Not signed in (Sign In)

Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeJun 19th 2013 edited
     
    one thing that I've seen in a lot of puzzles is a rotating circle linked to a rectangle to create a back and forth sliding motion.
    I came across this video and could only think to share it with you guys. I hope it inspires you to make some interesting mechanical designs.
    If it does, please post some stuff here guys :D !




    So I found some new inspiring videos that I thought would be great to make rube goldbergs out of with CIE!
    The element of magnetism I'm not fully sure how we would add, but I'd love to see it.




    Link to his youtube profile so you can check out his other 3D rube goldberg / dominos
    http://www.youtube.com/user/Kaplamin...
    •  
      CommentAuthorTheDudeFromCI (Advanced Member)
    • CommentTimeJun 19th 2013
     
    Nice. Very entertaining.-----------------
    Orange is my favorite number.
    • CommentAuthorXyuzhg (Moderator)
    • CommentTimeJun 19th 2013 edited
     
    Haha, I think I've seen this one before.
    It has some really cool concepts that could easily be incorporated into our 2D simulator.
    I'll actually get to building some of these in the editor...

    In the meantime, I found a few noncircular gears to gawk at.
    While the video itself is not the best in presentation, the concepts are rather amazing:
    -----------------
    Hopefully PA is inconsistent.
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeJun 20th 2013
     
    yes Xyuzhg this also has perked my interest.
    If I had the time I'd love to look into the relationship between the shape of the gear and the rotational speed of it.
    The oval gears especially demonstrate the variance of the gear speed!
    •  
      CommentAuthorMadball (Advanced Member)
    • CommentTimeJun 20th 2013 edited
     
    I might do something with it... But not now.

    .....missing.....
    -----------------
    LD35!
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeJun 20th 2013
     
    yeah I saw that too Madball! that's actually the first thing I saw that made me look up more about what it was and stuff lol

    it's called Geneva's Mechanism.
    here's the original vid I looked up.
    If you look to the right side you can see a link to the first vid I posted xD


    •  
      CommentAuthormeesman
    • CommentTimeJul 21st 2013 edited
     
    I've been away for a while, so here's my last try.
    It's a very basic (and I think it's terrible) design.
    If I had some more time, I would improve it, but I don't, so..
    -----------------
    life is life, now is now, but don't forget to live now.
    • CommentAuthorpuzzle geek (Advanced Member)
    • CommentTimeJul 21st 2013
     
    not the worst of things, but no that great eather. with improvment though, you could create a really cool design!-----------------
    puzzled
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeJul 22nd 2013
     
    that's really cool meesman!

    although your design is slightly off from how it actually works, it is an interesting idea as a break motion for the second side :) !
    •  
      CommentAuthorMadball (Advanced Member)
    • CommentTimeJul 24th 2013 edited
     
    Made a function that adds teeth to a polygon. It bugs sometimes, when the teeth cross somehow, making the shape not usable. Rotating by ~1 degree often fixes it.

    [edit]Made another couple of gears, shaped as Reuleaux triangles

    [/edit]-----------------
    LD35!
    • CommentAuthorXyuzhg (Moderator)
    • CommentTimeJul 24th 2013 edited
     
    That's a great way to build gears easily.
    I'll see if I can improve it, as it appears that the teeth often don't mesh well with each other.-----------------
    Hopefully PA is inconsistent.
    •  
      CommentAuthorTheDudeFromCI (Advanced Member)
    • CommentTimeJul 25th 2013
     
    Try using balls instead of spikes, and give them low friction. They mesh better, even under poor angles. Just an idea.-----------------
    Orange is my favorite number.
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeJul 25th 2013
     
    Are the teeth the same size? (they look to be..?) that seems to be the problem with the first one because there is a gap between the last added tooth and the first tooth.

    if you calculated the surface area of the shapes then you could easily divide that by any number of teeth to get the needed pitch for each tooth to cover the entire shape.

    (I didn't look at the code so if you already do this then ignore this comment xD )
    • CommentAuthorXyuzhg (Moderator)
    • CommentTimeJul 25th 2013
     
    Bio, I'm currently working on an improvement (with better teeth).
    The problem with directly using the perimeter and dividing is that it doesn't account for corners.
    Any tooth that ends up in between two corners will actually end up with a lower 'true' pitch.

    I'm seeing if I can devise an algorithm to fix this.-----------------
    Hopefully PA is inconsistent.
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeJul 25th 2013
     
    hmm... i see what you mean Xyuzhg, perhaps an algorithm then that would run a few times until the perfect pitch is determined ?
    • CommentAuthorXyuzhg (Moderator)
    • CommentTimeJul 27th 2013 edited
     
    Hey guys!
    I finally finished making my gear algorithm.
    Before I explain the function, just see here:

    The main function is (as Madball also called his) the Gearify function.
    What it does is basically locate ideal locations for the teeth to go and return an array of points that define the corresponding gear, in local coordinates.
    The inputs and outputs to the Gearify function are:

    Inputs
    Polygon: The reference polygon from which you want to build a gear.
    Tooth width: Note that is not the pitch; the width of each tooth is more convenient to use as you can have different shapes with matching sizes.
    Tooth height: The vertical size of each tooth.
    Offset: The offset of each tooth relative to the next tooth, counterclockwise. Must be between 0 and 1 and best with small teeth.
    Maximum iterations: This determines how many times the algorithm tries to find the best pitch. Note that higher values may not necessarily give better results. Also limit this to up to 100.
    Flip tooth orientation?: Setting to FALSE gives teeth pointing outward when the polygon is positively oriented (points are in counterclockwise order) and inward when the polygon is negatively oriented; TRUE gives the opposite orientation.

    Outputs
    Gear array: The optimized array of points that define a gear. Note that the input shape itself is not changed by the function.
    Number of teeth: Self-explanatory; the number of teeth in the optimized gear.
    Offset: The error between the first and last tooth compared to the normal pitch. This is usually below 0.005.

    I have written quite a bit of explanation in my code in the version above. For a clean version (smaller file size), use this:

    Currently, the optimization method is simply a brute-force check of different pitches between 2 and 2.5 times the width of each tooth.
    Any recommendations for a better optimization method would be appreciated.-----------------
    Hopefully PA is inconsistent.
    • CommentAuthorXyuzhg (Moderator)
    • CommentTimeJul 27th 2013 edited
     
    As my previous post is now getting a bit long, I'll keep it clean by sharing this as a new post.
    Using Madball's polygons and moving them closer (so the teeth can mesh well together):

    The main problem now is to actually make the polygons; some of the more interesting gears are difficult to precisely make.

    ---

    OK, I managed to quickly make two functions that generate an array of points for a polygon defined by either polar or parametric equations:

    As the functions are very simple and self-explanatory, I won't give much explanation.
    But here is a good application of the function with my Gearify function:

    Sadly, my example is not an ideal gear shape for meshing with itself, so there is a little error and considerable extra spacing between each gear.-----------------
    Hopefully PA is inconsistent.
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeJul 29th 2013 edited
     
    wow Xyuzhg that looks really cool!! ...but... now we need to find an actual application for this to apply it to CI xD

    I was thinking how the weird shapes allow for different moments of acceleration/deceleration could somehow be useful in a design, yet I cannot think of how...

    EDIT:

    also, I was playing with your design in the editor (for like 20 seconds) and tried to make the bottom or top gear power the three gears, and it didn't work :(
    •  
      CommentAuthorMadball (Advanced Member)
    • CommentTimeJul 29th 2013 edited
     
    Weeeeee

    I recommend to add "Offset" input in "Gearify" function, setting the position of the first tooth. In range of 0..1 maybe, with 1 being the distance between teeth.-----------------
    LD35!
    • CommentAuthorXyuzhg (Moderator)
    • CommentTimeJul 29th 2013 edited
     
    @BioManiac R2:
    Yeah, that's also why I put the powered gear the middle one.
    Sadly the three just don't mesh well enough together. :(

    @Madball:
    I've considered this already, but implementation is annoying to say the least. :(
    Nice gears by the way!

    ---

    Never mind, I still did end up implementing it.
    However, the implementation is not perfect, and expect weirdness when you set an offset below 0 or greater than 1.-----------------
    Hopefully PA is inconsistent.
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeAug 8th 2013 edited
     
    this is something I found interesting. I graphed the rotational speed of the joints.
    This function is very flexible if you want to add it to other levels, simply link the shape you want it to record to it

    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeMay 10th 2016 edited
     
    So I found some new inspiring videos that I thought would be great to make rube goldbergs out of with CIE!
    The element of magnetism I'm not fully sure how we would add, but I'd love to see it.





    Link to his youtube profile so you can check out his other 3D rube goldberg / dominos

    http://www.youtube.com/user/Kaplamin...
    •  
      CommentAuthorTheDudeFromCI (Advanced Member)
    • CommentTimeMay 11th 2016
     
    Lol, you mentioned me in that post Bio, but I didn't have permission to respond, or even like the post. xD
    I think the use of magnets was amazing. I really loved all of the different simple mechanics used in that design.-----------------
    Orange is my favorite number.
    • CommentAuthorXyuzhg (Moderator)
    • CommentTimeMay 11th 2016
     
    That video has been fairly popular across the internet lately.

    I suppose you could simulate magnetic balls via the Gilbert model (dipole approximated by two point magnetic charges), together with an efficient spatial lookup allowing for more complex systems. If you limit your attention to balls, then you could in principle model the interactions in three dimensions, allowing for the trick at 0:43.

    This wouldn't be easy though. :P-----------------
    Hopefully PA is inconsistent.
    •  
      CommentAuthorBioManiac R2 (Moderator)
    • CommentTimeMay 12th 2016
     
    using the existing equations we have for simulated gravity I'm sure we could get polarity to work on objects of a specific collision category or within a certain radius. Simply adding a small friction factor that would need to be overcome would prevent movement around the rest of the screen.