LinuxQuestions.org - [SOLVED] Why have three antennas on a wifi adapter?

Quote:

Originally Posted by slac-in-the-box (Post 4967357)

It comes with three omnidirectional antennas sticking out of the back... each antenna creates a donut shaped field...

Sort of. The typical antenna would create something like a doughnut-shaped field if it were working against a ground plane. That is, if there was a grounded copper sheet, and just one antenna poking through, then it would be pretty close. Putting three antennas as close on the back of a typical router would ruin the neatness of the radiation pattern (and on that they are way closer than on a typical router), rather than a plain (sorry) ground plane. (And, of course, you have to remember that any of this playing around with radiation pattern doesn't create extra power, it only squirts an amount of power in different directions, so, ab definitio, anything that gets you more antenna directivity gain gets you more sensitivity to misalignment, which is less of an advantage.)

Oh, and if the antennas are really omnidirectional, then that means that they radiate equally in every direction, x,y and z and that's not really a doughnut. Of course, a an antenna that is really omnidirectional in all three planes only exists a mathematical abstraction, but would be a 0 dBi antenna. But quoting gains dBi enables manufacturers to quote a bigger number than quoting dBd, so guess what antenna m/facts want to do...?

Quote:

Originally Posted by slac-in-the-box (Post 4967357)

Wouldn't one donut shaped field suffice?

For one data stream. What I don't quite understand is that usually the three stream devices are 3 x 150 Mbit (theoretical), and so claim 450 MBit (which they don't really achieve in any practical situation, but that doesn't stop it being claimed) and the three stream dual band devices claim double that, even though that is even less likely to be achieved in any situation in which a 1 meter bit of ethernet cable wouldn't work equally well. So, I'm not sure what this 3T3R device isn't doing from that theoretical 450 to give 'only' 300; maybe it can't dual channel bond on all channels.

Quote:

Originally Posted by slac-in-the-box (Post 4967357)

Short of buying different antennas and hooking them up and roaming around the land with wifi analyzer app on hand held devices and figuring it out the hard way, I was hoping some antenna guru could shed some light on this confounded conundrum of confusion, thereby saving me the cost of purchasing antennas I don't need or couldn't use.

Well, start off by playing around with what you've got. It wouldn't be atypical to see a 2:1 ratio between a good orientation and a bad one, so that might
a) work
b) give you a clue how far off you are.

If this suggests that a modest/substantial improvement will get you there, have a look into something like the 'cantenna' (pringles tube antenna); if you are prepared to put up with a specific alignment for this purpose, then you can get a lot of gain this way (but obviously,as alluded to earlier, with a lot of gain comes a more severe version of the alignment problem).

And, yet another thing, is that the more adventurous claimed speeds are often achieved with the manufacturers (ie, TP link, in this case) own wireless adapter at the other end. (And, sometimes, not.) the higher speeds depend on some of the flasjier bits of the n spec, and different m/facts adapters often behave differently, and the two different ends of the link don't have to misbehave much for you to lose you some data rate, and N is a totally standardised standard (well, even when it is fully standardised, it doesn't look as if this will be totally cured, because various bits will be optional and you won't know how well a pair of adapters will work together until you try a pair of adapters and their associated drivers). How well this works with the adaptor built into, eg, a typical laptop, with a different spec wireless adapter and different algorithms, may well be a different matter altogether.