Strange behaviour of revolve feature

[Naureka]

I've come to some very strange behaviour of revolve-cut feature.

I was modelling a part with some tapered round pins. I needed to have a full round domed shape at the top of an extruded circular pin. Since pin taper angle, pin diameter and height are variables I found using a constant radius fillet unsuitable. Istead I created a plane through the center axis of the pin and manually modeled a tangent fillet profile to revolve it. I used a projected-intersection profiles and 2 tangent constraints. There I found some unpredictable behaviour:

• For certain pin heights I'm unable to create a full 360 degrees revolve with an error message: "No target body found to cut or intersect". Changing the pin height from eg. 3mm to 4mm makes the problem dissappear. It seems to be completely random. There are no errors in the sketch itself.
• I can make a revolve-cut with partial angles with no problem. It only happens with full revolve.
• When I first somehow manage to create a 360-deg revolve-cut and change the pin height earlier in design history, the model would build correctly all the time, no matter what height value I use. However in certain (random) situations I'won't get the full round fillet and an additional edge appears (see the picture).
• Even when model rebuilds with no errors, when I try to edit the revolve feature I get the same error message again.

I attach a file with a simple geometry that reproduces this. Anyone knows what might cause this? Am I missing something? It looks like a software bug for me.

[johnptc314159]

how do i reproduce the error on the attached drawing ???

[johnptc314159]

Also why not draw a single half profile sketch and rotate it ??

[Naureka]

---

@johnptc314159 wrote:

how do i reproduce the error on the attached drawing ???

---

I attached an f3d file in my first post. You may take a look on it if you're interested in this issue.

I finally created a single profile as you suggested since it was the only workaround I came up to. It was possible in this particular situation. But I'm still curious why my previous method was giving such strage results.

[jeff_strater · Accepted solution]

just to be clear:  The bug is:  open the attached design, edit the last Revolve, and that edit will fail.  Is that correct?  I created FUS-78518 to track this issue.  Thanks for reporting it and for attaching the model - that always helps

[Naureka]

Not precisely. Revolve feature may fail or not. It seems to be quite random. That's why this bug is so elusive.

To reproduce this you need to:

1. Open the attached file
2. Edit the revolve feature - it should fail right afterwards ("ok" button will be grayed out and an error message appears)
3. If it does not fail, hit "cancel" and change the height of the extruded boss (use a custom "height" parameter i equations tab).
4. Repeat from step 2.

You shoud be able to crash this model after 1-3 attempts. This error seems to be randomly triggered only by recomputing geometry, but not its actual values. I usually get one successfull rebuild in 2-3 changes of boss height. For the same value (eg. 3.0mm) I may get an error or not. It only depends on how many times i change it.  If the rebuild is successful you'll get a visible seamline on the dome head (like in first picture to the left). If the dome is smooth all around, with no seamline, the feature would likely fail during edit.

I was also able to reproduce it without an additional plane and axis - just with a circle in the center origin and a sketch on a front plane. There seems to be an issue with computing "tangent" relations somwhere in the 3D engine. I've experienced some wild behaviour with them a few times already.