Regarding the reliability of the soc-wann2J data

[p7 idomeneo]

Hello，I have a series of structures.Calculating the DMI and exchange using Fleur's method is possible, but extracting all parameters seems difficult. The four-state method, on the other hand, has poor convergence and requires too much computation. TB2J is the most convenient method for extracting all anisotropic J and DMI. Since abacus2J also has difficulty converging in SOC calculations, wann2J appears to be the most stable method. Because the band structures of the systems are predominantly metallic, it is difficult to reduce the spreading to below 1like that in the official example when using Wannier projection. I would like to ask, for magnetic 3d, 4d, and 5d elements, what spreading should be reduced to respectively, and is the calculation reliable for publication? In my systems, it is perfectly fine for the spreading to be less than the lattice constant.

Regards，