Is it possible to precisly control a loft without excessive sketches in succession?

[ReeceEngineering]

I am designing a 3d printable RC airplane. All components have been weighed and drawn out on paper so I have a rough idea of where everything needs to be to maintain proper CG. This has provided me with the cross sectional area in terms of X & Z and the spacing between all the needed sketches in the Y direction.

While working through the nose of the plane I decided to try lofting the first few sketches to see what the result will be and I noticed that there is an odd buldge here. This is the sort of thing that I normally solve in Inventor by adding more sketches in very short intervals to more precisely control the shape of the loft. However this is also very time consuming and I wanted to ask if this is the best method or not?

The first sketch of the 4 I have so far is a single point, and I noticed that when I set the loft at that point to a tangent I can modify the weight of it so as to control how pointy the tip is. I am wondering if there is something simalar to this for the other sketches the loft passes through. The top edge of the nose is correct but the bottom has a buldge:

Is the best way to fix this just to have a few sketches there only a few mm apart? For frame of reference the nose section seen here is 250mm long.

[mango.freund]

Hi @ReeceEngineering you can get rid of this indentation with gradient lines. --- mango

[TrippyLighting]

One thing to mention is that the Fusion 360 an Inventor use the same geometric modeling kernel.

Lofts behave very similar between the two applications.

Usually such a loft can be controlled by using loft rails and not just profiles, however his loft already exhibits more than one problem. 

 

1. You are lofting into a singularity. Approaching the singularity - the point - curvature problems will usually effectively prevent shelling, , thickening or offsetting surfaces.

2. You are lofting this in one piece.

3. You are lofting filleted profiles.

Can you provide a photo of the paper sketches you created to help develop a more sound molding strategy ?

 

 

 

[TrippyLighting]

Here's a different approach providing much more control over the surfaces:

 

[TheCADWhisperer]

---

@ReeceEngineering wrote:

I can modify the weight of it so as to control how pointy the tip is.

---

When you Loft from Solids or Surfaces you have tangency and weight control at both ends...

I can only speculate on your true Design Intent.

[jeff_strater]

"Is the best way to fix this just to have a few sketches there only a few mm apart?"

No!  In fact, though it seems counter-intuitive, the more profiles you have, the worse your surface will be.  You goal should be to have as few profiles as possible in any loft.  Instead, as described here, control your surface with rails, tangent conditions, and weights.

[ReeceEngineering]

This looks like a good method, though I am having trouble replicating what you have done here. I see that in your loft you were able to get sketch 5's two lines to show up and rail one and rail two. When I try to select the 2 rails in my sketch 5 it selects the entire thing as one rail and the resulting loft isnt correct. Why is this happening?

Also why are the bottom rails in both your sketch 5 and mine undefined? No matter what I tried I could not get the line to turn black:

[ReeceEngineering]

What are these problems shown in color?

"1. You are lofting into a singularity. Approaching the singularity - the point - curvature problems will usually effectively prevent shelling, , thickening or offsetting surfaces."

True and I have had problems trying to do this in Inventor in the past. I did just try and shell this model and it worked with a thickness of 1.2mm which is what I was planning on using. I suppose I could model the tip of the nose with a miniscule circle rather than a single point.

"2. You are lofting this in one piece."

Yes my plan is to loft the entire fueslage, extrude cut it into 3 sections, save each of those as their own part file, and then shell the front and middle sections. The rear section will get a revolve cut to shape the thrust tube and some extruding to make room for the fan to sit in place.

After the center section is shelled I plan to loft the intakes as a solid, and then do a second loft cut to allow air to pass through them. I have the cross sectional area needed and was planning on proceeding with that alone. Though I do think the final design would benefit from some flow analysis as I'm pretty sure 2 hemispheres equal in cross sectional area wont flow the same as a ring will.

"3. You are lofting filleted profiles."

What is the problem with doing this? Is it better to loft square profiles and then fillet them in the 3d obeject once complete?

Here is the sketch, its a top view showing all the componets and the needed external dimensions. Im wanting to go with a form follows function approach. Its shown here in 1/3 scale.

[ReeceEngineering]

Very nice job here! I saw the curvature constraint and had never used it before. With a google search I found information on it and tried to use it myself. However when I try to replicate what you have done here its not working; when I click the spline line after selecting curvature constraint it wont allow me to select the adjacent sketch to give it the constraint. How did you pull this off?

[ReeceEngineering]

That is a good suggestion; I hadnt considered lofting between only a few sketches at a time and using common guide rails for the entire fueslage.

"I can only speculate on your true Design Intent."

I posted a top view sketch of the jet a few minutes ago if you were curious to take a look. It also included a basic plan of the commands I was planning on using. Im certianly open to feedback and suggestions on everything.

[davebYYPCU]

For rails in the same sketch, untick, chain select.

Shadow lines on this area - I would be checking the source document, fillet too large?

Discussing the surfacing is best left to Peter.

Perhaps the full fuselage can be framed up, seems like we are discussing a half job.

Might help...

[mango.freund]

[TrippyLighting]

It is had to describe "how I pulled this off". I'll make a screencast but it might take a few days.

Here are a few things to consider when lofting.

Most lofts create NURBS surfaces. The shape of a NURBS surface is governed by a qadrilateral grid of NURBS control points ((control vertices or in short CV) and their associated weights. There are two ways to create triangular NURBS surfaces. Trimming part of the surface  and lofting into a point, the so-called singularity. That means on one edge of that quadrilateral control grid, all CVs have to collapse into a single point. As a result the weights will go to infinity. So the software has three choices 1. to crash, 2. to not allow that lof, 3. to approximate.

Most CAD software will approximate, but as a result of that the curvature will suffer and often prevent shelling, thickening or offsetting surfaces.

However, lofting into a point can be helpful sometimes, so   trimming off the tip and capping it off separately is one way of dealing with that.

Another area is that too many folks try to create lofts that are better broken into several smaller lofts. In this case 2/3 of the body should basically be fairly straight surfaces with small or even no curvature, but then the front third of the object has areas with very high curvature. 

It is not uncommon to make separate, lofts and then blend them together with a third one.

The reason why I usually don't loft from filleted sketches is that it allows me much better control over every aspect of the object. While Fusion 360 can create nice lofts from such filleted sketches when the "keep tangent edges" option is enabled in the loft dialogue, it really depends on what is being lofted whether that actually works out.

[ReeceEngineering]

Here is the entire fuselage with rails ready to go and no filleted sketches. For some reason its telling me the guide rails do not intersect with the sketchs, however they do intersect. As the part file is right now the rails meet with the edge of the nose. I already tried moving them to the center point of the nose circle and it has the same error. How can this be fixed?

[davebYYPCU]

Close the file and reopen, because it worked first go for me.

That front former (Sketch 2) is still causing the bulge just behind it.

Behaves as expected with a separate loft for the nose block, but I would go back to tangent point for the front tip.

Might help....

[ReeceEngineering]

Same issue still, very Odd. Would rebooting my PC change this? 

I think that buldge is from the top view and will need the side rails to fix that part.

[davebYYPCU]

Red blob in your pic, at front of Sketches  7 / 1

my loft should have failed too, don't know why it didn't but I know how finicky Loft can be.

Your white points for the rails, is the dead giveaway, 

they should be snapped to the missing purple points in that sketch.

Edit Sketch 7, Project > Intersect the circle from Sketch 1.

Hide sketch 1, and snap your rails to the purple points.

Your getting the right error, I was just lucky.

The faces behind sketch 2 are concave, match the rail but are not flat across the fuse as expected, Fix it with a second loft from the first loft that finishes at sketch 2 and the circle, with same rails.

Might help....

[mango.freund]

hi @ReeceEngineering -- take this --there are a lot of white points in there --- white not good enough -- open face.

greetings mango 

[TrippyLighting]

Lofting from a circular to a rectangular profile takes some forethought, if the sharp edges are to be filleted.

Most people who have predominantly worked with sold moving consider fillets a 2nd step.

However, in this case it has to be taken into account right from the beginning.

As has been mentioned already, you've got to loose some of the profiles, or first make the rails and then size the profiles to match the rials.

The screenshot shows the curvature your top rail and because you are determined to match I to the profiles, the curvature looks bad. IF you really need these areas of high curvature, then this loft should be broken down into several individual segments.

Also, if the geometry between the 2nd (from left) circle and the the first rectangle should be straight, then this segment should be a straight loft and actually would not need rails, but if you do use a rail, I should be a straight line, not a spline with curvature.

 

[TrippyLighting · Accepted solution]

@ReeceEngineering hopefully my posts are not seen as just being overly critical on your designs and techniques.

Some of the things you'll observe in the timeline of this partial WIP model are simply the result of some problem solving.

Problems I did not think I'd run into! The reasons I am responding o those problems is because I am attempting to create a high quality surface. A quality that is unlikely to be discernible  in a 3D printed model from a model with lesser quality.

One of the "problems" you'll observe in the end result is that I loose G2 surface continuity across the center line. I know what to do about it, it's fairly simple (those were his last words) , but I don't think that should even be the case.

Anyway, here are some results. I'll keep working the model and post back something later.

Below is a screenshot of the current WIP with the curvature map tool applied :