Can You Build a Car With Square Wheels That Drives Smoothly? (Scrap Mechanic)
ฝัง
- เผยแพร่เมื่อ 28 ส.ค. 2024
- In this episode of Scrap Mechanic we see if it is possible to build a car with square wheels that can drive smoothly over a curved surface!
-------------------------------------------------------------------------------------------------------------
About Scrap Mechanic:
store.steampowe...
-------------------------------------------------------------------------------------------------------------
#scrapmechanic #kangaming
![ฮักเทื่อสุดท้าย - บอย ศิริชัย [OFFICIAL MV]](http://i.ytimg.com/vi/-1PoPoQ-B-k/mqdefault.jpg)
What other experiments should we try in Scrap Mechanic?
Try to make a vehicle that can fly, drive on road, off road, boat, and wall climb.
Swerve drive 😁
Make a one wheel car
@@VinignisGaming Yes
Try to do a multiplayer Monday on an additional day. Yeah… I like that idea.
I'm not a math major, but I managed to figure out the issue, and a potential solution. So basically, the issue is that the catenaries used in the video referenced here (0:20) match each other in 90 degree angles instead of the 180 degree angles used by Kan in this video. This 90 degree angle is meant to suit the 90 degree angles that sides of a square create. There are two potential solutions to this problem that I have come up with. The first is to use an appropriate 90 degree round piece in intervals of 1, instead of 2 lined up back to back. If these are rotated relative to the ground, ideally by using them to create a 45 degree "round staircase" and then rotating that entire plane 45 degrees to be parallel with the ground, one can create a surface of 90 degree arcs that sequence together perfectly. This would be ideal given its simplicity and the access Kan has to the glitch welder. If this is not an option, for whatever reason, the secondary solution would be to pretend like the sequence of 180 degree angles are actually only 90 degrees, because that is the only part a wheel is touching in practice. One would create two tangent lines off of each wheel and call one "unit" the distance between the intersection point of two tangent lines on each side of a given wheel. This "unit" must correspond with the side length of the squares used as wheels. While the square wheels will pass through this hypothetical surface, they will not pass through the real life (in scrap mechanic, anyways) surface created by the actual hitboxes of the wheels. One can create an equation that inputs the diameter of the curved surface and outputs the "unit" generated, or the side length of the wheel. Putting this equation through an excel spreadsheet essentially just creates a better version of the spreadsheet Kan used in the start of his video, which did not account for the discrepancy between the 180 degree angles of his wheels and the 90 degree angles of the wheels used in the first solution. This being said, I recommend the first solution, as it is substantially simpler.
Always remember:
Keep it simple, stupid.
(that's not an insult you're still way smarter than I am)
Also one thing that could work would maybe spacing the wheels (I talk about the one making the road) to be spaced by 1 (or more) to give the square wheel a bigger freedom of contact it probably won't make the ride as smooth as it would be if everything worked perfectly but it would make it at least kinda work maybe?
What I'm trying to say is having the wheels not on a perfect surface but at the angle that they would be if the curve of the road was possible, using friction to make up for the inability to make the curve perfect.
tbf he does seem to have an understanding of this when he talks about angles at 17:30 he just doesn't apply it. And it probably isn't the issue too given the square wheels will still grip the tires regardless.
I WAS WATCHING THE VIDEO WHILE READING THIS AND HE SAID "Keep it simple, stupid" RIGHT AS I READ IT
should be an episode where you test every vanilla blocks hit boxes for a spread sheet so it is never forgotten in future experiments after.
Who has the time for such a neich subject.
i can tell kan is an engineer cause he uses pi and 3.14 interchangeably and rounded 0.14 to 0.1
Fermi estimation. You only need to be as accurate as you need to be for the problem at hand.
in fact he rounded 0.28 to 0.2
Because technically there's still an inconsistency of measuring
@@vibaj16supposed to round 0.28 to 0.3
Math with Kan always makes me question math
what about using the 4 bearing trick for even number block wheels? (if something is connected to a square of bearings, it will rotate from the center of the 4 bearings.)
Wait is this a thing?!??!? How have I never heard of this in the 4 years of my scrap mechanic adventures?!?!?
@@agrisimfarming i think it was discovered recently
@@agrisimfarming yeah its a thing, go try it!
@@alexdacat7052 About 2 years ago it got some attention again, but it had been used before. I think it works well with steering, barely works with controllers. It doesn't work with engines, it can only rotate so far. For steering it's fine because it only has to do like 30 degrees, but you can't have it spin continuously. That would be way too much tension on the bearings. It's still really cool for symmetrical 2-wide-wheel steering, but that's about it as far as I can think of. And because of the limited use people keep forgetting about it.
if i'm not saying anything stupd and my math is correct, given a wheel diameter of X, you need to place them at a distance of X*0.707 (sin45) and need square of length Xπ/4
checked the first few sizes and i can't find anything really good, with diameter 9 you get a square of 7 but the distance has to be almost 6.4
if you want you could graph it in excel or something, like you already did (at least i think, for 2:41), but sizes likely have to be real big
I just did similar calculations under another comment but instead of having to worry about the distance, quarter circles can be used on a diagonal staircase, then rotated into place. This allows you to use a 5 "diameter" square wheel on a 4 diameter curve road. This is feasible as 1.25 is close enough to 1.273 that it shouldn't be too much of an issue.
I did the calculations the other way round, requiring a 4X/π radius road for an X/2 length square (starting out using a 2 length square for simplicity). Interesting to see how different people approach a problem from different perspectives
Oh, didn't think of tilting a staircase, was going more for weld-glitches, that's a great idea (for other people, this means allowing distances that are multiple of √2)
Yeah sometimes you just need to step back and reassess. Sometimes all you need is a triangle
With diagonals going on angles different than 45° we can get even more possible lenghts
@@sniper1a259 The 1:1 diagonal "staircase" is appealing because it produces an angle of exactly 90° between the curved surfaces. But that's not the criterion for success here; it's that the arc length and side length of the "wheel" line up. (Technically only true when the circles meet at 90° or shallower angles. For steeper touching angles it's the arc length between the points at +-45° on the curves plus the tangential distance to where the tangents of the 45° points meet.)
So your idea is actually really genius. I wouldn't have thought of it.
Mathematically, your best shot within scrap mechanic should be with 14-block cylinders (spaced 10 blocks apart from each other) and 11-block square wheels. Of course you can always go bigger. ;)
The issue with the rounded parts is that SM only supports a few hitbox types. You have basic things like boxes and cylinders, but for complex shapes, your only choice is to use a "convex hull", which can only connect via straight lines, so you can't get a mathematical curve that matches the visuals
technically it can match the visuals, because the visuals also just have straight lines, but convex hulls that complicated would be pretty expensive for the collision calculations
@@vibaj16 yeah
oversimplification
the convex hull the game uses is a grand total of 8 vertices I think though lol
Similar problems regarding gear meshing where overcome before when tremmel gears using component kits where used.
Also, get a half block division with 2 x 45 degree bearings to offset something 0.5 of a block.
Sign post tubes have sticky sides & ends BTW
You could also try adding parts with slimmer hit boxes for the wheels (like the license plate)
Ooh, this is a really cool idea! maybe some other shaped wheels could be cool to make too
So close! Fun video though, really like random experiments.
You should try wheels made out of wedges instead to get different values while looking the same.
But an equallateral triangle has angles of 60° not 45° so it shouldn't be possible to do in scrap mechanic.
Now, I'm wondering what kind of surface would allow for right triangle wheels to work🤔.
@@bungeetoons no you make a square using the wedges
because then it will be on the √2 scale not the 1 scale
I love these experiment videos. The math and just general messing about is fun, can't wait to see more!
at the end of the vid, maybe there's a possibility to spin (the road of) the wheels so slightly along when the car drives over them to manipulate the illusion of what you are after, I don't know.
2:08 Just use the fricking 7 diameter!
Did... did you not watch the video...?
@@justadude7455he could’ve done it first
@@T_rev-ud7zbthe point was to see if it was possible with smaller wheels as well, going straight for the one number that mathematically should work would ruin the experiment.
kAN I love videos like these! Please make more of this I'd gladly watch them.
kAN , I love your voice and joke delivery. Great videos, keep it up!
Now I want to try this in space engineers. Since can mix large and small grids to closer to the size needed
The square wheel has 90 degree corners. This means the sharpest angle any edge can point downwards at is 45 degrees (before the connecting edge becomes shallower). Assuming each curve is equidistant (and identical), this means the lowest point your square wheels can contact on each semicircular bump, is the point at which their tangents are perpendicular (at 90 degrees, each tangent being symmetrically 45 degrees from central axis).
Between a series of alternating +/- 45 degree angles, each segment covers 90 degrees. You're only contacting a quarter of the curve's circumference, not half. Then you'd have to account for things like slip, along with the questionable collisions, minor contact-bouncing, any imperfections in the alignment, and other factors.
Have you tried using a pattern of x+y for the wheels, and using a rotor to rotate the floor into a horizontal structure. That way you could space the wheels at distances of sqrt(x^2+y^2), it might allow you to get better wheel spacings. in example, placing wheels at (0,0), (1,1), (2,2) and rotating 45 degrees to get a spacing of 1.414 or any multiple of that by doing (0,0), (n,n), (2n,2n), etc. Distancing of (1,2) gives 1.7, (2,3) = 3.6, (1,3) = 3.16. Might be worth checking out, if rotors can be free floating, in Scrap mechanics (I can't remember), so it would let you get a flat surface
You need a road built with 7 block radius half circles where the circumference intersects the center point of the previous and next half circle and wheels with 11 block sides. I think the closest vanilla part there is is the generator coil segment, but I think that's only 6 block radius. That would require a wheel with 9.42 block sides.
Then you need to accept that scrap mechanic physics likely won't work the way you want it to.
If you can get a modded part that is a quarter circle with a radius of 7, you can build them in a stair step pattern then tip it over 45 degrees to get the road.
In theory, the maximum a square can fit between two curves would be a 90 degree angle. Since there are two curves, each would have to meet at 45 degrees to form this. Your square's side length should sweep out 90 degrees of the circle, from 45 to 135 degrees.
Thus, pi times D over 4 should give the desired side length.
The complicated part is how far apart the centers of each circle must be. Because we know a on a unit circle that sin of 45 is 1/root2, we know the arc we want goes from 1/root2 to -1/root2 on the x axis. This means that D/root2 gives us the distance required between the two centers of a given circle.
Unfortunately, none of this lines up very nice with the block grid, the best reasonable sizes being: diameter 13, side length 10.2, and distance between centers 9.19.
Graphing these two equations (pix/4 and x/root2) with x as the diameter should give a chart of values for both length of the square's sides and the distance between the circle's centers respectively.
(This is kind of quick and rough math, don't kill me please.)
But you *can* get a separation of √2 between the wheels if you place them diagonally on the build grid.
if D/√2 = separation, then on the diagonal grid, you can get it so that separation is a multiple of √2. thus separation=a√2
Substituting that in, D/√2=a√2 (*√2) gives D=2a
No irrationals in sight. All you need is to place the wheels to the side and up as many times as D/2. This of course only works for even values D
Aside from the 45 degree trick someone mentioned above, the real issue is that a catenary curve isn’t actually the arc of a circle.
I suggest, to subvert the block grid, to use a combination of several bearings to rotate the wheels at the center of an even number of blocks instead of an odd number of blocks, then after that start fine tuning it with parts that are barely less than a block until you get it right
if i remember the Mythbusters episode correctly, having all wheels offset by 22.5 degrees from each other, and you get a .... relatively smooth ride, kinda.
Add switches to the sides to get a pretty close .5 addition
The corners wouldn't quite be square. But I wonder if that matters in the creases between the circles...
Saw blades have questions as they were not used as a round surface to drive on. Thanks for the shows KAN.
I love when you do math in your videos
Logically the the semi circles on the road should only be a quarter of a circle, because then the ends come together at 90° to each other, which allows the tip of the square wheel to go in. If that works out perfectly for the height is another question, the road surface might need to be slightly elliptical.
You need your wheel edge length to be 1.2x the diameter of the arc you are trying to drive over. The problem is that, at that scaling factor, the only way to get a whole number gives you a multiple of 6 for the wheel edge length, for a 5 width rounded surface.
I am not familiar with glitch qelding and exactly how it works but if you can figure out a means of glitch welding a bearing into the center of a 6x6 square, you could use the large 5x5 wheels as your road surface. Alternatively, with some modded pipe pieces in the mod by Lord Pain, you might find what you need you center mount a bearing in the middle of the 6x6 square surface? I have messed with that mod to build a fairly accurate Wastelander from GTA in the Crashlander mod and there are a lot of wild and oddly shaped angled pipe pieces that might angle into the right spot that you need?
you can in fact also make a vehicle with square wheels that drives smoothly on flat surfaces, you just need a lot of axles.
you could also use pistons on the suspension, with distance sensors so even with 2 axles it is smooth, for the vehicle.
Another idea to get it to work is keeping the vent block back to back and changing the wheel base on the car to to get a smoother ride.
you’re actually moving 1/4 of the circumference of the floor wheel because thats where the floor wheels should merge so that their intersections from right angles
sorry if that doesnt make sense i’m high and cant convey this rn
Holy cats! I forgot this guy was a part of my childhood!
You could probably get an even numbered wheel to work with some mechanical linkages, but the hitboxes lining up would still be an issue in addition to finding an appropriate ratio regardless of if the linkages would even survive Scrap Mechanic's weirdo physics.
If you think about it, the entire concept of doing the square wheel and round road system is just a fancy and inefficient form of a rack and pinion. (Google to see.) You're building a gear system and attempting to match ratios.
I think I know how you can make an almost square wheel a little bit bigger than 3X3 blocs.
You use 5 blocs to make a cross, it will be the center of the wheel. you place bearings on one side on the 4 arms of the cross and attach 2 blocs arms extension. You lock those bearings with a controller and you force a few degrees rotation so the angles of the square stick a bit out.
Kan starting to sound like a crazy scientist in this video
Try covering the driving surface of the square wheel with surface panels form the mod pack. they are about a 1/10 of a block thick.
The square wheel can only roll around 90 degrees of arc until it encounters the next arc. this means for a 3 block wheel 90 degrees of arc must equal 3 blocks. This equates to a radius of 1.91 blocks with the centre points of the radius spaced at 2.70 block intervals. Good Luck
Spin really fast with a really good suspension
Try using the low friction blocks so that the square wheels have chance to slide in to that groove.
The math in this episode hurt my brain. Now I know how Kosmo feels all the time.
You know, having gone to a philology profile highschool class, stopping learning math in 10th grade and then to law school, didn't do me any favours understanding all the circle math, lol. Even though, I learned all this at some point in middle school 😂
Maybe you could try using the detail mod (idk if its called that but all the parts have the word detail in them so i think its called that). It adds 1/8th block parts, 1/6th parts, 1/4th parts etc. so you could make a 4 block square by just adding 1/2th blocks to all sides
what a weird train track concept, smart experimentation works too, random trial, large wheels and cusps, if you make a ramp then you can put the wheel distance to exact positions, ie grid stays on the ground but ramp is tilted up/down. also you can make a giga wheel and not have to hit every cusp. also angle joints make partial measured lengths pieces.
you could make an expandable square wheel using pistons and some mod pack logic to make precise adjustments
Maybe if you changed the length on the wheel base on the vehicle the squares can fit in the grooves more often
You could try taking it to a larger scale and then using larger modded wheels or make your own semi circles with wedges..
6:35 - Circumference is 2Pi*r
If the wheel is 3 blocks wide, r = 1.5, making the circumference 3Pi
or 9.42 blocks
You can make a square with wedges, making the length the sqrt(2) 1.4142135. perhaps this helps getting closer.
Make the wheels on the ground rotate too
This will push you a little bit forward or backward to compensate for the small/big square wheel.
Yep it would be cheating but..I mean you make the rules for your challenge ;)
Might wanna use friction-less blocks for the wheels so they can slip into the grooves better
The path it follows is not a semicircle. It is a more complicated shape, but a clear difference is the joining.
The square has a corner in where it joins. That is 90 degrees.
But a bunch of semicircles joins with a 0 degree angle.
Even if the hitbox was a perfect circle and the lengths matched exactly this doesnt work because two adjacent hemispheres meet at a highly acute angle while the corner of the square wheels are 90 degree angles, which blocks access to part of the surface of the hemispheres.
I wanna see more Demolition Evolution Trailmakers
The amount I wanted to scream when you didn't realise the size difference was killing me
kAN, you forgot about the small metal wheel used in the warehouse cranes. Probably still not the right size, but it might be close enough for the 3x3 squares.
Kan, you're looking at it the wrong way. The bumps don't need to be touching. You can use a singular line of the thinnest pipe pieces suspended in the air with a huge gap b4 the next set of monkey bars.
What you are looking at is gears and teeth. All that matters is that the square cog goes between the teeth. The size of the tooth, nor the distance between them, actually matter. All that matters is if the size of the cog is the right math comparatively to the distance between the teeth.
Again, try thin monkey bars. YOU CAN EVEN SET CUSTOM DISTANCES. You can use pulse pistons to set exact distances between the bars.
Can you add an arm to an even square and set it at an angle to get a centre point for the spinning bearing? I imagine it would have to be at a larger scale before you found a ratio that would allow it on the grid system
you say wheels have to be odd for pivot point but wouldnt using a bearing and controller allow even wheels with a centre point if the degrees are set correctly allowing basically double the choices of wheel size?
If you try something like this in the future i would like to see what happens with frictionless block wheels.
might be a stupid question but, could you use a controller with pipe pieces to adjust the Length by using angle by pushing the wheels slightly closer or further apart? Just a thought but i didnt do the math on it lol
6:02 it's 9.425 blocks for a 3 block wide wheel's radius as apposed to 12 blocks for the cube.
what about messing with the coordination of the wheels? they all have a point on the floor at the same time so maybe adjusting them to have two touching the floor with their 90° angle part at the time to avoid the bumping or even making them touch the floor at different points so there is always one corner touching the ground
You could play around with using screws for a different hit box on the squares
3x5=9
-kAN Gaming AKA the greatest mathematician to ever exist
Doesn't solve the block grid issue, but, you could try using frictionless blocks for the square wheels. What I imagine would happen is the wheel will slide until the corner drops into the 'seam' between the tires. It would be a bit of a cheez, but maybe? < He said, not watching until the end before posting a comment :D :D
support beam have an half collision so you could have used it with glitch welding doing half blocks instead of full blocks
What if you used the leap-year method to correct the offset? Where every, say, 2nd curve is a different radius so that the divergence is correct over the course of two rotations?
It is interesting how the problem can be summed up with pairity
Did you try using the small pipes for your square wheels? Then you can create a wheel that is slightly smaller.
did you try the 11 block wheels with the 7 diameter?
you could also try using pipes to make the wheels
Would it be possible to make an even width square wheel, attach a piston to one edge so it runs parallel to the wheel, put the bearing on the other end of the piston, and use the pulsing piston trick to move that bearing block to a half-position so it's actually in the center? It would impose a minimum size on the wheel, but hypothetically you might be able to get the bearing centering enough to work, assuming the pistons don't go all spaghetti trying to support the weight of the car.
very good attempt and good numbers. shame it didn't work 100% but really good try. hopefully you might find a mod that will do this trick justice.
I would of tried wedge/triangular patterns as you actually get all the way to the bottom of that shape unlike trying to fit your square edge in the arch and only getting down 2/3rds of the way.
Also, dont 1x2 bearing work for even creations, or is that just for turning?
someone needs to make a mod that just adds a block equivalent of the 2x2 lego jumper plate. then we could get even numbered wheels or center seats on odd chassis.
Seems like the curves are ramps. Makes sense but also never thought about it.
so we learn scrap science today :)
could you use bearings or controllers to get the slight like 1/10 ect you need? like make the wheel and then have bearings or controllers the outside of them to give it that tiny bit more block size? or do controllers and bearings have a full block hit box?
suspension- "am i a joke to you?" Edit: maybe u can use like a double pengulum mechanism to move the car body up and down so it creats the illusion that ur car is perfectly stable
Unfortunately with scrap mechanic's block grid the math just has to line up, and it doesn't seem to... :/
In theory if we have wheels with infinitely increasing diameters we should be able to make it work since we'll have more values to try _and_ more options for how closely they can be glitch welded the bigger the wheel gets, but we don't, so we can't.
I think Trailmakers is less confined to a block grid since pistons allow decimal values (and let you glitch parts into each other), so this should be more feasible there if the "road" pieces are connected with pistons, but I don't know if Scrap Mechanic has a way of breaking free of the block grid - other than bearings, which gets complicated.
could you use smart engines from the modpack to make pistons extend a decimal value and reach the ideal wheel size?
I dont know much about glitch welding in the game but cant you use that to offset a 4 Block wheel on bearing somehow?
you can use even number wheels by making a center point with elbow pipe and controler
Shouldn't it be possible to extend the "wheel" size in decimals by using something like a decorational sign or board piece at the outer side of the "wheel"?
What if you don't use blocks? Glitch welding and stressed pistons/bearings to make up for those tiny gaps between actually round blocks?
WHAT IF, u put a small panel like an exit sign on the square wheel so that it gives it the slight length for 3.1. And put the skinny panel on the 2 sides adjacent to each other. So two panels max. IDK
Important edit; your block wheels' widths is 4.24 not 3 blocks. You need to measure the longest length which is diagonal on squares.
There might have been some cobwebs but I was able to remember the math with paper. 😂😂😂😂😂
you should try to do a planetary gearbox
Glitch weld a piston into a block and use your number mod to fine tune your square wheels
Couldn't frictionless blocks for the square wheels help with this?
As in, in order to compensate for the grid, you have a square wheel that is just slightly to short for the "round" road that you have and with each step, you simply "slide" back into the groove between the two circles and thus get an approximation?
Like at 15:00 it looks like it could work as long as you would slide back into it every time.
he does that at 20:00
you can use 1x2 wedges to make different sized squares
Put switches on the 3x3 wheel to make it a little bigger
Can you make the square wheels slightly wider using something thin like four signs on the sides?
Hmm, did you try bigger space between curves box
What about a 2x2 square with the pivot point in the middle with pipes and bearings?
You should make spudnet 3.0 with full message storage
Use pistons to fine tune the space for the tiers