Steve Withers wrote: > Hi KLUG-ers > > This list has been quiet for a wee while, so I'm giving it a rev. > > We just moved house to the farm at Foxton (from a place we had been > renting while I renovated)....and the dearly beloved has banned my > server "farm" from the house (too small).....with the garage being the > only place left (big enough). > > The wireless network (mix of 54mb and 11mb) works perfectly in the > house. > > However....I was attempting to network the house and garage. > > I'm using wireless network cards (some of each - 802.11b and g) and an > access point (AP) from DSE. The AP is plugged into an ethernet port on > my DSE ADSL router. > > Problem was.....the effect metal cladding on the house and the garage > combined meant that the wireless AP in the house was all but invisible > from the garage - even with a 10Db directional antenna at one end and a > 7Db at the other. The signal was '15' (very low) and the link was zero. > > OK.....so the effect of one wall could be removed by drilling a hole in > it (done) and placing the antenna outside the wall. It is magnetic, so > mounting on the wall was easy. I ran the wire back through the wall. > > The signal level went from 15 to 18. > > Very small improvement..... > > Then....I noticed a bedroom window would be opposite a spot about 1 > metre to the left and a bit up the garage wall from the where I had > placed the antenna. > > So I moved the antenna over there....on the garage wall and opposite the > window on the house. > > My signal leapt from 18 to 30 and the link quality went from flatline 0 > up to 70! > > So I'm online in the garage.... :-) > > This is with the rt2400 802.11b card from DSE, optional antenna, and the > rt2400 drivers for Linux (v1.2.1) from sourceforge.net. > > By the way, I also got the 802.11g USB-connected wireless device from > DSE to work with Linux, using Ndiswrapper and the Windows XP drivers > that came with the device. Very stable. Works perfectly....but for a > small delay (1-2 mins?) at start-up while the device sorts itself out > and comes up. > > As I rarely re-boot, this is a very minor issue in my books......but if > you start/stop your system all the time, it might be annoying. > > Steve Hi Steve, Glad to hear you are settled in your proper home, although it seems you have some priorities inverted. Surely it is more important that the server farm is properly accommodated than other more mundane and less productive pursuits such as cooking and watching TV which are so naturally suited to the spartan environs of the garage. A few notes about antennae and radio, hopefully I am not preaching to the choir. Firstly, your observations about the metal cladding are spot on. you have a pair of Faraday cages. There are however a couple of other considerations that can vastly improve your wireless experience. Most vertical whip style antenna tend to radiate (and receive) predominately in the plane to which they are perpendicular, so if you have the antenna poking out sideways from the wall, it will be working mainly in the plane of the wall. Half of the signal will be going up into the sky, and the other half will be absorbed by the ground, especially the water in the ground. Remember that the frequency used by 802.11b and 802.11g gear is the same as that used by microwave ovens, and was chosen for that use because it is easily absorbed by food, living things and other organic matter. Anyway, for best results you want all the antennae on the network to be in the same plane, and perpendicular to it for best results. Where you have a number of stations, it's usually easiest to achieve this when all your antenna are vertical, so their maximum radiation is in a plane parallel to the ground. Vertical whip antennae generally produce vertical polarised signals. It's important that all the antenna in your network are operating in the same polarisation. Think about polaroid sunglasses, many people will have done the experiment with a couple of polarise sunglass lenses where you look through both lenses with one eye, and then rotate one of the lenses. When the polarisations of the lenses are at right angles, everything goes black. This is called Cross Polarisation. It can be useful if you want to minimise interference between two overlapping networks, but generally it's a Bad Thing. An analogy to using oppositely polarised antennae that are not radiating in the same plane would be looking for something on the wrong floor of a multistory building with a bag on your head. It's harder to make omnidirectional horizontally polarised antennae, so these generally get used for point to point links, where they can be aimed at each other. There is some advantage to using horizontal polarisation at distance over the ground too, so when you are linking two sites at a distance with directional antenna you may wish to orient your antennae this way. With you antennae mounted on the ridgeline of your roof you may get excellent results over a considerable distance, and find geeks with laptops parked in your vicinity enjoying your bandwidth ;) The results inside the house might not be so flash though. On another tangent, the domestic environment departs considerably from the theoretical model above. Reflections from the environment create an interference pattern of hots and dead spots. 2.4GHz has a wavelength of about 12cm, so the distance between hot and dead spots can be 3cm. Very small changes in position can make a huge difference to signal strength, as can any movement within the field. With this in mind, Diversity receivers are a considerable advantage. These have two antenna, and switch automatically to the one with the highest signal strength. This typically makes the difference between 90% reception and 98%. We don't like 10% packet loss. Pick access points with two antennae. Great work on making yet more DSE stuff work in Linux, I hope you are getting some kickbacks from them. regards Mark Robinson ------------- 27 Jan Auckland and Northland anniversaries. 27 Jan 1832 Author Lewis Carroll born. 27 Jan 1945 Russian troops liberate Auschwitz concentration camp. 27 Jan 1959 NASA selects 100 candidates for the first space flight. ------------------------------------------------------------------- To remove yourself from this list, email klug-request at linux.net.nz with "unsubscribe" in the body of the message.
| More information about the KLUG mailing list |
If you have any questions or comments about this page, email the
Webmaster Design Copyright © 1998-2005 Linux.net.nz |