Lofts twisting. Why?

[ccut93]

I am trying to build an aerodynamic(ish) canopy for my quad-copter, and am at a loss for how to do it. I tried a simple sphere, but I don't like how the camera sits inside of it. I envision something of an egg shape with the pointy end pointed towards the tail, that somehow blends into the square base.

I tried using lofts to do some cross-sections, but I can't seem to get it to work reliably. I get somewhere close, then somehow, there is twist induced between my sections. How do I control and eliminate this twist?

Design is my own.

 

[UGMENTALCASE]

Morning @ccut93 there's no way on this Earth any CAD software will be able to compute the change between the second, more square shaped ring, to the next ring up. There's way too much of a gap/expected shape change. Maybe make the bottom a basic square ish shape using a sketch and extrude, then start on top of that with the contour type lines?
I suspect some guidlines may help, but I don't do much lofting in Alibre and not sure it has an option for that or not. Even so I think you'd still see a similar thing with the twist/surface bumps.

 

[stepalibre]

Short answer:
No smooth transition. Those cross sections profiles are being stretched and pulled at the corners. The rotation of the base curves are going in the wrong direction for the loft to correctly transition. If you want the upper profiles to be @45 degrees you'll need to use another method or try rotating them 45deg instead of the base two sketches. There are many methods that can achieve good results.

 

[VECDESIGN]

I am trying to build an aerodynamic(ish) canopy for my quad-copter, and am at a loss for how to do it. I tried a simple sphere, but I don't like how the camera sits inside of it. I envision something of an egg shape with the pointy end pointed towards the tail, that somehow blends into the square base.

I tried using lofts to do some cross-sections, but I can't seem to get it to work reliably. I get somewhere close, then somehow, there is twist induced between my sections. How do I control and eliminate this twist?

Design is my own.

I can suggest you just add rails to the lofts, 2D or 3D. My experience is that are mandatory for getting symmetry or controlled asymmetry in most lofts in Alibre. An alternative, not so clean depending on the shape you aim, is to cut the asymmetric part you get from the loft, by removing the side you don't like, then mirror part and next make fillets in the joint if continuity in curvature is desired.

 

[stepalibre]

The problem with guide curves in Alibre is that they are not controllable enough or stable in complex geometries. They break too easily with the slightest dimensional change, and the loft settings are inconsistent unlike any other.

 

[VECDESIGN]

The problem with guide curves in Alibre is that they are not controllable enough or stable in complex geometries. They break too easily with the slightest dimensional change, and the loft settings are inconsistent unlike any other.

Agree that. I suggested to support team some way to force symmetry in the lofts but since with just a single rail can be fixed in most cases, I think they did not pay attention. My conclusion is that the root is that when the u v curves of the surface are created, the start point in each section is not selected with some symmetry. Probably the algorithm is just selection some first drawn element or closest, but not forcing to some in-code built rail. It would force to the usaly expected symmetry in our minds and avoiding the chaotic actual twists. The issue is mostly with closed section curves or when curves are created in each section with different quantity and type of elements. Close fully tangent section curves have no clear start point with symmetry, depends on how are built.

 

[Ken226]

I am trying to build an aerodynamic(ish) canopy for my quad-copter, and am at a loss for how to do it. I tried a simple sphere, but I don't like how the camera sits inside of it. I envision something of an egg shape with the pointy end pointed towards the tail, that somehow blends into the square base.

I tried using lofts to do some cross-sections, but I can't seem to get it to work reliably. I get somewhere close, then somehow, there is twist induced between my sections. How do I control and eliminate this twist?

Design is my own.

I couldn't get it symmetrical by playing with the loft settings, but came kinda close.

 

[ccut93]

Thank you for all your feedback. When I've done lofts in Solidworks a long time ago, the rule I learned for good lofts is to ensure that each sketch has the same number of segments. I tried that here, but seems unstable. I will look into learning rails in alibre.

Lofting itself seemed not to be well suited to this specific situation. Would you recommend any other solutions. I see design has surface modeling, but I haven't figured that out yet.

 

[VECDESIGN]

Just see how improves with just a single rail guide and no need for extra surfaces created. I would suggest to add three additional for this type of surface.

 

[stepalibre]

A seamless loft is often required for some geometries and modeling methods. I wouldn't add any additional, only if the design allow it or the topology difference isn't important. Guide curves and reference geometry can lock you into a shape or not allow you to make changes without surfacing or direct modeling tools. If a guide curve is used in this case I would make one side an mirror the other. That would be more flexible and easier to make equation driven.

 

[VECDESIGN]

I think the sketch view in the pictures can confuse. It is closed surface then both have a seam. The difference is that in the twisted loft the seam is not symmetric but in the one I attached it is just forced to be symmetric. This is what I tried to explain about the starting points of each section profile. Just a single rail guide forces a good symmetry in the seam then in the surface created.
twisted
non twisted

 

[stepalibre]

I think the sketch view in the pictures can confuse. It is closed surface then both have a seam. The difference is that in the twisted loft the seam is not symmetric but in the one I attached it is just forced to be symmetric. This is what I tried to explain about the starting points of each section profile. Just a single rail guide forces a good symmetry in the seam then in the surface created.
View attachment 43953twisted
View attachment 43954 non twisted

I forget Alibre doesn't have surfacing tools so you don't have many options.

 

[NateLiquidGravity]

Is it symmetrical? You might get better results doing half and then mirroring it.
You might also try creating temporary extrusions for each end to loft between the faces (selection the actual faces in the loft instead of the sketches). I've had that work great for making things tangent.

 

[kcoffield]

Haven't been on the forum in a while, and realize this is a couple month old topic, but am always interested in lofting techniques and improving my lofting skills.

I can suggest you just add rails to the lofts, 2D or 3D. My experience is that are mandatory for getting symmetry or controlled asymmetry in most lofts in Alibre. An alternative, not so clean depending on the shape you aim, is to cut the asymmetric part you get from the loft, by removing the side you don't like, then mirror part and next make fillets in the joint if continuity in curvature is desired.

Just see how improves with just a single rail guide and no need for extra surfaces created. I would suggest to add three additional for this type of surface.

What is meant by "Rails". Is it a feature in the Alibre lofting control function or a model feature?

The problem with guide curves in Alibre is that they are not controllable enough or stable in complex geometries. They break too easily with the slightest dimensional change, and the loft settings are inconsistent unlike any other.

I agree. I struggle mightily with centerline guided lofts along 3D splines. Lofting stringently requires tangency at the loft faces and even when this is achieved with the 3D splines, it's more often than not that the loft refuses to execute. The error codes that are cited are often so abstract or obscure, they offer little help to understanding the reason for failure if you are not someone steeped in the inner workings of Alibre. It's a bit ironic for me because centerline 3D spline guided lofting is probably the single most important feature for me in my modeling needs and IMO compared to other softwares, perhaps one of the weakest areas for Alibre........sighhhh......but I'm still trying to improve.

Thank you for all your feedback. When I've done lofts in Solidworks a long time ago, the rule I learned for good lofts is to ensure that each sketch has the same number of segments. I tried that here, but seems unstable. I will look into learning rails in alibre.

For sure this is helpful and good practice.

You might also try creating temporary extrusions for each end to loft between the faces (selection the actual faces in the loft instead of the sketches). I've had that work great for making things tangent.

I've found lofting between faces to be convenient but not sure how it helps with tangency of the loft, especially if it is a guideline controlled loft.

Best,
Kelly

 

[DavidJ]

I believe that 'rails' is an alternative name for Guide Curves.

 

[Max]

Guide Curves / Rails:



Also note that from the top view your 3rd from the top sketch is in some places smaller than the 2nd from the top - this forces the loft which is otherwise gently going outward to have a large convexity change in a small space causing squirrely results:


If you can try to make each profile larger:


Combing the 2 approaches gives a well behaved result:

 

[stepalibre]

Guide Curves / Rails:
View attachment 44404


Also note that from the top view your 3rd from the top sketch is in some places smaller than the 2nd from the top - this forces the loft which is otherwise gently going outward to have a large convexity change in a small space causing squirrely results:

View attachment 44405
If you can try to make each profile larger:
View attachment 44406

Combing the 2 approaches gives a well behaved result:
View attachment 44407

How did you make those rails? Share file?

 

[Max]

Just a spline a 3D sketch made in top-down view connecting the sketches together; did have to rotate view for the bottom most sketch connection since it is identical to the one above it. Created 3 splines in one 3D Sketch.

What I would also do, since I assume the bottom 2 sketches represent a regular extruded shape and not a loft, is only loft between 4 of the curves and extrude the bottom. This prevents bulge from the loft where we assume we want there to be planar faces.

Example part file attached.

 

[Max]

I've found lofting between faces to be convenient but not sure how it helps with tangency of the loft, especially if it is a guideline controlled loft.

It is a limitation of the kernel that you cannot use faces and also guide curves in a loft. However, see the example above.

 

[HaroldL]

Just a spline a 3D sketch made in top-down view connecting the sketches together;

It appears that you have three guide curves in the one 3D sketch. How is that possible, I thought multiple guide curve sketches had to be separate entities?