Jump to content


Monthly Contributor
  • Content Count

  • Joined

  • Last visited

  • Days Won


Everything posted by jed

  1. @Hrvoje I understand how to represent polar coords in cartesian in order to use formula spline. By 'differ' I meant you find it a good way to get a log spiral and I don't. Converting 1 equation with 2 variables into 2 equations with 3 variables - an extra unnecessary variable t - makes things more complex IMHO. Those 2 equations do not shout out 'spiral' to me, but the polar version does. I must admit that my approach to C4D is to use math, XPresso, Python etc and ignore built-in functions that I haven't bothered to learn about. A frequent comment on my stuff is 'there's a field that can do that'. I have, of course, never tried to learn fields...
  2. I used to think square cogwheels were just another fake YouTube thing, but I've almost got it sorted. Still WIP
  3. I think we're going to have to agree to differ on this.
  4. @Hrvoje can formula spline use polar coordinates ie radius + angle ? Spirals are easiest described (and understood) using polar vs cartesian. TBH I've never really understood what 't' is in formula spline. in math - y = f(x) ie y is a function of x in C4D x = f(t) ie x is a function of t y = f(t) ie y is a function of t
  5. @rasputin I just used the general log spiral equation where angle is in radians, e is the natural log base (aka Euler's number 2.71828) and tweaked the constants a + k until it looked right. I tried using a helix and radial bias but the gears were jumping. I guess it's the properties of logs that makes it work. Also, when I added the struts, with a helix I had to use a spline on the step effector to make things fit, but with the log curve the struts fitted without a spline. I usually animate gears with XPresso, but with the varying radii that would be beyond my math skills. Dynamics to the rescue ! The video above is version 2 - I didn't like how dynamics had introduced a small collision gap, so after baking for TR I increased the teeth size. The dynamics is Compound Collision Shape, but with all the geometry it was too slow in the viewport. I added some switched off rigid body tags to parts that don't contribute to the dynamics calculation. Here's the video I got the design from. The guy has some good stuff eg non-circular gears I'm still trying to figure out square gears...
  6. I saw a video on YouTube by a guy who makes kinetic art mobiles. He had an interesting one made out of toothed spirals, so I had a go at recreating it in C4D. I tried using a helix spline, but couldn't seem to get the objects to interact correctly, so I made a proper log spiral using tracer. The movement is generated using dynamics - there's one collider that's rotated in XPresso driving 2 rigid bodies. demo.c4d https://en.wikipedia.org/wiki/Logarithmic_spiral make log spiral.c4d
  7. Some more of my Reuleaux Triangle experiments - in this scene, the rotor is a collider rotated with XPresso. The fixed width parallel bars are a rigid body on a slider connector, and C4D dynamics does the rest RT parallel.c4d here's the same thing in a rigid body rectangle that is on a planar connector. The square moves, so the camera is parented to it making it appear stationary RT in square.c4d if you make the Reuleaux Triangle from the default size N-side, it fits exactly into a honeycomb cloner RT honeycomb.c4d this vehicle with Reuleaux wheels on a keyframed tilting surface rolls at constant height RT wheels.c4d I think that's enough wobbly wheels for this week
  8. Reuleaux triangles are weird https://en.wikipedia.org/wiki/Reuleaux_triangle well, it amused me reuleaux2.c4d
  9. Re teeth ratio - are you familiar with the Looney Gears ? It's a toy with asymmetric sun + planet gears that have prime numbers of teeth. The game is to work out how many times you have to rotate the sun gear to get all the gears lined up. There's also Somsky gears, with an extra cog. My version -
  10. The math problem I encountered was that for one cog to drive another they must have the same linear speed at the point of contact. For a given rotation speed, the edge speed is proportional to radius, and the cogs have varying radius. I found some equations on Wiki etc, but it was beyond my math ability
  11. I like to make stuff with gears and normally use XPresso or Python to get the required rotations. Recently I needed to use oval gears, but the math proved too hard. I thought I'd try C4D dynamics, not really expecting it to cope with non-circular gears. To my surprise everything worked - so I made some 3, 4, 5, 6 etc sided cogs and dynamics worked for those objects also (with a bit of help from constraints and XPresso). Using wheel suspension instead of hinges helps to keep odd shaped things things touching. Here's a video of my tests While I had my math hat on, I made this - no idea what it's supposed to be.
  12. Rendered quite fast on my Ryzen 3700X.
  13. It's fairly easy to use. You make the scene in AE - shapes etc, then click Newton which opens the scene in the app. Adjust gravity, velocity, friction etc and click render. That returns a copy of the scene back to AE with keyframes for position. In my clip, I made a circle shape bounce in Newton, and used the shape as a mask to reveal the alternate text (with distortion).
  14. I think Cairyn means something like this brake.c4d
  15. I had a tinker with Newton 3 recently. It's a Physics plugin for After Effects that can give AE shapes velocity, friction, bounce etc. Also has springs + levers, https://www.motionboutique.com/newton/ My first scene -
  16. It's not the kind of animation that can be made into a loop. Although the head is driven by a sine node (itself a looping function), the rest of the body reacts dynamically. The start position is never repeated. +1 to Turbosquid for the nice free model.
  17. jed

    Inverse Kinematics

    I'm not much of an IK expert, and my method may not be the best way - align all geometry as accurate as possible ctrl click with joint tool to make joints at rotation points (use snapping) select 1st + last joint > create IK chain (this makes a goal) click add pole and move to suitable spot drop geometry parts into joint hierarchy as children moving the goal then moves the model parts correctly I've used this method for robot arms etc
  18. Timeline, view, show animated. I see last + 1st frame are 0 deg - maybe set it to 1oop at 119F.
  19. You can get quite realistic semi-truck anim with dynamics. In this file the truck follows a target on a spline using XPresso steering - speed adjusts to keep up. You can keyframe the target to stop at road junction etc. Trailer is connected with ball + socket and 'trails' around corners. Since a heavy poly dynamic truck runs slow in viewport, my method was to use a dynamic primitive rig and make the poly parts children of the dynamic parts. You can re-use the dynamic rig for other trucks, using transfer tool to move the wheels + suspension to the real wheel locations. xp truck.c4d
  20. jed

    Inverse Kinematics

    I had a go using IK dropbox
  21. I think your cogs are a bit small for dynamics - the teeth are a comparable size to the collision margin. Dynamic cogs + motor works OK at the default cog size dynamic cogs.c4d does it have to be dynamic ? Your cogs work OK with XPresso rotation 2 xp cogs.c4d
  22. I think it has to be an even honeycomb, the same pitch as the channel spacing, to get a nice bell shape.
  23. Did you have a honeycomb array of colliders ? If you look at the math, each collider is a left or right decision and the number of possible paths to the channels is a bell curve. My model is a fair approximation, but I think maybe the balls colliding with each other affects the result - I should have had the balls more spaced out time-wise. Stats being what they are, you get the best results with a large number of objects. I made that scene in R14 - before honeycomb cloners
  24. A bit off topic, but related to balls and bell curves - There's a gadget called Galton's board (also Quincunx) where balls fall through a honeycomb of collider rods, and into some collection channels. The math says the distribution in the collectors should be a bell curve. It's related to Pascal's triangle. I made one a few years ago - it's just a glorified egg-timer... https://www.mathsisfun.com/data/quincunx.html https://en.wikipedia.org/wiki/Bean_machine

Latest Topics

Latest Comments

  • Create New...