Really no easy way to join tubes?

[JST]

There has to be an easier way to do this than I have discovered.....

To join two mitered tubes, as one might in a handrail, etc, there seems to be no simple way to proceed. None of the usual constraints seem to do what is waned, to get these two mitered tubes to line up face to face.

They can be mated, but they can slide around. They can be constrained to a plane, and then another plane put at right angles, at which time they can be constrained to the planes and line up. A system of axes can be created and they can line up with that.

But there seems to be no simple combination of constraints that will "grab the right points" to do it without creating an external structure of axes and planes.

Am I missing something?

 

[TylerDurden]

Like other smooth featured parts, some references are needed.

Another option is to add features expressly for alignment.

 

[Lew_Merrick]

It requires some additional reference geometry, but the attached Package should give you some ideas...

 

[HaroldL]

I would make use of the part's Reference Geometry to help get them aligned and, as has been mentioned, creating other reference planes and axes as needed will help. You can't always rely on the part surfaces as you have found.

In your example of mitered tubes, rather than using Draft to "miter" the ends of the tubes I would use offset Planes, Axes and Angled Planes to define the ends of the tube with an Extrude To Geometry (angled plane) function. Then those planes are available for constraining in the assembly.

 

[JST]

Hmmmm.

The goal here is not to properly canonically define the shapes, but to fit up parts and verify that they DO fit up. That does not WANT all sorts of external stuff, it wants to be able to fit the parts and see that they go.

I usually define parts from one assy, using all sorts of planes etc, but maybe not all parts at once. Then, when all the P/Ns are good, I build it up USING the faying surfaces and pin/bolt holes, in all the ways that are "poor CAD practice", strictly as a check. Basically building it with the same references that it physically fits together on.

I don't, at this stage, WANT to use fits of external planes, etc. I'll use axes of parts, defined points (if there is a way) where required, but I like to keep that down. It's not an idealized model, but a virtual physical assembly. Boeing etc do something similar, and even use it for training.

Some items are just very difficult or even nearly impossible to fit up that way. The tools do not exist within GMD to do it. To be fair, some of the physical parts can have many of the same issues, so fits using what are nearly "jigs" are allowable, I suppose.

 

[bigseb]

... it wants to be able to fit the parts and see that they go.

It also puts the lotion in the basket or else it gets the hose again

(sorry, couldn't resist)

OT: Sometime there's no way around it. Creating a plane/axis for reference purposes doesn't take long. The question is how do you expect to constrain the two parts. Obviously the mate but after that? There is nothing else for the software to constrain so reference geometry will be needed.

 

[cherkey]

LOL there was no "Like" button for that :lol:

 

[JST]

... it wants to be able to fit the parts and see that they go.

It also puts the lotion in the basket or else it gets the hose again

(sorry, couldn't resist)

OT: Sometime there's no way around it. Creating a plane/axis for reference purposes doesn't take long. The question is how do you expect to constrain the two parts. Obviously the mate but after that? There is nothing else for the software to constrain so reference geometry will be needed.

Precisely the problem.

Even reference geometry can be an issue with some sorts of form.....

All very well when you yourself make the part, but if you receive a STEP file, you may not get any reference geometry, and then you find you are still limited because the same issues that limit directly constraining, ALSO limit you in getting external geometry defined with reference to the existing STEP part.

For the pipes, the most direct method is to put a point at the intersection of the end plane and the axis. The points, if co-located will "stick the parts together", but it is the long way around.


There could be some more possible constraints, or options to existing.

1) Intersecting axes. This should not really be an issue, it is a matter of automatically using the axes which you would otherwise have to create, but which are "known". If you think about that, a "mate" for instance, is no different in its internal action, it finds/creates planes and mates them. For the pipes, a "mate" and an "intersect axes" would locate the miter.

It's not too different from aligning to a plane, but VERY different from aligning axes.


2) I do not think there is any direct way to put a sphere centered at a point, directly. I suppose this is, really, a subset of the intersecting axes.

3) Tangent is good, but I find I want some more control. I don;t quite know how to describe what I want, yet.

4) The angle option on some constraints is odd, it often does not act as expected. Some settable parameters to it would be good. Or knowing /showing the zero plane.

5) mating cones has been an issue forever. There really isn't any way to do it in a practical manner. Needs to happen.

6) a direct perpendicular constraint, possibly one that is not limited to perpendicular, but allows an enterable angle value, defaulting to 90 deg.


One way to look at the matter is that there are a lot of options in the sketch function, and many of those would be good if they could be done in 3D as well. That's not a complete set, but....

And, yes, I realize that the underlying ACIS is probably what limits many of the functions, and disallows others.