XPost: comp.mobile.android   
   From: rmblayrrroy@nlnet.nl   
      
   Jeff Liebermann  actually wrote:   
      
   > The antenna pattern charts are thoroughly confusing because of MIMO   
   > (also known as 802.11n).  In order to distinguish between the two   
   > independent signals or beams used by the MIMO in the NanoBeam,   
   > Ubiquiti calls one signal "vertical" and the other "horizontal" in   
   > honor of their relative polarization.  The direction, relative gain,   
   > and pattern are called "azimuth" and "elevation".  When used with   
   > 802.11a/b/g which use only one signal, I think (which means I'm too   
   > lazy to check) that only the "vertical azimuth" and "vertical   
   > elevation" patterns are used.  In other words, you're using the wrong   
   > pair of beam patterns.  Fortunately, the other two are sufficiently   
   > close so your numbers will be unchanged.   
      
   Now that's interesting because I never understood the polar charts.   
      
   Is this somewhat correct?   
   1. Vertical = Mimo #1   
   2. Horizontal = Mimo #2   
   3. Azimuth = radiation pattern when viewed from above   
   4. Vertical Elevation = radiation pattern when viewed from the side   
      
   >>The polar chart doesn't mention the distance to each of the six radiating   
   >>circles in the diagram, but I probably am in the first circle, which means   
   >>it's still at +/- thirty degrees (which is sixty degrees wide).   
   >   
   > There is no distance.  That's not how an antenna pattern works.   
      
   Um. OK. I guess I just proved what I said, which is that I had never really   
   understood those damn things! :)   
      
   > The   
   > outer ring is defined as 0dB, which is a reference level established   
   > by the strongest signal in any direction.   
   > Every other point on the   
   > pattern is at a signal level that is less than 0dB.   
      
   I did notice they were in negative decibels, but I had figured that that   
   was in dbM which is always (in our experience with our type of radios) in   
   negative decibels simply because of the large size of the 0dBm point of   
   1mW.   
      
   I didn't realize that the negative wasn't that, until you just said it.   
      
   > Doing it that way is highly convenient.  For example, let's pretend   
   > you aim the dish antenna at a distant receiver.  You maximize the   
   > signal by swinging the antenna.  If it's designed correctly, the peak   
   > signal should be when the dish is pointed directly at the distant   
   > receiver.  If not, you have what's called a "bore sight error".  You   
   > then scribble down the maximum signal level (in dBm).  You can then   
   > use the antenna pattern to calculate the signal level at any point of   
   > the compass (azimuth).  If the outer ring were a distance, which makes   
   > all the points on the plot absolute values instead of relative values,   
   > you couldn't do that.   
      
   I like that you provide a practical use example!   
      
      
   >>It works, without doing any math, but I thank you for bringing up the fact   
   >>that the area painted is a function of the distance to the radio at that   
   >>60-degree wide angle.   
   >   
   > Perhaps it would be helpful if I explained why I have a preference for   
   > calculation over "it just works".  One problem with my calculations   
   > are that they are the best case numbers.  Everything that brings the   
   > numbers closer to reality makes the range and coverage worse.  If I   
   > threw in reflections, Fresnel diffraction, interference, component   
   > variations, tolerances, and just plain lies on the data sheets, the   
   > range would be less and the coverage patterns smaller (or less   
   > coverage).  The question my calculations answer is "Under ideal   
   > conditions, does this thing have a chance of working"?  If it doesn't   
   > pass that test, I usually don't bother trying it and look for a better   
   > solution.   
      
   Makes all the sense in the world.   
   Especially if you're doing it for a business.   
   And if you're buying the equipment.   
      
   Luckily, for me, I had the radio equipment lying around so it was OK to   
   just try it and then worry if it worked or not.   
      
   > I forgot to calculate the fade   
   > margin going in the opposite direction.  The LocoM2 transmits at +23dB   
   > while the smartphone might manage +20dB.  In this case, a loss of -3dB   
   > of fade margin will have little effect on the rather large fade   
   > margin, but I should have calculated it anyway.   
      
   I hope Rod Speed has the courtesy of letting us know, in the future, if it   
   worked or not, and how well it worked (if he implements anything from   
   this).   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)