my head, and i dont care. The effort, and persistence of your journey exites me. I really love what you are doing and i hope you find a solution that can give both you, and the community peace of mind.

[@GordieGii]

I suggest having two separate drive systems. One for the music, and one to elevate the balls. I think that the vast majority of load variation will be the number of balls that have to be lifted back to the top of the machine per unit time as the number of notes played per unit time changes. This would put a high demand on the governor to compensate. The ball return mechanism wouldn't need to be very accurate so its governor wouldn't need to be very precise.

[@NotLegato]

at this rate, you'll be chasing perfect precision forever without ever actually creating music.

[@Oxtorayk]

This machine stuff became an end in itself.

[@NotLegato]

i can totally understand it. the engineering is (a large) part of the fun, but it'll be more satisfying if it actually... ever *exists* outside vague design ideas for more and more imperceptibly better precision.

[@NSGolova]

Uh oh, Martin is moving up technology tree. I wonder when he will discover the magic of rotating magnets in copper windings, and how incredibly tight they can be XD

[@robbyvansambeek]

You can also use your machine as the weight. Make a frame where the machine hangs in as the weight therefore you might not need as heavy a weight

[@lambo_drives]

Fantastic - I have dreamt of weight drive mechanics all my life - THIS is the device!! thanks Martin for sharing this nugget of knowledge

[@BjarneLindqvist]

Use a electrical drive with esc and governor. Problem solved. It won't be any tighter than that. If you still want it to look as if you are doing anything manually, put a crank on a generator and a supercapacitor.

[@eryktheviking3371]

I feel like the spirit of the project is to build a mechanical device that makes music. Sure, an electric drive would be the easy answer, but a mechanical drive fits the project better. He's dropping marbles to play songs...easy left the building a long time ago.

[@BjarneLindqvist]

@@eryktheviking3371 I was a little bit sarcastic really. I don't think he should go the governed way at all. The music should be tight enough with a flywheel. But anyway it's not up to us to decide. It's his project, he does what makes him happy I hope. 😊

[@MattGilliesCwnAnnwn]

Making a clock. Knew it

[@evanbarnes9984]

I love this idea! There's a rich world of clockwork mechanisms that I really think you should keep drawing from. They're well developed and well understood, and you could adapt them to your purposes rather than create them from first principles.

[@billkeithchannel]

LEGO has an actual _differential gear_ piece that transfers motion depending on where it is coming from. This is what winds a music box/clock without stopping the music.

[@YingwuUsagiri]

Every time I see that mock-up of the Marble Machine 3 I can't get over what a horrifying monstrosity it looks like. Instead of being a cool self contained instrument that can actually move relatively easily like a traditional street organ which is also a full band working on mechanical input it's looking more and more like an exploded monster that will take so much effort to world tour and enjoy on stage.

[@stuffiguess7680]

It’s was fan made, and if you look at the actual cad designs for his machine it’s much more modest

[@YingwuUsagiri]

@@stuffiguess7680 True! You're absolutely correct but the worrying part is that Martin does keep referring to it *and* keeps speaking of 'modular design like the example' and we all know how good he is with going overboard, leading to function creep leading to inevitable impossibility to be reasonably possible.

[@elindred]

In a sense this feels like reinventing the pillow blocks all over again; he could just attach a motor to the thing if he wants it to play tight for now, allowing him to chase down sources of looseness further down the mechanism, _which is what really matters._ I hope the machine will still be human-powerable in the end, with a constant drive as an option. I understand the appeal of chasing tightness to the extreme, but dropping manual power altogether would be sacrificing that special something about the human cranking the machine to play the music.

[@Robbedem]

I think he would still crank. But the crank would lift the weight that powers the flywheel. This will make the speed of the music independent of the speed he's cranking. How he will regulate the speed so he can play different tempo songs? No idea, probably changing the weight. But he won't be able to do speed changes in a song unless he adds a differential between the weight and flywheel. So, the machine is going to become complicated. (imho not worth it, but who am I?)

[@jeffreymach7741]

I am wondering if he's getting too fixated on a manual / mechanical energy source. He wants a weight drive but then realizes he needs to constantly reset the weight or it will be too large/dangerous. Unless you are willing to almost constantly crank the weight back up, you need a motor. But if you want to use a motor, you may as well let the motor power the machine and get rid of the unnecessarily complex falling weight. Worse, he realized after he fell in love with the Huygens mechanism that falling weights are only good if you want a single, unchanging power output. So, he needs to add a great deal of complexity to vary the speed. Or, you can adjust the motor speed control. If he wants to be creative, come up with an elegant input device for the motor speed control setpoint and let a tiny PID controller do the rest.

[@vosechu]

Thank you for sharing! I know the pain you talk about, and I feel you. Sometimes you can’t even express the pain adequately, so you just don’t. It’s okay to let it out, and it’s fun to see you develop the tears-to-rotational force mechanism that is Martin.

[@Tomcroese]

Hello again! Good to see your work!

[@ivoackermann15]

not to be a downer here...i told you about 3 years ago that you want to build a clock that spews out marbles. you're slowly getting there.

[@EdwinSteiner]

Don't you need variations of the tempo within a song for musical expression?

[@gitsurfer27]

Only in classical music

[@DaftFader]

Different weights?

[@martinlundin3106]

Oooh this reminds me of David C Roys early work with the gravity fed spinning art! Omg wouldnt that be the best colab ever, Wintergatan and David C Roy, both ears and eyes will have sensery overload! Its getting more and more awesome by the video, keep it up Martin you got this!

[@TheBookDoctor]

This is very cool. In your last video, I was thinking there had to be some clever way to make it continuous. I'm glad there actually is! Keeping the human-powered nature of the marble machine feels philosophically important to me.

[@Der_Dirk]

Didn't you already had a controlled electric motor for Marble Machine X? Just take that and get done with the project. ;)

[@ZapAndersson]

Go back to Speelklok museum, every clock there has solved this problem already.....

[@chuxmix65]

The bagpipes come to mind... you can squeeze somewhat inconsistently, the small aperture keeps the airflow to the pipes (somewhat) constant.

[@philiproyd6563]

For the past few years I have been pondering a device which will run off a descending weight, but I wanted it to run constantly for an hour, which would mean the cable the weight is attached to would be extremely long. However, thanks to you, when I saw this video, I knew my weight dilemma has been solved. You talked about a governor, which I knew was something that I would need, so I got my aunt's old wind-up Victrola Phonograph player to harvest one from it. Thank you for showing this. This was FASCINATING!

[@tcode3564]

Keep in mind that you want to go on tour with that thing. So you probably can't carry a to gigantic weight with you.

[@blueberry1c2]

Sandbag + locally sourced rocks perhaps?

[@Robbedem]

@@blueberry1c2 water is often used as a weight too.

[@blueberry1c2]

@@Robbedem tbh I would rather use rocks. Easier to fill up and it won't leak water all over the stage if it drops

[@Robbedem]

@@blueberry1c2 But the weight needs to be exact, since they will determine the tempo of the music (unless he puts in another contraption).

[@blueberry1c2]

@@Robbedem he talked about a governor, which makes it sooo much easier to change tempo than calibrating the mass, whatever mass is used

[@rexmcstiller4675]

2:33 To be honest. A 10m tall tower on stage where you can see the weight gowing down during the song would be cool.

[@martin-__-]

The advantage of manual power input is you can feel the resistans in the system and actively choose a bpm. With any weight drive, or other constant force drive, the force has to balance with the friction in the system, which can easily change if eg. a bolt loosens or you re-apply lubrication to som part of the system.

[@KnowArt]

love it, let's gooo

[@ordanarymods4990]

I imagine this is definitely due to my lack of experience in music production, but i genuinely cant tell a difference when you show back examples of tight and not-tight music. Obviously i know that it is a real issue because it was enough for you to cancel the mmx but I wonder if you might consider doing a video that really focuses on music tightness for those of us who try to hear the difference but cant

[@p3t3mit]

I'm a hobbyist musician and nowhere the calibre of musician like Martin, but I definitely can tell the difference in tightness. It feels like the difference between a high school band and a professional band.

[@ordanarymods4990]

@@p3t3mit i’m not saying there is no difference, i just meant to suggest it as a main topic for a video given that its playing such an important role in the design

[@p3t3mit]

@@ordanarymods4990 Makes sense.

[@Robbedem]

I am a musician, and the music doesn't need to be that tight. Plenty of succesfull songs on CD/spotify/LP/... are not tight at all.

[@p3t3mit]

@@Robbedem I think that's completely a style thing. Plenty of indie songs I listen to have pitchy vocals and sloppy mixing, but that's part of the aesthetic. I wouldn't say categorically that all music should be x or y ms tight or loose.

[@mielprovoost6049]

At this point just get a motor on it

[@AntaresVII]

I think the point is to do it purely mechanically. Otherwise could just open garage band and call it a day

[@DerrickJolicoeur]

I love seeing new mechanisms like the Huygen Drive

[@geraldillo]

Christiaan Huygens was a genius. He also developed an equal temperament tuning system that contains 31 tones. And he was working on the same problems like Martin; he created a marble run on which two marble with different track lengths end simultaneously.

[@chrisgoode2417]

How tight do the machines of old play? Is it worth the complexities to be computer tight?

[@thomasbecker9676]

Enough. No.

[@pacifico4999]

The original machine got 231 million views. Everyone loved it. I'd say it was good enough

[@JB_ea]

I dont get it, whats wrong with an electirc motor?

[@StuffandThings_]

I love how the channel has almost come full circle - the name Wintergatan of course inspired by astronomy, and now using a device originally designed for astronomy for the centerpiece of the channel!

[@madprophetus]

Flywheel + Transmission connected to the ratchet turns this into a means to provide both a lot of potential energy to keep lifting the weight as needed with the tightness of the weight drive.