Validating Sweep Trajectories

[nkpham]

I know that there is a glitch in Pro E that doesn't allow all 3D trajectories to be used in sweeps. Has anyone else encountered this and if you have, have you found any solutions?

 

[ProFishent]

I don't know the logic behind why PTC does not allow 3d curves in regular sweep's. But use VARIABLE SECTION SWEEP for 3d trajectories.

 

[miked]

The reason it doesnt work for a sinlge 3d curve is section orientation.

Thats why it works with var sec not the regular sweep.

This really p#sses me off.

Mike

 

[eServ]

I'm not sure what version you guys are using, but I just did a sweep today with a two projection datum curve. I did have some problems, but to solve it I swept the protrusion on one end of the curve and then swept it on the other end. It worked fine.

 

[nkpham]

I created a 3d curve using an equation. then I split the curve into multiple segments. On some of the segments, the sweep works fine. On the other segments it refuses to do it. I am actually making one segment then using copy to reproduce the other segments. It won't copy because some of the segment trajectories are termed invalid.



BTW, to solve my problem, I'm using the swept blend feature and am blending two identical cross-sections. Is using Var Sec Swp better? If so, how would you do this? I've never used that feature before.

 

[ProFishent]

I don't know what I was thinking earlier. A regular sweep does work on 3D curves. If you want to specify a different trajectory for the section to be normal to then you use Var Sec Sweep. Your problem is probably with the curve segments not being fully tangent to eachother. Maybe turning up your model accuracy would help. Why are you spliting the curve anyway? Why not just one curve and one protrusion?

 

[ProFishent]

BTW with variable section sweep your origin trajectory is what your section runs normal to. While your x-trjectory is the direction the feature actually follows.

 

[nkpham]

I am modelling a bridge with multiple cross-sections. The trajectories represent the different segments. The curve segments should be fully tangential since they are simply splits from the original cruve. My accuracy is generally maxed out at .0001 so that shouldn't be the problem.

 

[ProFishent]

Are the sections different for each segement? If so, why not a swept blend anyway. If not, why split the trajectory into segments?

Either way a regular blend should work. Sounds like you really have just hit a glitch like you originally stated!

 

[nkpham]

There are several different cross-sections that get reused later. Hence the need for the copy function. I originally was just going to split the curve and make separate sweeps for all the sections. When I got the invalid trajectory error, I tried to find a better solution. Someone suggested that I use the swept blend function. I'm not sure exactly what he meant, so I just created the same section twice to blend them. Can you explain exactly how this swept blend feature works? I'm not really sure if I am doing this correctly. I get discrepancies because I create the first sketch, then import it to blend it with itself.



I think your suggestion of using the var sec sweep mite be better. If I were to use this, what would I use as my x trajectory? The origin trajectory would just be the curve segment of the particular cross-section, rite?

 

[ProFishent]

If you want the cross section's to be parrallel, then you would use a Var Sec Sweep. Pick normal to origin traj. Sketch or select a straight line as your origin trajectory and The curve as your x-trajectory.



Otherwise use the swept blend. Select the entire curve chain as your trajectory. You will be prompted to sketch a section at each segment break. Import your section, and modify it to change accordingly at each sketch. Make sure your start arrow is in the same place in each sketch! You can skip segments if there is no geometry change, and even go back later and add more sections.

 

[nkpham]

will the swept blend mess up the places where the cross-sections join? i need them to be abrupt changes from being solid to hollow.



BTW, thank you for all your help. I appreciate it alot. =)

 

[ProFishent]

Probably! Sounds like an interesting project! I would suggest first blending the outside geometry, and then cuting out the hollow sections with individual swept cut's.

 

[nkpham]

Actually.... I can't create the whole bridge at once.... Pro E crashes when I have it at .0001 accuracy. I think it can't handle so much data at once. Besides, the cross-sections are really complicated. I say hollow because parts of them are. Each cross-section would need a lot of cuts to create the hollows.



For Var Sec Swp, does the sketch have to cross all the crosshairs? then I would have to create a plane tangential to the curve in order to make the origin trajectory.