How To Set Sweep Shape Corner Radius To 0

[Guest bbraho]

I am attempting to model the path created by a ball mill. If you wish, think of a router bit with a spherical end. The path of the mill takes a sharp corner. This will leave a sharp coner on the inside of the path and a radius on the outside of the path.

 

I swept a shape along a 3D guide curve. Initally, IC inserted an 'arbitrary' radius into the sharp corner. See attachment #2, the inside radius was automatically inserted. It needs to be 0.

 

The 'Allow Sweep Sharp Corner' option gives the physically correct inside corner, but gives a sharp outside corner. See attachment #1, the inside radius is correct, but the outside sharp is not.

 

How to I set a radius of 0? This model is going to FEA. "Close but no cigar" is not a good option.

 

(Maybe I should have 2 disconnected paths with a H-Sphere at the turning point?)

 

Thank you.

[Guest cdeveaux]

I don't know how to do it using a sweep, but...

 

This is one way to do it, it involves using surface reshaping properties to add a "cap" driven by a formula to make it exactly 1/2" the sizebox width (or height) of the circular cross-section. Would this get you close? I use this for ball-nozed end-mill style grease grooves.

 

 

Sharp_Ball_Noze_Corner.ics

[Guest bbraho]

Yes, thank you. That works well.

 

[tlehnhaeuser]

Yup, thats a great way to do .

I also just use simple modeling tools like extrude and Blends .

See attached

 

Ball_End_Mill_Machining.ics

[Guest bbraho]

Which is also good. But in this case, a .25 radius is only penetrating .125 deep. The blend operation fails.

 

Yup, thats a great way to do .

I also just use simple modeling tools like extrude and Blends .

See attached

 

Ball_End_Mill_Machining.ics

14062[/snapback]

 

[Marcus Bertilsson]

Doesen't this work in the real mathematical way?

 

If you shall have inside radius 0 you can't have the guidecurve in radius 0.

 

You shall have the guide curve radius at the corner as the radius of the circuar sweep profile.

 

You will then get inner R=0.

Outher R=diameter of circuar sweep profile.

 

It might be so that you need to put in a guidecurve radius that is something 0,001mm bigger so IC feels that the inner radius is bigger than 0.

 

/ Marcus

[tlehnhaeuser]

Which is also good.  But in this case, a .25 radius is only penetrating .125 deep.  The blend operation fails.

14071[/snapback]

 

I see what your saying Brian.

It only works on thru cuts but the minute I put a full arc on end of profile the blend fails.

 

I think this should turn iinto an ER.

[Guest bbraho]

For entertainment value (the immediate problem is solved), I tried filleting a sharp corner to see what happened with a small radius. IC responded with the message:

"Invalid radius. The radius must be greater than or equal to 0.0393701."

I work in inches, so IC is saying the radius must be >= 1mm.

 

I suggest an ER to allow 0 radius corners.

 

Marcus's suggestion of a small radius good, but 'tiny inaccuracies' in CAD can cause big headaches in finite element.

I am happy that my IC made, SAT exported models don't require cleaning for good FEA model. It is a moot point, but I dont' want to start in that direction.

 

 

 

It might be so that you need to put in a guidecurve radius that is something 0,001mm bigger so IC feels that the inner radius is bigger than 0.

 

/ Marcus

14078[/snapback]

 

[Mike Allen]

Hmm - I don't get the "invalid radius" message until I'm below .00005". (I can enter .00002 for the fillet, but when I dimension it, I get .00005). The message shows up when I try to enter .00001.