|
OK. I'll try to look into it. Can you give me a coordinate to try?
Hypothetically speaking the differences should be minimal. All of the tools are using exactly the same data sources (transmitters, terrain, etc.) and the same algorithms.
I won't rule out any possibilities until I've had a chance to investigate this, but off the top of my head, I'd guess that the most likely source of error is in slight round-off difference when going back-and-forth between the different tools.
In Google Maps, moving the pointer around lets you control the latitude and longitude very precisely (double precision floating point values), and the signal analysis is done using the full precision available. When you press the "Make Radar Plot" button or directly generate a radar plot report, the coordinates are rounded off to 6 decimal places. This might move the actual point being analyzed by up to about 1 foot.
In most situations, a small change like this would have almost no impact on the results. However, under the right circumstances (e.g., maybe steep terrain, very "pointy" peaks, or areas with sudden drop-offs), a small coordinate change might be enough to make a noticeable change in the results because some of the critical points in the terrain profile might have moved up or down a little.
Another thing to consider is that the Noise Margin is the one value that really matters. It does not matter whether the path type comes out as 1Edge or 2Edge as long as the estimated NM comes out right.
I hope that in all of your experiments, the Noise Margin numbers don't change by much. There is supposed to be a "soft" transition between 1Edge and 2Edge, so the variations in NM should be gradual. If there are sudden jumps in the NM, then this may be the symptom of a more serious problem (e.g., errors in the terrain database or the way in which terrain profiles are derived).
If the NM numbers are pretty consistent, then I wouldn't worry too much about whether it comes out as 1Edge or 2Edge.
|