Full Version: MINIMUM ANGLE DEVIATION BETWEEN 2 LINES TO APPLY A FILLET

Hi all,

Just wondering what is the tolerance for MOI to judge what is the minimum deviation angle from a straight line to which the mid point a fillet can or cannot be applied to - working in metric if that makes a difference?

I hope this makes sense? Thanks Alex.

Image Attachments:
continous FILLET.png 

---

Hi Alex,

re:
> Just wondering what is the tolerance for MOI to judge what is the minimum deviation angle
> from a straight line to which the mid point a fillet can or cannot be applied to

Filleting doesn't make a decision to bail out based on an angle measurement.

But as the angle gets smaller and smaller, the fillet arc will get very tiny and at a certain size level it will either fail to generate a minuscule arc or the solver mechanism will fail to converge.

The specific threshold where that will occur will likely depend on several factors like the size and location of the lines and the fillet radius.

I did a quick test over here and I was able to go down to an angle of 0.0001 degrees ok.

- Michael

---

Thanks Michael the explanation is great - with MOIv5 the fillet 3d shape engine is now controlled by acis is the plan to roll it out to the curve based filleting also and will this improve anything on that side? Here is what I am trying achieve - with the tube bending software it reads the fillets in the solid to gauge a change in direction for the pipe to bend I am finding if I make radius small it still works i can override the radius info in the cam software and apply the correct tooling radius which then corrects everything for bending. All the best Alex.

Image Attachments:
reconstructed curve.png  slight angle.png 

---

Hi Alex,

re:
> is the plan to roll it out to the curve based filleting also and will this improve anything on that side?

I don't know of any problems in curve fillets that would be improved by switching it to ACIS so there isn't currently a plan to do that.

Do you have some curve fillets that are not working?

- Michael

---

Hi Michael,

Thought that was going to be case with the curve fillets - im still working on the workflow its seems to be working 99% of the time very now and then I get an isolated node in the middle of a 3d polyline curve made from multi segments where the angles are large definitely works 100% of the time but in zones where its very slight rotations every now then I get the odd node that fails maybe its the slight angle in 3d space from 1 segment to next. Ill keep playing with it and see if I can isolate one and send it to you to look at. Currently its mostly working which is great.

Thanks Ill get back to you soon.

---

Hi Alex, one thing to look out for is that 2 adjacent segments will need to be coplanar for the curve fillet to be calculated.

- Michael

---

Hi Alex,

I don't think it's possible to fillet a curve corner where the angular deviation from smoothness is less than 1 degree. Construct > Fillet includes a step where a point object is drawn at each vertex and a prompt is shown "Select corners..." but corners less than 1 degree are excluded by that step.

- Peer

---

It looks like the angle cutoff is different depending on whether you are filleting 2 separate individual lines versus 2 lines that are segments within the same joined polyline.

For the individual lines there is not any angle cutoff, for the joined curve there is.

- Michael

---

Thanks for input appreciated here are a few tests I have carried out I created polyline with a vertical deviation of 0.7mm from which I could apply fillets and shell and the pipe software could still detect the bend info practically straight piece of pipe I have included a screenshot of the bend info. Once I put the curve deviation down to 0.6mm could no longer apply fillets which would back up the minimum 1 degree radius. In the reconstructed 3d polyline I have highlighted an out of range fillet node. Thanks for the info a getting a better idea of the limits to which I can push the pipe bending machine. I will play with applying 1 at a time rather then group application.

Image Attachments:
20240319_153347.jpg  3d polycurve.png  fillet limits.png 

---

Hi Alex, so you can probably get the missing one by using Edit > Separate on the polyline and then selecting the 2 separate individual lines there as the input to Fillet.

When fillet is given 2 separate curves that's when there isn't an angle limit used.

The fillet piece will be very small though when the the lines get close to colinear.

- Michael

---

Thanks Michael - the information provided has been very insightful into the process of converting these 3d spline curves back into usable polylines for tube bending process and understanding the limits ill post some pictures of bent pipe once we begin just been sorting files.

All the best
Alex.

---