Heavy Flywheel plays Tighter music! - DAY 11 - Marble Machine Flywheel Prototype
Vložit
- čas přidán 11. 08. 2023
- Support Wintergatan:
- Patreon ► / wintergatan
- CZcams membership ► bit.ly/4cQVM7C
Marble Machine Engineering Discord Server:
/ discord
Video edited By Martin and Hannes from the Trainerds CZcams Channel:
/ trainerds
-
PATREON ► / wintergatan
CZcams MEMBERSHIP ► bit.ly/4cQVM7C
WINTERGATAN RECORDS ► www.wintergatan.net/#/shop
SPOTIFY ► bit.ly/2oKxXWd
ITUNES ► apple.co/2ntWNsZ
ENGINEERING DISCORD ► / discord
COMMUNITY DISCORD ► / discord
-
- Follow the Trainerds adventure: / @trainerds
The prototype is designed to answer the age old question: Can I play tight music using a very strong flywheel? The new flywheel have 20x more moment of inertia compared to the MMX Flywheel.
We are recording at the wonderful location of Siegfrieds Mechanical Music Museum in Rudesheim Am Rhein, Germany. Check out their youtube channel: / @musikkabinett
Edited By the Glorious Hannes Knutsson From the @Trainerds CZcams Channel
Trainerds: / @trainerds
-
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
- - Hudba
Not only does everyone need a friend like Hannes, everyone needs to *be* a friend like Hannes.
+.
he is gentle giant - solid friend group archetype
TT^TT +1
Hannes is sooooo cool!
I lost 80lbs due to illness. I don't think I can be the hannes I once could.
Here's an idea: instead of trying to sync up to the beat, just try to play at a consistent tempo. I feel like focusing on syncing to the perfect beat is causing the oscillation of up and down. If you just play without any reference, it might make the music tighter, but not necessarily the perfect tempo. I don't know if you're prioritizing tempo or tightness, but you can sacrifice perfect tempo for better tightness.
Yes, but I think this comes into conflict with the idea of using the machine in a live/stage setting with a band. There you usually play with a common metronome click, and everybody has to align to that.
However I still agree, that aligning to a fixed click is unnecessary. Maybe just let the machine give the click and everybody else aligns to it.
In order to keep time, everything must be in sync. Every beat, every click, every vibration.
Yeah I think this will be key! Maybe the issue isn’t the machine at all but that there is always human inconsistency in input, even in the best drummers
@@Steiger08 Isn't the fix for this to just not play with a click, and rather have proper monitor sound on stage, so everyone can hear each other and can play together as musicians?
This. Trying to sort out this issue would be a nightmare, and would probably require a PLL of some sort. Either that or a motor that is capable of keeping a very consistent speed, probably a servo.
Also, more moment of inertia ≠ tighter music. More moment of inertia = more energy in the control loop = more energy required to correct errors = lower frequency oscillations if the control system isn't properly tuned. In other words, martin is just more consistent with the petal.
You have a catch 22. The heavier the flywheel, the harder it is to change the speed, so it will stay at a set speed better, but when you need to catch up to another beat, the harder it is to catch up or adjust your speed, and you will have a tendency to overshoot. I think a governor of some sort is needed.
This is what PID control is designed for. Adjusting the D coefficient when adding more weight will compensate for the overshoot.
Over/under shooting isn't an issue with live music. Musicians are rarely on a rock solid tempo in live performances but no one notices because they keep in sync with each other. So the tempo waving around a certain BPM is to be expected anyway. What's important is that it's easy to keep it close.
I was thinking this for some time now. The Flywheel was supposed to smooth out inaccurate timing within stomping the pedal, while for one, the time keeping of many bands is centered around someone stomping a pedal without such aid, and seccondly, it robs martin of the chance to quickly act upon inaccuracies himself. The machine should rather be as _low_ moment of inertia as possible. But how then add a power storage? Is one even needed?
@@thoaxm6687I don’t think it is needed (or it only needs to be tiny - smaller than the MMX flywheel). There was too much friction in the MMX. I would go for a much smaller flywheel and then add a governor in v2 once I had an MVP.
Could use something like a bicycle gear to quickly change tempo.
You should use a speed governor and a freewheel on the output of the flywheel, that way your cranking/pumping is decoupled and doesn’t need to match.
It's definitely a cool idea, but I thought the point was that Martin had to be the one doing the timing, just as playing any other manual instrument requires you to do, and the street organs he's featured in previous videos
He mentioned he wanted this way so he can match the crowds energy with his playing. aka, go slower or fast on a whim, which a governor is made to prevent
@@MegaGadgetdude Yeah, exactly!
An adjustable speed governor could be a perfect solution. I have no clue how they work though, can you explain?
@@Dougieknight With a CVT belt drive controlled by a centrifugal governor, this should be doable. But all in all, I have the feeling that it is not the fault of the mechanism that there is a deviation, but rather a systematic error. Of course, as the flywheel mass gets larger and larger, the pulsating input torque of the pedal gets more and more damped. This should make it easier to maintain a speed - at the same time, of course, the system becomes more and more lethargic, so it can no longer respond quickly to tempo changes. I am not a musician - but what is the tightness of human musicians anyway? Is there such a thing as tightness muscle memory? 🙂
This is not at all an exercise in how tight music can be played on this machine. It's an exercise in how Martin can match an arbitrary spot on a wheel for a few spins before it inevitably goes out of sync again.
Either, after dozens of comments, he keeps confusing bpm with synching, or we miss something. It is beyond me that he does not simply use an rpm gun as a realtime guide on whether to slow down or speed up, ditching the unecessary, useless and complicated software analysis that reality is completely based on biased flawed (trying to always sync) user input that neglects what is important: stable rpm...
To me it seems like he is just testing the physical mechanics of the operator trying to keep a consistent temp. This makes sense to me from a user perspective. I do think adding a speed governor and a gear cassette to shift between predefined tempos (as well as a "free" run tempo mode") would be much better solution, if the goal is to satisfy the "must have" requirement of tighter music - which given the context includes matching tempo perfectly
It's beyond my comprehension. MM could play tight music with a sloppy rpm. Tightness is in parallel events lag or I miss something? It's like a real drummer with a bass player could be super tight but they don't stick to perfect consistent bpm.
At first glance you might think that his test is simply a measurement of whether "HE" can hit the right beat. However, what Martin is testing is the consistency of whether the flywheel can consistently match HIS rhythm. Think of how practiced Martin is at playing music. HE wants to be able to play tight music without worrying about whether the machine will copy his beat. Essentially, its an input/output test. If I put in the right amount of force, do I get the right tempo out of it consistently? That's what's being tested here.
Needs a governor then
Since you are planning on going on tour with this thing at some point, I would suggest increasing the radius of the fly wheel instead of the mass. That way the machine can be assembled with regular man power and not a forklift or Hannes.
Yesterday you commented that the hand-crank felt "tighter" than the pedal but actually objectively wasn't.
Today I noticed something that might explain your perception: the timing of the click.
When handcranking, you're clicking when doing the most physical effort, yet while stepping, the click seems to be when you are stepping off -- ie when you've just *reduced* the effort. What might be worth experimenting with is rotating thetiming wheel andjust experimenting with what feels good to you.
REALLY good point. Timing doesnt land on the foot stomp, it ends in the middle of the leg. The midpoint of the leg is harder to time.
@@mrlithium69 Exactly. However, the problem is more nuanced, because there's no straightforward way to land the timing on the foot stomp, because the pedal moves in a sinusoidal motion, so is very slow when it starts moving down, and is therefore not accurate.
Players of stringed instruments need to start their bowing motion slightly ahead of the beat so that they have time for the note to reach full volume on the beat. Compare with a piano player (or even a xylophonist!) who has an instantaneous response and has to hit the note absolutely bang on time. The exact offset for the treadle to feel right is something that absolutely needs trial and error.
That flywheel is both terrifying and impressive. I'm scared of it, actually, but I can't look away!
Reminds me of Adam Savage's Panjandrum :D
I'm more terrified - or rather, I have the utmost respect for having and machining such a lump of steel on a large lathe. I wouldn't want to see a spinning 50 kg steel disk whizzing through the shop - not to mention a face plate with workpiece doing the same...!
Oh, no thank you, sir. lol @@JonaasK
Ey… ey. Stick yer hand in there.
@@JonaasK Brings a new meaning to the word "flywheel" - now I can't unsee it.
Can't wait to see the fork lift come on stage to install the flywheel on your world tour
A car engine hoist would do. They can take 1000kg
Watching this new creation become the most unrealistic contraption for touring. We’ll see it reach 95% completion when Martin realizes it didn't check his first box (tour-ability) from the first step (center frame & flywheel) so he abandons it completely to start another marble machine.
Poor man has chased perfection so hard he sees it as flawed when he finally nears it.....
I can see 1 just coming on stage to do some donuts
This is absolutely the flaw in this build.
I wish he’d realize that what made the first marble machine build amazing to watch was its imperfections and how it still worked despite them. His mad rush to keep it all moving was the part of the performance that made the music all the more special.
If we wanted perfection we’d watch a CD player.
@@dwightpries8330While I do agree to some extent, I also have to say that while the MMX is an amazing contraption it wouldn't have done the job. They got it to play, but very slowly+ it had to be repaired constantly. It needs to be reliable for touring. And let's not forget they're still prototypes
From day one, which is quite a few years ago now, I've been so charmed by Wintergatan's natural, almost organic musical style... This is why I've been very surprised by Martin's obsession with the quantisation requirements he has been searching for ever since half-way through the MMX. Surely, anything plus or minus 2 milliseconds is ok? Watching the tests with this prototype confirms once again that 'tight' music will not come from your muscles, Martin. The tightness you're looking for will require a mechanistic, electronic approach. I'm ok with that, actually, and I'm a little surprised you're still trying to pedal your way through.
He's addressed this many times. The goal is to be as precise and tight as possible to create a robust machine that is usable in as many settings as possible. The precision needs to be there or martin cannot trust the machine. He did not design the MMX this way and spent years chasing his own tail.
It has nothing to do with the relative 'naturalistic' or quantization of his musical style.
@@awogbob At the moment the issue with 'tightness' is not on the machine side, but on Martin's ability to crank.
@@VulpeculaJoy at the moment, the issue with tightness is with the machine reacting (by design!) very slowly to Martins impulses. He should try a flywheel just large enough to mimic the rest of the machine (I reckon, about 1/10th of the current) and fiddle with lever ratios of the pedal/crank, then increasing the inertia of the flywheel. I wouldn't be surprised if the outcome of this would be that the flywheel has negative affects on timing after a certain size.
@@thoaxm6687 This whole thing of matching speed with a click track is essentially a feed back loop. Martin is the controller, the flywheel is the system. Martin is trying to tune the whole feedback loop by making changes to the inertia of the system.
@@VulpeculaJoy yes. analog to a PID controller he's making the "I" portion heavier.
I noticed how the slip bushings were actually giving way during cranking. They will suffice for testing, but critical shafts like the input shaft and linkages should be keyed.
He should check if the bushings are on spec (able to handle the specific forces of acceleration and inertia) and if the torque applied to the bushing bolts is on spec too, with a torque tool. It may just be an issue with under tighting.
@@constantinosschinas4503 If you overtighten, you risk deformation.
@@SoranoGuardias yes, that is why i said on spec.
If you really want precise music, you can use a mechanical speed governor linked to a clutch, like old record players had. Would look very steampunk with those two rotating orbs.
Though I think you should just set the spee by ear, the original marble machine worked fine in that regard. No-one was like "oh the music isn't exactly x bpm"
Martin: "This is the whole idea of being in Germany at all!"
What Germans hear: "Let's overengineer the crap out of it!"
Given the choice between making something of higher quality and making something more complex, Germans will choose complexity every time. The entire country is shitty Rube Goldberg machines
How are we so priveleged to see another marble machine video so soon? Thank you Martin.
+.
That flywheel design is legitimately scary. There's so much power and force behind that thing's rotational mass.
I'm concerned how hard it is to start up and stop. He may put two sided brakes like car brakes
I'm a mechanical engineering professor and I refer my students to these videos because they are a fantastic example of how to approach mechanism design
+.
Also a fantastic example of progressively eliminating unnecessary design requirements. He's slowly but surely moved towards simpler more robust designs, and I have a feeling he's going to eventually eliminate either the self-powered requirement or the extremely tight music requirement.
Are they ? Cause so far it's been 6 years and the project just keep getting abandoned, delivering nothing promised. In a company you'd get fired quickly lol
@@TigreDemon This kind of shows the difference between the academic world and industry. If your goal is to develop a new idea, it often helps to "play" around for quite a while before eventually discovering a solution. If you have to earn money by developing and selling a working product, you can't experiment that much but will heavily rely on known and proven designs.
If this project is a good example or not depends on what you want to teach the students. If it is the creative process, then it fits (although it is not a good example for scientific methods as the way Martin measures and evaluates different things has been sketchy lots of times in the past). If it is "how to get things done" it is definitely not :P
@@TigreDemon it hasn't been 6 years on this machine, MMX is abandoned and this is a whole new machine.
If you add a second smaller flywheel controlled by an electronic clutch you can get a perfect tempo. If the power flywheel spins faster the clutch disengages and slows the music playing flywheel, and vice versa
the idea is it to be 100% analog and have no electronics, from what i've gathered
@@imblackmagic1209 That's where the centrifugal governor comes in. Virtually every music box in history has either a wind-friction governor or a centrifugal clutch for this reason. Completely analog.
@@frollard Centrifugal clutch indeed. It would be wonderful to see a marble machine running literally balls-out!
@@frollardI wonder if he'll try the governor he designed while he's in Germany
Suggestion: Add an "air scoop" wheel to the setup. Now the flywheel isn't constantly providing any energy to the marble machine. In the 'real' machine you'll have to put in more energy into the flywheel to keep the same tempo and this can have a huge impact. Keep up the good work!!!
That's a great point! the machine will slow down a lot faster than this when it's being used. That might make it easier to match the speed, too, since it'll be easier for it to slow down when overshooting the tempo.
The crank is tighter than the pedal because you can input power more continuously throughout every 360 degree turn, whereas the pedal can only input power 180 degrees out of 360, sapping power from the flywheel every half revolution, which introduces an inherent energy oscillation in the system. Some sort of bicycle-type crank mechanism would probably give you the best of both worlds: more power from legs and steadier 360 degree power input.
With a bike mechanism, he could also set it up so a semi-turn would correspond to a quarter note. So to get to 140 bpm, for instance, that would be 70 rotations per minute on the bike. For each beat, that's a foot down, so it's still intuitive.
Just be careful with that wheel. It seams really powerful and if you add more you should look safety first and then ergonomics. I love this project!
The wheel is fine. There's no way it's going to escape the bar even if the things catastrophically blow up. In a production machine, yeah you'd want extra protection but that's more to save things falling into the wheel. The way they're testing though, there's very little risk of anything snagging in the wheel.
I agree. I was more afraid of getting his hands or clothes to the moving parts cause it won't gonna stop if it finds resistant. When the machine start to get more complex he should be more careful. Mainly because yesterday he hit his hand
Take also care to resonance effect. Check (if it's not already done) thé proper mode of thé all système. If you exited it a thé wrong fréquence, you Can destroy everything
We need to add a PID algorithm to Martin’s next software update. 😂
Love seeing your videos, Martin!! I know it's gotta be so much extra work to make these, so I just want you and Hannes to know how grateful I am you let us come along on this journey. It's always incredible watching an artist chase his passion, and yours is so unique and wondrous.
Martin I love you and your content, but I think you’ve gone crazy
this is the case from the start, but especially at the middle/last phase of MMX.
my guy could be writing the fourth wintergatan album by now...
I appreciate you doing this series on putting together the prototype for the next marble machine. I've been enjoying your content since your original big marble machine video. Through the MMX and till today. I've always enjoyed your content and its just nice to have more of it.
+.
Always excited for a new video!
Another thing: testing against an already playing, electronic beat will always be worse than just having the MM3 *be* the beat and set a slightly variable BPM in concert. It sounds very tight already, definitely good enough as it is for a live performance, and it will sound even better once Martin doesn’t have to be constantly chasing after an already playing click track (I’m guessing he could and should just install a real-time, numerical beat counter and keep an eye on it). The band will just have to learn to play along with it.
I've been enjoying these videos. I like the level of precision you are working towards.
have you considered how a clock or watch works? assuming proper gearing (think about the Lathe you used to turn the flywheel) could give you the precision and variability in a far more accurate manner than the current setup gives.
Variable shaped Turbillons for different tempos 😅
I love what you're doing and where this project has come from. Nicely done!
Would be interesting to apply this software and measure the tightness of a good drummer playing with a good band live. I would assume that it’s not much tighter than this, but I also might be terribly wrong 😂 to me, the heavy flighwheel definitely sounds tight enough for a world tour because all the other musicians are not machines and can therefore easily adjust as the heavy mass doesn’t change speed to quickly!
humans are at 5-15 ms "tightness". we play early mostly. In a live band each foot you hear sound 1ms later then it was played. with how big the MMX2 is sound from one side will travel to you 12ms delayed then the closet side to you. the heavy fly wheel is unnecessary. The ability the change tempos quickly is more beneficial. Like to quickly sync to a click track Vs slowly sync.
@@BensMiniToons oh that’s an interesting thought! Thank you! So you would say that he should use less mass? Or would you recommend a different power input?
@@markuskaiengel the flywheel only needs to have the mass to smooth out the machine once built. The big rotating programming wheels may have enough mass that a flywheel is redundant. Turning a wooden handle with out flywheel could have perfect tempo with practice.
I can draw circles in the air around my ear with my finger with his click track perfect. Try it your self while watching the test. Remember you have no flywheel. Just drawing circles in the air around your ear with your finger.
Whichever is "better", I think your tour version should include both a crank and pedal. Over the course of a show, you will get fatigued. Having the ability to switch between the two input methods will be a huge advantage, even if one is slightly less than optimum. I also second the idea of @Hrotti, and think you need one or more clutch mechanisms attached to each input on the tour version. One manual clutch so you can disconnect the pedal or crank when you are just using the other, and one freewheel or other one-way clutch (like on a bicycle) that prevents the flywheel from "driving" the pedal or crank with any significant force.
i honestly feel like the huygen drive would be a perfect blend between constant power output, but with the changing speeds of the crank/pedal
The hand crank plays tighter for several reasons, but one of them is that both positive and negative power input is possible 100% of the time, while with pedal it's much more restricted - when pedal goes up it's possible to only add negative power (to slow it down), when pedal goes down - only positive (speed up), and in end positions the power transfer is zero, making for a much more uneven power input.
Hannes the forklift is in da house! The collab with Hannes The Crane was smooth.
The crank might be getting tighter results because you can "see" the beat with the dial movement. The pedal is more of a linear movement but with it slowing as it reaches either end so it might be harder to match.
Time to turn it into a pedal marble machine instead of a pump
Can a centrifugal clutch be used to keep the machine spinning at a specific RPM? Depending on how accurate they are when disengaging and engaging the power, it might be a simple way to control the at what BPM you are running. But a problem would be that you might need to switch the clutch every time you want to run at a different BPM.
Keep it up! Let's gooooo!
I believe the control you have with the crank is higher than with the pedal, as your hands might react quicker to the feedback than your feet. You are able to increase or decrease your input faster and more accurately
doesn't make a difference
Tight music or not the fact is that the mm3 is the drummer which is important to keep in mind. An adaptive gear would be an interesting addition to this
Trainerds is awesome, Hannes should be so proud of what he's created. His videos gave me a lot of positivity to help me get through the down time I had when I grade 1 tore my A2 pulley
Hannes 3500 invoking his inner Thor.
Martin, the crack axle is bending when you turn the crank.
Looking forward to seeing the governor added.
What if the flywheel simply charges a dynamo that powers an electric motor? That way you remain engaged with the machine, it plays perfectly tight, and you can relax and enjoy playing the instruments
Or you could power electronic synth or even PC!
Thx for keeping the magic!
I believe!!
Great work Martin! So happy to see it coming together! Just a note, experimental evidence requires repetition! It is not enough to test once and call it a day, do multiple tests with each configuration and average the results for more thorough conclusions :)
You can see it is better at 120bpm as Martin is getting a bit of a groove on!
Feels like the faster you go more the bpm is stable, this is giving me hope again 😮
The faster it goes the more energy it has and the more resistant to change it becomes.
@@AnonyMous-jf4lc ke=1/2mv^2, boi howdy that square is op
@@AnonyMous-jf4lc yes of course simple law of physics, what I ment is that it’s giving me hope again for the music tightness perhaps he can achieve tight music with more speed of the flywheel and tweaking the gear ratio, and as up to now the standard deviations results seems promising !😉
Hannes the Dream Bear... Woofs!
Tomorrow he's gonna have 20 manhole covers on this rig
It might be useful to differentiate 1-handed from 2-handed cranking, as well as considering what that means regarding additional instrument manipulation limits.
Also, a bicycle pedal input could be more similar to the crank, especially with a toe loop. Following that, a similarly offset 2 handle hand crank vs 2 bike pedals.
Bicycles often come with a clutch and derailleur. The one-way clutch may be good for safety, and the derailleur might be helpful when charging and changing tempo.
As the weight and gear ratio further favour the flywheel, getting up to speed and changing tempo could become more difficult and time consuming during a live show.
Some covers, even if they are clear plexiglass, for the moving parts that shouldn't be touched, may become even more of an important addition as you increase MoI.
A remotely triggered emergency stop could be good, if a loose shirt, cargo pocket, or headphone wire gets caught. Good luck, Martin, Hannes, and the rest of the team!
Martin, don't forget that this is only the flywheel, without the friction and energy bleed of the rest of the mechanism; you will need to work way harder than this to keep up the tempo, so you might need an electric motor to assist? A roller cam clutch/freewheel on your input will enable you to interrupt cranking. Also consider; the faster the flywheel rotation, the greater the chances of vibration, which may compromise the operation of the ball escapements, for example. But great results from these tests!
I think I agree with everyone else. The machines ability to be tight is not the same as the machines ability to match a metronome.
I know at the beginning you were debating having a click track and I think this core question should be revisited
He could use the metronome to get to the target bpm, then turn off the drum track when he think's he's up to tempo. then just try to keep time to ear.
Then measure deviation of intervals between subsequent beats.
But i think he needs to set a clear target of some sort - so that his experiment has at least a fixed start point.
Interesting to benchmark vs a decent human drummer doing the same test as a control . . . .?
Hey Martyn! I really do think you should go with a hybrid solution with a weight-based clock actually moving the machine, but the pedal keeps the weight off the ground. I would definitely start each song with the weight all the way up as well, so that the inertia is already built when you put your foot down.
This way, the consistency of your foot doesn’t matter to timing, but it is still you who is creating the power and driving the machine. As long as your foot power is greater than or equal to the weight falling, you’ll have consistent music.
I agree and can't see what the downsides would be other than adding more weight to the machine. Should be possible to rig a system which allows for quick changing of the weight also.
It makes good sense. More mass in the flywheel means higher angular momentum, which means the wheel will accelerate/decelerate more slowly, so the BPM will change more slowly.
It will be interesting to see how the addition of the instruments and all of those linkages will affect the tightness of the music.
This reminds me sooo much of when I used to DJ back in the day on Vinyl decks! xD
You should add a visual cue for the sync. Like a disk that turns with the arrow with holes in it and a pattern on the circular board. ✳︎
Yes or like a string of lights, middel is on beat lef is under right is over. Kind of like a rev meter. That way you can see much quicker if you are giving it too much power or not enough
Day 2 of asking Martin to put a display that reads out the machines tempo live.
Imagine designing a whole live show all centered around how long it takes the flywheel to arrive at the correct bpm!!!
Another thing, you can see that you're getting "lag" in the system when you add power, you can see that the belt is sagging on the top when you press down on the pedal. This means that no matter how smooth you're trying to add power to the system, the rubber belt is causing you issues.
Hydraulic Bike disk brake to stop this crazy fast death wheel will be absolutely necessary. But I'm glad to see Martin happy with the tightness, it means the project may keep going!
really love you can even see the sound input on the preamp flashing green (I've got the solo)
Have you considered a sprag clutch bearing to transmit power to the flywheel?
It’s an off-the-shelf one way clutch that acts as a roller bearing.
It would certainly add a safety measure by allowing the flywheel to rotate freely when not being driven, whilst allowing the driving mechanism to be stationary.
When driven at any greater rate than the load is rotating it transmits 100% of the input power.
The final marble machine is going to apply angular momentum on the flywheel.
Now you're testing the flywheel under no load condition. You should really test how this system will perform with load applied to the flywheel.
You can try simulating the load with a weighted break or even better find a way to apply varying angular momentum.
I love this project and I wish you Martin best luck on your journey.
bro if the "Marble Machine 3" video with a cool song comes out, it will trending so hard!!!
"where do you go?"
"world tour, yes"
Martin, have you considered to play around with the pedal height position vs the downbeat of the click track? If your prototype is set up to match the click track when the pedal is in its lowest position (which from a firts glance makes sense to mimic eg the kick pedal of a drumset), that’s however where you have the worst spacial resolution in your prototype due to the way the mechanical connection is made between your pedal and the flywheel. You see, I think it follows a sinewave, and the top and bottom positions are where you have that lowest spacial resolution. So with this setup you want your click to instead be set on the halfway down (or halfway up ) travel position to have the best resolution and give you the best chance to play on time, ie where the derivative of the sine curve is the steepest.
Kind of difficult to explain my thoughts in the comments section, would be easier to draw a schematic. But you probably get my point. In my opinion these details are the most limiting parts right now (although I don’t know exactly how you have matched your pedal position vs the click right now, I must admit).
You should even think of a second gear ratio just for the hand crank (which I thought should be twice as fast as now -> you could feel the off-beat more -> more precise for slower rythmn)
Proof that it is harder to play slower than to play faster.
You should consider using a “oneway bearing” on the hand crank, so when it’s free-wheeling its not swinging around dangerously :)
We use them a lot in RC helicopters so that we can AutoRotate when the engine or motor is stopped.
This machine isn't finding its own synchronized tightness, it's finding _yours_ and responding in kind.
As you find equilibrium in the forces you apply to either pedal or crank, the machine responds with tighter music.
You realllllly need that governor to make it so much tighter
So great to see those results,
Would you add more discs in the future ?
Now that's some heavy metal
I am starting to get very concerned for the safety of the Operators. If you increase speed on an already fast spinning heavy flywheel, the increase of kinetic energy may be catastrophic
just dont let any kids nearby, they are fascinated by spinning gears and will be curious to touch something and get mangled.
Adults? put a sign and a guardrail, and make sure its not gonna fall apart. thats as best you can do.
@@mrlithium69 It's irresponsible to design a machine that is INTENDED to draw an audience closer (to watch all the marble chase around), but yet also make it a death-trap for anyone that get's close.
If Martin wanted to test consistent tempo but not use electric motor(maybe design is pending addition of it), he should do a rotation deviation check.
It consist of rotating input to x amount of revolution of the output and see if the output rotation deviates from input.
Usually belt slippage in their use case(because a smooth belt is used) and belt stretching/tension can cause deviation to accumulate over time.
I think it is not time to say it but if electrical motor is added, feedback sensor to detect rpm and PID would be a good addition so Martin can just set a rpm and the Marble Machine will be able to quickly ramp motor output to reach correct tempo in short time.
Its nice that you found a friendly giant to help you lift
@Martin, two things I noticed which I think had an effect on the consistency of your timing: 1) the second half of your flywheel I believe you had the opportunity to machine to correct specifications, that is to say, you lathed it after correcting for the loose chuck, which means you probably had the second flywheel half being more properly aligned than the first. When you bolted the second half to the first this probably had a corrective effect to the misalignment of the first half. There was barely any noticable wobble to the flywheel in this test. 2) I believe you mentioned this in a previous video, but the axles for the flywheel need to be stronger. Whereas you corrected the flywheel wobble, the axles began to noticably wobble once you got up to speed with the heavier weight on the system. Rebuilding the flywheel with proper alignment and increasing the strength or diameter of the axles should help your machine perform better. Lastly, if manual effort is a concern, consider replacing the mechanical pedal with an electronic one powering a motor which actuates the flywheel. This will allow you the same foot control while leaving you free to manipulate the remainder of the machine. If the marble machine needs power, you need simply step on the pedal till the necessary energy is back in the system. You had a similar motor actuator in the MMX and I believe this gave you a measure of autonomy while the MMX ran it's program, allowing you to accompany it by playing other instruments.
I wonder if a dual mass flywheel would be enough to iron out the the power pulses of the crank or pedal.
An outside the box thought here: Wouldn't it be good for the whole stability of the machine / fly wheel if you would have each half of the flywheel on each side of the machine ? That shaft would be under less torque. Can't help but to imagine gym stepper machine being attached to the marble machine, that would probably make the powering half as tiring, and probably synchronising to the beat easier.
One thing I consistently notice is that you naturally pump the pedal in line with the beats, which often creates a tempo different from what is intended. Meanwhile, the crank allows your motion to always sync up perfectly with the beats. More so than they flywheel, I think the natural tendency to follow the beat is the issue. There's got to be a way to make it so that the pedal will sync up with the intended BPM, perhaps a transmission or even CVT (continuously variable transmission) connection between the pedal and the main set of shafts to allow you to achieve this.
Lets go.
I pressed the Like button before even pushing Play. That’s how strong my Wintergatan blood flows.
Who needs a forklift when you have an editor?
I recently was recommended the wintergatan video on CZcams The Prototype is finished day 10. I thought I might recommend the potential to include a simple freewheel mechanism into the crank arm to increase the safety margin.This would allow the power train to accept input only in one direction and freely ratchet in the opposite direction of applied force or “coast”. I think the easiest way would be freewheel directly mounted to crank arm shaft and a crank arm made of wood designed to clip directly to it. All in all it might cost $20 and an hour or two if an adapter is needed. Depending on implementation this could also allow for simple gearing reductions by using off the shelf bicycle components with little efficiency loss. I would chose a belt drive system as it could offer even more potential safety margin over use of chain drives. Whether or not this information is useful to your project or overlooked entirely I wish you luck in your ventures and success.
Inertia is essentially an objects resistance to rotation around a defined axis. In the case of your flywheels (that axis goes through their center). Inertia of a wheel is only linearly related to mass but is defined by the square of the radius. So, easier to get more inertia with less mass by increasing the wheel diameter ( but that of course makes machining and material costs go up).
Since inertia is resistance to a change in rotation around an axis, and your goal is to maintain a certain RPM, then the more inertia your flywheel has the tighter it will play.
Can I suggest a test for the effect of the wheel on keeping the RPMs. Run various BPM and measure the decay (ideal time minus actual time) and measure it over time. The measurements of the decay start when you have the timing and release all forces from pedal or crank on the system and let it run.
The data is collected for multiple starting BPMs and collected over 5-10 times for each BPM. The average decay (assuming linearity over time) for all BPMs are calculated and you then plot the Average decay for all BPMs. Use BPMs in intervals of 5.
It’s a lot of work and test, but you should get a nice graph that tells at what BPMs the systems is playing tightest. If the graph shows same avarage decay (flat line) it would be same tightness for all BPMs (it would be surprising to say the least).
Hi you can opimize the weight of the flywheel by the design. You can make it half the wheight and still store the same energy…
The mass of the flywheel need to be as far from center as possible, make the fw center as light Possible and the weight gatered. as far out as possible
I wonder if the disconnect between when the petal is pushed and when the timer registers the beat is throwing Martin off. I know I’d find it almost impossible to keep time if I tapped my foot to the beat, but there was a delay before hearing my foot tapping the ground. Might also explain why it’s so much easier to keep time with the handle as opposed to the petal.
Regardless, great work Martin! Keep it up.
*crank, not handle. I’m guessing the variable delay seen with the petal is due to elastic losses in the mechanism (I.e. the petal flexing and not transferring all the energy). Making the entire petal mechanism much stiffer might help!
The pedals lowest point has been calibrated to match the tap of the mic so should hopefully mitigate this
Martin for your final version you could add decorations to the flywheel to make it look cooler spinning, not that it needs it. Maybe some little wilsons or something
Keep in mind that this is the minimum amount of energie you will have to spend. Adding all the mechanisms of the machine will make it only harder to turn. And you have to spend that amount for at least a song, ideally for a whole concert. So not only think about the tightness of hand vs. foot, think also about sustaining that level of energie.
Also, how do you plan on engage the programming wheel? Is it constantly powered and you have to wait until it has been turned to the start of the programm to beginn the song. Or can you engage it when you have enough power in the system, but then loose some of that power and therefore the achieved BBMs.
Add speedometer !!! Set the speed of the crank for each BPM on the speedometer it will be much more simple to get the perfect speed of the wheel for the operator. You will be able to focus more on the music.
Nice job so far ❤
That would work. RPM to BPM he would have a gage to read and a much lighter flywheel.
would it be possible to keep the input constant and maybe use gears to change tempo?
I want to see the hyugen drive and governer. :) Love your work!
My knee hurts just listening to this🥵
that shaft was in pain before, now I think it is screaming for help. xD
A benefit of the crank is that you can adjust the power added through the full revolution of the crank whereas with the pedal you are really only adjusting power on the down stroke. You could probably increase the level of control of the pedal by adding a toe cup so that you can also control the pedal on the up stroke.
Big concern is to be able to keep that momentum up during playing the rest of the machine as well as the fact that as you turn the crank your slightly accelerating and decelerating the crank which might affect the tightness of all the other beats either side of the crochet.
I don't know if anyone has pointed this out (I don't typically read the comments) but syncing up with the metronome will actually be more difficult than getting back into rhythm without it. Without it you just need to get the frequency back to 80BPM - with it, if you fall behind, you actually need to first go faster than 80bpm to catch up, and then slow back down to match it.
It seems like it would be interesting to have some kind of wind-up or hyugen drive that lets you break the energy of the flywheel into either a spring or to lifting a weight, then you can let that weight fall to restart the device.
A bonus would be that you can use more or less weight to roughly control the tempo you accelerate to