[ltp] Wifi range: Cisco Aironet 350 vs IBM a/b/g combo

[Ivarsson, Torbjorn (T)]

> -----Original Message-----
> From: honey@gneek.com [mailto:honey@gneek.com]
> Sent: Tuesday, January 06, 2004 3:55 PM
> 
> I recently replaced my miniPCI card in my T40: it was shipped with
> the 802.11b Cisco Aironet 350, but I wanted to try 802.11g, so bought
> the IBM a/b/g combo miniPCI cards.  After faithfully following the
> instructions and movies on IBM's website, I successfully replaced the
> card and reattached the two antennae via those fiddly little plugs,
> but am finding the range *greatly* reduced compared to the Cisco, and
> compared to a PCMCIA Orinoco card.

I am not an expert in WiFi-technology, but I am a working professional in the field of RF (Radio Frequencies). Your statement does not surprise me. Yes, the hardware (and perhaps software) play a role in the coverage of WiFi. For example, a better receiver yields better coverage (or throughput). That's why some FM radios are better than others...

It is also a function of the technology you use - 802.11a, b, or g. Comparing a with b and g should give lower coverage for a. 802.11a operates in the 5.3GHz frequency spectrum, whereas 802.11b/g operates in 2.4GHz. Because of physics, the higher the frequency, the smaller the coverage. Though not accurate, a rough rule of thumb would estimate that twice the frequency yields half the coverage.

Also, the modulation technique of the technologies plays a part in coverage. Both 802.11a and g uses OFDM (Orthogonal Frequency Division Multiplex). This is a more efficient modulation technique (placing your data onto an RF signal) than used in 802.11b (whose modulation I've temporarily forgotten). More efficient modulation technique means more coverage.

Therefore, you *should* get the best coverage with 802.11g (efficient modulation, 2.4GHz). Assuming identical hardware of course... Again, the hardware may not be as good as your other cards.

> Secondly, even though the IBM card's connecting as a G card to my
> Linksys WRT54G router (proven by it still working when the router's
> in "G-Only" mode and all nearby B clients turned off), and iwconfig
> reports 54 or 48Mb/s, I can only get a pretty consistent maximum of
> 7Mb/s when uploading a file to a server on my network.  iwlist rate
> reports that rates from 1Mb/s to 18Mb/s are available, and that I'm
> connected at 48Mb/s right now :)  Can someone make some sense of this
> for me?

Ahh, the beauty of marketing. First of all, the datarates, or more accurately throughput, you get from WiFi depends on the amount of interference in the frequecy spectrum. More interference, less throughput. In the 2.4GHz spectrum you'll have interference from other WiFi users, cordless phones, Bluetooth, etc. Advertised rates are often in very favorable conditions (i.e., labs withouth interference).

Having two antennas (in the lid) is usually only beneficial when receiving from the AP (access ponint). Transmission (in wireless technologies) is most often done using just one antenna. Therefore, expect higher throughput from AP to laptop than from laptop to AP.

Also, the 54Mbps (or 11Mbps) can be achieved when there is no error-correcting overhead information in the bitstream. In a practical situation (real RF environment) you always have error-correcting overhead info. Believe me, you need it. Taking overhead information into account, the (for you) useable throughput is reduced by 40-50% (ideally, no interference). Of course, if you only have a DSL/cable/T1 connection from your AP, the max rate you'll be able to use is somewhere between 0.5-1.5Mbps...

Note: It is possible perhaps to run compression on the PPP protocol layer, something that cellular/PCS operators are doing, to increase the user-perceived throughput. I don't think a WiFi AP is doing that.

The 2.4GHz spectrum is getting crowded (lots of interference), that's why 5.3GHz is being used.

BTW: Antennaes are for bugs, antennas are for humans... but who cares?

Hope this helps you understand the technology a little bit better. Sorry for any spelling errors.

T.