Wow, another distraction... Do you really think someone like me would make the mistake of putting said radials below this point? Actually, I put them slightly higher than this point as a buffer, and to account for things like radial diameter. Yet still, according to every resource, including Cebic who I quoted above in this matter, the Sommerfeld-Norton ground results are still more accurate than the Mininec Groung results with no radials. I know you want to believe otherwise, but this is simply the case.
I can tell you exactly what happens when you exceed this point, its the same when you exceed the limits of the Mininec and Fast grounds that are also available to me, although those grounds do loose accuracy much quicker than the Sommerfeld-Norton ground seems to. Yep, I've searched for information on this and did experiments myself. As an experienced modeler, I would be lacking if I didn't do at least this much.
However, this limitation is limited to the Nec2 engine. If your using the Nec4 engine, the Sommerfeld-Norton ground allows you to put radials below this point, and yields accurate results even if the wires are underground. I have considered getting the Nec4 engine, I have software that will use it, but as of yet I couldn't justify the price for what little additional I get at the frequencies I tend to model...
Yep, you were done when you suggested that you could accurately model an antenna that requires radials without said radials. After that, all you had were distractions...
The DB