Weekly Challenge Group

Nice challenge. Unfortunately, i have done couple of variants and printed them too. So won't post my version until the others have. Maybe something new comes up. :D

Conceptual Design of Dovetail Joint | GrabCAD Tutorials

Its goal is not to focus on the solution to this problem, but rather to convey the idea that it is useful to look for the rule behind ideas that seem different to the naked eye (this is a common practice in conceptual engineering).

This was the version I modeled and printed.

You have turned it into a practical item, I see!

I printed a version and showed my wife and she said:

"But it doesn't do anything".

And she is absolutely right!

But it should make you THINK !

And it certainly makes us think... but when it comes to our wives, it's better if you turn this into a perfume bottle and tell her you designed it with her in mind!

实际上在我们国家，这种结构的小盒子通常被称为“机关盒”通过在里面添加小钢珠和磁铁可以锁住它，需要一定技巧才能打开，类似于“鲁班锁”，如果你们知道“鲁班锁”的话如果你了解它的话一定会震惊的；现在这种结构通常用于古典家具。

It’s easy to just make a straight diagonal cut. It becomes a little more interesting if the cuts pivot around an offset axis.

But what if the offset axis isn’t vertical.

or maybe make it horizontal below the part

I found a new trick with Inventor doing all of this: sketches for revolutions do not have to be coplaner with the revolution axis.

https://grabcad.com/library/dovetail-blocks-1

Apparently in SolidWorks they must be (or at least I did not find a way to make a revolution surface from non -coplanares sketch).

In fact, to achieve the forms that you showed, with horizontal and inclined axis, I must resort to a loft with a guide curve generated around said axis.

However, what I could do are two "Sweeps" using non -coplanar sketches as sections and a circle as a journey, generated from the inclined axis:

I could not choose both sketches simultaneously as sections for the Sweep, nor a curve that composes them (accepts composite curves as paths, but apparently not as sections).

PD: in case it really solidworks cannot make revolutions from non -coplanar sketch ... the "Sweep + Circular route" provides equivalent functionality, although somewhat uncomfortable.

I could not choose two sketches for one revolution. It’s two separate revolutions merged together.

You guys have taken this to a whole new level. It took me a while to figure out what the coplaner-problem is.

I managed to do the split with a surface made in one revolution, of a planar sketch constructed with the help of a 3d-sketch.

Interestingly, the geometry breaks down at some angle and/or radius...!

The straight lines of the dove joint are no longer straight.

I tried that same method to begin with. None of the straight lines are actually straight if the axis is at any angle. They are not simple arcs. I think they are either a section of parabola or an ellipse depending upon the angle.

Oh, so you think that it is only this method that has issues? Must look into it; it is far from obvious what is going on?!?

Using the sketches on the faces forces the revolved profiles to have a straight line intersection on the face of the block which is what we're after. But the revolved surface forms something like a paraboloid because the straight lines are on sketches (planes) that are not coplanar with the axis.

Conversely , if straight sections that are coplanar with the axis are used as in your method then the surfaces are a simple conical shape. Because the block faces (planes) are not coplanar with the cone axis, the intersection always forms something like parabola. I say like a parabola but it could also be an ellipse or hyperbola. it all depends upon the relative positions of the axis and block faces.

It sounds rather convoluted but it breaks down into simple geometry.

If you rotate a straight segment around an axis contained in its same plane, you will generate cylindrical or conical surfaces.

On the other hand, if the segment is in a plane that does not contain the axis and, moreover, is not parallel or perpendicular to the axis, then you will obtain hyperbolic surfaces.

Some cases are tricky and appear to look like conical surfaces.

But if you exaggerate the steepness of the segment (as in the following figures) you will find that the surface is hyperbolic. If you cut it with a plane that contains the axis, you can see the curve perfectly, but if you see it in the plane of the generating segment, it is a straight line.

Cool explanations and demonstrations! Thank you, gentlemen.

Looking at some of your work and ability I feel I like I know absolutely nothing in SW.

Looks different than the other solutions here. How can they be separated? Can you post the 2 pieces side by side?

In CADD my skills are limited because until now I have used it for industrial models with simple shapes and without the need for rendering or photorealism. Unlike other colleagues, such as Steen, Bob, Hans or Ivo (among many others), who exhibit very elaborate models in their galleries and know CADD functionalities that I have never explored (these challenges help a lot in that). What I really enjoy about engineering is "figuring out" some not very intuitive things through analysis which is made much easier by using CADD and CAE-Motion.

Hello Nathanael,

It would be good for me to see the two pieces turned upside down to finish understanding the idea... although I think it has a similarity with the one with straight movement in two directions, is it so?:

Yes

Fun challenge