nan values in the central part of the sphere

[inzhevatkin]

Hello.
I am using Windows 10. I have compiled scattnlay and I use standalone program fieldnlay.exe to calculate the electric field inside the sphere.
I use the command:
fieldnlay.exe -l 1 250 1.1 0 -p 0 0 1 -250 250 800 -250 250 800.
Calculations are successful but I have several elements with nan values in center of the sphere. The output table is as follows (for brevity, I do not list all the lines with nan elements):
X, Y, Z, Ex.r, Ex.i, Ey.r, Ey.i, Ez.r, Ez.i, Hx.r, Hx.i, Hy.r, Hy.i, Hz.r, Hz.i
...
-17.5000000, -2.5000000, -2.5000000, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan
-17.5000000, -2.5000000, 2.5000000, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan
-17.5000000, 2.5000000, -2.5000000, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan, +nan
...
Could you help me with this issue.
Thank you in advance.

[kostyfisik]

This is most likely due to the way, how we evaluate vector spherical harmonics, it all has 1\Rho terms. https://github.com/ovidiopr/scattnlay/blob/master/src/nmie-impl.cc#L539-L553
To fix this we need to implement it (and may be a couple of other places) with asympthotic values for Rho<<1.

[kostyfisik]

A related dirty hack https://github.com/ovidiopr/scattnlay/blob/master/src/nmie-impl.cc#L1105

[inzhevatkin]

Thanks for the info. I also wanted to add that with a further increase in the size of the sphere, the area with nan values also grows.

[kostyfisik]

It may also be related to #3