Cross section properties in Karamba

Hi everyone,

I have attached a photo result in table format obtained from Karamba. (fig3)

First I selected the cross section as in fig (1), and tried to obtain the properties result as seen in (fig 3) first part of the table.

Secondly I defined the same cross section with the exact same geometrical parameters of the section as in fig (2), and tried to obtain the properties result as seen in (fig 3) second part of the table.

Question : The geometrical parameters provided in karamba for the cross section are same , but the properties obtained are different . Why is that so?

Please let me know.

Thanks in advance,


Veronika Phadtare


could anyone please help? I am very curious to know.

I’m not sure if the Fillet Raduis size is the right one.
EN 10210 (Fench version) standards do not precisely define these dimensions, only R <= 3t is indicated.
That means the fillet radius are 36mm thickness

With these dimensions, the results of your properties are very close

Hello Jacques,

Thank you for answer however, this doesn’t seem correct because external r = 1.5t for sure in hollow rectangular beams. I tried using various beams. For small beams the values match and for large sized beams they gradually differ. There is something different and I really can’t find what is the exact reason. As you mentioned the results come same, but it coincidentally came, for other cross-sections they do not.

Hello @veronika.phadtare,
it seems the reason for the deviation is the inner fillet radius. For a RHS300x200x16 EN10210-2 defines an outer fillet radius of 24mm and an inner fillet radius of 16mm (see e.g. the tables at
However with the cross section component of Karamba3D allows only to define the outer fillet radius. The inner filet radius gets calculated as described in the attached definition.
It is however possible to use a 2-line script to create the box cross section directly defining both fillet radii: (20.5 KB)
Then the properties of the EN10210-2 and the Karamba3D cross section coincide.