Suggestion: Visit
http://www.emclab.umr.edu/csoft.html
etc. for more info
What about a visualization technique for 3D?

I am still interested in this topic. Where can I find theoretical information about extrapolating, interpolating.., or some other technique to calculate the gain at any point (azimuth, elevation) based on the 2 fundamental antenna pattern cuts ?

Thanks

Y26Rin

Have a look at:
IEEE Antennnas and propagation magazine vol43 no2 April 2001
for an article by Lius M Correia, Francisco Gil et al.
This is a good start but is still a linear interpolation process.

Pavlos

The basic problem is that you're attempting to "create" information.  If the antenna has a complicated pattern, then the two cuts will provide no information about the complexities that may lie between the cuts (and you should know that the manufacturer will probably select the cut angles to show the best parts of the overall antenna pattern).

If the antenna is relatively simple (and has a 'simple' pattern), then the best approach might be to model the antenna using NEC and then attempt to match the model to the provided cuts.  In other words, use the provide cuts to calibrate the NEC model.  The possibility exists that the NEC model will reveal that the provided cuts are not representative of the overall pattern.

Either way, the overall pattern should integrate (over the entire sphere) to something just under 0 dBi average gain (just under 100% efficiency) - a very useful final 'sanity check'.

I don't believe there is any practical way to project the 2D pattern to 3D.  Many antennas have complex patterns in one axis or the other and sometimes both.  You'd need to compute it from 1st principles using NEC or similar.

I do agree that information is missing for more complex patterns but had a feeling that for less complex patterns and specific antenna configurations there might exist maybe an emperical approach to approximate the patterns to a reasonable extend.
Simulation would be the better option if the antenna dimensions and configuration were known.

In the simplest case, you might be able to interpolate assuming that the pattern is related to a cardiod pattern (or something similar), but that is such an extreme assumption that the predictions would be useless (and possibly even dangerous).  I don't really recommend it.

Besides, for a simple antenna, you could easily use NEC to make a better prediction.  www.eznec.com has a free demo that can handle 10 segments (enough for the simplest case).  It is very easy to use.

Suggestion: Visit
http://borneo.mp.tudelft.nl/~jvdkruk/stw/pdf/progress_report99_1.pdf
For: Forward modeling is carried out for a multi-component 3D survey in which a buried point scatterer is
present. 3D imaging is carried out using this multi-component dataset for different source-receiver setups.
Using the finite bandwidth of the received signal and a limited aperture this 3D imaging scheme is efficiently implemented.