Is Gravity Really Constant? I Built this machine to find out.

Support Wintergatan:
Patreon ► / wintergatan
KZbin membership ► bit.ly/4cQVM7C
Marble Machine Engineering Discord Server:
/ discord
Video edited By Martin and Hannes from the Trainerds KZbin Channel:
/ trainerds
-
PATREON ► / wintergatan
KZbin MEMBERSHIP ► bit.ly/4cQVM7C
WINTERGATAN RECORDS ► www.wintergatan...
SPOTIFY ► bit.ly/2oKxXWd
ITUNES ► apple.co/2ntWNsZ
ENGINEERING DISCORD ► / discord
COMMUNITY DISCORD ► / discord
-
Yes, yes it is.
Huge Thanks to Tom and Yan for the www.tightinator.fun tool!
This tool has really been such a help in my process to find a way to accurately rotate a shaft using mechanics only.
The Tightinator is measuring the tightness of transients by measuring the standard deviation of a set number of consecutive transients. When we are looking at the "Best 300" its the tightest 300 consecutive transients of an audio recording we are measuring.
-
MUSIC DOWNLOADS ► wintergatan.ba...
WINTERGATAN RECORDS ► www.wintergatan...
SPOTIFY ► bit.ly/2oKxXWd
ITUNES ► apple.co/2ntWNsZ
MERCH ► teespring.com/...
DISCORD ► / discord
SECOND OFFICIAL CHANNEL:► / wintergatan2021
-

[@craftedbyorre]

Martin is becoming a 17th century physicist, and I’m all here for it.

[@alexbennie]

17th century scientist WHITH access to a 3D Printer! This journey has been awesome so far!

[@TheGreatAtario]

Some of me isn't here for it, like my left knee

[@Pootie_Tang]

​@@TheGreatAtarioHey, maaaan, that's my knee, get you hand off me! 😡 😂

[@djb5320]

in a few hundred years, he'll realize that if he wants perfect timing, he'll have to invent some kind of digital technology

[@Pasci234]

He is what Newton whould have bin, if he had a Calculator and Access to Redit.

[@antoineboillet2288]

Hi Martin, here are my 2 cents from a Mechanical engineer's point of view. Don't forget that your driving mechanism will not only have to be tight in a constant load condition (ex: a tuned watch, or air fan at equilibrium), but will also have to play tight with highly variable load as the marble machine plays. If you are not controlling the speed and power input by direct human power, you will need a very reactive control system to maintain constant speed under variable load especially if this system is not compatible with high moment of inertia. You don't want it to end up with more variation than with a high moment of inertia human-driven system.

[@thistemba]

Very good point

[@JorisKofman]

Agreed he will want a governor that actually governors rather than an air brake

[@NeonNijahn]

​@@JorisKofmanI also think he could use a frictionless magnetic breaking system instead like brake fins on roller coasters so that it doesn't wear down over time.

[@SoraHjort]

Also to add on: Martin, you also did the test on a scaled down version, and tested it against a fullsized unit. There is going to be differences between the scales. And for an example that is a similar problem, is that people have had with battlebots between the beetleweight bots being scaled up to full size, where they would run into problems where the new size and weight cause their own problems. So just because something works at the small test scale doesn't mean it'll perform as well at full scale. Different stresses, different energy required to be put in, and so on.

[@irwinkevincarpio]

Exactly what I have been thinking. All this test proves is that steady input (force of gravity) vs a variable input (human power) naturally proves the constancy of gravity. This is not a test to prove how tight you can play music with a variable load. @Martin you did summarize this basically at the end, however, I feel you might have intentionally glossed over the scale of the weight and flywheel you need to build for the marble machine. You will need a flywheel with a heavy enough moment of inertia to keep music tight whether you have 36 channels all playing at once or just 1. You only tested the energy source when you compared them in the sheet. My hypothesis is you will arrive at the conclusion that you will need the flywheel you built in Germany and couple it to a Huygen drive (if that is your chosen method of propulsion) and to a speed governor as your "engine" of sorts. Then it will connect to a clutch and gearbox to output to the machine so that you can select different BPMs without having to change the revolutions per minute of your flywheel.

[@jcagle31337]

Martin - "No more scope creep" Also Martin- "Let's reprove science"

[@_ata_3]

It's not scope creep it's going back to the roots!! 😂

[@brianbaker2759]

No, it’s definitely scope creep.

[@thatguyStrike]

i've given up hope, i'm just here to see what happens.

[@_ata_3]

No, you are just cheaply promoting your junk channel.

[@thomasbecker9676]

@@thatguyStrike I predict this version of the machine being abandoned once views start tapering off, and a new new one being started. Rinse and repeat.

[@AsbjornOlling]

Martin: I need to simplify and reduce scope, so I can have a marble machine that actually works. Also martin: let me design a mechanical instrument to check if gravity works.

[@Rocknoob49]

wait this is still about the marble machine? I feel like i started watching him build that 10 years ago :'D

[@michaelmauch7914]

And ALL these crazy awesome accurate things he builds... but that wind resistance??😂

[@some_other_guy]

Trust but verify

[@lesalmin]

What you have built, is basically an old time weight driven wall clock, so it's not a big surprise it's tight as a clock. 😃

[@brucehooke7535]

Except that the critical part of a weight-driven clock that's missing here is a pendulum and escapement. That's the very precise speed governor on a typical weight-driven clock. I suspect a key engineering challenge still to be figured out with this idea is how to create a practical governor.

[@LinusNil]

@@brucehooke7535 Indeed. The force caused by the mass of the suitcase along with the acceleration of gravity (g) will lead to a constant force pulling on the chain, but the speed is very dependent on other factors where friction is one of them. Without a governor, this will be very unstable depending on varying parameters such as friction and load (from the musical machine).

[@waffle8364]

depends on the gear ratios

[@CyclesAreSingularities]

@@user-ip3zi5mb6l XD good one

[@tylerchiu7065]

@@brucehooke7535that is the purpose of the cardboard flaps on the back, used as an air governor

[@dalegray934]

Fast forward 10 years and Martin is experimenting with atomic clocks to create the ultimate tight music!

[@markgriz]

and that world tour will be right around the corner

[@alanhilder1883]

But there will still be that 0.000000000000001 msec variation, can he make it any tighter? ( I don't think I put enough 0's in there )

[@valiakosilla2413]

He has to invent world's first mechanical atomic clock.

[@alanhilder1883]

@@valiakosilla2413 I read this and all I could think about was in an extremely old movie ( called young Einstein ) he was splitting a beer atom ( to make it frothy ) with a hammer and chisel.

[@Oxtorayk]

just wanted a new album :(

[@mattcy6591]

Oh no now his music wont be tight because of gravitational waves.

[@benr1286]

Lol

[@LordDragox412]

Don't worry, I'm sure he'll solve the issue with the next MM. This time, it will be put outside of Earth's gravity! It will include an artificial gravity generator to have true constant, uniform gravity. Tight music shall be achieved, one way or another, in this life or the next one!

[@Avetho]

@@LordDragox412 But then what about quantum fluctuations causing the mass and energy of the system to flutter minutely?

[@mattcy6591]

​@@Avethois that greater or less than blackholes colliding 13 billion light years away?

[@Avetho]

@@mattcy6591 It has a little less of an effect, but once the gravitational wave inaccuracy problem is taken care of, it'll become noticeable, a dozen femtoseconds of variation is too loose for proper modern space-age music, not like the ultra archaic "techno" or "dubstep" or "orchestrion" music types.

[@evanramee796]

I would love to see you use bike drive chain components, like cassettes and derailleurs, to change tempos. They're readily available with a wide range of gear ratios.

[@EV-wp1fj]

The freewheel for the ratcheting, the derailleur for chain tensioning, and a cassette (+ derailleur) to select gears. He's practically there. But just lose the governor and build a proper clock escapement.

[@AlejandroCaicedoPUJ]

That wouldn't allow for gradual tempo changes tho I'm very curious of how he's gonna handle that problem

[@WispyFrost371]

@@AlejandroCaicedoPUJ CVTs!

[@RikKoedoot]

He'd have to build in a kind of clutch, as derailleurs don't like shifting under full power, though I'd love to see this happen aswell

[@Phriedah]

@@AlejandroCaicedoPUJi think one simple way to handle tempo is directly by changing the programming wheel. Same rpm for all songs, but some songs have more/less beats per rpm

[@litterbox019]

can't wait for the one where he truly builds a marble machine from scratch, by first reinventing the universe

[@nanamacapagal8342]

[You feel Carl Sagan smiling from afar.]

[@fijnman3813]

7:00 Martin casually making a graph in CAD as an easy solution just explains a lot haha

[@DoodleChaos]

Anybody else chuckle when they read the title? I love Martin's contraptions

[@colorscream]

Wait till he really goes into space with the MM3.

[@OtherTheDave]

I almost dismissed it as clickbait until I saw who posted it and realized what it must be about. Edit: I was thinking about G, not the strength of the gravitational attraction between two things, which varies with their masses and the distance between them.

[@garysmcdermott]

My big laugh was when he said, "Physics is the Law, everything else is recommendations" simply beautiful!

[@gorasuhl97]

​@@OtherTheDave, why should it be clickbait? gravity on earth isn't really constant. It actually depends on your position. NASA startet the Grace mission back in 2002. They launched 2 Satellites (Tom and Jerry) into space to map the earth's gravitational anomalies. It don't change that much, but it's not constant ;-)

[@OtherTheDave]

@@gorasuhl97 Correct. That’s so obvious, though, that my mind went to the gravitational constant, G. Poor wording on my part, and thanks for pointing it out.

[@americanbaldguy]

I think Martin is trying to build a clock that can play music!

[@WeArePharmers]

Yup, it's as if he's really trying to reinvent the wheel

[@BensMiniToons]

A clock could never be precise enough to play tight music. 0.03 ms to shave before we can make mmx3

[@henlostinky273]

which is ironic because clocks are where he got the idea in the first place. but yeah it seems to be evolving into a clock you re-wind with a foot pedal and the chimes are always playing.

[@DemsW]

Is this ironic ? i'm never sure with comments@@BensMiniToons

[@BensMiniToons]

​@@DemsW It's more sarcasm. Honestly, everything Martin has made could play music tighter than any human (less than 5 ms). Nothing seems good enough. But I hope he pushes through. "Even the most perfect timing system isn't perfect." (Irony)

[@StuffandThings_]

The story of the Marble Machine reminds me a *lot* of the story of the Marine Chronometer. John Harrison spent over 30 years working on his designs, along with many unsatisfactory prototypes, and many others improved upon his work later to achieve even better results. Don't give up!

[@tobyCornish]

AGREE! I have seen strong a Harrison-like thread developing since he rejected MMX

[@scaredyfish]

I’ve had that thought myself. Perhaps that means that MMX4 will be a tiny desk-sized contraption.

[@tobyCornish]

@@scaredyfish If he proposes taking it on sea trials to the Caribbean, we should get concerned

[@Ezox2408]

I thought the same thing! i actually posted a comment a bit like this on his last video ahaha

[@constantinosschinas4503]

big difference: MM is a useless piece of sheet, whether it works or not.

[@XIIchiron78]

The mechanism you built is almost identical to a lot of really old clock towers, so you're in great company in terms of realizing that gravity can be used for precise timing

[@nickrl113]

my favorite feature of all of these elaborate testing and tuning setups is the constant use of contact microphones for everything. no light-based rpm sensors, no angle measurements, just *tak tak tak tak tak tak* for everything. there's always more than one way to skin these particular cats.

[@hsfinlayson]

While I absolutely find enjoyment in this ongoing series - and have for the past few years - I must profess that I am curious as to the reasoning for recently obsessing so much on milisecond-level "tightness" of the system. Some of the best music we've all ever enjoyed was made before click tracks were used in the studio. Many of our favorite rock & top-40 hits from the '60s, '70s, and '80s have tempos that fluctuate all throughout... or ramp-up across the length of the tune... and nobody cared(cares). If your next version of the Marble Machine has tempo fluctuations, who cares...? The live musicians can adapt. Unless you plan on sync'ing to backing tracks, it seems like overkill - I presume I am missing something... (?)

[@KingHarkinianZC]

I second this. Also, I still feel like part of the appeal was it being powered directly through Martin's actions, stepping on a pedal or rotating a crank and directly playing the machine instead of just winding it up. Dunno why he's changing the whole power input system it seemed fine on MMX

[@hsfinlayson]

@@KingHarkinianZC I can totally understand that one might want to minimize variance (as he mentioned in the video) at all points of linkage in the chain, since it will compound by the time input makes its way to output (marbles striking playing surfaces)... but even the previous iteration was *more than good enough* to play along to in a live performance. If the performance is too perfect, then one might as well just build a machine that strikes the instrument surfaces based on MIDI data and bypass the Rube Goldberg aspect of the design. I think the variance and "will it work?" aspect is part of the charm and appeal of this endeavor - not precision... but that's just me. I would vote for lasting durability WAY over "tightness" of performance. :D

[@djgummikuh8895]

The Problem I see (and I believe is also what Martin tried to convey in the end of this video) is that the marble machine is an immensely complex machine. If you were to make music electronically, you have the CPU timing and then the tone, and then that's it. But with the Marble Machine, THIS is just the base signal. It will mechanically move a lot of parts, which in turn interact with a lot of OTHER parts, which in turn will create notes which will presumably still interact with even more parts. Each of these steps incurs further deviation. By nailing down the timing of the first element, he is essentially buying himself enough headroom later down the road when moving parts inevitably will drive down tightness. If you already start with a sloppy timing at the first stage, it only CAN go downhill from there.

[@stephenkeen6044]

On the other hand, he keeps forgetting to take a number of factors into account, so will inevitably end up with a mahcine that doesn't have "perfect" timing anyway... The irony.

[@boryswwa]

I have a theory, just a theory, backed by some observations, which i’m not having enough time to explain (although I tried under one MMX vid looong time ago), that Martin has found an issue with MMX’s timing, but he looks for a solution in a totally wrong spot. And that is why he keeps failing in finding that solution, but he thinks that’s it’s because the mechanisms he invents aren’t “tight” enough, and this drove him obsessed by it. But the real problem is, that he’s not looking in a right spot. He simply misdiagnosed the issue. I think that the marble machine has a fundamental design flaw which prevents it playing tight music in a different tempos, unless fall time of marbles from each channel is exactly the same. Adding a drum machine that is not based on marble fall time at all, but on direct interaction with contact microphones, is only introducing further issues (and actually highlighting the main issue). It can basically be tuned to work tight for one specific tempo, but the moment that flywheel speed changes, all goes out of sync. That is my feeling, but again - have no time, nor will, to prove it.

[@Troglobitten]

for the love of god, can you just set a goal and work towards it instead of moving goalposts. This endless quest for precision past the decimal point will never be good enough. This is exactly what Margret told you in Germany. It's a marble machine, the core idea behind it is to play music in an elaborate and visual way. If you want tight music to the millisecond, get a midi sequencer.

[@blackssalad245]

Martin, please, pleaseeeee, use two separate power inputs in the machine, one for the programing wheel/info of the machine, and another one for the marble lifting, you will get way better results

[@angellestat2730]

what would be the point? this also mean he will have two different pedals? I am an advocate that for this first version he should use an electric motor, because he still has a lot of issues to solve ahead, manual power input could be included in a next world tour. But trying to include this design with the huge weights that would require and all the power lost by friction from all marble machine components, would be a nightmare which would make him quit.

[@imblackmagic1209]

@@angellestat2730 that would defeat the point of the marble machine, being all mechanical

[@angellestat2730]

​@@imblackmagic1209 having an electric motor defeat the purpose of all church organs? Which use the motor to store compressed air? Besides, you are wrong, the marble machine is not just mechanical, it would use microphones, amps, sound filters, etc. The same as an electric guitar. The mechanism of the marble machine will remain, the only thing that you are changing is the power input.. You were fine when he used an electric motor in his previous design? why not in this one?

[@kevinr.9733]

​@@angellestat2730 Nowhere in the _design_ of either previous Marble Machine did he use a motor. He may have attached one, but only on a temporary basis for testing purposes. In performance, everything between the power input and the marble striking the instruments would have been purely mechanical. And organs used hand-pumped bellows for over a millennium before the introduction of electric motors. Besides which, the purpose of church organs is not to be "all mechanical"; it's to be _loud._

[@angellestat2730]

​@@kevinr.9733 BUT NOBODY IS HAND PUMPING NOW DONT YOU? This remove some of the magic, sound quality or merit to the player or the instrument? No.. because is not related to the mechanism on how those instruments produce the sound. It is just the power input. Then if you said.. with an electric motor you are removing the merit of the player to keep the music tight, but in that case, by adding a Huygen Chain Drive with a governor and a flywheel you are doing the same thing! Not to mention the COMPLEXITY and issues that this kind of feature adds to the machine and the logistic. Many of these mechanism does not scale well, because some things depend on the square of the dimensions and others by the cube. If he continue this path, he may end chasing his own tail because he still doesn't know the power lost that the marble machine will have, this mean he cant calculate which would be the required power input, weights and pieces of the whole power mechanism. Still, with an electric motor that can be adjust to the needs. The risk is that he can build all this and start to design the rest of the machine to later find out that he needs to redesign all the power input, this is when he may resign. Instead, he can start simple (with an electric motor and solve the important mechanism related to the instrument it self, and then in a second world tour, he can try to do it without the electric motor. *Besides which, the purpose of church organs is not to be "all mechanical"; it's to be loud.* This phrase does not have sense. The sound produced by church organs is all mechanical and requires pressure to work. That part of the volume depends on the air flow does not change the mechanic of the instrument. You also need electric amps on the marble machine to amplify the sound.

[@CelticKnight2004]

The fact that its that close… and theres no compensation for friction is absolutely amazing…

[@AlEbnereza]

Maaaaaaaaaarrrtttiiiiiiiiiinnnnnnnnnnn!!!!!!!! MAN I missed you and your inventions!! 🎉❤

[@metern]

You really need to install a Governor. It is a reason they use it on a machine with a heavy flywheel. It really helps to keep a steady speed. 😉

[@lucasherrera1989]

Very true

[@StevenHarnish]

Is there a way to have the governor tied loosely to the input? Martin cranks at ~100bpm and the drive evens out the output to keep it constant. If he changes to 80, the governor adjusts to the new rhythm.

[@ptitnhane]

Yes ! Without a governor, this idea with the weight won't work. The weight (and force) will be constant, but the machine will go slower in parts of the music with more notes. When there are less notes to play, the machine will accelerate. Martin, you need a governor !

[@mm-hl7gh]

But aren't these things controlling the power ? Like the vent from a steam machine.. (?). So they don't make any sense on a hand-or pedal driven system ..

[@metern]

The Governor combind wit Huygens drive is probably the best solution.

[@doyowan]

My assumption was that if we’ve been building clocks on that mechanism for centuries, it’s probably tight. Well done on the prototype Martin. I still look forward to the day I see you live on stage, with or without Marble Machine!

[@lasskinn474]

an escapement clockwork doesn't depend on gravity being a constant as much as a freefall clock with a break needs it to be. into an escapement system you can also add load - like relatives of this mechanism have been used for thousands of years rather than hundreds of years like proper clockworks.

[@waveclaw]

It won't effect the flywheel at the scale that Martin is operating, but yes, the local gravity does very along the surface of the Earth. This variation in mass distribution is on the scale of how the moon, Sun and Jupiter affect your weight as the Earth rotates. The "force of gravity" is constant but the mass underneath you locally is not distributed evenly and you are affected by the gravity of everything else in the Universe. Unless you are playing music with the tides, differences are so small that the music on the frequencies humans can hear won't be impacted.

[@giakobbo]

imagine if clocks were handcranked

[@abcdefghijkl123454]

@@waveclaw it also varies with the distance between centers of mass, so if the weight is allowed to change height the force will change by a miniscule amount

[@whippy-dz1bc]

my big concern is that clocks don't change their BPM. I can unfortunately foresee some issues in the future with this. changing the BPM would surely mean changing the friction in the machine, which would affect the rate at which the weight falls. I may be wrong but I FEEL like there's a problem to overcome with that

[@eosfuto]

When perfection takes over, I understand the concept you are chasing and why, I just hope you will be able to enjoy and appreciate The just enough

[@Tubeytime]

My lifelong question: "If you make precise things using precise instruments, how do you make the instruments themselves precise?" The answer is: "You don't. Physics does."

[@Doktor_Vem]

We've known for a fact that gravity is constant for over 350 years now, why would you doubt such a renowned physicist as Sir Isaac Newton?

[@Araye]

humans aren't that tight... music is human.

[@user-ex6xc5ox3k]

Guys, I think Martin is actually losing it at this point.

[@Coaster105]

100%

[@WeArePharmers]

You could say he's...losing his marbles.

[@LordDragox412]

From musician to engineer to scientist to mental facility patient.

[@adobedoug2564]

Definitely lost the thread 😢

[@GamingDad63]

@@LordDragox412 - It's the same old story... 😛

[@Oeli_Maurer]

I think a govenor will work much better in the actual machine since you don‘t want constant force, but constant speed.

[@extracted225]

Just combine them. Like he did in the video we’re watching right here

[@rvlougdon9500]

Govenor means you wouldn't need the ratchet and sound that goes with it.

[@woosix7735]

i wonder if a direct drive version would preform somewhat better with a gouverner. But Yeah you cant argue with the results of the huygen drive I guess

[@viniciusvbf22]

Martin, you're great. I love your videos, and I think you are a magician of musical notes. I really do. That being said, I dare anyone to say that they can hear a 1ms difference on any music whatsoever. Please don't overthink this. I'm not as educated as you on this, but I can say for sure that "tightness" is not what makes great songs. My real fear is that, by investing your time on this pursuit, you're losing time that you could be doing music! Please be well! Cheers!

[@dansegelov305]

Yes, yes yes!!! I don't know if you saw my constant whining comments on your Germany videos, but this is exactly the direction I was begging for you to take! There are ways to make a 'programmable' governor to vary your bpm, and more complex multiple weight setups that allow a much longer drop time. Great stuff Martin!

[@LDVSoft]

Martin, Gravity while you're using the machine won't change. Moving around the world, however, will have a very small effect!

[@adamrak7560]

Depends on the sensitivity of the instrument. You can count the large boats travelling on Danube with a torsion balance in the basement right next to the river.

[@pstrap1311]

Yeah, gravitational acceleration will vary depending on the distance to the Earth's center of gravity. Another minor factor is that different atmospheric pressure changes the density of air which changes the buoyant force air pressure provides and that effects the perceived weight of an object. Probably neither of these factors would be very noticeable though.

[@MikeRI170]

And the biggest downside is that it won't work in space!

[@SystemX1983]

​​@@MikeRI170 so there will be no moon tour? 😱😧 oh, I forgot ... moon has gravity! 😂

[@JimmyThomasDev]

My main concern is that this drive seems very dependent on load, this would means if you're to keep this tightness, you'd have to make sure load is also constant

[@framegrace1]

Yep, it may need some sort of torque converter linked to a governor. But I think, at large, the load will be almost constant during the song. It will just depend on the number of simultaneous notes played.

[@frazbombe]

I can imagine times when, for example some tracks are muted. The load will drop and without a governor, it will begin to speed up

[@JimmyThomasDev]

@@frazbombe yep muting was my first thought, drum sections on the MMX seems to take quite some load, flywheel will dampen somehow. But yeah, without a governor, no way to get back on timing

[@MikeRI170]

Governor would add resistance when going over desired rpm. When load makes the machine slower governor will reduce its resistance to compensate. There are also governor designs with integration component that work like PI controller and can achieve exact RPM as desired with variable load

[@sabata2]

He can swap loads to change the BPM. That's why this was the ONLY way for him to get the production tightness he really wanted. He just needs to build the final version and test the BPM at varying weights.

[@HolySerega]

Such title deserves VSauce music

[@giacomomosele2221]

I've never seen your channel but KZbin recommended it. I have absolutely no clue what you are talking about, but the fact that you used a DIY machine to calculate all of that stuff is... it's pretty fucking amazing

[@vlastasusak5673]

Sheogorath, the patron of madmen and artists was once Jyggalag, the prince of perfect order, but has been cursed into his current form by the other princes for fear of his power. Once per era Jyggalag returns and overtakes Sheogorath in an event known as the Greymarch to try and reestablish his order, and reconquer the land through his logic, probability, and knowledge of all possible outcomes through determinism. However, when his time is up, Sheogorath takes back control, and Jyggalag is forced to watch from within as the Madgod tears down and burns everything he has built. Now, since Martin - once a man filled with hopes and dreams of slaying dragons, a man who used to view lifting marbles as a game, rather than a chore - now literally checks if the fucking planet even runs good enough for his Bearing Processor™, I fear the Greymarch is once again upon us.

[@Septimius]

Wasn’t the idea after MMX to not get bogged down into perfection? I’m here for this content, so long as we all agree that this won’t make a marble machine

[@thatguyStrike]

yeah exactly ... I'm treating this as "excursions into acoustics physics" now and not the building of a marble machine. too scared to be disappointed.

[@slevlonderco3198]

The questions get more questionable every day

[@awkwardsaxon9418]

love how Martin casually prints extremely durable 3D-print parts but never as much as mentions it. it's a tool to be used like any other

[@pancakelegend]

They look to be carbon fiber nylon prints. I can take a guess at what printer he's using. He might as well not draw attention to it unless they want to pay him for the privilege.

[@markbooth3066]

That 3d printed ratchet won't work for long, we can already see it missing latches in the video, but it only needs to last long enough to do these tests.

[@silentferret1049]

Now you know why its used for clocks. The fairly constant load by the weight is simple but it is heavily based on the usage load by the marble machine that will dictate the Flywheel and weight. That will be the fun part for you to learn.

[@lese39]

My man here has gone from musician, to craftsman, to engineer to scientist.

[@WeArePharmers]

Next up: another side attachment mechanism onto MM3 to account for variability in gravity due to elevation to play tighter music

[@fffmpeg]

that's not even a joke because Martin said he's going to install a governor

[@noone-ez6on]

He absolutelly lost it after MMX didn't He?

[@xonor13]

Can we get the group who finished the MMX to go on tour? Martin is looking for absolute perfection, which he'll never find, not even with a real band performed by humans

[@haydenlinkel3258]

@@xonor13 I'd like to see him use the tightinator on a person with a drum pad and a stick

[@MikeRI170]

@@xonor13 MMX going on tour would kill it. There's a reason it's being placed in a museum.

[@prodigalsorcerer1415]

The new marble machine will have a 10 m tall tower with a comfy chair suspended from a chain. A lucky member of the audience within the appropriate weight range will be selected to act as the energy bomb for a given song. Song is over, audience member has reached the floor and can rejoin their friends, rinse and repeat. The Wintergatan Tower of Power.

[@PhilipStubbs]

I see this working fine with a constant load, just as it has for clocks for years. But your marble machine is going to have varying loads. Maybe the varying loads are small in relation to the size of the flywheel, but I still think it needs to be considered. A centrifugal governor that moves the wind vanes in and out may help. It would certainly look cool!

[@PloppySonofPloppy11]

I agree with the centrifugal governor for adapting to the varying loads. Please consider pendulums rather than air.

[@evanrhildreth]

I would suggest an escapement. In a metronome or clock, the escapement isolates the load (e.g. a metronome's clicker or clock's geartrain) from the movement of the pendulum.

[@GerryG91100]

I am so glad that all these years later you are as excited today as you were at the beginning if this project all those years ago.

[@opinionshurt2905]

When the best musician from back in high school ends up becoming a mechanical engineer that hasn’t slept in 96 hours.

[@squirrelwood8008]

You are going to need to do some testing with that flywheel and the prototype in Germoney under actual load, driving something that puts up varying resistance. That should give you more reliable/useful measurements.

[@bmotik]

Germoney? Is that like someone from Harry Potter?

[@Deplissee]

Yes, I don't understand how this design could work with varying torque as it needs a constant resistance, no?

[@cactus445566]

I think it comes down to how the governor ends up being designed, because the governor will end up adjusting how much load it puts on the flywheel in response to everything else that's being powered by the flywheel. Like this is two major improvements, creating a constant source of power and the governor to use that constant power to regulate the speed of the flywheel. I'm interested to see how much more effort it takes to keep this going, since it's basically wasting a lot of power in the form of air drag to keep the speed constant

[@foldionepapyrus3441]

Indeed, a 'tight' power train when it isn't doing useful work really doesn't mean anything. You will need a really really massive flywheel to overwhelm the tempo changes caused by variable forces of actuating all those gates to be a tight as Martin desires (though worth pointing out the drum full of pins is also a flywheel of some reasonable merit in its own right, so once you get both the common flywheel and the music box drum spun up you should get even better results). Not to mention in the modular machine context its possible whole extra modules are also only sometimes providing resistance at all... But Martin does seem to barely notice comments such as this... And that Hyugen will probably not scale up practically for a world tour, if you can even make it run on one man power at all - bench size test is not the same as lifesize!

[@ps.2]

@@foldionepapyrus3441 Nah it's fine. The Huygens drive will scale fine, so long as the size of the MM3 can be 2m×3m×35m. Who wouldn't want to see Martin's suitcase dangling from a chain 30 meters above the stage?

[@IanZainea1990]

Martin, I think that you push your prototypes beyond what they are capable of measuring. Rough prototypes can give a direction, but testing prototypes have to be close to the final design, it's not a prototype because its "rough" or "half-finished" it's a prototype because it's not the final. But to run exacting tests, you need the prototype to be full fleshed out, that way any changes you make toward the final product will actually be realized in your final product. The way you are doing it, you're building a rough model, and then going straight to final based on the rough model, and then getting frustrated when the final doesn't work... because really it's just the first prototype. What you've built here isn't a prototype, it's a practical model. Talk to your engineers about it, I'm sure they know more.

[@John_Weiss]

Louder! I don't think Martin can hear you. Martin does seem to make things worse for himself…

[@gary1anderson]

The weight of the chain will vary based on its length and the impulse of raising the weight will also contribute to the variance. Glad you decided on the more consistent choice. Speed can be adjusted from the governor so you still have the ability to add your feelings on the fly. It might be fun to have someone from the audience use a bicycle type contraption for raising the weight so you can preserve the use of a pedal.

[@mafiosomax7423]

He mentioned that the weight needs to be kept at a constant height which would negate the influence of the chain length. Either way he should have used an escapement, if he wanted a clock.

[@JayChuckOnFire]

This is why I’ve been following you since the very beginning. I can’t believe you did this with such ease!! And it’s only going to get tighter! I’m so very excited to be on this journey of discovery with you

[@JayChuckOnFire]

You’re building a machine that plays live music. Better than anyone else in history. Your instruments will be enjoyed and studied in the world of art and engendering for as long as the conscious mind prevails.

[@br52685]

@@JayChuckOnFireThe hype-machine is real

[@kentslocum]

There's a reason why clocks and lighthouses used dropping weights to regulate rotation. I wouldn't mind a Marble Machine powered by a falling weight.

[@JaredJanhsen]

The cool part with this mechanism is you still get to be the power input. To me that's the coolest part of a mechanical music machine, the human provides the input. Whether you put it into a spring or a weight to be metered out later.

[@George_vv]

Fun fact: About every 4 1/2 hours gravity pauses for about 1ms to cool down.

[@saxus]

I think Martin yet again lost a very important engineering principles and forgot to ask that do we *really* need 66x times more tight music?

[@khazaddum6570]

I don't think he actually did. It seems to me like he has a very good understanding of what tight music means to him. And that the huygens drive fulfilled that expectation. The comparison between the two prototypes is just way more informative and interesting that him just saying "wasn't tight, now it is"

[@saxus]

​@@khazaddum6570 I don't agree. If you're a mechanical engineer (or any kind of engineer) and you design something you first want to set up your goals: in this case it should be something like "it should have n ms standard deviation with X loads and having y and z features". Then you compare your results to that. Instead of that he just compare the two prototype to each. That's more like how an marketing team works instead of an engineering project. Best part is no part also applies here: don't do more than it's absolutely necessary. If 1ms std deviation is enough then it's no reason to add extra complexity just to have 0.1ms. One of my colleague joked about that "you can't let engineers without supervision". Because they will lost the focus and will invest time into fine-tuning less important details instead of delivering a viable product. One of the hardest thing what an engineer (or a software developer or whatever) have to learn is when is the point when a solution is good enough and investing more resources into development doesn't moves the project forward. Also as others have pointed out: he tested this prototypes without any significant load.

[@zx3215]

Just keep in mind gravity constant is different in different locations on Earth :) Keep this in mind when travelling with the Marble Machine :))

[@impromptu_ninja]

Ah I was looking for this sort of comment; it's not just that, the gravitational constant is only related to gravity, but gravity is never actually constant. We have gravitational wave detectors including pulsar timing arrays... they are big in the news these days, they pick up the continuous symphony of gravity ripples as they wash over us and other parts of the universe.

[@spmpf]

Here I would say gear/belt friction and air resistance (induced in the flywheel) should be the major factor

[@nanamacapagal8342]

not to mention the gravity of literally every single other celestial body in the universe is affecting the gravity on Earth (albeit not by much)

[@SergioGugliandolo]

I think that a heavier wheel will compensate for the difference force applied while making the weight go up and when not. A heavy wheel + this mechanism will produce a sound tighter than all the ones you have ever produced. Btw extremely cool video, research is the best part of every job in my opinion :)

[@lucidmoses]

For all intents and purposes gravity is a constant for you. It's other forces in your mechanics that can make it very. But again, If you check popular songs you will see that the rhythm is not as constant as you would think. In fact it's the variations in rhythm that gives it that human feeling.

[@jtorola]

It’s been a long slow road for Martin but I think he’s fully lost the thread by this point

[@bmotik]

Very what?

[@lucidmoses]

@@jtorola It does kind of seem like he's after sounding like a robot or drum machine or something. Imho, the foot peddle was the best option. It's consistent enough and allows him slow/speed sections to follow the music's feel.

[@seemlesslies]

​@@jtorolaMartin went down a rabbit hole years ago and has never come back. He's violated so many engineering principals along the way.

[@seemlesslies]

​@@bmotikI assume he means that gravity based on the setup he made could have variations, but those variations are from the machine testing it. He worded it very poorly. You have wind currents, wind resistance, gear resistants, and I could go on for pages about factors.

[@LightProgRock]

If you want to change the tempo analogically (not just 100bpm to 120bpm), you can use the mechanism for the motor speed on old vibraphones ! The belt is on a conical shaft and a command moves the belt on the small or big diameter for tempo changes. My vibraphonette has that, works like a charm, even though you have to ensure perfect friction

[@Leeengold]

Short answer to your intro question: No, it's not (a force). It is an artifact from curvatures in space-time.

[@PADARM]

I think there is a bit of confusion in the title. Gravitational Acceleration on Earth is constant because it depends of the mass and radius of planet Earth, 9.8m/s2.C at equator. This means for every second that you are falling, your speed will increase by 9.8m/s On the other hand, Gravitational Force is variable because it depends on the mass and distance between two objects.

[@jameswilshaw8143]

Have you considered putting your albums/drum stems through the tightenator? Playing tight music is really important but human musicians aren’t that tight it may be σ = 10 ms over a track gives you character whereas σ = 0.1 ms over a track makes your masterpiece into a convoluted midi machine. It might become more sterile and lose character. Playing tight music as an engineering requirement is great but you may want to decide both an acceptable upper and lower bound for tightness. More tighter isn’t necessarily more better. Just like auto-pitch correction can remove character from sublime micro-tonal performances if not applied delicately.

[@EV-wp1fj]

My background is in watchmaking and I see what's obvious going on here. Another poster here used bicycle components as a metaphor. And what's interesting is that you could use a stock bicycle freewheel and a derailleur to act both as the ratcheting / chain tension mechanism AND a bicycle gear cluster + derailleur + bar shifter as a transmission selector. The only part that is missing here is the escapement. Forget governors, use an escape wheel, a pallet fork, and a balance wheel. You're 90% there but this is settled engineering. Why in heavens would you want to use something inferior like a governor?

[@Robbedem]

He would still need a system to keep the tempo stable. It's rather difficult to bike at a fixed speed on a hilly road. ;)

[@EV-wp1fj]

@@Robbedem That system is called an escapement.

[@hamarana]

That graphic visualization of the variation within the second is mind blowing!

[@bystandard239]

Thanks!

[@timetu]

Isn't gravity inconsistent across the planet? Maybe something to consider when planning a world tour?

[@travishancock9120]

As you turn the different parts of the machine on and off you may find that changes the tightness. Just something to think about so you can potentially have a solution by the time you get there.

[@EnlightenedSavage]

He's says that in the video

[@SystemX1983]

Well, yes and no... It is pretty constant as long as you don't change location. I also remember having heard, that it also varies slightly over time. But this variance is probably too small for Martin to be measured 😁

[@Tvngsten]

Exactly, and for some stuff it that variance can have an impact, but for the sake of this marble machine, it's safe to assume it is constant

[@MGlBlaze]

It also varies based on altitude and geographical location, though also slightly. You need to be flying 100km above sea level for earth's gravity to go from 1g to 0.96g. Standing on the top of Mt. Everest you're still experiencing 0.99g.

[@allfunnydogsstories2129]

7 years ago you made a video that has over 165 Million views. It was the best music. The best machine. The most interesting video. I loved the flaws. Don’t ever forget 😊

[@NickShabazz]

I eagerly await Martin's watchmaking arc. It is inevitable.

[@zorod5475]

That metal prototype was a very expensive test.

[@THarSul]

Lmao, i posted on the reddit about this when you were making the prototype in germany, and was confused why you were making it at all when you were so keen on the chain drive in the first place. I now recognize that you were doing it for the workshop experience and to gather the data used in this experiment. Glad to see things are moving forward in a sensible direction rather than fighting an uphill battle against rotational inertia by hand.

[@metern]

A spinning gravity fed Governor is the best version. Like the one used on old steam engine and clock mechanisms.

[@thorvalar]

You are so going to win this round! Push on with the new marble machine! We can wait for it =)

[@eddievanbasten1751]

I used to watch this guy before he started milking it. The Marble Machine will never be built, his world tour will never happen.

[@phantis7064]

I took a moment to calculate estimate slopes for the new design for the deviation over the time, and the slopes almost show a SLOWING DOWN of Growth as the length increases, causing the fast growing difference to the old design, with an almost exponential growth.

[@akivaharker7865]

I can see this with a gearbox that would allow for changing tempo on the fly

[@Chris-rn9zx]

Exactly what I was thinking

[@Nevir202]

That would have nothing to do with this part, you'd want a tempo selector AFTER the flywheel where you are taking out power, and this would be before it, inputting power. You don't really want to try and modify flywheel speed, you want that to be as close to constant as possible, at all times. It's a cool idea, my biggest concern with it is how much parasitic drag would the speed selector add? And also doubling the tempo would likely double all the other losses as well. Remember, the goal is to have the machine human powered. Anything that adds too much drag would either make playing exhausting, and said exhaustion could make the music sloppy, or literally make it impossible to get the machine up to speed with the amount of power he can add in the first place.

[@Robbedem]

@@Nevir202 I fear he is testing the wrong thing though. He has now tested that the energy input of this system is much more constant than cranking or pedalling. Which imho he didn't need to test at all. We all could have told him from the start. (pretty sure he knew himself as well) What he actually needs to test is what he needs to make a machine that can keep a steady tempo on varying loads (less vs more notes). What he needs: [energy source] -> [energy buffer] -> [tempo regulator] Also, the energy buffer can be a flywheel , pressure tank, weight on a height,... The hyghens drive is another option for the flywheel. There is no need to use both. The tempo regulator will need to consist of: - energy 'valve' (something that can change the amount of energy taken from the energy buffer). - an option to choose a tempo (dial, buttons, lever,...) - a feedback loop going from the tempo regulator output back to the energy valve.

[@Nevir202]

@@Robbedem You're definitely right, in principle. On the other hand, such a big flywheel should mean that if you can regulate its speed accurately enough, the upstream loads can almost be treated as a constant, even though they won't be. Playing a handful of notes by opening gates vs not doing so, will be such a tiny difference in energy being extracted from the system, in comparison to the friction of moving the marble elevators, spinning the programming wheel, etc which ARE more or less constants, as to be insignificant.

[@youareliedtobythemedia]

Better, use a Vario transmission.

[@Chaisz3r0]

I'd assume for the tightest music possible you'll need a Huygens drive for the power input and a speed governor for the output. Something like a centrifugal governor (also invented by Christiaan Huygens, btw) might still be reasonably interesting to look at and still allow you to adjust the speed manually, allowing for audience interaction.

[@IvoTichelaar]

A beautiful mechanism and such a fast and successful prototype. I bet this would work well when combined with a barndoor star tracker. In a post-apocalyptic steampunk kind of way.

[@rtbeerzi]

7:20 Your explanation was perfect and it blew my mind

[@lortzo573]

i wanted to build a machine like that for power generating and now u gave me the perfect example of how to build it :D thanks

[@HowlinnWolf]

Okay, I take my words back. This is amazing! If it will be constantly wound up by the crank to keep playing, it's still satisfying to see it being played. That precision is worth it.

[@alextivey31]

I think despite the huygen drive being more consistent and tighter, I think the pedal drive gives the music more personality and makes it feel like you're actually playing instead of just winding up a machine

[@brentwheels]

But then you won’t be able to know it’s playing such tight music, within 0.09ms. That’s sarcasm, I agree with you. Didn’t Martin say in an earlier video the human ear can only detect what was it 0.3ms or something? Martin is just going down another rabbit hole.

[@rev_dude]

​@@brentwheels He's explained this in this video and multiple times in the past: once he starts integrating all the different systems, errors will compound. If each mechanism is only ok, then the whole instrument will not be tight at all. By getting every individual component; marble gates, power input, drivetrain, programming wheel, etc. to be super tight on their own, Martin keeps the timing tolerances from getting out of hand.

[@brentwheels]

​@@rev_dude Does it really matter when there are no stated goals? Does he have a goal for each part of the machine? Not that I see. The MMX had an electric motor to assist and keep the flywheel powered without pedaling. Now Martin has to figure out another mechanism to keep a falling mass well, falling. Another rabbit hole, another step against making the machine less dumb. When he said "is gravity really a constant force? let's find out" it just tells me he will toil with this project for years and decide one day to start over, again.

[@StuffandThings_]

You'll still need a way to change tempo in the song - I'd suggest a CVT (continuously variable transmission) to achieve this, you could adjust the speed with a dial or something.

[@EV-wp1fj]

An escapement mechanism, and a stepped transmission wheel is all he needs.

[@zookaru]

haha, he really is just making a clock with extra steps 🤣@@EV-wp1fj

[@TypewritR]

this the man who made my childhood... messing with gravity. how far we have come.

[@falconproductions2708]

Yea that must be why old coo coo clocks used gravity to keep time accurately

[@esalehtismaki]

Yes, it is very accurate as long as resistance doesn't change. But it will change, when more programming pins make contact, more marbles need to be lifted etc. And you can't compensate by using more force, the force is constant. Air governor balances that, but you can't make that more effective freely either, because it works by wasting energy. Another difficult thing is tempo changes. You need a gearbox, one gear per tempo. That is doable of course.

[@Azeazezar]

A gravity wave detector. Cool!

[@Khantia]

Just keep the load in mind. If you change the load (engage a new instrument) mid-song, then the tempo will also change a bit. Hell, even playing a single note can create "unevenness" in the tempo, however miniscule it may be. The heavier your flywheel is - the less noticeable such changes will be, but the load would also need to be heavier. I can't think of any solutions on top of my head to deal with this problem, although there may (or may not) be such.

[@sly-shot]

> the load would also need to be heavier why?

[@Khantia]

@@sly-shot bigger inertia/momentum. If you double its mass, you will halve the effect each 'instrument' has on the tempo.

[@sly-shot]

@@Khantia I get that a bigger mass is harder to stop, so the instruments won't stop it as easily. My question is why you said the load (instruments) would also have to be made heavier.

[@Khantia]

@@sly-shot I guess my wording was confusing. When I used "load" that time, I meant the load that's actuating the flywheel, not the load of the instruments.

[@KnowArt]

huygen drive is absolutely stunning. Visually, mechanically, musically. It really is the perfect fit

[@deetsteve00]

His mission to make another marble machine will lead him to graduate on physics just to figure out where the marbles will fall to

[@steverobb5360]

Excellent results! We'll turn you into a clock maker yet! Keep it up!

[@RegulationTyler]

I’m interested to see how you’ll control the tempo

[@EaglePicking]

With an adjustable governor and a gauge to read the speed, it could be very easy and accurate to set a fixed speed.

[@EV-wp1fj]

It's so painfully obvious: An escapement mechanism. Martin just built a clock.

[@shavono8402]

You've built a clock! That's so cool!! Having a clock on the marble machine to govern its tempo is actually kind of poetic - humans were the only form of moment-to-moment timekeeping for a long time, and music was a key application of that skill. A clock to keep time is a really cool way of sort of playing with that concept of human-kept time, using machines as extensions of ourselves.

[@adri__2000]

Attention, the tightness not only depends on the tightness of the input but also the output one, meaning that that inertia of the fly has to overcome change in output. For instance when changing load from one instrument to all instrument. Your test now is without any output load changes. Thanks for all your inspirational video and positivism.

[@audirealan6496]

This is going on three of my playlists, Music, Craft, and Science. Excellent work as always.

[@philvogelfilms]

I wonder what a human player’s standard deviation would be over the same time span… 🤔

[@mitchellsteindler]

Would be interesting to see how much worse

[@thomasbecker9676]

This seems like it's more procrastinating under the guise of testing.

[@Culpride]

Martin, you have to ask yourself the question: Do you want the timing PERFECT or do you want it YOU?

[@marcelleuenberger4648]

Problem with perfection is.. there's no limit as soon as het gets too close to 0.01ms it will suck him in and the new goal will be 1us Not that a human can actually tell a difference.. which is a shame, because what's the difference to an electric drive? Increasing stability will decrease being able to change the tempo flexibly as you add emotion to the song

[@awkwardsaxon9418]

9 seconds over the course of a song (Germany flywheel) doesn't sound insignificant to me

[@bignashi]

It's 9 miliseconds. Which would be pretty close to a very tight musician I think, probably better.

[@bignashi]

I looked into it and professional drummers have around 15ms of standard deviation in a 5 minute interval. So I think Martin's journey for tightness is a bit overkill.

[@steampunkscientist]

6:29 - Literally my brain when I find out the machine/program I've been dismantling for 3 weeks didn't work because I forgot to plug something in.

[@glasslinger]

dog: "If it wasn't for the generous portions of dog food I WOULD BE OUTA HERE!"

[@Svelix]

Next question: Is a marble machine playing music *using* a constant force?