Full Version: Why cannot do a clean, solid sweep

Hello everybody.

I have my 2 profiles (slightly different) and my rail in the form of a drop. Somehow It was built by a 3 points arc (stage 1), mirror the arc (stage 2), and then again a 3-points arc (stage 3). I have built the shape also as continuous arc, first doing the bottom arc, then one continuous arc up to the pointy side, then mirror it, tried otherways too.....numerous. did some tiny continuous arc to the pointy tip, because this is where the problem is (rather than a clean intersection).
Anyway, can somebody enlighten me where is the mistake here for a clean, solid sweep?

Thank you for your time.
Rudy

Attachments:
Sweep 2 profiles + 1 drop rail.3dm

Image Attachments:
Drop not complete.jpg  Stage 2 drop.jpg  stage 3 drop.jpg  stage1 drop.jpg 

---

Hi,
Cut the drop in the middle and do a Sweep on the half curve
After that, mirror it and join. You have a clean Solid

---

Great, Thank you Finema.

Still I would like to know the reason. But yes, excellent.

---

Hi Rudy, I think it has something to do with your upper profile curve, it's slanted in relation to the curve's perpendicular and that means that things will try to transition to that type of slant as it moves from the lower profile which is oriented nicely perpendicular to the curve.

So the solution is to not use the upper profile at all, only select just the lower profile alone and do the sweep:





That will avoid having things morph to a slanted shape as they are traveling along the path, instead the shape will be a nice regular tube with a cross section perpendicular to the curve. Then the tip is automatically formed by corner mitering, that's where the one rail sweep automatically extends and intersects junctures at sharp corners. If you turn on the control points of this result you'll see the miter extensions:




Hope that helps!

- Michael

Image Attachments:
rudy_sweep1.jpg  rudy_sweep2.jpg  rudy_sweep3.jpg 

---

And if you do only do half like Finema's method that will avoid mitering extensions and intersections from happening since there isn't any sharp corner in the rail in that case.

Your original problem has to do with mitering trying to happen on your slanted profile I think.

But the half method will not yield a totally uniform tube, if you turn on control points you'll see the slanted shapes as it travels along the rail:



- Michael

Image Attachments:
rudy_sweep4.jpg 

---

Hi Michael
yes but in the Rudy's 3dm , the 2 profiles are different

---

Hi Finema, oops I didn't notice that part. So yes in that case it won't work to set up a single rail with a sharp corner in it as in the original, it would need to be done in 2 pieces as you showed. To try for no slanting you'd need to position the end shape to be perpendicular to the curve's end tangent and extend the curve by a straight line, then sweep along those and intersect them with a cutting plane.

@Rudy, a little more precise explanation for your original result, it's due to the pivoting at sharp corners that happens as part of the mitering process.

Here's a visual explanation - here is a circle being swept along a rail with a sharp corner in it:


It starts by building this piece here:


Then for the next piece, it's going to take the ending shape and rotate it by the difference between the rail segment's end tangents like this:




Then it will build this next piece like so:


Then to form the mitered corner, the pieces are extended and then trimmed to produce this result:



So your odd initial shape comes from the rotation step shown here. With your setup you wouldn't want any rotation to happen, however MoI's sweep does not have any concept of this type of setup with no rotation being desired at a sharp corner. Your profile there is going to be taken as belonging to either the side before or after the sharp corner and then rotated.

- Michael

Image Attachments:
rudy_miter1.jpg  rudy_miter2.jpg  rudy_miter3.jpg  rudy_miter4.jpg  rudy_miter5.jpg  rudy_miter6.jpg 

---

Hi Michael, Thank you for your valuable time. Yes, the problem is the tip.
Yes, Finema is correct. I need both profiles as this is a real life piece and the solid needs to have both profiles, which are a bit different.

---

it's not OK with this solid in attachment ?

---

Thank you Finema, a great trick! And thank you Michael. V4 when?

---

Hi Rudy, I don't know yet when the final release of v4 will be, but the next beta should be ready in a few days I think.

- Michael

---