Chamfer (Distance x Angle) problem

[bmcwilliam]

Hi all,

Running Inventor Pro 2019.

I’m having an issue with a chamfer on a pipe elbow that doesn’t seem to be behaving as I would expect, I’m not sure if I am missing something or if it is working as intended.

I have defined a chamfer on the end of a pipe elbow by distance and angle, image 1.

I would expect this to bring a chamfer 4mm in from the OD, and at 40 degrees to the end face of the pipe.

When I dimension the drawing of this part, I am getting varying angles, see images 2 and 3.

The section view in image 2 was down the centreline of the part.

Any ideas why this would be? Is this working as intended or some sort of bug?

[johnsonshiue]

Hi! I recall the behavior has something to do with the section plane. I would like to understand it better. Please share the files here. 

[bmcwilliam]

The third image also shows a different angle to the defined chamfer angle and is of an orthogonal view, not a section.

I've attached the part, not sure if you will need other files as well to open it up.

[johnsonshiue · Accepted solution]

Hi! Inventor Chamfer is based on Distance-Distance (offset). The three different options essentially interpolate the distance values required to make the Chamfer (see the sectioned picture below). When you assign the distance and the angle, Inventor calculates the offset distances. Next, find the loops based on the offset distances. Lastly, loft from one loop to the other and trim off the solid body.

Since the algorithm is Distance-Distance based, the angle is not strictly maintained through the full length of the chamfer (depending on side faces). If you need strict angle-based Chamfer, you will need use Ruled Surface command. But, the downside is the Chamfer may not look even from the side.

I have attached a part to demonstrate the behavior (see attached part).

[bmcwilliam]

Thank you. Just to check, I am thinking that the approximation is accurate and produces consistent results for simple side face geometry, such as a prism, but if the side face gets more complex then the chamfer varies around the edge, and the DxD approximated chamfer will not match the desired DxA chamfer. For example if the object was straight pipe (cylinder) then there would be no error?

[IgorMir · Accepted solution]

Hi Ben;

In cases like that it is cheaper to use a Leader Text with the chamfer parameters in it. In your case - they are d10 and d12. The result will be just as it is shown on the attached picture.

Cheers,

Igor.

---

@bmcwilliam wrote:

Thank you. Just to check, I am thinking that the approximation is accurate and produces consistent results for simple side face geometry, such as a prism, but if the side face gets more complex then the chamfer varies around the edge, and the DxD approximated chamfer will not match the desired DxA chamfer. For example if the object was straight pipe (cylinder) then there would be no error?

---

[johnsonshiue]

Hi! Yes, if the adjacent faces are ruled faces ( U and V direction are consistent), I believe chamfer algorithm will create consistent width and angle.

Many thanks!