Cable chain assembly movements

Answers

• Ste_Wilson Member Posts: 629 EDU

  May 28

  Not able to look at your doc at the moment but a quick fix might be changing some some rotate mates to cylindrical might help. Sometimes slight rounding errors mean your rotates aren't as aligned as you think they should be.

  https://hwh.stephenteacher.com/txtbook.html

  0
• rich_north Member Posts: 21 PRO

  May 28

  Tangent mates can also be more reliable as a point on a surface rather than cylindrical surface to planar surface (for example) in my experience.

  1
• nicholas_witham Member Posts: 23 ✭✭

  May 29

  @Ste_Wilson Hmmmm, I'll double check my rotate mates again, then try cylinder mates.

  0
• nicholas_witham Member Posts: 23 ✭✭

  May 29

  @rich_north

  That's a strange anomoly, the example I followed to create my model used tangent mates on the cylinder surface aligned to the curved surface of the path.
  

  0
• rich_north Member Posts: 21 PRO

  May 29

  @nicholas_witham I have had situations where the cylinder will switch sides on the path surface. That ambiguity can sometimes be a problem as you drag things around. If it works, that's all that matters really.

  The energy chain example you provided seems to run through all the allowable movement based on your constraints.

  0
• martin_kopplow Member Posts: 1,434 PRO

  May 29

  Usually, one cylindrical mate in the chain (EG on the last link) will be enough, just to allow a minimum 'freedom' in axial direction that makes the math easier for the system. If that fails, a ball mate that rides on a slider mate might also be a good option. It's all about avoiding overconstrainism.

  0
• wille_j Member Posts: 72 ✭✭✭

  May 29

  Here's one from @NeilCooke

  Link

  0
• S1mon Member Posts: 4,234 PRO

  May 29

  I haven't tried this yet, but I wonder if one way to avoid creating so many mates would be to make small subassemblies of a series of links and then make a next level assembly with multiples of those… Obviously you have to be diligent about the final number you need for the top level, and there's no shortcut for the tangent path mates.
  
  It would be really nice to have a better way to replicate mates along a path… sounds like I need to create an improvement request.

  https://forum.onshape.com/discussion/31088/replicate-along-curve

  Simon Gatrall | Product Development, Engineering, Design, Onshape | Ex- IDEO, PCH, Unagi, Carbon | LinkedIn

  0
• nicholas_witham Member Posts: 23 ✭✭

  May 30

  I've tried changing the tangent mates from "face to face", to now face (of path) to edge of chain circle
  Seems to be the same, quite glitchy, especially once the number of chains gets too much.
  At times i can get it to work, but then sometimes i get to the same point and it overconstrains itself….quite frustrating. Maybe the movement on the tangent gets mixed up and tries to flip around.
  
  Perhaps constraining the tangents as "surface of path" to "point of chain", as @rich_north has suggested.
  I wonder how I can create a point on the centerline of the chain link shafts….a point, maybe a line. I don't think a mate will work as a tangent mate

  0
• nicholas_witham Member Posts: 23 ✭✭

  May 30

  I added a surface to the link assembly, tried to tangent mate the link surface to the path surface, but that method allows the tangent to go beyond the actual surface as long as the extension of the surfaces still tangents.

  

  So I tried tangent mating the vertex of the link surface to the face of the path surface, and that seems to be a more reliable mate, as it can't get locked trying to move one way instead of the other.

  

  Now I think i'm suffering from some lag issues. The cable chain rolling will work well by using the slider mate variable value, but it is extremely glitchy just trying to drag the chain end.

  

  This might be as good as it gets for me

  image.png 279.9K

  image.png 336.9K

  image.png 251.8K

  0
• martin_kopplow Member Posts: 1,434 PRO

  May 30 edited May 30

  I'm not sure what everybody expects from an energy chain. If we just model those conditions it has, the thing moves as chains do and behaves as expected, running pretty smooth for me.

  

  https://cad.onshape.com/documents/67fd4b93c70ca60f6a3b003a/w/8a36287dc8b351903f517cba/e/81fa9942f4b32a00f48e7d58?renderMode=0&rightPanel=variableTablePanel&uiState=6a1ab0d1dfcd8f502b38afbc

  Video (realtime):

  https://nextcloud.webo.hosting/s/xd4DZRZ82KYGdpD

  The free end will always have more distance from the fixed end than 2x minimum bend radius, so there'll always be a little wiggle in the chain loop, but that's just typical for chains.

  If we want to introduce a lot of conditions to make it act as if it was under the influence of gravity, that's a whole different story, though. But OS has no gravity simulation built in, so we end up with a really complex set of conditions and slow solving. If I have to model energy chains, I usually go the simple way and just let the chains be chains, instead of making them move like they never do but in an idealized sales video. ;0)

  (Edit: I made the max angle between links a varibale. It is 40° by design, but setting the variable to someting just below the point where it hangs (38.5°), helps to take out most of the wiggle:)

  

  Screenshot 2026-05-30 111101.png 584K

  Screenshot 2026-05-30 114418.png 494.8K

  1
• nicholas_witham Member Posts: 23 ✭✭

  May 31 edited June 1

  Thanks @martin_kopplow

  I don't believe im expecting it to do more than it should…i've essentially copied your method except for a couple of small differences..

  I have 40 revolute mates, and 40 tangent mates…all trying to recalculate a new position on every movement.

  Is it possible there are too many mates in the assembly for the system to compute dynamically (hence impossible to drag), and probably right on the limit for it to compute with the provided variable (hence the time it takes to move after entering the new variable).

  I can move the chain along the path using the variable value on the slider mate, but only if I make small increments of maybe 25mm each time (however the change is extremely slow). If I make a change too great the mates become over constrained. For example if I change the slider length from 650mm to 100mm in a single move it will fault, but if i make small 25mm increments from 650-100mm it will be fine.
  If I try to drag the path sketch nothing will happen at all.

  

  Here's an explanation of my method…………..

  In document 1 I have modelled the chain link and the chain ends (A and B).

  Each chain link is made of 4 individual parts, and combined in an assembly.

  https://cad.onshape.com/documents/fe88b93161d123511a04d621/w/45c9eb5be90db7384713f0fd/e/a24e88528bcd26b26540e673

  https://cad.onshape.com/documents/fe88b93161d123511a04d621/w/45c9eb5be90db7384713f0fd/e/abb90b6a05d3d9f6107b6d90

  https://cad.onshape.com/documents/fe88b93161d123511a04d621/w/45c9eb5be90db7384713f0fd/e/9e724aca900aebf80183a328

  In document 2, I have a part studio with the path…made from a 3d fit spline extruded to a surface.

  https://cad.onshape.com/documents/9fcd0c80829f11bb770a03ae/w/5d3ce76f78241bd88e22f475/e/9ac5e191c8a19ae1a5809f31

  Also in document 2 I have an assembly that references 1xchain end A, 41xchain links and 1xchain end B.
  This assembly also references the path sketch and surface.

  The chain end A mated to the path sketch with a slider…without any limits:

  

  Each chain end and link is mated with revolute mates…with minimum 0° and max 45°, which is more than required to get around the bend radius of the path. I've done this as a configuration instead of a variable so I can change it once and have it update to all instances. A variable would make me have to update every instance.

  

  Then each link is mated with a tangent from the vertex of a surface within the chain link assembly, to the path surface.

  Using the vertex instead of using the cylinder surface means the system calculation doesn't get confused and bound up.

  

  image.png 89.4K

  image.png 282.4K

  image.png 156.3K

  image.png 288.9K

  image.png 76.1K

  0