A Better Flywheel Design - Marble Machine 3 Ep.4
Вставка
- Опубліковано 21 лип 2024
- Support Wintergatan:
- Patreon ► / wintergatan
- UA-cam membership ► bit.ly/4cQVM7C
Marble Machine Engineering Discord Server:
/ discord
Video edited By Martin and Hannes from the Trainerds UA-cam Channel:
/ trainerds
-
PATREON ► / wintergatan
UA-cam MEMBERSHIP ► bit.ly/4cQVM7C
WINTERGATAN RECORDS ► www.wintergatan.net/#/shop
SPOTIFY ► bit.ly/2oKxXWd
ITUNES ► apple.co/2ntWNsZ
ENGINEERING DISCORD ► / discord
COMMUNITY DISCORD ► / discord
-
- Do you want to machine three pulleys for the Marble Machine 3?
Fill in the form here:
www.smmk.de/en/task-listing/m...
Leave your feedback on the latest design at the reddit page here:
www.reddit.com/r/MarbleMachin...
A Better Flywheel Design Using Proven Solutions - Marble Machine 3 Ep.4
Lucas Filmed by Don Carlos:
doncarlos-medienproduktion.com/
Check out Lucas UA-cam Channel:
@Musikkabinett
Check Out Hannes UA-cam Channel:
@Trainerds
#4k #wintergatan
-
MUSIC DOWNLOADS ► wintergatan.bandcamp.com
WINTERGATAN RECORDS ► www.wintergatan.net/#/shop
SPOTIFY ► bit.ly/2oKxXWd
ITUNES ► apple.co/2ntWNsZ
MERCH ► teespring.com/stores/wintergatan
DISCORD ► / discord
SECOND OFFICIAL CHANNEL:► / wintergatan2021
-
Thank you for listening to the feedback. This gives me so much more hope for the finalised machine
To be honest, I am kinda impressed by his ability to still extract constructive things from the explosion of not-so-constructive criticism he received since deciding to go for MM3.
@@mateusbmedeiros Good point. He must have a whole team of comment-readers, organized and sharing information in multiple levels, where he himself is at the top level, only reading the valuable, actually helpful ones.
(or Martin himself might be up till 3 every night, reading all the comments and making notes)
@@EaglePicking Good point as well.
@@mateusbmedeiros A lot of that was driven by emotion; I don't think the disappointment was unjustified. I'm enjoying the progress on MM3 but the reasons for liking it are different to the reasons for liking the first two. Martin keeps saying not to aim for perfection, but that's kinda how I'm viewing this. No longer are we witnessing an eccentric artist construct something utterly crazy, we're witnessing a budding engineer construct something crazy. It loses something, but it's still fun.
I disagree on this leading to more hope for the machine. I dont see any move towards truely being more humble.
"I'm sure my bearing design would work flawlessly" - 0:54
This is just ego talking. There is no years of history of coming up with new designs and having them having 0 problems... in what world, given that, can someone logically say "I'm sure my new design is flawless".
"I'm sure" + "flawless", it's crazy to still hold such a prideful stance.
Humbleness is the virtue, not pride. I want to see him let his ego go and be more humble.
Yes! Thank you for using the stock parts. I have wasted hours of my life over complicating designs and then switched to something extremely simple and better than what I had made.
Don’t feel bad about not designing your own bearing mount. Engineering is all about trying a design and fixing the mistakes you make. A good design on paper is just that, a good design on paper. Just because the math and logic makes sense on paper does not mean it will work in real life. That’s why people spend millions testing.
Try, fail, try again. Good engineers learn from mistakes and that’s the whole point of marble machine 3.
Martin isn't an engineer.
@@thomasbecker9676 I don’t mean engineer in education, I mean casual engineer. He is designing mechanisms that need to interact to a precision, that’s engineering to me.
@@dylandreisbach1986 I call it tinkering.
The actual answer is to use samples and program the whole thing. This could be done with software that existed 20 years ago.
The very premise of the marble machine is over engineering.
It depends on what you are building... before he was building art, now he is building peace.
Using standard parts as opposed to custom has another major benefit you may not have realized: if a part fails you can go out and simply buy a replacement, as opposed to lugging around a trunk full of spares on the world tour.
Hi Martin!
At around 4:30, you talked about reaming. As a machinist (or soon-to-be, I'm still at school), I would tell you that the hole that you plan on cutting at 19.9 millimeters, I would resize them at 19.5 millimeters. If you plan on using a machine (a press drill, hand drill, or a milling machine), I'll give you the equation to calculate the spindle speed :
(90 (the Surface feet per minute, or SFM for short, of steel) * 1000) / (pi * tool diameter) = Machine RPM
Then, divide all this by 4 because you're reaming (it puts a lot of stress on the tool). Also, make sure you're using A LOT of cutting oil, because reamers tend to generate a lot of heat.
In this case, ((90*1000)/(3.1416*20))/4 equals 358 RPM.
Why am I telling you all of this? Because I don't think hand reaming in your situation is doing you any good. You cannot, and I repeat, cannot drill, ream, or even tap by hand being 100% straight, it's impossible. Imagine reaming all of your holes in your flywheel, then assembling them and realizing that none of your holes is truly perpendicular to the flywheel. What a disaster it will be. At that point, I would ask a nearby machine shop to ream them for you, as it'll take them minutes to do.
If you have any interrogations about machining steel parts, you can always ask me, and if I don't know the answer, I'll ask my teacher, who has been building extreme precision parts at Pratt & Whitney and at Bombardier, making aircraft parts.
You should probably copy paste this to the subreddit, that's where they take feedback
Post this on the MM3 subreddit. This is where feedback is given
@@sqbuilder1328 Thanks guys, after weeks of procrastination, I finally posted it on r/MarbleMachine3. Without you two, I wouldn't have thought about it.
Reaming will create accurately sized holes with a good finish, but wouldn't necessarily make the flywheel concentric. This is because the ream will cut exactly where the previous hole was made as you touched on in your discussion of using a hand reamer. This is why preparatory spot drilling in CNC machines is so important, you have to have a positionally accurate starting location for your hole if you want positional accuracy (i.e. concentricity). Laser cutting should still be plenty fine for creating a concentric enough starting hole, and reaming is still probably necessary to get the accurate size and surface finish, but it is important to note that reaming is NOT helping the concentricity itself.
One other thing to consider with the flywheel is that weight in the center of the wheel doesn't do much for you. An equivalent flywheel can be made by removing some weight from the center so that it is concentrated on the rim. You can get some good ideas from 19th century stationary engine flywheels.
Good point. This will make the overall flywheel lighter in comparison to it's storage capacity. That's less weight to carry around while touring and no downsides because the operational properties of the machine will be the same.
Say you have a solid flywheel. Can you make it a better flywheel by removing center mass? or is it that you can make it a slightly worse but much lighter flywheel by removing center mass?
I came to the comments to say this too. Adding some lightening holes will give a very slight reduction in rotational inertia, for a significant reduction in overall weight.
With 6 plates, he gets a little over 1 kg m^2 of MMOI for 60 kg of mass. For the same mass and wheel that gets wider at the edge (100 mm instead of 60 mm), he can get upwards of 1.5 kg m^2, just to put some ballpark numbers out there. In the same way that the same mass can get better tempo stability, Martin could also get the same tempo stability as 6 solid discs with the same weight as 4 solid discs. Reducing weight means more margin for the bearings which means more safety.
@@GordonWrigley By removing material, even from the center where it has less effect on the energy storing capacity, you will make it "slightly worse". However, you can make a perfectly equivalent flywheel by increasing the diameter slightly, while still having lower mass than a full cylinder.
He's LEARNING!
Martin, I think that bearing housing was an interesting idea but I agree with the community. Use proven designs. This whole process will be a feat of design and engineering all its own even if you use off the shelf parts. I just finished going over a custom skid we had built for my work in biotech. The vendor most certainly didn't custom design every part on this machine. It's a beast in its own and custom designing things that can already be bought would only have increased the timeframe and increased the opportunity for failure.
Don't worry about leaving your identity and fingerprint on every part. This build will 100% be a part of you even if you bought every part of the shelf
Custom designing parts is a good idea whe' the solution doesn't exist for your problem or the tue a'd tested solution doesn't work because of the specifics of a project
It made me smile to see that you are using clamping collars on a straight shaft (I had made the suggestion in an earlier episode 🤓)
This is in fact the way servo motors are connected to ballscrews on large industrial cnc machines.
regarding the pillow bearings, they can have a fair amount of turning resistance, so make sure to order metal ,not rubber, shields.
they have less dust protection but there is no contact between between the inner and outer race . also consider flushing out the
original grease and replace it with a much thinner ,but appropriate for the task, grease .If you are already consulting with SKF they would
have exactly the correct answer.
To made assembly and disassembly as easy as possible ,ream the clamp collars same as you are doing the flywheel segments, that way
the shaft is never scratched .
lastly do not screw the bearing setscrews directly to the shaft , place soft copper pieces into the setscrew holes and tighten down on that
Oh and waterjet cutting works very well.
I'm so happy that you listen to feedback at such a large scale. I think the biggest advantage you have on any other non-professionnal engineer is the huge number of qualified and experienced people willing to give you feedback and advice.
Of course you are much much more hard working than most of us which is the second key to your future success, which I'm now sure will come some day. (I'd already consider everything you've achieved as a big success even if you don't have a working machine yet)
And a huge number of unqualified people trying to give advice and messing up the signal to noise ratio hehehe. Must drive anyone insane when like only 25% of the advice makes any sense hehehe. But yeah it seems like this video is more sane than the last two. Really hoping Martin wins the battle against the machine, then again I'm attached to these videos hahaha.
@@minerscale Let's hope the bin expert answers are different enough that a majority votum is fine for the correct answer.
Low SNR? Need more data !
@@minerscale Even those who give "bad" advice, because they understand little of engineering, may sometimes come up with cool new brainstorms of original ideas, which in turn might help the engineers among us to be enlightened and think of better stuf to actually build.
I use this technique with my wife all the time. I ask her opinion on engineering stuff, not because I'm expecting an expert answer, but because she occasionally comes up with great stuff without knowing it ;-)
Belt tension:
Your design will require you to carefully adjust both sides to maintain the shaft perpendicular to the belt since failure to do so could lead to premature belt wear or 'throwing' the belt.
An idler wheel pressing on the outer surface of the belt would give a single adjustment and would also increase the wrap around the smaller wheel, which could be beneficial.
Next time you make a drawing for any sort of part, try and position the dimensions outside of the part (dimension lines still going into it) this makes the dimensions much easier to read since there is nothing obstructing them then! I've found it really helpful to dimension and label my drawings like I would never talk to the person who would be machining it, which would lead me to more practical, readable and helpful drawings
He should read ASME Y14.5
Congratulations, the design is much improved and glad to see you excited for the next machine. Bowden cables and pillow blocks seem to spell success.
Some suggestions (others have also mentioned):
1. Move the pulley to the outside of the pillow block. This is a huge deal for maintenance and will make assesmbley/dissassembley much much easier.
2. Rigidly mount the pillow blocks to the frame and make sure that that part of the frame is secure as possible, no slots or extra complexity in the part of the frame tasked with holding on to the flywheel, it already has a very important job.
2.5a Extra triangulation will be easy to add in the fabrication phase and do wonders for keeping this part of the machine behaving as expected. I'd guess that the design of the frame at the moment will flex like a parallelogram. Cheap insurance.
2.5b Maybe even some removable concrete slabs could be added to the machine to help dampen vibrations and keeping the flywheel under control but transportable separately for touring?
3. (related) Move the belt tensioning to an external mechanism.
4. You can buy shaft with key slot already in it, it's not the cheapest thing in the world but a great option for this part of the machine. The tighter you keep mechanisms like this, the less you will need.
5. Incorporate an independant brake (off the shelf) for the flywheel that is directly coupled to the flywheel shaft. Could be something you never need, could be something that you thank the stars for including. (Breakdown lever for artistic use implemented separately the otherside of the clutch)
Bowden cables stretch and need adjusting, and when they break, they fail completely.
Vibrations aren't "dampened," they're "damped."
@@thomasbecker9676 bowden cables are relatively off the shelf parts, so seemingly a non-issue
@@thomasbecker9676 hes not using them for precision motion so its no biggie
@@stevensimpson3252 It's still a part that when it fails, it fails suddenly and completely.
@@jackradzelovage6961 "Precision motion" made me laugh.
Hi Martin, forget about laser cutting then reaming. Simply CNC mill the entire wheel. This will ensure concentricity to within the machine precision. An additional advantage is that you can face both sides of the wheel to avoid thickness variations, that will provide an excellent result in one step
SKF also makes standardized pulleys
Yeah it's a bit weird how Martin has come to the conclusion that standard axle bearings are a good solution, but he is now designing his own pulley.
Surely if he buys an axle, buys fitting bearings to hold it and then buys a belt, he can simply order a pulley to fit that belt to that axle. That will make sure that it's perfect and you can just order the part.
SKF uses ISO 5294 pulleys, but Martin uses ISO 9982 pulleys
Just use a standard pulley. The power input module should be able to use 95% off the shelf parts.
Nice to see some good progress Martin! Not sure what the weight of the flywheel will be, but it's better practice to keep the bearing blocks as close to the flywheel as possible, and the pulley on the opposite side to avoid gravitational stress on the shaft during storage & transport. A flywheel in that size and potential speed can make the whole machine vibrate and affect the rest of the functions (and sound). So from my point of view, you should have bearing block, flywheel, bearing block pulley, bearing block, that way you keep the whole system as rigid as possible. Keep it up, love your progress!
Yeah it was looking a little wide to me, I wonder if he was giving himself a wide clearance for adding plates if needed...
@@xxportalxx. True :) On the other hand, since he uses modular design, it should be fairly easy to move the bearing block stretcher in/out when he have found the amount of flywheels he need.
I felt a massive weight taken off my soul when Martin finally caved to the tried and tested design..
You should look into using 20/20 or 40/40 aluminium extrusion for the frame parts, one less thing to have custom machined except the lengths, which most places already do.
I often build things out of 45/45 and I would not trust it to remain 100% rigid. One good wack and the joints will move. Aluminum is also a relatively soft material, which doesn't help any.
Martin, I’m so proud of you for going with an existing proven solution for this bearing situation!
And yet, you support him cobbling together a flywheel?
Martin, for added rigidity and stability, your flywheel should be separated from the pulley with a bearing block mounted in between. So it should look like bearing block mounted to frame --> flywheel --> bearing block mounted to frame --> pulley --> bearing block mounted to frame. You don't want your flywheel and pulley to be beside each other without some form of support in between.
Long time listener first time commentator.
I love watching your determination Martin and thank you for continuing this project. We all love watching
if you want to keep changing the the pully you may want it at one end of the shaft. At the present design you would need to take apart the whole flywheel just to change the pully. but if the pully was at end of the shaft you change the pully without taking it totally part.
I suggest putting this comment in the subreddit, if it's not already there
If the flywheel is running the music machine, wouldn’t it be easier to adjust pulley tension from the input pulley side and leave the flywheel shaft stationary?
Honestly I'd imagine that the power input side of the design is going to end up more complex than a simple flywheel
ty ty ty for listening to feedback. ❤
I would recommend keying your pulleys and weights to the shaft, rather than relying on a friction fit, so as to avoid the pulley or weight just spinning on the shaft. The shaft collars will prevent the axle from shifting longitudinally along the axis of rotation, but won't do much for countering the angular momentum difference between the shaft, weights, and pulley.
Yes, but not being keyed that could also be a safety feature. I don't think he is going to spin it up fast enough by hand that the inertia is going to create a problem with slippage. But, being able to slip if something jams might just be worth the slippage risk
You can get an insane clamping force with a split cone pulley. In the region of a couple thousand newtonmeters.
One good that came out of this was: I knew the design you originally proposed with the threaded bolts was bad practice, but I got to learn from many (armchair) engineers the technical reasons as to WHY it was bad practice
It's Wednesday, my dudes.
it indeed is
aaaaaaAAAAAAAAAAAAH
AAaaaaaaaaaaaaaaaaaaaaaAAAAAAAAAAAAAAAA!!!!!!
Ja
Indeed it is
Hi Martin,
There's a lot of good things here but I do have a concern about your tension adjustment method.
It may be that there's details and features I can't see just from the screen shots of your design.
However looking at the screen shot at 0:47 & 2:52 it looks like the bolts you need to loosen, to allow the sled to slide back and forth, are also the bolts which hold the pillow blocks in place on the sled allowing the bearings to move independently of the pillow blocks while you are adjusting the sled position.
This will make making accurate adjustment of the tension a nightmare.
My recomendation for a quick solution would be to thread the exising holes and use them to fix the pillowblock directly to your sled.
Then extend the sled, and add two new holes in the sled, in conjunction with shoulder bolts (Which will locate your sled more accurately in the slots than the current threaded bolts).
I'd also recommend some sort of reinforcing plate on the box section where your adjustment nuts are located to prevent accidental pinching of the box section (Ive seen it done with large rbox section than you are using).
If you really want accuracy and ease of alignment, rather than the slots in your box section, you might want to consider using linear rails and bearings such as those made by Matara or THK (Other brands available), these aren't particularly expensive (though ) but they massively help with keeping the bearing sleds parallel and also mean that you have less bolts/nuts to loosen when you want to adjust the tension.
If you see this I hope you find it helpful.
I really do admire the work you've put into this project.
You have done an amazing job and as a mechanical desing engineer watching the progress in your design skills come on from marble machine x to now is brilliant.
I also don't think you give yourself enough credit for what you achieved with the first two marble machines.
Even the most experienced designers will find flaws in their designs when they are actually built, and finding soutions to those problems within the restrictions of what has been built is a skill all in itself which you have shown a great talent for In my humble opinion.
So bravo and I wish you all the best as you progress with th enew marble machine.
I once worked with a very fine gentleman who was a carpenter by trade, but at the time we were carpet fitters, I was his apprentice. He had a great saying, "That will do, won't do, Perfect will do but only just". I had a great respect for him and we ended up great friends.
My friend, the best design can be made, but always remember that the assembly and adjustment is the final and essential part so that everything works perfectly later, misaligned pulleys will cause vibrations and efforts that can compromise the machine in the medium term. This is from someone who has assembled and maintained different types of very precise machinery. Go on whit it!!
Hallelujah Martin has seen the light!
Martin, top tech tip for you from an old High school machinist building stacked parts like your flywheel. Get your stack as close to aligned then ream all the center holes at one shot. This will give precision reference point to do all the other machining for the bolt holes that will give precision fits you want then before balancing spin it at a low speed and sand it out to make it concentric.
THIS WILL WORK. I've been using this shaft and bearing setup with 60" dia. plywood wheels on my tricycle parade float for several years.
"Tricycle parade float"? I was like WTF?
Then I looked it up and now I get it, but it's one of those typical regional traditions that if you're not from around there, you've never heard of it.
Your design really gave me flashbacks to MMX... Thanks for listening to the community!
Hell yeah! Good to have you back, Martin. I've been with you since the Detektivbyrån days, you've come so far!!
I am loving your new series Martin! I really hope you carry this energy and see the project till the end, and I hope even more to catch the MM3 live.
Hello Martin,
Ever since my childhood, woodworking has been my passion. This love led me to discover your
channel and the unique blend of music and craftsmanship in your Marble Machine X project. It
intrigued me and introduced me to the world of music as a new hobby.
When I had to relocate for my studies, I had to leave behind my woodworking workshop. This left a
void in my creative pursuits, but your work inspired me to fill that void with music.
No matter what the future holds for
the Marble Machine, your music, or your UA-cam videos, I want to thank you for reigniting a creative
spark within me. I hope you never lose yours.
Thank you for all your work!
Jérémie.
Well done going with a proven design. A tip for the belt tensioning (proven design) - use an idler pulley. 1 additional pivoting wheel with a spring for tensioning. An example would be the tensioning on the drive belt for a ride-on lawn mower.
WE'RE SO PROUD OF YOU! MM3 is totally going to work!
Hey Martin, tensioning two side of the flywheel bearing will make the task harder than necessary. If one side is more than the other, the belt will slip off your pulleys. You need really good alignment. Add a tensioner pulley instead and its way easier to maintain alignment while tensioning
Good choice on the pillow blocks. Excited to see a big piece of metal spinning very fast!
That is good, i'm glad that you're going for a center support bearing. Those deal with incredible pressure in trucks, they will work amazing.
So I gather from this video that you decided to build it!
Very happy to hear that.
remember that you can still work on the flywheel when it is in place. So my suggestion would be to make it a bit oversized and then grind it down to spec when it is assembled .
Happy building Martin! Looking forward to the world tour!
for the pulleys, again, off the shelf... you use a bushing machined to shaft diameter.
Wow, this video was miles better than the previous ones.
Even the tensioning design is more than acceptable. :o
Martin shows class and dignity.
Using known solutions for known problems allows more time and resources for the more unique issues.
There will be plenty of time for experimentation on MM4, but only AFTER the world tour.
What if it is perfect and there won't have to be a mm4?
Tip for dimensioning a part for machining: if you have a large number of features in line with each other (such as the pulley grooves), try to dimension from a single point rather than incramentally. Compounding rounding, measuring, and tolerance errors can stack up to give you a part that's too long or short, even if each individual feature is in spec.
concentricity can be achieved after the wheel portions have been crafted. Once it's mounted in the machine, a temporary brace on the nearby strut at the rear of the frame, and a few minutes with a lathe tool, and you can true up wheel to be nearly perfectly concentric. But as you said, balance is more important than concentricity, particularly if you're using a material like plywood which is not entirely uniform across the sheet.
Great news you managed to get your head back straight. Standard parts might be 'boring', but once you start going down the rabbit hole of designing basic parts, you are only a small misstep away from designing a new thread form to make better bolts, and then you'll never get anything done.
Keyed shaft would be simpler, than that friction clamp, no?
You actually still end up with a clamp when using a key to keep the parts from moving axially along the shaft.
A keyed shaft doesn't prevent axial movement.
keyway would result in the low tolerance lasercut parts sliding off the keys. Would still need a clamp!
So much relief the proven design was choosen
That was a close one. Other than the kinetic fingers, I've never seen the community get so riled up against a design.
Hey Martin, good on you for keeping on going against all odds. Your rediculous dedication is inspiring after being kicked in the head multiple times by your machine. Awesome stuff.
I really enjoy your videos. I savor them like a fine wine, on a Friday evening. I'm so glad that you've moved on from that bearing housing design -- there are better hills to die on. It takes a very strong and experienced man to admit that an earlier stance is wrong. I salute you Martin. Godspeed.
Box tube looks and feels really stiff but when you apply the mechanical advantage of a bolt thread it's surprisingly easy to deform them.
One solution is to offset the attachment points to cleat plates although that adds eccentricity.
A second more time consuming option is to sleeve the hole with hollow tube slightly larger than the bolt. That way all bolt forces are taken in compression but the sleeve and the box is isolated from the bolt tightening forces.
For most applications this is not a big deal but where you are loosening and tightening bolts often or looking for precision it could save you a lot of trouble.
Next when you get a bit closer to finalizing your base frame don't just check the stiffness look at its harmonic frequencies and compare them to all the rotational frequencies that are in the range of the machine. A precision marble drop is only precise if the target an the drop location are steady.
Hope this can avoid rework the enemy of all engineers.
Best of luck.
You are well on your way to becoming an engineer, doing technical drawings now too! It's been great following your progress over the years! Also, thank you for addressing the hardening effect of laser cutting, I was getting ready to write a comment about it before you brought it up haha
He's not. Signed, an actual engineer.
@@thomasbecker9676 I'm an actual engineer too... I'm not saying he's about to work as a professional engineer or anything, just saying that I see a huge improvement in his thinking and design process compared to when he was making the first MM out of plywood and rubber bands
@@thomasbecker9676 A real engineer doesn't call themselves "an actual engineer", but nice try.
@@andrewgrzanka "Improvement" is a meaningless term if he's still going down rabbit holes and ignoring what is good and common practice. There's a reason why companies have been selling the same mechanical parts for decades. He supposedly had an entire "engineering team," including a project manager for some reason," and still couldn't get out of the weeds.
@@LordDragox412 ERAU Daytona, BSAE (aero) with minors in Mathematics and Flight Test & Simulation.
I blushed when Martin said ”shaft clamp” 🥰🥰
Yay! This video shared some extra smiles and encouragement around the world. Keep up the great work! 😃👍
Im not entirely convinced youre 100% happy about using a proven method here but Im glad you did
What a time to be alive!
Cool to see how even though you accomplished so much but you're still humble enough to take what was probably pretty hard feedback
One consideration on the tensioner design you have is that there are two adjustment points to adjust a single thing. My suggestion would be to take a page from automotive design: an idler pulley on the outside of the belt loop which is spring mounted with an adjustment for how springy the spring is. The makes it have a single adjustment and this adjustment isn't a friction mount which may need to be periodically adjusted due to vibration causing stuff to move.
Advice/experience comment from working with machines that use belt drive.
Belt tension was spoken of. Here. I see it is being solved by moving the WHOLE flywheel assembly to get the belt tension needed, which after the adjustment, could lead to belt walk to one side or the other and wearing the belt out sooner should the adjustment not being true/square. That belt is still just a flat belt with groves that can slip.
Belt tension Idler wheel/roller. Why was that not considered? I work on copy machines. They do have adjustment for the belt by moving the whole motor assembly for small motor drives. On the big power drive motors? ( like here driving this major flywheel ). It's anchored down. Most often they have tension roller adjustments. Plus the belt and the drive / destination pulleys are ribbed. Zero slippage. Perfect timing. Unless the idler is loose and it allows the belt to skips a tooth on the gears. That's very rare. No moving of the motor mount or the other end. Just retention the Idler to fix it.
One thing you also need to remember is the tolerancing of the laser, as just because you set something to be 19.9mm doesn't mean it will end up that way. Depending on things, it may be too small or too large.
At my place of work, we have +/- 1/16" tolerance on all plasma-cut parts, and I actually had a good example of why tolerancing was important to happen a few weeks ago: I noticed excessive rework being done to upsize a hole that was called out for exact dimensions on the print, not accounting for the tolerancing, nor the hardware to be used. Despite the clear measurement, the hole was 1/32 too small...still within tolerancing spec but added to the cycle time, not to mention affecting the part's surface finish.
with this, it's now looking more likely we'll get to see the MM3 actually built! Great progress
With regards to the 'feedback on things you edited out' thing, I don't think it's a coincidence; completely editing out a topic gives the impression that it hasn't been given much (or indeed any) thought, so there'll be more discussion about it, including points you had come up with already
Well done, Martin.
I would add as I did last week that you might design a rubber vibration dampening block between the bearings and the frame. Fairly early in the life of the bearings, they will develop internal vibrations that your metal frame is likely to amplify. Those are also commercially available. They’ll have something like a 95A shore harness rating. Also, be mindful of the precessional forces, making sure the frame can’t bend in directions you might not expect (like twisting as opposed to sagging). Also, consider a longer belt with an idler pulley for tension. That may be more robust in the long haul as the spring tension will automatically take up the slack from belt stretching. Also, it will dampen the belt’s tendency to flap on the low tension side of the system.
You may want to use separate bolts to fix the pulley to the rest of the flywheel assembly. That way you aren't loosening the heavy plates which could disturb the balance when swapping pulleys.
I think you'll also want to be able to easily remove the flywheel from the machine without loosening the through bolts. This may make balancing much easier depending on exactly how precise it needs to be, as well as improving serviceability in general.
Funny that picutre you showed earlier of "Rüdesheim" is actually a view from Rüdesheim across the Rhine towards "Bingen am Rhein" thats where i lived a few years and still work at :D
Hello Martin.
Consider adding holes for alignment pins for all the flywheel disks. This way, concentricity and balancing will be maintained after disassembly and reassembly of flywheel
Machine the holes after laser cutting them.
I'm sure you well already understand the flywheel variability. The answer involves a balance between 2 specific factors:
A) how much time & effort (energy) you wish to have to put into it, to get the flywheel up to operational speed.
B) how much output you wish to have of the flywheel's momentum to both stabilize rotation speed, and also assist in output when it is not being driven.
In short, the answer is mathematical, but the real questions is the current ambiguity of how much work energy the total machine requires to run.
Martin, this feels really good!
So few open variables in your Design.
All variable Parameters adjustable.
man that will work I think
Third time’s the charm they say, still believing in you Martin I hope we see the third machine working in a near future!
Awesome. Suggest using linear slide rails instead of laser cut slots in the square tube. It's much more reliable over time, and the square tube clamping surface will wear and create "indents".
Yes .. also slots dramatically weaken the tubing in the perpendicular plane. The slide doesn't need to be a precision linear bearing, a metal clamp around the tube would be fine
What we're trying to teach martin is that it's cool to experiment but don't redesign the wheel unless necessary and well if you have the time and funding to do so while guaranteeing the part would work. I'm glad.
It's great to finally get excited about a new marble machine again.
Martin are you sure that those clamps are going to have enough clamping force to stop the fly wheel from slipping on the axle.
The blueprint at 6:51 is good enough and will make sense to a machinist, but convention is to place dimensions outside the part whenever possible. for the 50mm a 6:1 scale factor would have been enough, and 4:1 would have been a tad small but probably workable.
@ Martin:
Why are you using sleds to tension the belt? You would need to ensure they are parallel at all times. Why not keep the flywheel adjusted and fixed all the time, and just tension the belt? You could even have it self tensioning when using a strong spring.
Hello! Good to see you again, and can se you building again! It nice to see you proses!
it cant feel good to show a design for something you worked so hard to be torn apart on such a public stage, especially on details like the bearings where what you came up with would have still worked just fine, but just wasnt as proven as other designs. but it shows a lot of character and I have massive respect for you doing it!
Isn't it far easier to control belt tensioning by having a third small wheel that you can move around? That way once the flywheel is balanced, it doesn't have to move anymore.
This is a much better design. I think this will be stronger and more reliable. It also helps that you'll be able to get new bearings and holders if an issue pops up while on tour.
Pay attention to the length of the 20mm shaft between the two pillow blocks. This shaft deflection goes up by the cube of this distance.
All the best! Watching from Vancouver Island, Canada, by a 77 yr. old woman who loves building and making things. Recently an automata with 7 junk journals as waves that rock up and down in wave patterns. If I was younger I’d be working with 3D printer and competing with Ruben Goldberg.
For a long time instead of Wintergatan I read... Wintergarden ! 🏡🌴🌳😁😊
I am relieved
Big Love Stay Awesome Brudda!
THANK YOU SO MUUCH for not going forward with that MADNESS (altough it was pretty ingenious)
I'm curious why the flywheel assembly need to move to tension the belt, instead of the other side of the belt. Is there a reason for that? I'm concerned that an adjustable flywheel (even with those bolts) might effect the balancing or unwanted movement of the flywheel?
Viva la revolución! Great to hear youre listing to the feedback. Feels like MM3 is picking up speed, i have high hopes for this project :)
Martin, thank you for taking everyone's feedback into consideration.
Rubber suspended bearings may help against resonant vibrations caused by the fast spinning flywheel. De Laval used very thin shafts instead, for steam turbines, as the springy elements.
this video sounds top in the right direction
I work on a chaff cutting machine down in Tasmania. We use the exact same setup for our fly wheel which has our cutting blades clamped on it. It also runs all our driving belts, rollers and 2nds elevator too. We run 4 pulleys in sequence from our driver motor. Our biggest issue is getting everything stable as we have a lot of vibration, we go through a lot of grease and oil lol.
3:02 wouldn't that cut off blood flow and hurt like hell tho?
I was lucky to see nr 1 and 2 at Nerdfestival in Belgium this weekend. They look very impressive in real life.
If you're going to tour with this, you want to optimise the flywheel for shipping - that means maximising the rotational inertia while minimising the dead mass. That means a heavy rim with spokes. In order to prevent spoke flex with high rotational inertia, the spokes, whether tensioned wire or solid, should be designed with this in mind. With the current design probably 60%+ of the mass of the flywheel isn't contributing effectively to the rotational inertia.
You bring up a good point-and in addition to the cost of shipping, the thing also needs to be built to handle the rigors of shipping. It will be subject to lots of shocks. It needs to be built to handle a small drop without damage. Minimizing dead weight (especially dead weight suspended off the floor) is key. It’s going to have to go up and down ramps, cross door thresholds and roll over cable raceways. It might have to be lugged up stairs. These margins can be enlarged either by reducing weight or by stiffening the frame or both. But reducing weight has the added benefit of saving Martin’s back (and the backs of his roadies).
*2023 Wintergatan:* "Proven design."
*2025 Wintergatan:* Giant music box with hand crank on side.
May This Old Tony bless this video and machine the parts and create a video in the process. Amen.
Greetings from Wiesbaden!
There is a weird double audio glitch at 2:11. Looking forward to this project and love watching this videos.
i will continue watching every single video.