Assistance with a linksys setup

connected to

throughput of

Normally stock wrt54gs dont talk to each other, are you running a different firmware? Make sure the cables, connectors and anything else in the path are

50ohm and are rated for 2.4ghz. If all is well there then I would suggest high gain grid parobolic antennas. Also get netstumbler and check to see if there are other SSID broadcasting networks in the area (doesnt mean there are no non-broadcasting SSID). Also, clear LOS does not necessarily mean a good signal, you need clear RF LOS which means that 60 to 80% of the signal (called the Fresnel zone) should not be blocked by trees, buildings or what have ya.

I guess you've installed alternative firmware in the WRT54G to act as a transparent bridge. Wireless routers normally do not talk to each other unless you're using WDS.

Ok, let's play with the numbers. My guess(tm) is that the 18" long Yagi antennas are about 12dBi gain. I'm also guessing that the 20ft of coax is LMR-400 and that you have a pigtail somewhere in the system. From: * 54Mbps OFDM, 10% PER, -68dBm) * 48Mbps OFDM, 10% PER, -68dBm) * 36Mbps OFDM, 10% PER, -75dBm) * 24Mbps OFDM, 10% PER, -79dBm) * 18Mbps OFDM, 10% PER, -82dBm) * 12Mbps OFDM, 10% PER, -84dBm) * 11Mbps CCK, 8% PER, -82dBm) * 9Mbps OFDM, 10% PER, -87dBm) * 6Mbps OFDM, 10% PER, -88dBm) * 5.5Mbps CCK, 8% PER, -85dBm) * 2Mbps QPSK, 8% PER, -86dBm) * 1Mbps BPSK, 8% PER, -89dBm) I'll guess that you're connected at about 12Mbits/sec (or your users would also be complaining about the speed) for a sensitivity of

-84dBm.

Plugging values of: Distance = 0.25 miles (1500ft) TX power = +15dBm RX sens = -84dBm (at 12Mbits/sec OFDM) TX ant gain = 12dBi (yagi) RX ant gain = 12dBi (yagi) TX coax loss = 4dB (including connectors and pigtail) RX coax loss = 4dB (same as tx) into:

get a fade margin (SOM) of 23dB. Ok, this thing will work as specified.

The built in diagnostics are close to useless. I like to use ping. Well, actually I use fping, which is a replacement for ping that numbers the packets. Download a copy of fping 2.09 at:

Here's the results of pinging a neighbors wireless link (slightly edited):

fping 192.168.1.50 -c -t 1000 -b- Pinging 192.168.1.50 with 32 bytes of data every 1000 ms: Reply[1] from 192.168.1.50: bytes=32 time=0 ms TTL=127 Reply[2] from 192.168.1.50: bytes=32 time=10 ms TTL=127 Reply[3] from 192.168.1.50: bytes=32 time=0 ms TTL=127 Reply[4] from 192.168.1.50: bytes=32 time=0 ms TTL=127 Reply[5] from 192.168.1.50: bytes=32 time=0 ms TTL=127 Reply[6] from 192.168.1.50: bytes=32 time=0 ms TTL=127 Reply[7] from 192.168.1.50: bytes=32 time=10 ms TTL=127 Reply[8] from 192.168.1.50: bytes=32 time=0 ms TTL=127 Reply[9] from 192.168.1.50: bytes=32 time=0 ms TTL=127 Reply[10] from 192.168.1.50: bytes=32 time=10 ms TTL=127 Reply[11] from 192.168.1.50: bytes=32 time=0 ms TTL=127

Note that the latency is normally zero, but every once in a while, it creeps up to 10msec. That's because I've got some interference from another access point in the neighborhood. (Actually, a wireless TIVO abomination that downloads TV junk all day long). Whenever there's a collision, the packet has to be resent, and the latency increases. Under ideal conditions, the latency should be constant. If you're getting interference, have a marginal signal, or have a sick radio, it will show up here.

Of course not. If it did, people would call support and demand to know why the signal strength doth suck. Give the customer as little information as possible seems to be the trend among the cheapo wireless vendors. Diagnostic output only breeds complaints.

You can extract the signal strength and s/n ratio by sniffing the wireless traffic with a portable machine using Netstumbler or Kismet. I'm not sure how well it will work with a transparent bridge setup. Since such a bridge does not respond to probes, I'm not sure it will work at all. However, it's worth a try.

Oh, so they are having speed problems. Well, try this simple experiment. Block *HALF* the antenna with a piece of RF absorbent material. A wet towel will work. Black carbon doped intergrated circuit foam is better. No sheet metal please. Effectively, you've reduced the signal about 3dB. If you lose connection, you have a marginal system with some type of excessive loss. You can also do this by rotating the antenna the number of degrees specified on the datasheet for half the 3degree beamwidth, again reducing the signal by about 3dB. If the signal just dies, you again have a marginal RF signal. What you're measuring is the fade margin. You should have at least 10dB of fade margin. The calcs show that you have 20dB so you should be able to do cruel and unusual things to the antenna system and still have it work. It's it's very critical, then there's something wrong with the RF system.

Offhand, my guess is that you have either:

1. Interference from other radios. Change the channel.
2. Water in the coax cable or connectors. This is fatal and difficult to remove.
3. A crappy antenna. I hate Yagi's for 2.4Ghz.

If you want more speed, and don't mine spending real money, I suggest:

rid of the coax cable is a big plus for reliability and weather resistance. Getting rid of the yagi will make me happy.

If you don't wanna spend quite that much money, one of these:

work. Again, no coax cable and no yagi.