[Aironet] Diversity -> check the Howto ;-)

[Jean T. <jt...@bo...>]

Bob Edwards wrote :
> Elmer Joandi wrote:
> > 
> >  Could anyone reply that question ?

	I hate to repeat myself, but my Wireless Howto is always a
good place to look for that kind of information. In particular, the
answer to your question is at :
http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/Linux.Wireless.modem.html#diversity

> I don't know _exactly_ how it works, but the basic idea is to have two
> antennas working together in a mobile environment (where one machine is
> being used whilst in motion) to avoid "dead-spots" - presumably due to
> standing wave interference or similar. For a fixed set up (both ends
> in fixed locations) it is not as necessary. So, most PCMCIA cards use
> diversity these days.

	It is also worthwile in fixed setups if you use omni antennas
(hint : the environment is not static).

> In the Aironet cards, the two antennas can be set up to receive or
> transmit. So, for example, you can have Tx and Rx on both, Tx on one
> and Rx on the other, Tx and Rx both on the same one.

	Nope. There is a little switch between the antennas and the
rest of the system (you can sometime see that on the PCB), so that
only one antenna is active at at time, controlled by this switch.
	You can't have both Antennas active at the same time,
especially not Tx-Tx and Tx-Rx. In theory, you could Rx on both and
apply interferometry or other complex algo, but that's too expensive
(and that's the way you do beam forming antennas).
	Antenna diversity is most effective in Rx (because you sample
signal). For antenna diversity in Tx, the system usually use the
antenna that was used for the last Rx, but you can also alternate
antennas for eachr Tx retries or use more complex algo.

> I know of a project
> associated with the Olympic games where Tx and Rx are on both, but one
> is connected to a patch antennae pointing one way and the other to a
> higher gain Yagi pointing the other way. The system happily transmits
> and receives on both antennas.

	This doesn't contradict my explanation, but you don't use both
antennas simulaneously, just sequentially.

> I hope this might help you a bit.
> 
> Cheers,
> 
> Bob Edwards.

	Have fun...

	Jean