Mates frustration

dan_engerer

Hi. I have been experiencing a lot of trouble with Onshape's new mate system. I realize that it is supposed to save work by eliminating steps. But sometimes, you just want to create a simple constraint, period. For example, why does the planar mate select anything other than planes? I wanted to create JUST a planar mate while still allowing full Z rotation, but Onshape now asks for a rotational input (default is 0 degrees). 

I don't know if I'm using the system wrong, but I was not able to create a simple mechanism where an object moves up and down in a track while being driven by tangent pin/slot mechanism. See image. The grey piece is supposed to move up and down in the magenta channel when the orange piece plunges left and right. This would be really simple to do in solidworks: 2 planar mates to keep the grey piece in the channel, and 1 tangent mate to keep the pin in the orange slot. Boom. Onshape won't let me do this: As you can see, the back face of the grey part is not parallel to the channel. I keep getting overconstrained mates or mates that APPEAR to work until you realize that the part is not truly behaving in a planar fashion but following some other plane. 



Question: How do I create JUST a planar mate that ensures that two surfaces touch eachother, no more and no less? 

owen_sparks

Hi @dan_engerer, sorry can't see an image here.
Just some text stating "Inline Image 2"
Owen S.

billy2

Dan you need to start out simple and build upon small successes:



Start with a 4 bar linkage.

Ben_Misegades

Since I can't see what exactly you're working on, I can't try to figure out your exact issue, but here is something I quickly modeled that has a block sitting inside of a track, and you can slide it up and down. 

https://cad.onshape.com/documents/d5019ed3841e8ed2e01c3608/w/86fffcfc6f9a288002d6aa5f/e/56e4ecfbd0267075d133a8fc

Does that help any?

Research_01

Dan

You bing up a interesting point. Now when you do a planer mate it is only constrained to the face and can rotate if you do not enter a rotational input. But just moving the part with the mouse I see no way to rotate the part just translate it. However if you make two boxes and do a planer mate then add a revolute mate on a corner you can then rotate the part not translate it. So in the end it may be a feature request to ask for a ability to move and also to rotate a part via the U.I. 

https://cad.onshape.com/documents/f7c473d48d416cee23bcacbb/w/76ef25994d82a876497f48fb/e/3f3fb48335fe5d6058759b6f

In the part studio 2 and assembly two I have a pin and slot mated for you to dissect as well. It allows the pin to follow the slot and the pin can rotate. However when you drag the pin it can flip the tangent mate alignment in some odd behaviour. I'll send that in as a problem to Onshape. 

IN the third assy I am trying the pin slot mate to work things out but not sure this will serve your purpose. It only allows translation along one of the main axii so if the slot is canted off like in my  example the pin will drift out of the slot. also it will go past the end points of the slot if you do not input limits into the slot mate command.

Research_01

Ok one thing solved here by @lougallo the tangent mate I used a edge and it allowed the flipping of the pin in the slot. However when one picks the surfaces of the slot this goes away. 

brucebartlett

Ben_ said:

Ok one thing solved here by @lougallo the tangent mate I used a edge and it allowed the flipping of the pin in the slot. However when one picks the surfaces of the slot this goes away. 

That assembly2 slot mate is really neat. Initially, I went looking for limits but none needed as it is controlled by the tangent mate. I'll have to use this in the future. 

billy2

Wrong mate connector, use this one:


In 3D you have 6 degrees of freedom (DOF). So, how do you lock them down? How do describe the movement between 2 objects?

Use mates (sw)? Lock one down at a time, requires minimum 3 constraints to secure. You have to understand all the constraints to understand the behavior between 2 parts. 

Use mate connectors (os)? Use one to describe the DOF between 2 parts.

Typically OS will have 1/3 the mates as SW. Looking at the mate connectors in OS is easier to understand than the mates in SW.

Please understand that mates(pro/e) were conceived in the last hour just before a release. They wanted to remove DOF because people don't think in terms of DOF. Their idea wasn't a perfect solution.


Now here's one for you, fastened mate connectors should be frowned upon. Why? Because if parts aren't moving relative to one another, put'm in a part studio. Mate connectors should be used between part studios with multiple bodies that aren't moving relative to each other. Why have a bunch of part studios with one body only to fasten them in the assembly? This seems silly.


And another thing, please version your document before sharing, then give me edit rights to your document. You can always get back to the original version and branch to continue. When you share with read only rights I have to make a copy to see what's going on. I promise I'll branch using my name before making any changes to your design and hopefully others will branch before changing with their ideas. In the end we can have a document with Ben's ideas, my ideas and anyone else who wants to contribute. Right now copying documents is the old fashion way and nothing accumulates. Instead we just end up with a bunch of documents with no continuity.

Ben_Misegades

@billy2 took me a while to figure out how to make that Slider mate work. Wouldn't mind seeing some more functionality added to it...rather than setting explicit numerical limits, how about letting us pick a vertex or face or similar to set the limit of movement?

coleman

Coming from SW myself....I can guarantee you guys that you will come to realize the onshape mating methodology is much better than SW.  
It just takes sometime to get acclimated to.  

3dcad

Yep, with traditional mates it was all about how fast you can click dozens (hundreds) of mates in and after that crawl through them all to solve why you lost certain movement. With Onshape it needs more thinking but many things can be solved with a single mate. This thinking should start when creating parts into studio(s). 

And like @billy2 mentioned static parts should be created in their desired place in first place and just use group mate for fixing them in place. This felt a bit weird at first since I tried to create assemblies like I would do in real world (this face needs to connect with that face etc..) but with Ons this happens when creating parts using constraints.

One thing that I would change is possibility to add 'live' part studios into assembly so that it follows any changes made to that studio. In this case assembly should follow part studio constraints ie. create hidden mates in background. This could also allow movement according to constraints. 
Now adding complete studio is just 'shortcut' to adding all parts one by one and parts created after insertion needs to be added separately. 

Ben_Misegades

@3dcad I second that motion. I think in order to fully utilize the "power" of part studios, they ought to be "live" in assemblies, i.e. you add new bodies to your part studio, it is reflected in the assembly, etc.

billy2

Tangential mates, gotta love'm.

Thank for sharing Ben,


Understand that a tangent mate connector is a weird mate connector. It can be added to other mates. It doesn't follow the one mate/ one motion construct. Oh well, we need it and sometimes the math gets a little ugly. 

billy2

"add 'live' part studios into assembly "   &  "they ought to be \"live\" in assemblies" aren't these incontext or designing in the assembly? we have that now and it works really, really well.

The problem with groups in an assembly is: if a part changes it can interfere in the assembly. Groups won't space the parts out. Groups are great if the components are off the shelf items that'll never change. Otherwise, move the groups into a partstudio.

billy2

dan-

I should have shown the mate connectors:


mate connectors are really easy, after you understand them. Figuring them out, takes a little effort. I finally got'm when adding derived components and positioning them inside a part studio. Maybe this might be a better way to figuring out mate connectors?