Posted: Mon Dec 15, 2008 10:30 Post subject: Help... long distance link: 20 km (12,42 miles)
I am doing a long distance WIFI link, 20 km (12,4 miles ).
Hardwares and positions:
Point A: Repeater bridge: Linksys WRT54GL (DD-WRT v24-sp1 (07/27/0 std) 100 mW, antenna 24 dBi, altitude 1060 meters (1053 ground + 7 roof)
Point B: AP Router: Linksys WRT54GL (DD-WRT v24-sp1 (07/27/0 std) 100 mW, antenna 24 dBi, altitude 200 meters (191 ground + 9 roof)
Point C-D... : Clients but they are not the problem.
Result:
Point A Signal: -69
Point A Noise: -96
Point A SNR: 27
Point A Signal Quality: 30%
Point B Signal: -70
Point B Noise: -96
Point B SNR: 26
Point B Signal Quality: 29%
Average Transfer Rate - 80 Kbps !!!!!!!!!!!
My problem is the slow bandwidth, 10 KByte/s
Now my configuration for the point A and B is,
Wireless Network Mode : B-Only
Sensitivity Range (ACK Timing) : 42000 meters
Authentication Type : Auto
Basic Rate : 1-2 Mbps
Transmission Fixed Rate : 11 Mbps (I can to configure 1 Mbps but the I have always 80 Kbps)
CTS Protection Mode : Disable
Frame Burst : Disable
Beacon Interval : 100 ms
DTIM Interval : 1
Fragmentation Threshold : 2346 (I have tried o lower but the situation does not change)
RTS Threshold : 2347 (I have tried o lower but the situation does not change)
TX Antenna : Right
RX Antenna : Right
Preamble : Short
Shortslot Override : auto
TX Power : 100
Afterburner : Disable
Bluetooth Coexistence Mode : Disable
The signal is sufficient, why I have a slow bandwidth?
Anyone have some ideas as if I should mess with the RTS Threshold, Fragmentation Threshold, DTIM Interval, etc.?
how I can to increase the bandwidth?
Joined: 13 Nov 2008 Posts: 5266 Location: CENTRAL Midnowhere
Posted: Mon Dec 15, 2008 13:18 Post subject: Re: Help... long distance link: 20 km (12,42 miles)
carlonifabio wrote:
I am doing a long distance WIFI link, 20 km (12,4 miles ).
NOW we're talking!
Your signal strength looks good. You might have to do some research on long distance wifi links. I haven't heard of anyone here doing as long a link as that.
Joined: 30 Nov 2008 Posts: 206 Location: Wherever the wind blows.
Posted: Mon Dec 15, 2008 14:27 Post subject:
First, enable the antenna for tx n rx on the one that you are using, not both.
That is one heck of a distance !.
IF you are running B only you can set the tx power at 200 mW, NOT at G which is limited to 80 mW to keep lineraity and sideband supprssion within limits.
Small gain, yes. but every bit helps.
What kind of antenna are you using, and is it a clear line of sight ?
You ARE asking a lot from such a distance. You should be using the narrowest beamwidth antenni possible. I would look into parabolic dishes at this sort of range.
Rudds _________________ I would rather have a bottle in front of me than a frontal labotomy
20km...Sounds like fun!
Some tips you might want to try:
-As previously mentioned, ACK Timing can hugely affect your performance. Try incrementing the ACK timing on both ends in 500 meter intervals starting from 0. It may help to have someone on both ends in case the link goes down completely.
-Definitely set the Fixed rate to like 2 Mbps. I know you said that it doesn't make things better but start slow and when you achieve good rate at 2 Mbps you can slowly turn it up.
-Enabling frame burst can improve things a bit as it allows the wireless to transmit multiple frames and group acknowledge. Less back-and-forth communication is better.
-Increase your beacon interval to like 5000 ms (or even higher). This is a point-to-point like between two nodes. No need to announce your SSID.
-A long preamble may be necessary for the nodes to sync when dealing with long distance links
It might help to post a screen shot of your Status>Wireless page. It has TX/RX Packet info that can be useful...
I don't know if B-only is a good way to go. He's looking to keep up bandwidth. That's an insane distance you're going, but very much possible. Beam direction will be extremely important, so you may need to work on dish positioning with guiding devices to help get them as directly pointed as possible.
Do you have to use the equipment you have? It's nice to be able to use existing equipment, but I'd recommend using something in the 5GHz range over such a distance. I'd have to review my info over Band-A, but not too much is used in the band-A anyway. Band-N is probably your best bet if you want to keep bandwidth high. If you can't use other equipment, I wouldn't worry too much, although that would be a preference.
If you need pictures or guides for setting up a parabolic system, let me know. I have quite a bit.
Joined: 30 Jun 2008 Posts: 825 Location: near Toronto, Canada
Posted: Mon Dec 15, 2008 19:28 Post subject:
I think it would be worthwhile testing with B-mode, preferably locked at 11mbps. If successful, it would be almost an order of magnitude better performance than what he is currently able to achieve.
I also agree that hiding the SSID is a good idea, and it might also be worth enabling the MAC filter to ignore unwanted devices.
EDIT: re-reading the original post... it appears that he is already using B-only, locked at 11mbps. So much for that suggestion!
Joined: 13 Nov 2008 Posts: 5266 Location: CENTRAL Midnowhere
Posted: Mon Dec 15, 2008 20:15 Post subject:
I think many people are missing the point here. He has a good signal. Great strength. There is nothing that needs tweaking in his hardware setup. 5ghz will provide LESS range than 2.4. He has it set to B only, and he should have it set there. G will not provide the range he needs....though I would still try G only on both devices, to see the impact. But in theory, it should be B.
The problem that is being asked here is, GIVEN that he has a great signal connection, WHY is he getting no real throughput on this signal. He's reporting 80Kbs!
It's not the frequency that gives it the shorter range. 802.11n has 2x the range of 802.11g. If he's using parabolic dishes, then the range may not necessarily be important if he can't get the signal properly. 802.11g, would be well capable of the bandwidth desired. 802.11n, even more bandwidth available. It would also be outside of common appliance & and devices that normally use the 2.4GHz range which could be causing signal attenuation over that length of distance.
802.11n would be an ideal band to use. However, the current equipment he's using does not support it.
Joined: 30 Jun 2008 Posts: 825 Location: near Toronto, Canada
Posted: Mon Dec 15, 2008 21:42 Post subject:
Long-range links using 2.4Ghz devices are certainly achievable. I personally don't have experience with links exceeding 6 -7 km, but the critical success factor at those distances is generally the selection and installation of appropriate antennas.
The signal strength being reported certainly suggests that his existing equipment should be capable of sustaining this link.
The altitude information provided in the original post also suggests good line-of-sight, but there is no detail provided about any possible line-of-sight obstructions.
To achieve a clear fresnel zone over a 20Km distance, the line-of-sight "beam", needs an unobstructed path almost 50m wide at the midpoint of the link. If there are tall buildings or other geographic features between the two antennas that may encroach on the fresnel zone, this could be a problem.
In addition to understanding how the bandwidth is being measured, I would be very interested to know what kind of latency and jitter is being encountered over this link.
Given that there is almost a km of distance in height of the two antennas, I don't think fresnel is a real concern.
So what causes slow throughput with a good connection. Perhaps solving this will solve other issues for people. _________________ Warning: I'm "out of my element!"
http://www.youtube.com/watch?v=MjYJ7zZ9BRw&NR=1
Joined: 30 Jun 2008 Posts: 825 Location: near Toronto, Canada
Posted: Mon Dec 15, 2008 22:25 Post subject:
Donny wrote:
Given that there is almost a km of distance in height of the two antennas, I don't think fresnel is a real concern.
Hard to say... Are we dealing with tall antenna towers in the middle of nowhere... or rooftops of highrise office towers in a dense urban environment??
20Km across an open plain, a large valley, or through a mountain range?
Without details, we have no idea what obstructions might exist between the two antennas.
Although I suspect you are right, we have to be careful about making assumptions.
Joined: 13 Nov 2008 Posts: 5266 Location: CENTRAL Midnowhere
Posted: Mon Dec 15, 2008 22:37 Post subject:
hpsenicka wrote:
Donny wrote:
Given that there is almost a km of distance in height of the two antennas, I don't think fresnel is a real concern.
Hard to say... Are we dealing with tall antenna towers in the middle of nowhere... or rooftops of highrise office towers in a dense urban environment??
20Km across an open plain, a large valley, or through a mountain range?
Without details, we have no idea what obstructions might exist between the two antennas.
Although I suspect you are right, we have to be careful about making assumptions.
Those are not fresnel issues....those are geographic line of sight issues. Different.
AND....He has a GREAT connection signal strength. The issue is WHY can't he get throughput with the settings he has. C'mon guys....this has nothing to do with signal strength! His is better than mine! _________________ Warning: I'm "out of my element!"
http://www.youtube.com/watch?v=MjYJ7zZ9BRw&NR=1
Joined: 30 Jun 2008 Posts: 825 Location: near Toronto, Canada
Posted: Mon Dec 15, 2008 22:42 Post subject:
The original post indicates that one of the devices is configured in "repeater bridge" mode.
I would recommend trying to establish a stable link between the 2 devices in a simple point-to-point client bridge configuation...
Only once a satisfactory basic connection is established would I consider experimenting with a more complex setup.
Also, I would respectfully suggest that the original poster consider using additional devices to support the repeater functions. I'm expecting that optimal performance over a long-distance link like this will only be achieved by dedicating 2 devices exclusively to this function.