how-to block ads
|
bcbuie
join:2005-07-20
| reply to nwn
Re: -90 on AP, -62 on Client
I think he was saying the trees would not be the REASON for the difference in signal readings. When you moved the CPE and picked up a signal on the CPE, the signal probably increased equally on both ends. It just turns out that the signal at the AP was not higher than the AP's sensitivity, so it showed no signal.
The reason for the difference would most likely be higher amplification at the AP end than at the CPE end, it is a pretty common problem.
Remember, amplification is one-way. All things being equal, if you have 30db output power at the AP, but only 23db output power at the CPE, then you would see 7db less signal at the AP (assuming equal antenna gains).
The easiest solution is to just use a higher gain antenna at either the AP or the CPE. Antenna gain is bi-directional and is typically preferred over increasing output power.
Having said that, -90 at the AP and -62 on the client is a big gap. What is the output power and antenna gain of the AP (and the CPE)?
Ben | |
nwn
Premium
join:2004-03-05
Centerville, IN
| reply to Antennaguy
said by Antennaguy  :Antennas, trees and multipath will not create a difference between these units because they affect the total path in both directions. Until 2 weeks ago I would agree, but...
In my lab I had a client set up pointing directly at my POP 2.5 miles away. Wasn't trying to connect, just that is the direction it was laying. No signal either direction, trees are in the way. I moved some things and the CPE antenna moved 90* AWAY from the POP. It was now pointing directly at the woods 70' from the lab. Client showed -80 RSSI from the AP. Don't remember the LQ. AP saw nothing from this client.
The only explanation I have is that enough of the AP signal bounced of the trees for the CPE to see it, but because of distance, direction, antenna gain, etc. the AP could not see the CPE.
I would think the same thing could apply here, the CPE is seeing the AP on a path that does not work the other way.
--
Scott | |
|
