This is the result if a point to point link obstruct with trees.
I have this tested with omnitik.
http://www.wirelessinfo.be/index.php/li ... nitike.php

Nice demonstration. Really illustrative.

Interesting, but no more than informative. We could do with more info on the link parameters.

I have a 130 meter link from SXT towards OmniTik shooting through dense leaf tree and around a building;
SXT-5nD r2 power set to 27dBm and directed towards Omnitik (=AP) power set to 26dBm; signals at around -55 between both... Connection rate 130Mbps 20Mhz channel width, 5580Mhz. 'n' protocol with NV2 (+encrypt) enabled.
Bandwidth test shows almost 70Mbps tcp traffic and at the same time a ping of 7ms average. (At the same time 2nd station connected to OmniTik) (Test done just on the units themselves. Probably from outside attached units we could go better.)

To give your test any better value you'd have to tell us all the configs of the both radio's. Only than can we see if your results are indeed influenced so much about the line of tree or that we have some other factors playing a role. Because I have a 'feeling' your second test could produce better results....

Are both chains enabled? (By default 2nd chain is not enabled.)
What protocol and NV2 or nstreme or just legacy 802.11? (By showing "Ack-timing" I'll guess you use 802.11?)
What test (udp/tcp?) and from where are they running. (I made the test between the units themselves. Both show 100% cpu so limit here is more to gain if test would be run from external attached units.
About the power; How come you use 30dBm? Omnitiks are default at 26dBm and it is very NOT-advisable to use higher power outputs in the manual power setting. Radio can become very unstable and over modulate the signal which is actually worse than using a slightly weaker signal. So try other power settings too to see if you can get better results. I'd presume there in the woods the background noise is very low....

Some remarks:

In trees it is usually the leaves killing the signal. Its the HO2 (water) in the leaves actually consuming the radio energy. Off course the wood of the trunk is also full of water but compared to the surface the leaves have compared to the trunk the leaves are the killers.
(I have somewhere a link that in winter gives better results than in summer. It shoots through one line of trees.........

So, although your test setup is nice, and gives an idea in the right direction, it would be nice if we can see more details of the configs.
You should also try some other 5Ghz frequencies. It can make a difference.

Great article. Would you be able to translate your other articles to English, too please? I also agree with the other posts - it would be good if you could provide the wireless config or full config exports of each device used in the articles too.



Tony

@wirelessrudy, When I speak of 30 dBm is that it really radiated power. (ERP)
Not that the radio 30 dB employment.
I will put more details online, glad you got this fair have said.
So I know what the people want to know about these tests and the tests become more interesting. In this case, it is about 802 .11n. thanks for your honest opinion.
Yes i know that water is the bottleneck.

@TonyJr, I work on the translation but leave the translation done by someone who can good English I'm not such a good translator.

I have edit the website and hope this is much better.
Please tell me what your missing on this test.

http://www.wirelessinfo.be/index.php/li ... nitike.php

Ok, more info;
But you client site says; mode=ap-bridge. That can't be true, because it is a client. It should be 'station' or 'station-bridge'. (Probably a 'copy and paste' error?)
Also; in wireless section, is 'frequency mode' set to "regulatory domain"? I presume so but it looks better to print that too. (And to be complete you could print the 'current power' outcome.)

Can you also show a 'scan' outcome test, to see if the spectrum is clear or used. (So if interference is, or is not, around)?
If there is, you than also should try wireless-protocol "NV2".

To find a more realistic throughput result I would than also try a bandwidth test from attached devices. (So leave the radio to do the work only they are supposed to do, sending and receiving and processing the wireless data. Bandwidth tests consume quit some cpu power.)
And do a tcp test. Sometimes udp test are very unreliable or not even possible where a tcp test gives more realistic results. (Show the routing/bgp config. I don't use BGP so don't know a lot about it but someone could find it interesting/poor config in this setup?)

Your remark of the pingtest in the NoLOS test (showing the 17ms) is not showing valuable information. An occasional high ping is always possible, even on the best links. The rest of the ping test still shows mainly <1ms. If you look carefully at the both tests, actually the second is better. The first one shows two "=" results which are more reliable than the "17ms" of the second test...... (The "17ms" could be filtered out as "exceptional and thus unreal for the test")

Overall a nice test that can still be improved a bit here and there....


(And, if you'd need some help in translations, 'we kunnen ook in het Nederlands communiceren. Ik help je dan wel verder. Waar in Belgie getest? I woonde vroeger in Roosendaal, net over de grens.')

@wirelessrudy, can you mail me on email on my website?
In dutch is a must of cource.

Still heard nothing from you Rudy??

Still heard nothing from you Rudy??

Send you a mail right now....

well, typical 5GHz link behaviour in NLOS situation when hitting trees.
2GHz or even 900MHz would yield better results in this environment.

well, typical 5GHz link behaviour in NLOS situation when hitting trees.
2GHz or even 900MHz would yield better results in this environment.

Yeah, 900Mhz would be nice in the forest.... but very not allowed in Europe.

The higher the frequency the less the effect of deflection and this is clearly visible.

I have a 3KM link with a SXT Lite2 on one end and a 912 on the other with an 11.3 dB antenna. Settings are wireless-N @ 2.4GHz.
Goes through about 20M of thick bush - you can't see anything through it + the neither end are particularly aligned well.

For the home end I just stood behind it whilst squinting and the remote end I just pointed to where I though it should. I get 40mbps.

Permanentley?
What are the ping times and are the stable?

@Plisken: you referring to me?

Ping is below. Very stable link. 100% uptime for months at a time. The only time the link drops is when a bird sits on the enclosure in front of the antenna (the small shit pile under the antenna suggests they do).

Hello RogerWilco indeed the question was for you.

Good result tnx for share that with us

No worries, mate.

In a week or so, I'm going to try and align the antennas better and see what happens then.
The link is also WDS with non-MIMO antennas.

This looks good. 300 meter with obstruction is not bad at all.

The only time the link drops is when a bird sits on the enclosure in front of the antenna (the small **it pile under the antenna suggests they do).

What I usually do is to put some long tieraps (rapties, bridas, plastic tensioners, whatever people call them) on the pole or any other point to prevent birds sitting (and sh*tting) at my antennas. Works perfectly...
Although you might not want to drill a tiny hole in the enclosure the pole that holds the antenna can be used to horizontally have on of these tie-raps sweeping over the antenna. It keeps the birds away...