Weak or fuzzy mates

You can do it this way but it is tedious.... especially because you must iterate each step for each window but it will properly position your windows.

Create 3d sketch - convert edge of window openiing into sketch. This will be a spline with 1 start/endpoint. Create 2 more points on the spline. Close sketch. Use these 3 points to create a plane. Create a

2d sketch on this plane, sketch a circle and create relations between the 3dsketch points and the cirle to center it over the opening. Then you can mate your window to the centerpoint or edge of the sketched circle. (note: there are many variations instead of sketching the circle - find what you like the best)

Told you it was tedious and I would love to see another better way posted.!

If you can place ONE point precisely on the spline edge of the holes, and if that can be at the topmost or bottommost point, then I have a solution for you . . . I think.

What you do is to create an axis which is vertical -- maybe using the Front and Right planes if that is how your assembly is oriented. Hopefully your window parts are symmetrical to their own Front/Top/Right planes and one of those axes (for each window part) can be made Coincident to the point that you place one the spline edge of each hole. Then you can make that same axis also Parallel to the Assembly axis that you created. Finally you can adjust the tilt of the windows and their proximity to the part with holes in it so that they fairly closely match the curvature of the outside of the part. You MIGHT be able to do that by making one of the window planes Tangent to the outside curvature.

'Sporky'

Chebeba wrote:

Chebeba,

No offence, but if you find yourself stuck in this situation, you probably aren't using optimal modeling methods. Your inserted (window) piece should have it's base geometry extracted from the part it fits in. This is especially important if the main part is a non-analytic (swept, lofted) shape. You can either model the window into the main part as a seperate, fully asociated, body, or in the context of an assembly.

Modeling two parts seperately and trying to work around the inconsistencies is a kludge.

Regards

Mark

Mark,

You make an interesting observation, but it doesn't necessarily apply to all situations. Chebeba describes the need to mate a standardized window (apparently flat) to a slightly curved lofted surface. We obviously don't have the complete story, but its easy to imagine that the models in question accurately represent the parts as they will be manufactured (without undue complexity and cost). Modeling the reality of a situation should not be summarily classified as a "kludge". You can easily imagine methods to model parts that mate very cleanly, but the resulting designs need to have good manufacturability.

Are the holes for the windows non-spline geometry except where they intersect the lofted surface? If so, I assume you have already constrained the windows in two directions and just need to establish the mate where the window frame touches the highest point on the lofted surface. In such cases, I would suggest creating a 3D sketch and placing a point freehand on the edge of the hole approximately at the high spot of the lofted surface. This point can then be mated to the back of the window frame.

Mr Voltin is exactly on target here. The application is a aluminium boat hull. The window is an off the shelf purchased part. It is welded to an opening in the sheet metal making up the hull shell. The hull is fully curved in all dimensions, but to the welder this curvature will only cause a gap between the window and the hull an order of magnitude smaller than the weld itself, and it is thus totally insignificant.

I used to model things the way Mr Mossberg describes for the conceptual model, but I am now doing the engineering model, and need a single part with correct BOM data for all 20 windows. Unfortunatly there are still changes in the hull shape expected from the hygrodynamic analysis, so I need the model to be fully associative as well.

Thanks for your kind input! /C

If you are using SW2006, there is a new width mate that allows you to position something to stay half way between two other things. Might help you.

WT

Good suggestion! It turns out it still wants to be EXACTLY half way, so it didn't help :-( /C

No, then it would be easier ;-) The hull is curved in all dimensions, and the windows need to orient themselves towards the hole as well as possible, while not being constrained any other dimension except that the x axis of the window is paralell to the top plane of the assembly.

The holes are created by a cut extrude of the shape of the window from a plane tanget to the hull curvature at the center of the window. Thus all edges of the hole are splines.

The way I figure I will do it now is by mating each a fixed distance from that plane. The drawback of that is I have to carry all the planes over from the hull part to the assembly. (I'm trying to cut down on relations between parts, since this model is already almost out of hand in complexity :-) )

/C

I don't understand the problem well enough to provide a complete solution, but you should be able to mate the windows to the hull in two directions. Whatever geometry was used to define the cut extrude features for the window holes should provide enough to constrain the windows in two directions. It would be helpful if I could see the assembly in question.