This content has been marked as final. Show 11 replies
You will have to create (either as an underlying sketch, or the one used to create the loft) a 3D sketch that contains both the section curves and the guide curves. You can use the "on surface" constraint to keep the splines on the plane, so it will behave similar to a 2D spline. This is the only way I know to keep both sections and guides with dynamic relations between both that solve simultaneously.
If you use guide curves, then the connectors should be irrelevant.
"The shape seems to depend on artistic arrangement of connectors for the loft. I am concerned that this is a non-precison and probably asymmetric geometry"
So model half of the shape and mirror it. If you are using a loft with GCs, why are you also using connectors?
"I would like to define the fuselage using a sweep."
Why? How do you know that another feature type is not a better way to do this? Why a sweep? I'm not saying a sweep is a bad thing, I just want to make sure you have a reason for choosing it.
"The section would be defined as a spline. I would like to have the section change width with the left and right guide curves and height with the top and bottom guide curves. That means that I want the shape to scale in X independently of Y direction. How can I make a sketch that will have this property? "
One way would be to use a 4 point spline, piercing each spline point with a path or GC. You could use a spline with more points, but it would need to have relations to control the additional points. Another way would be to use an ellipse, but an ellipse needs less control, and might not work, especially if the GCs on the sides are 3D rather than planar. You might be able to make 2 half ellipses (top and bottom) and make an arrangement with 3D side GCs work. It really depends on the cross sectional shape, which I'm guessing will not allow a spline.
Another option might be to use a sweep with a proportional spline, but that doesn't scale X and Y independently.
Really, a loft or boundary are your best options. Sweeps are limited in how the shape is allowed to change. Looking at this, you're gonna need more change than a sweep allows.
Matt you are correct that a real shape will need much more complex variations in the sections. For the purpose of a conceptual airplane one section that has independent scaling in X and Y would be adequate. The conceptual model is to package the major components, spars, passengers, and engines. The resulting design is used to determine weight and aerodynamic properties.
I do not know why the loft with 5 sections, each with 4 pierce contraints needs to have connectors edited. The loft otherwise develops some nasty asymmetrical wrinkles. The sections that I used are all different; some are just mirrored splines, others have straight lines, circles and splines, none have the same number of "points".
I tried the half fuselage approach. The centerline surface refused to be perfectly planar.
The 4 point spline as a section for a sweep produced spectacular results. The ends became 4 lobed. Apparently the spline handles did not scale with the spacing of the points. I have not yet tried adding equation relations to control the spline handles.
The sweep with a proportional spline makes some very nice shapes for a smooth strake to vertical fin airfoil contour. The relative direction of the path and guide curves gives beautiful control of the thickness of the airfoil. It is not very handy for a fuselage.
Is it still required that all loft section have the same number of "points"?
Lofts don't HAVE to have the same number of sections, but it makes it much easier if they do. The software doesn't have to guess where to connect in that case, and it's not very good at guessing. Helping it along by your own best guess on the connectors makes life much easier for the software and the designer.
You might try breaking up your fuselage into a couple of sections instead of making one big loft or sweep. One for the main body, one for the nose and one for the tail.
A couple of more points I forgot. You might want to use a boundary surface instead of a loft or sweep. You get more control over tangencies on the edges than with a loft. When trying the half fuselage approach, make an extruded surface on the centerline that is perpendicular to the center plane, then make your loft or boundary surface tangent to that surface. That should get you a better join when you mirror your half fuselage.
If you haven't already done so, you might want to download Ed Eaton's Curvy Stuff tutorials at http://www.dimontegroup.com/Tu...idWorks_Tutorials.htm
Get a copy of Matt Lombard's SolidWorks Surfacing and Complex Shape Modeling Bible. Amazon had a good pre-publication deal where you could get it along with his SolidWorks 2007 Bible for less than $32 each. I don't know if they still have such a good deal, but both books are well worth list price.
Since you already are looking at using sweeps vs. lofts, also consider simplifying your section geometry. Most of the time, an ellipse or a partial ellipse works better than anything. An ellipse usually has fewer DOF's than a spline and is easier to control end conditions. Also, an ellipse will not develop unexpected inflections.
The ellipse works wonderfully in a sweep but it does not look much like the airplane I had in mind.
I will try breaking each section into the same number of segments. Maybe I can avoid the connector editing. Breaking the fuselage into smaller sections and applying some end conditions speeds up the loft. It is difficult to have the transitions smooth in all respects. The parts are conspicuous in the highlights.
Features like a bubble canopy that have an intentional break in curvature are very nicely done by a cut into the fuselage solid and then a loft from the front and rear faces and a centerline guide line. I then color the loft somewhat transparent. The solid fuselage then appears hollow when looking through the canopy.
The extruded surface at centerline may be a good solution. That shape will be very controllable from a simple inboard profile sketch. There will be a small single curved surface which is almost always present on completed designs.
I really like Matts srufacing book. I have not tried the boundary surface on this kind of shape. It will be a good exercise to learn more about boundary surfaces. I also want to try some of the less used types of fillets. Matts book definitely provides me with some homework.
Rick, this is off topic, but where did you get your the data for your 'people' that you have in your fuselage?
I find that when you have a complex shape, try and split up the lofts/sweeps/revolutions into shape dependent features. Also, if you have any complex shape, always model half then mirror - solidworks will be much more successful at giving u what u r looking for. As for the fuesalage, I would:
1. Develop the main center section using a loft, be very careful as to direct all sketches and sufaces tangent across the mirror plane from the get go.
2. Develop the end sections using the main center loft for tangent or curvature directors, including skectches for influence.
Remember, solidworks complex surfacing is driven by what you use for influence. Influence can be a multitude of things. Simplify as much as possible, use everything you can. 3dsketches, sketches and especially other surfaces. If you dont get what you want, look at the main loft for ways to remedy problems.
Solidworks can surface ok, but not professionally - but with the right manipulation - will provide what you are looking for. Garbage in, garbage out. Question - When is solidworks going to implement T Splines and such based surfacing - this is the question to ask if you want professional control for a kindegartener