Bug when trying to create a normal to a face

Hello guys,

I hope someone can help me.

I need to compute normals to different faces, but when I try  the measure tool "Normal to a face"

and use it to compute the normal to a sphere at the point at the bottom of it (in z-direction) i get

the error massage: "Normal vector of a face has null magnitude"

The same happens if I write a script and use GetNormal.

The strange thing is, that this only happens if i want a normal in the negative z-direction.

I use version 5.1.3 - so if  someone of u uses the newest version and he could check if this bug has been fixed

that would be nice.

Thank you,

Richard

Previously Richard S. wrote:

I use version 5.1.3 - so if  someone of u uses the newest version and he could check if this bug has been fixed that would be nice. 

Hello Richard,

I am using 5.1.4 and the problem occurs in both the +ve and -ve directions of Z.  My GUI procedure was:

1. Create a sphere radius 100

2. Explode its face

3. Create 4 vertices at (200,0,0), (0,200,0), (0,0,200) & (0,0,-200)

4. The create normal works for the first 2, but errors on the last two with the message you reported.

If you can post the python script here, I am willing to test it for you in 5.1.4.

Regards,

JMB

Hello JMB,

you´re right, this problem has nothing to do with the direction. In my particular case i had a radius of 20 i think and the problem was that there is an "edge"  at the point  (0, 0, -20) on the surface of the sphere, which came from approximation of the ideal sphere. This is the same in 5.1.4. The getNormal funktion can´t handle the points on those "edges".

So I tried to approximate the normal at this point by two normals near it.  But this didn´t  work  perfectly, as i don´t know where these "edges are, until i get the error message.

Perhaps you have an idea how to solve that one.

Thanks a lot.

Regards,

Richard

Hello Richard,

Try a normal to the point (1,0,200), then it worked for me.

Regards.

Previously Richard S. wrote:

Hello JMB,

you´re right, this problem has nothing to do with the direction. In my particular case i had a radius of 20 i think and the problem was that there is an "edge"  at the point  (0, 0, -20) on the surface of the sphere, which came from approximation of the ideal sphere. This is the same in 5.1.4. The getNormal funktion can´t handle the points on those "edges".

So I tried to approximate the normal at this point by two normals near it.  But this didn´t  work  perfectly, as i don´t know where these "edges are, until i get the error message.

Perhaps you have an idea how to solve that one.

Thanks a lot.

Regards,

Richard

This bug is present in development version too ... we are going to see how to fix it for next version. Regards, Erwan.

Previously Erwan ADAM wrote:

Previously Richard S. wrote:

Hello JMB,

you´re right, this problem has nothing to do with the direction. In my particular case i had a radius of 20 i think and the problem was that there is an "edge"  at the point  (0, 0, -20) on the surface of the sphere, which came from approximation of the ideal sphere. This is the same in 5.1.4. The getNormal funktion can´t handle the points on those "edges".

So I tried to approximate the normal at this point by two normals near it.  But this didn´t  work  perfectly, as i don´t know where these "edges are, until i get the error message.

Perhaps you have an idea how to solve that one.

Thanks a lot.

Regards,

Richard

This bug is present in development version too ... we are going to see how to fix it for next version. Regards, Erwan.

Fixed in 5.1.5

Hello Erwan,

thanks for the great work!

Just for curiosity I´m interested in the way you solved the problem, perhaps you could give me an impression of how you fixed it.

Now I have another problem:

The make-common algorithm seems to  need an improvement too.

First problem is the execution time which seems to be very long (compared to the other boolean operations). I wrote a tui script which needs to execute a lot of make-common operations and  sometimes one operation takes several seconds, which is way to slow to make the script usable.

Secondly, I often get an error message that says the operation can`t be done on the given shapes. most of the time this occurs if I try to do the operation on a small solid like a sphere and a big face.

Hopefully this could be fixed as fast as the last one!

Best regards,

Richard

Is there any enhancement planned for the make-common algorithm?

Regards,

Richard