official place for engineering feedback, please post and help upvote the best feedback! www.reddit.com/r/MarbleMachine3/comments/137ew6l/cheap_bearing_housing_marble_machine_3_episode_2/?
It will not work, as the tangential power will drag the disks of center. Any imperfections in centering the discs will drag the disks of center when you start to spin them. And they force pulling the discs of center will increase dramatically with rotation speed.
You should really make a automatically centering mechanism that can be tightened with a bolt. Something like in a mechanical camera has to change the hole size for light to pass through. Beefed up.
😂 I’m only halfway through this video, but I recall from a previous video he said that these bolts are already needed for the mechanism, he is just making them do double-duty by also locating the bearing
I'm an industrial draftsman, and the vision of a bearing inside a bolt cage will haunt me tonight !! I'm sure you could use so much more reliable and compact design.
Wish he'd hire a competent engineer, even just part time, on-site. This project would have long been on the road. Martin is masterful in making music, with a senior engineer he'd have the talent which this 1-man team lacks.
@@axiom1650 I would take that job :) , but yeah this is basic engineering design, you could even ask for a student to work for you as a project for his school.
At least use shoulder bolts instead of the entire thing threaded. SMH. His idea will start to chatter in a short period of time. This needs to have the bearing press fit and the outside turned to be perfectly true to that center.
Another Engineer here. This may work for slow lightweight applications in the machine. But for your high RPM high inertia flywheel, DON"T DO IT!!!!!! You may only need properly made bearing housings for the flywheel. Bite the bullet and move on. Save your nifty bearing mount for things that are rotating at 2 RPM and weigh a couple kilos. You don't want to rename your band "Catching Flywheels". Also think about how you are ever going to balance the flywheel, remember trying to do that on the MMX? This bearing capture method is likely to slip and move on you. I have seen a 7000 RPM flywheel break off a 100hp snowmobile engine and dance around a dynamometer room until it came to rest. By the marks in the concrete floor, I am very happy I was in another room. Just glad it didn't catch on a solid part and launch itself through the triple pane window!
Maybe this was covered in another video, but as anecdote about the snowmobile shows, it's not like flywheels and spinning disk shaped objects in general (wheels) are uncommon. Is there some reason he isn't using off-the-shelf components and modifying them for his purposes? It seems an obvious alternative to designing a cheap ineffective solution or ordering an expensive custom solution.
As an engineer I would say there are so many better solutions, if you want to stick to laser cutting, make a third circular plate inbetween the plates in which the bearing would sit. You could also use platforms like hubs to custom machine the parts for you based on step files. It seems to me that a 3D printed bearing holder would also be sufficient for your case if you'd want to go cheap. The best (and probably cheapest) option would be to just use standard flange bearings. They will self align, are easy to mount, and are stock items which are easy to re-order without custom parts.
...Possibly also to design parts using different material thicknesses, that can be combined to replicate parts that would otherwise have to be machined.
Using as many stock parts as possible will be important for a machine that's expected to tour around the world - if anything breaks and the maintenance team runs low on spares, a replacement can be easily ordered without having to rely on **that one guy with a 3-D printer who happens to have the files for this specific piece.**
@@InventorZahran would require it to be touring first - something that's been put-off for a decade now. This marble machine will probably end like the last one, in great success after years of effort just to be shut-down from a tiny glimpse of concern.
You will suffer from two things. First is that the outer diameter tolerance of threated rods is way higher than you expect and not even constant on thread rods of the same batch. Second is, that the outer bearing needs full support on the outside for good lifetime and smooth running.
He is almost certainly using 6205s which have a static rating of 7.8kN and dynamic of 14.8kN. I don't see him get over 1kN even with the flywheel. It should be fine
I think the bigger problem will be that the fine points of the thread will get squashed over time as the shaft vibrates. He could start with a perfect fit, and find it sloppy once it's been used.
@@robinbennett5994 Possible but my gut instinct and back of the envelope calculation is that the clamping force should keep it in place and that it takes quite a lot of force to plasticly deform say a 12.9 bolt.
This bearing mount is a terrible idea. But Martin is going to figure it out himself no ? He will order it, mount it, make a few turns of flywheel and after a few times, IMMENSE VIBRATIONS will totally make him think again about his mount 😅
You forgot about how he's going to tweak the design over a year to try and dial it in before he gives up and does what he should have done in the first place.
The good thing about the standard solution to this problem is that the problem stays solved. It's also quick to solve and cheap to solve. The bolt cage will be an expensive time and attention drainer that will seem to solve your problem several times but it will never stay solved. (Tolerances, threads getting worn/bent.)
So from an engineering background I'd say this *might* work but its likely to be pretty unreliable which was the whole point of making this new machine. Mostly that the tolerance on a bolt is far more than the acceptable tolerance on a bearing meaning you're unlikely to get a consistent fit, some will be far too loose, some will be far too tight. If you do want to do it this way then a shoulder bolt is probably the way to go, still relatively cheap but comes with a precision machined shaft so they'd be more reliable. Personally though creating a simple standardised 3D printed/machined bearing mount with locating features that you can just batch produce and bolt on would be about as simple and *far* more reliable.
what about the tolerance in the holes?! Are they being milled, drilled, drilled and bored to size? Then he's got an 8 bolt pattern so the stackup is going to be murder!
@@huhnesser thats the obvious choice, this is just possible alternatives. Something with low forces on it would probably be fine with a 3D printed housing and a skateboard bearing, maybe it'd even act as a built in failure point if something goes wrong.
I have two comments, as a mechanical engineer. One is caustic, one is helpful. Helpful: You are trying to find a solution to center the part on the bearing. If it is not centered, you will have an eccentric flywheel. This is what I would consider a critical part. You are also trying to use “simple” manufacturing: laser cut flat parts. These parts have good tolerance but it’s not perfect enough for a tight (an interference) fit with a bearing. Everyone here is telling you there is a better way to mount the bearings. Consider finding a bearing that works with the tolerances available to you, rather than this very bad solution of using the bolt threads as an interface. Caustic: You are using a bearing which is 100x more expensive than no bearing. Oh, and it works. Oh, and it’s the only real solution. Now, use the bearing correctly.
Do you think he forgot about the whole "reduce parts" idea? The new bearing has so many parts, and there's a reason we don't see this ANYWHERE else used, so I wonder why he even thinks it might be a good idea :/
@@bl4cksp1d3r Because he doesn't count standard parts in his part total, it would only increase from 1 part to 2. Is that assumption right? Well he doesn't have to manufacture any standard parts, and if one breaks a replacement is easily available, so in that sense it is logical to neglect standard parts. Also, the bolts would have been used anyway to mount a conventional bearing housing, so the part count does not differ as much as you show. As a further note, while no part is the best part, A simpler part is also better than a complex part. in that sense the design Martin came up with is pretty logical. I myself would advocate to go the safe route and just use a proper bearing housing, but I do get Martin's drive to try to do this.
There are commercial bearing housings on the internet... he could just buy some of them and test to see which one he likes the most... they aren't even that expensive
I'm an Aerospace Engineering student. Martin, you have been a big inspiration in choosing my career pathway and I've always admired what you do! This design relies too much on specific tolerances that just aren't available from outer thread radius of bolts. There's a lot of other options for this part, ones that require a lot less R and D too.
Yep! They even sell a single off-the-shelf component that accomplishes the entire goal he is setting out to try and find a unique solution to. All he needs to do is purchase a flange bearing. It's aligned, and has mounting bolts holes to bring it concentric. And it's 1 part, not 40. Also glad to see another Aerospace Engineer! I'm currently in school for it to with a focus in Astronautics!
Balancing the flywheel is SUPER CRITICAL!! Flywheels are death traps if you downt have them really well balanced. A machined beari g housing will locate the bearing with the required accuracy to make the flywheel safe to operate. How does that not make it 100x better?? You can get away with a plate clamped bearing in lots of spots. But not with the flywheel. If you want a bearing that is easier to mount they exist. They are called pillow block and flange bearings. They are already mounted in machined housings that are easy to mount. You the question isn't your plate mounted design vs a fully custom machined bearing housing it is using a ready made mounted bearing. However for the flywheel you want as integrated a solution as you can possibly have. If you don't you could really hurt yourself or other people.
A concentric, true-running flywheel is 100x better than an eccentric flywheel. And designing this system with threaded fasteners to locate the bearing WILL result in an eccentric flywheel. Therefore, a machined bearing housing is at least 100x better than what you're contemplating.
@DarkGrisen your comment makes me think of the Joker explaining when there is a plan, people are happy to go along with it, even if the plan is horrible...
1. the Bearing housing is designed to hold a Bearing evenly at the rotational plane - your Bearing holder with cause pressure points and failure over time 2. your Bearing Housing design will cause shear forces at the plane of rotation because all the holding pressure will be above and below the plane of rotation 3. most Bearing are designed for the outer casing to be under a specific pressure - as in the Bearing is meant to be forced into a specific sized housing dimension there is an old question in engineering that you need to think about "Will i trust my life to this?"
As a becoming mechanical engineer there are a few things that come to my mind regarding this solution: The thread surface of the bolts contacting the bearing is very small which means there is a high load on the very tip of the threats probably resulting in fast wear of the tips. This can quickly make a snug fit into a lose bearing fit. For a bearing that is holding a heavy, fast spinning flywheel it is very important for it to be dead center and well balanced. Hell breaks loose if this thing has a chance to get out of centre. I´m aware that the load of the flywheel doesn´t go through the bolts but rather through the mounting plate of your bearing but this only holds it by friction and in my eyes is not that reliable. As other comments also pointed out there´s also something as error propagation and tolerance stacking which could become a big problem. It is well possible that this solution could work but it begs a big potential to failure. P.S. A few Lasercut metal plates with a big, precise hole for the bearing in the middle and holes for the screws on the outside could be a cheap more reliable solution or just use a pre mounted bearing.
The PS is my first thought. Why not just make the pre-existing hole slightly larger and press fit it? He could even double up on plates to keep it locked in place. This solution is asinine.
@@JZStudiosonline Hard to get that precision from laser cutting. Really need .0005" accuracy or better for press fits, which is very close to the limits of that method. Might be hit or miss, which isn't good for 'production' as he intends.
Martin literally reinventing the wheel. 😄 No doubt everything I am tempted to say has been well covered by others, but honestly I do wish you well Martin.
Still really hoping that the extra complexity of different marble sizes will be at least postponed. I really want to see this machine finished and scope creep scares me!
He didn't even mention the issues basically everyone sees and were telling him with the scope creep in the last video. I don't think he cares to build a functioning machine. As long as he has a project to build people will pay for it on Patreon or do free work for him. So doing something overly complicated without getting his results just gives him more time to work on the project and get more money.
Yeah you are thinking in the old paradigm. He is thinking modular. If the snare module is too complex and gets scaled back it has no impact on the other modules of the machine.
That whole video I was shaking my head. He's *really* letting the scope creep start building up again. He wants an instrument that can do everything instead of embracing the limitations and making it be really good at doing what it can do.
Martin!!! First, a part doesn't need to be 100x better to justify 100x more cost. If the lesser part fails to meet certain minimums, you go with the better part. Remember how you kept putting in more and more effort to fix cheap solutions on your past two projects and they just got so bogged in being doomed to fail from the start that you ditched the whole machine? Don't let that happen here! Second, on dynamic sound, since I missed last week: running multiple sizes of marble through the machine screams Bad Idea to me. Have you tried simply dropping from different heights? That should affect the sound similarly, and you're already trying something like that with the roll.
Martin, I am so sad to see this after all the great progress you made as a designer. You don't have to use custom housings I am sure there is a mass produced housing that works for you. There is literally no good reason to reinvent the bearing housing, the outer ring of the bearing is not strong enough to support the load.
I'm a software engineer and I've learned to be very skeptical of complete "rewrites". First it was "I believe", now it's "I verify". I have a feeling it will be "I iterate" next.
No kidding, this is "second system syndrome" to the next degree. He should have made a second music video with the MMX, called it done and moved on to new projects.
LOL! But the experience is fairly normal for designing something completely new....it's just that most designers don't declare victory until they have a working solution!
Them storing your password in plain text and then emailing you your password sounds like an absolute security nightmare and is just asking for a massive breach with huge consequences lol
I didn't think anyone would even consider doing that since the 2000s, least of all a company that must have developed the software that he was using, and so should know better.
hasn't LTT made a 16k monster like a few years ago already? (I'm guessing he either has a very old PC, or he uses crap screen recording software (USE OBS!!!), or his monitor/PC/Capture card is somehow not all compatible, OR, he just didn't set it in the settings)
Hey Wintergatan. I don't expect you'll read this because you've got so many followers and comments, but I want to get something off my chest. A long time ago I've watched your music video with the very first marble machine. I remember how I was very impressed by it. I enjoyed how an impossible machine creates such beautiful music. I know it couldn't have actually worked, and you must've put a ton of editing in to make it look like it did, but you could tell you put a lot of thought in the machine. It wasn't all smoke and mirrors and worked for at least 60%. Just enough to make people believe that against all odds, this flimsy handmade machine, with marbles spilling all out over the floor, actually worked. A few days ago I re-discovered your channel. I found out you made this machine not just for a video clip, but how you were serious about making it actually work. That you spend years trying to create a second marble machine that you could take with you on tour and use to play live music. I was even more impressed than before. It wasn't just a one time gimmick that only worked with editing tricks, but something that will eventually work in real life. I liked the design of the second machine even better than the first one. Even though it didn't look as boutique as the first one, it was wild yet elegant, with the organic materials like wood and different coloured metals. I wasn't a huge fan of the transparant plastic tubes, I don't like plastic, but it didn't look too ugly and I liked how you can see the marbles rolling through it. At the same time I felt you were chasing an impossible goal. Not only because the machine probably would never work well enough to perform live, but mostly because the better it became, the further it was removed from what I liked about the first machine. The first machine was an underdog. It looked like it could never work, but you were rooting for it because of that. The better the second machine performed, the more solid it looked and the less impressive it became. I felt sorry to learn that despite your years of work, you didn't get to finish the second machine. I hope the people at the museum where it is now will finish it. The third machine on the other hand... I'm not convinced you're heading in the right direction with it. You've mentioned putting "function" over "form", but are you sure you know what the "function" is? You seem to think the main function is to play music, but I think you're wrong. You mention Elon Musk and his five points. How with the first 2 points are to question if a machine should exist at all. If the function would be to play music the answer would be no. There are more efficient ways to play music than a machine that uses marbles. A more effective way is to put a cd player, with and amp and speakers on a stage and just turn it on in front of an audience. It's less likely to break down as an instrument or machine and less likely to make mistakes like musicians do. However, that wouldn't work at all. No one wants to see that. The audience wants to be amazed by live musicians, with real instruments or instead maybe some outlandish machine that plays music with marbles. I believe your machine could succeed, but be conscious of the fact the "function" is to amaze people, not simply play music. I see a lot of this fallacy in your new designs. Where you are influenced by people who design gym equipment that is solid but ugly looking, and how you order mass produced parts from third parties that are practical and cheap but very generic. I understand your desire to finally create something that actually works and it's too early for me to tell what I will think about the final design, but please; Don't put "function" over "form". Be aware that the "form" IS the entire "function" of this machine. Once you've got the design down so it's efficient and works well, make sure to execute it in a way that's aesthetically pleasing. Use beautiful shapes, good looking materials, mask ugly parts, maybe even add things that aren't functional but just for show, etc. I really hope you'll succeed this time around. That you'll manage to create something that works, that you can perform live with and that will amaze people the same way the first marble machine did. Good luck.
There's just so many things wrong with this concept it's hard to know where to start with ripping it apart - there's reasons why bearing housings are the way they are!
I remember the discussion about the flywheel being incredibly dangerous So I would like the discussion about anything around the fly wheel to not be about expense, but rather to be about making it as secure as possible, so the audience don’t have their heads chopped off So I hope the “cheaper” way works as well as what all the engineers on the forums are saying you need to do
Tolerance stackups are a thing: How tight are the tolerances on those normal all-threaded bolts (shoulder bolts will be more precise)? How big are you making the holes for the bolts? How tight are the manufacturer's tolerances part-to-part? CAD is the only thing that is infinitely precise, and that sucks when you're chasing accuracy.
I don't think it is a big deal. Those bolts aren't a wear surface. Worse case he has a precision jig to locate the bearing and if some bolt is way out of whack you just toss it.
@@Ammoniummetavanadate gravity, imbalances, and vibrations will all cause the outer bearing race to hammer against the threads of the bolts. The points of the threads will flatten over time and the bearing will get loose. Using only 8 bolts, if the bolts are too tight at the start it will force the bearing to deform. This will add drag, noise and vibration to the machine. He really needs to go ahead and make a proper bearing housing.
@@billtheunjust What load would you expect each of these bearings to see? Why are you assuming this will be running imbalanced? How much force is needed to plasticly deform the the threads of a 12.9 M6 Machine screw? Now do a little back of the envelope math and you will see why this is worth trying.
One thought, since you are concerned about costs. Why are you prototyping in stainless steel and not in mild steel? It's a much cheaper material and you will only see differences in performance over the long term. On a personal note, I suspect that you will have longevity problems with your design since you are entirely reliant on the clamping force of the two plates to keep the bearing from moving. If the axial load is enough to overcome the clamping force, then you are relying entirely on the ability of the thread points to take up the force. Threads can be deformed just by pressing on them with a hammer using body weight alone. Where this is really worrying is that it may take time for the problem to develop and you will have committed to a design that will, ultimately, fail completely but will appear to be viable initially. To give this the beat chance of success, I would suggest not bothering with trying to use the bolts to act as a cage, but lean into the clamping force of the plates. Make the plates heavier and use stronger bolts. (M8 10.9) and maximise that clamp force. (Use a torque wrench) this is me speaking as a heavy duty mechanic that has seen what happens when engineers try to cut corners on bearings and their mountings. I want you to succeed and weaknesses in design at this foundational level are multiplicative. Honestly, why are you messing around with this and not just getting your flywheel machined to accept a bearing at its centre? Get it machined with an interference fit and call it a day. If your flywheel is made of wood, give it another .04mm of crush and glue it in place. You're trying to make a very simple thing very complicated.
Yes. He has some strange design assumptions that seem to come from nowhere. I thought I was missing some very important fundamentals watching this. Until I read other comments that convinced me I'm not delusional!
I really just hope that Martin doesn't fall into the same pitfalls he encountered in the last machine. The first one was a proof on concept, the second was the trial-and-error phase, and now that he has ample information on what works and what doesn't, I hope to see MM3 becoming reality.
If he had wheels he'd be a wagon. He is absolutely not the right person to make the machine - he's an artist and musician, a perfectionist without the experience to design the machine quickly and reliably. He's already admitted that even if the machine worked, it would sound lifeless. Wish he'd go back to making mostly human music with a music box in the band.
The whole point of a flange bearing: mount a shaft or axle to a not perfect housing or stand and let the flange pivot and shift to correct it. The mounting holes have big tolerances and they are supposed to. But Martin is mounting the flywheel to it. So, it has tho be centered very well with low axial run out. Doing this with flange bearings is trouble waiting to happen...
There are detailed 3d models for bearings, for other stuff, a special soft for "todo list" and etc. This guy has definitely a lot of free time... for developing anything, except the marble machine itself.
I can really empathize with the emotional response to barriers when you're trying new things. I try to remind myself that those barriers are effectively distilled instances of learning. They're painful because learning/growing can sometimes be painful. It's the mental equivalent of sore muscles. So instead of being hard on yourself for it, be proud of yourself for pushing through.
i don't follow the logic that something has to be 100x better to value 100x the cost. If something is 10% better, but you need those 10% for it to work, it's worth 100x
The irony of a company called 247 Steel only working 9-to-5 on weekdays 😂 Also, the company absolutely should not be sending you passwords via email. They should have sent you a web link to reset it and choose your own. Also also, the bearing cage is a bad idea. Going with a normal bearing housing is going to save you a lot of time in debugging and maintenance, and time is precious.
Any smaller (non-tech) company will send passwords by email and have terrible overall security hygiene. It's fine. Being security conscious is a luxury many smaller shops can't afford. Shaming them for assisting their customer in the most straightforward way possible is counter-productive.
@@AlexandreMacabies Security is every company's responsibility, no matter the size. There are strict rules around the handling of personal data, covered by GDPR. There are no excuses.
@@daverayment I agree. Before GDPR small companies could just wing it and hope for the best. When GDPR was introduced, it set new rules where you can be fined up to 10 million EUR for willful ignorance of protecting customer data, no matter how small your business is (and for business making more than 500 million a year, the fine can be up to 2% of their entire global turnover so that e.g. Google and Microsoft cannot just take the 10 million Euro fine and keep breaking the law).
Martin, you had me worried for a long time that you were giving up on this. I'm so happy you are going for the MM3. Yet, this design scares me. I am just about to graduate with a Mechanical Engineering degree. I have a lot to learn, but I have seen quite a few bearings fail. If you cannot get your design down to absolutely perfect tolerances then your flywheel may have a lot of vibration involved. The bearing may be able to handle the vibration for a long while, but the rest of the machine may be impacted by the vibrations. I also worry about wear on the machine itself. If the machine is assembled too many times, the extremely tight tolerances on the bolt holes will become sloppy and lead to a loose bearing fit. The way I see it, this design may work with some very fine tuning at first, but it will only degrade over time and lead to problems later. I know you mention that the cost of a fully machined housing is very high, but it may be your only option here if your bolt design does not meet your expectations. I do think you will figure it out better than I can. After all, I am not working on the machine like you are. I only wanted to warn you of possible problems in your future.
*@Wintergatan* 2:22 This is a VERY BAD idea, because real life bolts are NOT perfect, there are small variations even between bolts that "should" be identical, they will not be identical. The best you can do (I think) is to machine a press-fit "sleeve" for the bearing with (threaded or smooth?) holes in for the bolts, make the sleeve LESS wide than the thickness of the bearing, if you want the clamping force from the side (but I think that is a bad idea too, material flex, part-slipping & such). Also: *High tolerances = EXPENSIVE* (exponentially) 7:17 .
Martin I wrote you in 2016 asking you when I would have the privilege to listen to you performing in Italy... well, I was almost 13, now I'm almost 20... this project is awesome ❤
when so many people are telling you how bad an idea it is to reinvent the wheel here, and you decide to dedicate so much time to it anyway, its hard for me to not think you are just milking content at this point.
Unfortunately I have less hope for mm3 than I had for mmx. Extra features ike different size marbles and also not listening to engineers when they say using bolts to hold a bearing is a bad idea. When you have this large and educated community, it is worth listening to them.
Yep. I feel Martin is suffering from some sort of expression of the Dunning-Kruger effect, and is now in a very dangerous zone. You don't screw around with dubious, custom-built parts for critical applications, period. You just don't skimp on security or on part quality, especially if there's an audience involved, and no amount of reinforced steel cages can replace proper care in designing systems which won't predictably fail in catastrophic fashion. OSHA would have a field day with this machine, yikes! Here's the thing that I don't get; I'm just a graphic designer, and even I can sense there's something clearly wrong and dangerous with this part in particular. Must've been all those videos I've seen of ball bearings exploding, or something…
Martin, That intense emotional reaction to adverse events has a name. I hope that being able to give it a name and understand it better will make it a little easier to manage. This intense reaction can take on many forms from frustration to anxiety to depression and hopelessness. It is referred to as Rejection Sensitive Dysphoria (RSD) and there are myriad methods for managing it. Most of them are rooted in behavioral therapy and mirror mindfulness exercises. My preferred poison for this behavioral bug is dialectical behavioral therapy or DBT. Thank you so much for sharing your generous, patient and persistent spirit with the world through your Content and your indomitable creative drive.
@@andymouseits always better to know you have issues so you understand why you feel a certain way than it is to think you're broken or something. I know from experience.
I immediately felt what martin experienced. My desk (and a whole workshop) is filled with projects that have stalled because one small thing went wrong and I felt it was such a catastrophe I could not continue. I have a lotus speedo on my desk I was rewiring and the wire ferrules I fitted are too big to fit inside the connectors on the control board I designed and had made. Its a tiny fix in reality and yet the thing has sat on my desk for 6 months, staring me down and reminding me what an abject failure I am. I discard any notion of the hard won achievements up to this point and only see the failure. Then one random day I will just decide to cut off the ferrules and jam the wire into the connectors and carry on. Sometimes years later. Rejection Sensitive Dysphoria is a thing and I didn't know I had it until recently. For me, like many other its rooted in the breakup of my parents when I was a kid. I think a lot of ADHD typical behaviours are basically coping mechanisms for RSD.
One of the main issues I can see with the threaded bolts holding the bearing in place is that as the bearing is rotating and loads are imparted radially through the bearing and onto the bolts, the threads of the bolts may start deforming (depending on what loads you are looking at) or potentially wearing away material from the threads or the outside diameter of the bearing casing. In either case, the more the threads bend or if the threads are cutting into the bearing after many cycles, your tolerances will be getting looser and your outside bearing housing will be more susceptible to spinning freely and causing more wear as it rubs up against the bolt threads. The main proponent of having machined bearing liners or housings is to keep uniform pressure via a pressure fit all around the bearing to prevent the bearing from shifting around and causing wear and to keep the bearing precisely where you want it.
This video encapsulated soo many of the issues and thoughts I have in day-to-day work. I hope you keep up this project! Show us the struggles and victories ❤ An expert, someone masterful in bearing housing, might be able to quote the ‘perfect answer’, but the act of designing while learning can result in truly creative results! I wish you inspiration and good luck 👍
Martin. Don't go insane in design hell. Get this part of the project just done. You derp. (I say this with love) And start building stuff. A designer is always going to run into design conflicts. Essentially what you're training yourself to do is how to take these digital drawing's form CAD and turn them into physical objects. That can be incredibly difficult to do especially if you're just doing it for the first time. With a fresh start. You got this. My advice is hold yourself more accountable in the design process and get something done and finalized and start building. I know the amount of work you have in your plan can be daunting at times but like you shouldn't chew yourself up because you failed to build the machine. It's not really about the machine or the tour. We are here to watch the process of building machine. GL
In this case, perfection has already conquered and destroyed all that good, completely. I think this reign of perfection has been going on the 3rd year now.
Martin. If you are worried about the PM workload for your bearing part, maybe you should consider closing the scope creep hole instead. I can hardly believe that ordering the correct bearing mount would be that hard.
though we should focus on making a machine that works, i really hope he doesn't scrap the previous designs visually. The design is what started it all and grew this great community around it. That design is kind of the soul of this channel and is what we hope will play the same music in that very first video. If he really goes with the "concept art", it just wont feel the same, even if it works. It'd kinda be like another singer singing your song, and not you, the one who originally sang it.
at a MINIMUM you should be using bolts with a shank for this or the bolts WILL break eventually. With a non-rotating shaft like you have ideally the bearings would be pressed into the flywheel. Though I'm not really sure why you didn't just go for a rotating shaft so you could use pillow blocks and just not worry about this particular issue.
I would question whether you're calculating all the costs into the non-machined bearing housing. Is the pre-fabbed part 100x the material cost only, or 100x the total cost, including the opportunity costs of your design time, prototyping, evaluating, iteration, etc?
@Ammoniummetavanadate I cannot accept that there is no off the shelf bearing housing that he could not adapt to fit his design or adapt his design to fit, that would be less expensive than ordering a bunch of prototypes, trying them, spending hours fiddling with the idea in CAD, ordering more prototypes, realizing it isn't going to work, and then ordering proper housings.
@@RhynoD2 There absolutely are, they are also like $100 USD each from McMaster, Maybe down to $50 each for sketchy ones. This will be like $10 each and built directly into the part vs needing mounting bolts and then locating them.
Martin continues to make the same mistakes one mm3 that he made on mm1 and mm2: Not listening to the experience of people who know what they're talking about.
I have used "Split Shaft Collars" as bearing housings. The collars have a tightening screw, I attach the collar to the frame by drilling and bolting through the collars flat sides
Mechanical engineer here. This solution will create more problem than it solves. You're dealing with small loading here, especially compared to what those bearing are capable of. Just 3d print a housing. 90x cheaper than machine, 90x better than your bolt arrangement. Plastic bearing housing are used everywhere. Just look at powertools : drills and angle grinder motors (not gearboxes) use plastic housing for their bearing. Don't try to reinvent the wheel, i don't want to hear you announce Marble Machine 4
The flywheel is basically the heart of your machine, I know you know it but, these things store an enormous amount of kinetic energy. If you test your idea, test it like crazy, because I'm pretty sure it's the kind of idea that will look to work great at first and eventually fail on you at the worst moment. You said 100x better, I say what would be 100x worse: break 20 or more parts on stage and having to cancel 5 shows and lose whatever revenue that means.
NGL, my first reaction to the idea of making a bolted bearing cage was WTF. But on further reflection, I don't think this application will be under high enough loads to suffer from having a poorly supported bearing. I do, however, not like the idea of threads being used as load bearing members. If I had to do something like this for my a client design, I would either A) use a stack of cut plates so that the bearing still has full support. or B) use shoulder screws instead of fully threaded screws, so that there is at least a nice precise smooth surface supporting the bearing, not the sharp edges of the screw thread. Also, screw threads are not toleranced to the necessary level of concentiricity, unlike shoulder screws which are designed to be used for load bearing and alignment.
A lot of people have commented on how this solution probably won't work. I am also a mechanical engineering student, and I think you could get the tolerances correct by using precision ground shoulder bolts instead of normal M6 bolts. M6 bolts have allowable major diameters between 5.974 (upper) mm and 5.794 mm (lower). That's an allowable variation of 0.18mm. For holding a bearing in place accurately, you probably want at least a locational clearance fit, if not a press fit. The maximum variation of size for a 52mm diameter hole (as would fit the bearing in a locational clearance fit) is +0.03 mm. You might be able to get close to this by using precision ground shoulder bolts and reaming or broaching the holes to precisely the diameter of the shoulder bolts, but the tolerance stackup on that means that it would probably be easier and cheaper to just have a bearing holder machined or use off-the-shelf insert bearings.
It is good that you experiment with this idea and will put it to test, so there is no more - what ifs, but I also agree to I guess many viewers that it won't be reliable, it will be nightmare to center, bolt threads are weak and will probably wear out because there will be movement from all the loads on flywheel and axle. Also if you have watched the episode of Sandy Munro saying, that they did a reasearch at Ford on what was the cause of most of their manufacturing and realiabilty problems and the answer was bolts. So again I think it's good that you test this idea, but it won't be easiest or best solution.
3:59 having a radial force on a Bolt is the worst type of stress you can put on a screw. The only way that this method could work is by having clamping forces high enough that the friction between the plate and the bearing holds the flywheel.
Hello Martin. I do structural design in Canada. All structural (not mechanical mind), bolt holes are oversized by 1/16", or about 1.5mm. To keep things simplified, torqueing the bolt eliminates the ability for the structure to slide around in the hole. My point is don't get hung up on making extremely tight holes for your bolts. Your loads are light for the number of fasteners you have, just use washers and torque properly and you can get away with larger holes.
This design with the bolts is tremendously cursed. Specially if you plan with store bought ones. Save yourself the trouble, search for existing solutions such as "insert bearings", or "bearing houses". PS: Also an industrial draftsman/mech eng here. This is the second most cursed thing i've seen today.
rule #1 - the comment section is always right. rule #2 - first learn rule #1. there's a reason people are telling you to go with something that already works, but if you *MUST* do things the hard way for engineering/comparison sakes...
As a rocket engineer working on bearings I have to say, 100x more expensive is the best deal of your life. Although I think there are other options designed around a flat plate with holes you put between the two plates. Very hopeful you can make it work!
Already pointed this out in the previous video, will repeat it again here against the general criticism. I love that you're experimenting and trying stuff out. That's how you learn, but also how we learn. Please keep your creative mindset. (I'm also an engineer, as this seems to be a necessary disclaimer to comment here nowadays. :D)
As a retired jet machanic I've seen bolts with a shank long enough to go through both clamping plates, if the bolt holes are too tight just shave a flat spot on the bolt shank. I'm with the engineers on this design. A custom bearing housing is the correct means to support the rotating assembly.
*FAILED???* *_No way did you fail!_* You created some *_amazing_* musical instruments that are just as beautiful to the *eyes* as they are to the *ears!* Each new iteration is a new expression of your artistry, and we, your fans, enjoy being able to share in your joys and struggles! Did the inventor of the guitar fail simply because people continue to modify and improve on it, _to this day?_ No! The fact that people continue to tinker with ideas for new expressions of it is validation for just how successful it was! I applaud your humility as it only serves to further express your success! God bless you! _"Let another man praise thee, and not thine own mouth; A stranger, and not thine own lips."_ *--PROVERBS 27:2*
Your reaction to minor setbacks like the place being closed when you called is absolutely something I've come to realise I have! The problem has gotten noticeable because I started making basic rules for myself to avoid those setbacks that started becoming "basically don't start something that could possibly result in that" meaning I started doing nothing much at all
As is said in the previous video of the bearing housing, as stubborn as you are, at least try to use partially threaded bolt. The bearing will be guided by a real cylinder, a bit more precise than threads.
He cancelled the last version because it was too complex...and so his solution was to instantly make this one stupidly overly complicated surpassing the last one before he's even begun construction this time... It's baffling
a few seconds into this video im am thinking: how can a so clever person like martin come up with such stupid idea? this must be the supidest idea of all he had so far.
Let's spend 14 hours engineering readily available, mass produced inexpensive part that's critical to the operation of the machine, have it fail completely, and do it all in the name of saving money. He's not an engineer. He's an artist with a romanticized vision of what engineering is stuck in his mind, finding new and innovative solutions to things that aren't even problems.
Well his hero is Elon Musk, who made the world's largest rocket. Said rocket was, at best, expected to clear the launch pad, experienced failures in 1 out of 6 engines, and then exploded.
I'm like were your heading with this New Marble Machine , giving it a different set up may to the Trick in making it work , cant wait to see what you got coming next 👍👍👍👍👍😎 . SKK .
You keep talking about the machine failures, how awful past designs were, and maybe from a perspective of engineering I could see why, but... The 4 minute 32 second one that I first learned about you from, the one titled "Wintergatan - Marble Machine (music instrument using 2000 marbles)", with what is apparently the first marble machine, is still the best video of yours I've seen. I keep watching your stuff anticipating that you'll make another song from one big system machine again. I'm just waiting for the next great song that matches that first track I ever heard. I still rewatch that 7-year-old video (v=IvUU8joBb1Q) now & then, and it's amazing. Still. All your engineering is interesting to my nerdy side, but I keep subscribed *just in case* you upload an actual new song. That's what I'm waiting for now. Another Marble Machine song.
official place for engineering feedback, please post and help upvote the best feedback!
www.reddit.com/r/MarbleMachine3/comments/137ew6l/cheap_bearing_housing_marble_machine_3_episode_2/?
It will not work, as the tangential power will drag the disks of center. Any imperfections in centering the discs will drag the disks of center when you start to spin them. And they force pulling the discs of center will increase dramatically with rotation speed.
A PC has more resolution than 1080p.
Its not only Mac that go 4K. A PC can also go to 4K. 😁
You should really make a automatically centering mechanism that can be tightened with a bolt. Something like in a mechanical camera has to change the hole size for light to pass through. Beefed up.
@@metern what, pc's only go 4k?
@@dbtest117 Im not sure on which one has only 4K. But i do think most of the newes and most powerful pc gan go 4K. Maby you have to overclock them.
Martin: "I've reduced the parts count in this new design"
Also Martin: "I wanna hold this bearing with 40 parts"
😂 I’m only halfway through this video, but I recall from a previous video he said that these bolts are already needed for the mechanism, he is just making them do double-duty by also locating the bearing
Also Also Martin: "No Compromises"
Also Also Also Martin: "*waves hands* ItS OnE hUnDrEd TiMeS bEtTeR"
Exactly what I was thinking hahaha
I think maybe he needs to open a window to let some fresh air in😵💫 hahaha
i say forget the machine thing, call the rest of the band a make a new album 😢
I'm an industrial draftsman, and the vision of a bearing inside a bolt cage will haunt me tonight !! I'm sure you could use so much more reliable and compact design.
Wish he'd hire a competent engineer, even just part time, on-site. This project would have long been on the road. Martin is masterful in making music, with a senior engineer he'd have the talent which this 1-man team lacks.
@@axiom1650 I would take that job :) , but yeah this is basic engineering design, you could even ask for a student to work for you as a project for his school.
It's perfect comment bait for the engagement algorithm. Obviously the threads will compress and run out of tolerance, but live and learn.
I think that you can buy bearing houses so no custom parts needed
At least use shoulder bolts instead of the entire thing threaded. SMH. His idea will start to chatter in a short period of time. This needs to have the bearing press fit and the outside turned to be perfectly true to that center.
Another Engineer here. This may work for slow lightweight applications in the machine. But for your high RPM high inertia flywheel, DON"T DO IT!!!!!! You may only need properly made bearing housings for the flywheel. Bite the bullet and move on. Save your nifty bearing mount for things that are rotating at 2 RPM and weigh a couple kilos. You don't want to rename your band "Catching Flywheels".
Also think about how you are ever going to balance the flywheel, remember trying to do that on the MMX? This bearing capture method is likely to slip and move on you.
I have seen a 7000 RPM flywheel break off a 100hp snowmobile engine and dance around a dynamometer room until it came to rest. By the marks in the concrete floor, I am very happy I was in another room. Just glad it didn't catch on a solid part and launch itself through the triple pane window!
Hahahahahahahaha catching flywheels. Youbgot me good there XD
Maybe this was covered in another video, but as anecdote about the snowmobile shows, it's not like flywheels and spinning disk shaped objects in general (wheels) are uncommon. Is there some reason he isn't using off-the-shelf components and modifying them for his purposes? It seems an obvious alternative to designing a cheap ineffective solution or ordering an expensive custom solution.
You need to use a type of bearing called “Insert Bearing” which has simple stamped plate flanges that you can bolt to either side of your flywheel.
This man is talking with the real talk here.
I was thinking: ok there must be pre-made parts for this purpose, I hope Martin reads your post.
Most all those types of flange mount bearings are self aligning. And i don't think his construction would be compatible with that.
Still say he should just get the flywheel machined to accept a press fit bearing and call it a day.
Are these also called flange mount bearings?
As an engineer I would say there are so many better solutions, if you want to stick to laser cutting, make a third circular plate inbetween the plates in which the bearing would sit.
You could also use platforms like hubs to custom machine the parts for you based on step files.
It seems to me that a 3D printed bearing holder would also be sufficient for your case if you'd want to go cheap.
The best (and probably cheapest) option would be to just use standard flange bearings. They will self align, are easy to mount, and are stock items which are easy to re-order without custom parts.
...Possibly also to design parts using different material thicknesses, that can be combined to replicate parts that would otherwise have to be machined.
Using as many stock parts as possible will be important for a machine that's expected to tour around the world - if anything breaks and the maintenance team runs low on spares, a replacement can be easily ordered without having to rely on **that one guy with a 3-D printer who happens to have the files for this specific piece.**
@@InventorZahran would require it to be touring first - something that's been put-off for a decade now. This marble machine will probably end like the last one, in great success after years of effort just to be shut-down from a tiny glimpse of concern.
A company called "24/7" that is only open Monday through Friday during normal business hours is extremely funny.
It was worth it for hearing him say "Dankjewel", though haha
I have a 24 hour pizza open 8/11 pm.
It's not even open later than other local pizza places
Im guessing you can order parts 24/7, customer service is just not open 24/7
@@sjege Correct, And the automated machines run 24/7.
Martin builds a marble machine except each time he moves on to the next one he reaches insanity in half the time.
I would recommend listening to the engineers in the comments. As you said, the word of a battle worn engineer is a thing of beauty
You will suffer from two things. First is that the outer diameter tolerance of threated rods is way higher than you expect and not even constant on thread rods of the same batch. Second is, that the outer bearing needs full support on the outside for good lifetime and smooth running.
He is almost certainly using 6205s which have a static rating of 7.8kN and dynamic of 14.8kN.
I don't see him get over 1kN even with the flywheel.
It should be fine
I think the bigger problem will be that the fine points of the thread will get squashed over time as the shaft vibrates. He could start with a perfect fit, and find it sloppy once it's been used.
@@robinbennett5994 Possible but my gut instinct and back of the envelope calculation is that the clamping force should keep it in place and that it takes quite a lot of force to plasticly deform say a 12.9 bolt.
The outer diameter of threaded rod isn't even consistent along the length of the same rod.
If Martin is hell bent on this idea he needs to use hex head shoulder bolts so only the end is threaded.
This bearing mount is a terrible idea. But Martin is going to figure it out himself no ? He will order it, mount it, make a few turns of flywheel and after a few times, IMMENSE VIBRATIONS will totally make him think again about his mount 😅
You forgot about how he's going to tweak the design over a year to try and dial it in before he gives up and does what he should have done in the first place.
yeah and anothey total fail with anothey stupid ligma balls machines. He should stop ruining his time with this one stupid idea
We are here for the journey, not the destination.
@@To1ne The destination is death, so yeah: better enjoy the journey.
I hope he also learns the importance of axial pre-load. I can't see any of that in this current design
The good thing about the standard solution to this problem is that the problem stays solved. It's also quick to solve and cheap to solve. The bolt cage will be an expensive time and attention drainer that will seem to solve your problem several times but it will never stay solved.
(Tolerances, threads getting worn/bent.)
Hot Take: you should go to the scrapyard and get a clutch off a motorbike.
So from an engineering background I'd say this *might* work but its likely to be pretty unreliable which was the whole point of making this new machine.
Mostly that the tolerance on a bolt is far more than the acceptable tolerance on a bearing meaning you're unlikely to get a consistent fit, some will be far too loose, some will be far too tight.
If you do want to do it this way then a shoulder bolt is probably the way to go, still relatively cheap but comes with a precision machined shaft so they'd be more reliable.
Personally though creating a simple standardised 3D printed/machined bearing mount with locating features that you can just batch produce and bolt on would be about as simple and *far* more reliable.
what about the tolerance in the holes?! Are they being milled, drilled, drilled and bored to size? Then he's got an 8 bolt pattern so the stackup is going to be murder!
@@nickgoodall578 yeah, that too, I just assumed it was going to be laser cut sheet so not great but consistent tolerances that could be fudged a bit.
Flangebearings are like 20$, not that expensive
And they surely will work with perfect tolerances
@@huhnesser thats the obvious choice, this is just possible alternatives.
Something with low forces on it would probably be fine with a 3D printed housing and a skateboard bearing, maybe it'd even act as a built in failure point if something goes wrong.
I have two comments, as a mechanical engineer. One is caustic, one is helpful.
Helpful:
You are trying to find a solution to center the part on the bearing. If it is not centered, you will have an eccentric flywheel. This is what I would consider a critical part.
You are also trying to use “simple” manufacturing: laser cut flat parts. These parts have good tolerance but it’s not perfect enough for a tight (an interference) fit with a bearing.
Everyone here is telling you there is a better way to mount the bearings.
Consider finding a bearing that works with the tolerances available to you, rather than this very bad solution of using the bolt threads as an interface.
Caustic:
You are using a bearing which is 100x more expensive than no bearing. Oh, and it works. Oh, and it’s the only real solution.
Now, use the bearing correctly.
Do you think he forgot about the whole "reduce parts" idea?
The new bearing has so many parts, and there's a reason we don't see this ANYWHERE else used, so I wonder why he even thinks it might be a good idea :/
Do you think there's a way to make a good meal in decent time with 2.5 million chefs?
@@bl4cksp1d3r Because he doesn't count standard parts in his part total, it would only increase from 1 part to 2. Is that assumption right? Well he doesn't have to manufacture any standard parts, and if one breaks a replacement is easily available, so in that sense it is logical to neglect standard parts.
Also, the bolts would have been used anyway to mount a conventional bearing housing, so the part count does not differ as much as you show.
As a further note, while no part is the best part, A simpler part is also better than a complex part. in that sense the design Martin came up with is pretty logical.
I myself would advocate to go the safe route and just use a proper bearing housing, but I do get Martin's drive to try to do this.
There are commercial bearing housings on the internet... he could just buy some of them and test to see which one he likes the most... they aren't even that expensive
He’s literally “reinventing the wheel” here?
Standard Bearing housing are already available, buy one as per your requirement.
I agree with your sentiment, but "as per" is redundant.
I'm an Aerospace Engineering student. Martin, you have been a big inspiration in choosing my career pathway and I've always admired what you do! This design relies too much on specific tolerances that just aren't available from outer thread radius of bolts. There's a lot of other options for this part, ones that require a lot less R and D too.
Yep! They even sell a single off-the-shelf component that accomplishes the entire goal he is setting out to try and find a unique solution to. All he needs to do is purchase a flange bearing. It's aligned, and has mounting bolts holes to bring it concentric. And it's 1 part, not 40.
Also glad to see another Aerospace Engineer! I'm currently in school for it to with a focus in Astronautics!
Balancing the flywheel is SUPER CRITICAL!! Flywheels are death traps if you downt have them really well balanced. A machined beari g housing will locate the bearing with the required accuracy to make the flywheel safe to operate. How does that not make it 100x better?? You can get away with a plate clamped bearing in lots of spots. But not with the flywheel. If you want a bearing that is easier to mount they exist. They are called pillow block and flange bearings. They are already mounted in machined housings that are easy to mount. You the question isn't your plate mounted design vs a fully custom machined bearing housing it is using a ready made mounted bearing. However for the flywheel you want as integrated a solution as you can possibly have. If you don't you could really hurt yourself or other people.
When the flywheel breaks free and crushes the skull of someone in the front row he will rethink the cost savings when the lawsuit is filed.
There will never be somebody seated in a front row. Not with the scope creep seen from the beginning of the project.
"I've been wrong before." you currently have a track record of 2 machines, get ready to add bearing houses to that list
A concentric, true-running flywheel is 100x better than an eccentric flywheel. And designing this system with threaded fasteners to locate the bearing WILL result in an eccentric flywheel. Therefore, a machined bearing housing is at least 100x better than what you're contemplating.
The fact that so many viewers did not approove your design may actually be a hint to revise it
@DarkGrisen explaining it to us doesn't stop it being bad
@DarkGrisen your comment makes me think of the Joker explaining when there is a plan, people are happy to go along with it, even if the plan is horrible...
@DarkGrisen he's still doing it though lol
i just wanted a new album 😢
He's testing it, like a good engineer... This isn't the final product smh
1. the Bearing housing is designed to hold a Bearing evenly at the rotational plane - your Bearing holder with cause pressure points and failure over time
2. your Bearing Housing design will cause shear forces at the plane of rotation because all the holding pressure will be above and below the plane of rotation
3. most Bearing are designed for the outer casing to be under a specific pressure - as in the Bearing is meant to be forced into a specific sized housing dimension
there is an old question in engineering that you need to think about
"Will i trust my life to this?"
As a becoming mechanical engineer there are a few things that come to my mind regarding this solution:
The thread surface of the bolts contacting the bearing is very small which means there is a high load on the very tip of the threats probably resulting in fast wear of the tips. This can quickly make a snug fit into a lose bearing fit.
For a bearing that is holding a heavy, fast spinning flywheel it is very important for it to be dead center and well balanced.
Hell breaks loose if this thing has a chance to get out of centre.
I´m aware that the load of the flywheel doesn´t go through the bolts but rather through the mounting plate of your bearing but this only holds it by friction and in my eyes is not that reliable.
As other comments also pointed out there´s also something as error propagation and tolerance stacking which could become a big problem.
It is well possible that this solution could work but it begs a big potential to failure.
P.S. A few Lasercut metal plates with a big, precise hole for the bearing in the middle and holes for the screws on the outside could be a cheap more reliable solution or just use a pre mounted bearing.
Hex head shoulder bolts! At least the contact point is smooth and only the end is threaded.
The PS is my first thought. Why not just make the pre-existing hole slightly larger and press fit it? He could even double up on plates to keep it locked in place. This solution is asinine.
@@JZStudiosonline Hard to get that precision from laser cutting. Really need .0005" accuracy or better for press fits, which is very close to the limits of that method. Might be hit or miss, which isn't good for 'production' as he intends.
@@t0kinl3lunts It just needs to be a softer material. I'd expect a chunky 3D print or even Martin's favourite CNC plywood to work fine
@@t0kinl3lunts I'd assumed he was getting them CNC cut. Doing this method requires both the bolts and bolt hole pattern to be precise as well.
Martin literally reinventing the wheel. 😄 No doubt everything I am tempted to say has been well covered by others, but honestly I do wish you well Martin.
Still really hoping that the extra complexity of different marble sizes will be at least postponed. I really want to see this machine finished and scope creep scares me!
He didn't even mention the issues basically everyone sees and were telling him with the scope creep in the last video. I don't think he cares to build a functioning machine. As long as he has a project to build people will pay for it on Patreon or do free work for him.
So doing something overly complicated without getting his results just gives him more time to work on the project and get more money.
Yeah you are thinking in the old paradigm.
He is thinking modular.
If the snare module is too complex and gets scaled back it has no impact on the other modules of the machine.
@@Ammoniummetavanadate You're giving him too much credit.
@@private1177 he closed his Patreon
That whole video I was shaking my head. He's *really* letting the scope creep start building up again. He wants an instrument that can do everything instead of embracing the limitations and making it be really good at doing what it can do.
Martin!!!
First, a part doesn't need to be 100x better to justify 100x more cost. If the lesser part fails to meet certain minimums, you go with the better part. Remember how you kept putting in more and more effort to fix cheap solutions on your past two projects and they just got so bogged in being doomed to fail from the start that you ditched the whole machine? Don't let that happen here!
Second, on dynamic sound, since I missed last week: running multiple sizes of marble through the machine screams Bad Idea to me. Have you tried simply dropping from different heights? That should affect the sound similarly, and you're already trying something like that with the roll.
He needs to understand a really simple principle too: Time=Money. But even more important Time>>>Money in terms of scarcity.
Omg! I didn't realize he's 2 using different size marbles! How's he going to keep them sorted out. Two separate machines?
Martin, I am so sad to see this after all the great progress you made as a designer. You don't have to use custom housings I am sure there is a mass produced housing that works for you. There is literally no good reason to reinvent the bearing housing, the outer ring of the bearing is not strong enough to support the load.
Windows should be able to do 4k just fine. The limitation is probably from your graphics card.
Yeah, i am running 4k 120Hz HDR on my OLED screen right now.
I can't remember, was he using an Apple monitor?
Limited to 1080...does he even have graphics drivers installed?
He was making a joke. It's his pc that can't. Not all PC's.
Yeah, I don’t think there are any “production” OSs that can’t do 4k at this point.
I'm a software engineer and I've learned to be very skeptical of complete "rewrites". First it was "I believe", now it's "I verify". I have a feeling it will be "I iterate" next.
No kidding, this is "second system syndrome" to the next degree. He should have made a second music video with the MMX, called it done and moved on to new projects.
LOL! But the experience is fairly normal for designing something completely new....it's just that most designers don't declare victory until they have a working solution!
Never let perfect be the enemy of good.
Them storing your password in plain text and then emailing you your password sounds like an absolute security nightmare and is just asking for a massive breach with huge consequences lol
I didn't think anyone would even consider doing that since the 2000s, least of all a company that must have developed the software that he was using, and so should know better.
I'm exited to see the marble machine 4. In ten years so much can happen. And then the marble machine 5. 2051 would be a great year!
and it still wont work then
What do you mean PCs only go to 1080p? 😂
hasn't LTT made a 16k monster like a few years ago already?
(I'm guessing he either has a very old PC, or he uses crap screen recording software (USE OBS!!!), or his monitor/PC/Capture card is somehow not all compatible, OR, he just didn't set it in the settings)
I suspect it's because of Apple's insane locking down of their monitors.
Where you can't use the 4K mode unless you're using a mac
I assumed he was just trolling here, to see how many people would comment about it
I guess the whatever screen recording software or hardware he's using only supports 1080p on Windows for some reason.
Hey Wintergatan. I don't expect you'll read this because you've got so many followers and comments, but I want to get something off my chest.
A long time ago I've watched your music video with the very first marble machine. I remember how I was very impressed by it. I enjoyed how an impossible machine creates such beautiful music. I know it couldn't have actually worked, and you must've put a ton of editing in to make it look like it did, but you could tell you put a lot of thought in the machine. It wasn't all smoke and mirrors and worked for at least 60%. Just enough to make people believe that against all odds, this flimsy handmade machine, with marbles spilling all out over the floor, actually worked.
A few days ago I re-discovered your channel. I found out you made this machine not just for a video clip, but how you were serious about making it actually work. That you spend years trying to create a second marble machine that you could take with you on tour and use to play live music. I was even more impressed than before. It wasn't just a one time gimmick that only worked with editing tricks, but something that will eventually work in real life. I liked the design of the second machine even better than the first one. Even though it didn't look as boutique as the first one, it was wild yet elegant, with the organic materials like wood and different coloured metals. I wasn't a huge fan of the transparant plastic tubes, I don't like plastic, but it didn't look too ugly and I liked how you can see the marbles rolling through it. At the same time I felt you were chasing an impossible goal. Not only because the machine probably would never work well enough to perform live, but mostly because the better it became, the further it was removed from what I liked about the first machine. The first machine was an underdog. It looked like it could never work, but you were rooting for it because of that. The better the second machine performed, the more solid it looked and the less impressive it became. I felt sorry to learn that despite your years of work, you didn't get to finish the second machine. I hope the people at the museum where it is now will finish it.
The third machine on the other hand... I'm not convinced you're heading in the right direction with it. You've mentioned putting "function" over "form", but are you sure you know what the "function" is? You seem to think the main function is to play music, but I think you're wrong. You mention Elon Musk and his five points. How with the first 2 points are to question if a machine should exist at all. If the function would be to play music the answer would be no. There are more efficient ways to play music than a machine that uses marbles. A more effective way is to put a cd player, with and amp and speakers on a stage and just turn it on in front of an audience. It's less likely to break down as an instrument or machine and less likely to make mistakes like musicians do. However, that wouldn't work at all. No one wants to see that. The audience wants to be amazed by live musicians, with real instruments or instead maybe some outlandish machine that plays music with marbles.
I believe your machine could succeed, but be conscious of the fact the "function" is to amaze people, not simply play music. I see a lot of this fallacy in your new designs. Where you are influenced by people who design gym equipment that is solid but ugly looking, and how you order mass produced parts from third parties that are practical and cheap but very generic. I understand your desire to finally create something that actually works and it's too early for me to tell what I will think about the final design, but please; Don't put "function" over "form". Be aware that the "form" IS the entire "function" of this machine. Once you've got the design down so it's efficient and works well, make sure to execute it in a way that's aesthetically pleasing. Use beautiful shapes, good looking materials, mask ugly parts, maybe even add things that aren't functional but just for show, etc. I really hope you'll succeed this time around. That you'll manage to create something that works, that you can perform live with and that will amaze people the same way the first marble machine did. Good luck.
There's just so many things wrong with this concept it's hard to know where to start with ripping it apart - there's reasons why bearing housings are the way they are!
I remember the discussion about the flywheel being incredibly dangerous
So I would like the discussion about anything around the fly wheel to not be about expense, but rather to be about making it as secure as possible, so the audience don’t have their heads chopped off
So I hope the “cheaper” way works as well as what all the engineers on the forums are saying you need to do
I've been following this for 6 years now
Yessir, it's been quite the trek.
Cool! I've been following this for 4 years
I've joined the journey when the first mm was published. Still my favorite band and favorite ingeneer
Tolerance stackups are a thing: How tight are the tolerances on those normal all-threaded bolts (shoulder bolts will be more precise)? How big are you making the holes for the bolts? How tight are the manufacturer's tolerances part-to-part? CAD is the only thing that is infinitely precise, and that sucks when you're chasing accuracy.
I don't think it is a big deal.
Those bolts aren't a wear surface.
Worse case he has a precision jig to locate the bearing and if some bolt is way out of whack you just toss it.
@@Ammoniummetavanadate I don't think you understand what precision actually means. Neither does Martin.
@@Ammoniummetavanadate gravity, imbalances, and vibrations will all cause the outer bearing race to hammer against the threads of the bolts. The points of the threads will flatten over time and the bearing will get loose.
Using only 8 bolts, if the bolts are too tight at the start it will force the bearing to deform. This will add drag, noise and vibration to the machine. He really needs to go ahead and make a proper bearing housing.
@@thomasbecker9676 Lol.
What shaft fit should he be using for this project?
@@billtheunjust What load would you expect each of these bearings to see?
Why are you assuming this will be running imbalanced?
How much force is needed to plasticly deform the the threads of a 12.9 M6 Machine screw?
Now do a little back of the envelope math and you will see why this is worth trying.
One thought, since you are concerned about costs. Why are you prototyping in stainless steel and not in mild steel? It's a much cheaper material and you will only see differences in performance over the long term. On a personal note, I suspect that you will have longevity problems with your design since you are entirely reliant on the clamping force of the two plates to keep the bearing from moving. If the axial load is enough to overcome the clamping force, then you are relying entirely on the ability of the thread points to take up the force. Threads can be deformed just by pressing on them with a hammer using body weight alone. Where this is really worrying is that it may take time for the problem to develop and you will have committed to a design that will, ultimately, fail completely but will appear to be viable initially. To give this the beat chance of success, I would suggest not bothering with trying to use the bolts to act as a cage, but lean into the clamping force of the plates. Make the plates heavier and use stronger bolts. (M8 10.9) and maximise that clamp force. (Use a torque wrench) this is me speaking as a heavy duty mechanic that has seen what happens when engineers try to cut corners on bearings and their mountings. I want you to succeed and weaknesses in design at this foundational level are multiplicative. Honestly, why are you messing around with this and not just getting your flywheel machined to accept a bearing at its centre? Get it machined with an interference fit and call it a day. If your flywheel is made of wood, give it another .04mm of crush and glue it in place. You're trying to make a very simple thing very complicated.
Yes. He has some strange design assumptions that seem to come from nowhere. I thought I was missing some very important fundamentals watching this. Until I read other comments that convinced me I'm not delusional!
As other engineers have pointed out, the bolts and general structure appear substandard for a heavy flywheel.
Literally reinventing the wheel.
I really just hope that Martin doesn't fall into the same pitfalls he encountered in the last machine. The first one was a proof on concept, the second was the trial-and-error phase, and now that he has ample information on what works and what doesn't, I hope to see MM3 becoming reality.
There's always Third System Syndrome for those who didn't learn from their Second System Syndrome ;)
Right now, I think that Martin will work on this until he dies. And sometime before that he will become a great engineer.
he mostly likes his patreon income to be alive
If he had wheels he'd be a wagon. He is absolutely not the right person to make the machine - he's an artist and musician, a perfectionist without the experience to design the machine quickly and reliably. He's already admitted that even if the machine worked, it would sound lifeless. Wish he'd go back to making mostly human music with a music box in the band.
All new pitfalls, he's learned so much about how to goof around with it forever.
I have asked before. Why not use off the shelf flange bearings?
The whole point of a flange bearing: mount a shaft or axle to a not perfect housing or stand and let the flange pivot and shift to correct it. The mounting holes have big tolerances and they are supposed to.
But Martin is mounting the flywheel to it. So, it has tho be centered very well with low axial run out. Doing this with flange bearings is trouble waiting to happen...
@@DrKlausTrophobie There're both variants with oval holes for adjustment and round holes for no adjustment.
@@nikolayyurchenko5075 Even the round holes have big enough tolerances to screw your alignment up in an application like this.
@@DrKlausTrophobie In this case bushings can be added.
And shoulder bolts if necessary
*@Wintergatan*
3:00 Part counts: (just ONE of many arguments for why right side is better)
Left: 18 + bearing.
Right: 1 + bearing (maybe + 8 screws?).
There are detailed 3d models for bearings, for other stuff, a special soft for "todo list" and etc. This guy has definitely a lot of free time... for developing anything, except the marble machine itself.
I can really empathize with the emotional response to barriers when you're trying new things. I try to remind myself that those barriers are effectively distilled instances of learning. They're painful because learning/growing can sometimes be painful. It's the mental equivalent of sore muscles. So instead of being hard on yourself for it, be proud of yourself for pushing through.
i don't follow the logic that something has to be 100x better to value 100x the cost. If something is 10% better, but you need those 10% for it to work, it's worth 100x
The irony of a company called 247 Steel only working 9-to-5 on weekdays 😂
Also, the company absolutely should not be sending you passwords via email. They should have sent you a web link to reset it and choose your own.
Also also, the bearing cage is a bad idea. Going with a normal bearing housing is going to save you a lot of time in debugging and maintenance, and time is precious.
Any smaller (non-tech) company will send passwords by email and have terrible overall security hygiene. It's fine. Being security conscious is a luxury many smaller shops can't afford. Shaming them for assisting their customer in the most straightforward way possible is counter-productive.
@@AlexandreMacabies Security is every company's responsibility, no matter the size. There are strict rules around the handling of personal data, covered by GDPR. There are no excuses.
@@daverayment Yep. Company in Europe can get fined if their cyber security is lacking.
@@daverayment I agree. Before GDPR small companies could just wing it and hope for the best. When GDPR was introduced, it set new rules where you can be fined up to 10 million EUR for willful ignorance of protecting customer data, no matter how small your business is (and for business making more than 500 million a year, the fine can be up to 2% of their entire global turnover so that e.g. Google and Microsoft cannot just take the 10 million Euro fine and keep breaking the law).
Martin, you had me worried for a long time that you were giving up on this. I'm so happy you are going for the MM3. Yet, this design scares me. I am just about to graduate with a Mechanical Engineering degree. I have a lot to learn, but I have seen quite a few bearings fail. If you cannot get your design down to absolutely perfect tolerances then your flywheel may have a lot of vibration involved. The bearing may be able to handle the vibration for a long while, but the rest of the machine may be impacted by the vibrations. I also worry about wear on the machine itself. If the machine is assembled too many times, the extremely tight tolerances on the bolt holes will become sloppy and lead to a loose bearing fit. The way I see it, this design may work with some very fine tuning at first, but it will only degrade over time and lead to problems later. I know you mention that the cost of a fully machined housing is very high, but it may be your only option here if your bolt design does not meet your expectations.
I do think you will figure it out better than I can. After all, I am not working on the machine like you are. I only wanted to warn you of possible problems in your future.
Martin is exactly why engineers have project managers
*@Wintergatan*
2:22 This is a VERY BAD idea, because real life bolts are NOT perfect, there are small variations even between bolts that "should" be identical, they will not be identical.
The best you can do (I think) is to machine a press-fit "sleeve" for the bearing with (threaded or smooth?) holes in for the bolts, make the sleeve LESS wide than the thickness of the bearing, if you want the clamping force from the side (but I think that is a bad idea too, material flex, part-slipping & such). Also: *High tolerances = EXPENSIVE* (exponentially) 7:17 .
The thumbnail CAD drawing looks like some joke a Chinese life-hack channel would cook up.
Martin I wrote you in 2016 asking you when I would have the privilege to listen to you performing in Italy... well, I was almost 13, now I'm almost 20... this project is awesome ❤
If You lucky it will be before your 50s
@@coin777 let's hope I'll see it even if in my 50s ahah
when so many people are telling you how bad an idea it is to reinvent the wheel here, and you decide to dedicate so much time to it anyway, its hard for me to not think you are just milking content at this point.
Unfortunately I have less hope for mm3 than I had for mmx. Extra features ike different size marbles and also not listening to engineers when they say using bolts to hold a bearing is a bad idea.
When you have this large and educated community, it is worth listening to them.
Yep. I feel Martin is suffering from some sort of expression of the Dunning-Kruger effect, and is now in a very dangerous zone. You don't screw around with dubious, custom-built parts for critical applications, period.
You just don't skimp on security or on part quality, especially if there's an audience involved, and no amount of reinforced steel cages can replace proper care in designing systems which won't predictably fail in catastrophic fashion. OSHA would have a field day with this machine, yikes!
Here's the thing that I don't get; I'm just a graphic designer, and even I can sense there's something clearly wrong and dangerous with this part in particular. Must've been all those videos I've seen of ball bearings exploding, or something…
I fear no bearing housings, but that thing... It scares me.
PS: I love what you do man
Martin,
That intense emotional reaction to adverse events has a name. I hope that being able to give it a name and understand it better will make it a little easier to manage.
This intense reaction can take on many forms from frustration to anxiety to depression and hopelessness. It is referred to as Rejection Sensitive Dysphoria (RSD) and there are myriad methods for managing it. Most of them are rooted in behavioral therapy and mirror mindfulness exercises. My preferred poison for this behavioral bug is dialectical behavioral therapy or DBT.
Thank you so much for sharing your generous, patient and persistent spirit with the world through your Content and your indomitable creative drive.
FFS the man has enough issues.
@@andymouseits always better to know you have issues so you understand why you feel a certain way than it is to think you're broken or something. I know from experience.
I immediately felt what martin experienced. My desk (and a whole workshop) is filled with projects that have stalled because one small thing went wrong and I felt it was such a catastrophe I could not continue. I have a lotus speedo on my desk I was rewiring and the wire ferrules I fitted are too big to fit inside the connectors on the control board I designed and had made. Its a tiny fix in reality and yet the thing has sat on my desk for 6 months, staring me down and reminding me what an abject failure I am. I discard any notion of the hard won achievements up to this point and only see the failure. Then one random day I will just decide to cut off the ferrules and jam the wire into the connectors and carry on. Sometimes years later.
Rejection Sensitive Dysphoria is a thing and I didn't know I had it until recently. For me, like many other its rooted in the breakup of my parents when I was a kid. I think a lot of ADHD typical behaviours are basically coping mechanisms for RSD.
You didn't fail, you learned.
Yes, Mrs. Johnson
facts
Did he though?
Nah, he failed.
And now he is about to learn again, but it may take a couple of fails and a month. Sigh.
One of the main issues I can see with the threaded bolts holding the bearing in place is that as the bearing is rotating and loads are imparted radially through the bearing and onto the bolts, the threads of the bolts may start deforming (depending on what loads you are looking at) or potentially wearing away material from the threads or the outside diameter of the bearing casing. In either case, the more the threads bend or if the threads are cutting into the bearing after many cycles, your tolerances will be getting looser and your outside bearing housing will be more susceptible to spinning freely and causing more wear as it rubs up against the bolt threads. The main proponent of having machined bearing liners or housings is to keep uniform pressure via a pressure fit all around the bearing to prevent the bearing from shifting around and causing wear and to keep the bearing precisely where you want it.
Yep. Use friction of the mating surfaces through preload vs shear forces on the bolts.
This video encapsulated soo many of the issues and thoughts I have in day-to-day work. I hope you keep up this project!
Show us the struggles and victories ❤
An expert, someone masterful in bearing housing, might be able to quote the ‘perfect answer’, but the act of designing while learning can result in truly creative results! I wish you inspiration and good luck 👍
Martin. Don't go insane in design hell. Get this part of the project just done. You derp. (I say this with love) And start building stuff. A designer is always going to run into design conflicts. Essentially what you're training yourself to do is how to take these digital drawing's form CAD and turn them into physical objects. That can be incredibly difficult to do especially if you're just doing it for the first time. With a fresh start. You got this. My advice is hold yourself more accountable in the design process and get something done and finalized and start building. I know the amount of work you have in your plan can be daunting at times but like you shouldn't chew yourself up because you failed to build the machine. It's not really about the machine or the tour. We are here to watch the process of building machine. GL
All the mechanical engineers in the audience collectively winced when that bearing housing was on screen
Perfection is the enemy of the good
In this case, perfection has already conquered and destroyed all that good, completely. I think this reign of perfection has been going on the 3rd year now.
@@mishuha You aren’t an engineer are you
@@andylewis7360 Actually yes
@@mishuha Then you should know better
6:14
Martin: "PC's apparently only go to 1080p. 4K is only for Mac, apparently."
Me: watching this video on my 4K monitor being confused...
Martin.
If you are worried about the PM workload for your bearing part, maybe you should consider closing the scope creep hole instead. I can hardly believe that ordering the correct bearing mount would be that hard.
The desire of wanting to build the perfect marble machine
Fills you with determination
Thank God i found Look Mum No Computer so i can see a project get actually finished and engineered properly.
though we should focus on making a machine that works, i really hope he doesn't scrap the previous designs visually. The design is what started it all and grew this great community around it. That design is kind of the soul of this channel and is what we hope will play the same music in that very first video. If he really goes with the "concept art", it just wont feel the same, even if it works. It'd kinda be like another singer singing your song, and not you, the one who originally sang it.
at a MINIMUM you should be using bolts with a shank for this or the bolts WILL break eventually. With a non-rotating shaft like you have ideally the bearings would be pressed into the flywheel. Though I'm not really sure why you didn't just go for a rotating shaft so you could use pillow blocks and just not worry about this particular issue.
@Wintergatan Flanged Bearings is what you are needing. There are 4 bolt designs.
I would question whether you're calculating all the costs into the non-machined bearing housing. Is the pre-fabbed part 100x the material cost only, or 100x the total cost, including the opportunity costs of your design time, prototyping, evaluating, iteration, etc?
Turning is super expensive.
Laser cutting at scale is unbelievably cheap.
@Ammoniummetavanadate I cannot accept that there is no off the shelf bearing housing that he could not adapt to fit his design or adapt his design to fit, that would be less expensive than ordering a bunch of prototypes, trying them, spending hours fiddling with the idea in CAD, ordering more prototypes, realizing it isn't going to work, and then ordering proper housings.
@@RhynoD2 There absolutely are, they are also like $100 USD each from McMaster, Maybe down to $50 each for sketchy ones.
This will be like $10 each and built directly into the part vs needing mounting bolts and then locating them.
@@Ammoniummetavanadate For a flywheel I would absolutely spend the extra $180 to make sure it's safe
@@Ammoniummetavanadate It's already been €40+ just by paying for the prototypes! He will save so much time by doing it right the first time.
MARTIN!!! You did not fail on those previous marble machines!!! You just found 2 ways how NOT to build a marble machine!
Martin continues to make the same mistakes one mm3 that he made on mm1 and mm2: Not listening to the experience of people who know what they're talking about.
I have used "Split Shaft Collars" as bearing housings. The collars have a tightening screw, I attach the collar to the frame by drilling and bolting through the collars flat sides
Mechanical engineer here.
This solution will create more problem than it solves.
You're dealing with small loading here, especially compared to what those bearing are capable of. Just 3d print a housing. 90x cheaper than machine, 90x better than your bolt arrangement.
Plastic bearing housing are used everywhere. Just look at powertools : drills and angle grinder motors (not gearboxes) use plastic housing for their bearing.
Don't try to reinvent the wheel, i don't want to hear you announce Marble Machine 4
The flywheel is basically the heart of your machine, I know you know it but, these things store an enormous amount of kinetic energy. If you test your idea, test it like crazy, because I'm pretty sure it's the kind of idea that will look to work great at first and eventually fail on you at the worst moment. You said 100x better, I say what would be 100x worse: break 20 or more parts on stage and having to cancel 5 shows and lose whatever revenue that means.
And then have to do it properly.
NGL, my first reaction to the idea of making a bolted bearing cage was WTF. But on further reflection, I don't think this application will be under high enough loads to suffer from having a poorly supported bearing. I do, however, not like the idea of threads being used as load bearing members. If I had to do something like this for my a client design, I would either A) use a stack of cut plates so that the bearing still has full support. or B) use shoulder screws instead of fully threaded screws, so that there is at least a nice precise smooth surface supporting the bearing, not the sharp edges of the screw thread. Also, screw threads are not toleranced to the necessary level of concentiricity, unlike shoulder screws which are designed to be used for load bearing and alignment.
A lot of people have commented on how this solution probably won't work. I am also a mechanical engineering student, and I think you could get the tolerances correct by using precision ground shoulder bolts instead of normal M6 bolts. M6 bolts have allowable major diameters between 5.974 (upper) mm and 5.794 mm (lower). That's an allowable variation of 0.18mm. For holding a bearing in place accurately, you probably want at least a locational clearance fit, if not a press fit. The maximum variation of size for a 52mm diameter hole (as would fit the bearing in a locational clearance fit) is +0.03 mm. You might be able to get close to this by using precision ground shoulder bolts and reaming or broaching the holes to precisely the diameter of the shoulder bolts, but the tolerance stackup on that means that it would probably be easier and cheaper to just have a bearing holder machined or use off-the-shelf insert bearings.
It is good that you experiment with this idea and will put it to test, so there is no more - what ifs, but I also agree to I guess many viewers that it won't be reliable, it will be nightmare to center, bolt threads are weak and will probably wear out because there will be movement from all the loads on flywheel and axle. Also if you have watched the episode of Sandy Munro saying, that they did a reasearch at Ford on what was the cause of most of their manufacturing and realiabilty problems and the answer was bolts.
So again I think it's good that you test this idea, but it won't be easiest or best solution.
3:59 having a radial force on a Bolt is the worst type of stress you can put on a screw. The only way that this method could work is by having clamping forces high enough that the friction between the plate and the bearing holds the flywheel.
Hello Martin. I do structural design in Canada. All structural (not mechanical mind), bolt holes are oversized by 1/16", or about 1.5mm. To keep things simplified, torqueing the bolt eliminates the ability for the structure to slide around in the hole. My point is don't get hung up on making extremely tight holes for your bolts. Your loads are light for the number of fasteners you have, just use washers and torque properly and you can get away with larger holes.
The issue here is that Martin wants to use the bolts to align the bearing in the center, not just for transmitting the force.
This design with the bolts is tremendously cursed. Specially if you plan with store bought ones.
Save yourself the trouble, search for existing solutions such as "insert bearings", or "bearing houses".
PS: Also an industrial draftsman/mech eng here. This is the second most cursed thing i've seen today.
rule #1 - the comment section is always right.
rule #2 - first learn rule #1.
there's a reason people are telling you to go with something that already works, but if you *MUST* do things the hard way for engineering/comparison sakes...
As a rocket engineer working on bearings I have to say, 100x more expensive is the best deal of your life. Although I think there are other options designed around a flat plate with holes you put between the two plates. Very hopeful you can make it work!
Already pointed this out in the previous video, will repeat it again here against the general criticism.
I love that you're experimenting and trying stuff out. That's how you learn, but also how we learn.
Please keep your creative mindset.
(I'm also an engineer, as this seems to be a necessary disclaimer to comment here nowadays. :D)
the bearing housing will grind on the threads, if nothing else eventually the fitment will loosen.
Yep. The best solution is a 4 hole flanged bearing. He can bolt the bearing right to the flyweel.
Good to see that you will be testing this design 👏
As a retired jet machanic I've seen bolts with a shank long enough to go through both clamping plates, if the bolt holes are too tight just shave a flat spot on the bolt shank. I'm with the engineers on this design. A custom bearing housing is the correct means to support the rotating assembly.
The MMX was totally functional and tour-able. We are witnessing feature-creep on a massive scale.
I tune in week after week specifically for your stupid and glorious battles....and the music, it makes them even more glorious.
*FAILED???*
*_No way did you fail!_* You created some *_amazing_* musical instruments that are just as beautiful to the *eyes* as they are to the *ears!*
Each new iteration is a new expression of your artistry, and we, your fans, enjoy being able to share in your joys and struggles!
Did the inventor of the guitar fail simply because people continue to modify and improve on it, _to this day?_
No!
The fact that people continue to tinker with ideas for new expressions of it is validation for just how successful it was!
I applaud your humility as it only serves to further express your success!
God bless you!
_"Let another man praise thee, and not thine own mouth; A stranger, and not thine own lips."_
*--PROVERBS 27:2*
Preach!
Well, he did fail but, not the way he thinks... He failed in learning from his past mistakes... Because hes making them again
Your reaction to minor setbacks like the place being closed when you called is absolutely something I've come to realise I have! The problem has gotten noticeable because I started making basic rules for myself to avoid those setbacks that started becoming "basically don't start something that could possibly result in that" meaning I started doing nothing much at all
As is said in the previous video of the bearing housing, as stubborn as you are, at least try to use partially threaded bolt. The bearing will be guided by a real cylinder, a bit more precise than threads.
Martin is so good at inventing new things that are amazing and reinventing old things badly
More procrastination?
He cancelled the last version because it was too complex...and so his solution was to instantly make this one stupidly overly complicated surpassing the last one before he's even begun construction this time...
It's baffling
Motivational Poster Idea >> - Great things are never perfect -
a few seconds into this video im am thinking: how can a so clever person like martin come up with such stupid idea? this must be the supidest idea of all he had so far.
What compelled you to make such a caustic comment today?
@@cartanfan-youtube i was just in the mood to do so.
Let's spend 14 hours engineering readily available, mass produced inexpensive part that's critical to the operation of the machine, have it fail completely, and do it all in the name of saving money.
He's not an engineer. He's an artist with a romanticized vision of what engineering is stuck in his mind, finding new and innovative solutions to things that aren't even problems.
Well his hero is Elon Musk, who made the world's largest rocket. Said rocket was, at best, expected to clear the launch pad, experienced failures in 1 out of 6 engines, and then exploded.
I'm like were your heading with this New Marble Machine , giving it a different set up may to the Trick in making it work , cant wait to see what you got coming next 👍👍👍👍👍😎 . SKK .
Amazing to see the advancement in machine design since v1. Always look forward to each video.
The future failure modes of this bearing housing will be fun to watch!
🤣
You keep talking about the machine failures, how awful past designs were, and maybe from a perspective of engineering I could see why, but... The 4 minute 32 second one that I first learned about you from, the one titled "Wintergatan - Marble Machine (music instrument using 2000 marbles)", with what is apparently the first marble machine, is still the best video of yours I've seen. I keep watching your stuff anticipating that you'll make another song from one big system machine again. I'm just waiting for the next great song that matches that first track I ever heard. I still rewatch that 7-year-old video (v=IvUU8joBb1Q) now & then, and it's amazing. Still. All your engineering is interesting to my nerdy side, but I keep subscribed *just in case* you upload an actual new song.
That's what I'm waiting for now. Another Marble Machine song.
We all are...