Managed to break my previous WIFI record

[gibsonlp]

Hey guys.
I am very excited - 2 days ago I managed to break my previous (50Km) WIFI link record (in a single hop) with no less than 220Km (2 hops, 135Km & 95Km) (220Km is the direct line between the first and the last hop).
According to the information available online this is currently the 3rd longest wifi link in the world for a multi-hop link and one of the longest single hops link.
The speed is actually pretty good, getting around 26Mbps on the 100Km link (802.11g hardware in 10MHZ spectral width) and 36Mbps on the 135Km link (802.11n hardware - in 5MHz spectral width).
This link will let us surf the net in our upcoming yearly convention which is held somewhere in the middle of the desert. Our internet provider this time will be - Haifa university (located 220Km away from our desert venue).
If anyone wants more technical information - I will be glad to provide it.

[cjj]

Cool! I assume you've got some specialized antenna setup for this?

I'd be interested in more technical info!


Yeah, I'm a Techno-geek...

[gellkeller]

Nice work. Sounds like you have a pretty clear spectrum there as well.

[johnallg]

You did it Gil! All that planning really paid off. The special help was a plus too. Congrats!

[winston]

Congrats Gil,

That sounds like it was a fun project. As I mentioned a while back, we have a radio based internet system at our lake property. It works extremely well.

[gellkeller]

Out of interest, how tall are the towers at each end to cater for the earth's curvature?

[gibsonlp]

Thanks guys.
Some details:
The link is from Haifa university to Ashkelon's power plant's south chimney and from there to our location (2 different links).
Haifa university has a tower 100m high (32 stories) on top of the carmel mountain which is 482m high so my dish is 585m high (3m pole) on the university side.
Ashkelon power plant chimney is 260m high.
86% of my long link is above water, this height was enough to get around earth curvature.
The other link is from the chimneys to our conference location which is 660m above sea level (Ramon area) deep inland.

For the link from Haifa to the chimney I used a pair of Ubiquiti 3ft dual polarity dishes and a pair of Ubiquiti RocketM5 wifi units operating at 5.2GHz (which is illegal to use outdoor in Israel without a permit, which I have). Total EIRP is ~61dBm (34dBi dishes + 27dBm TX power). Spectral width is only 5MHz due to the fact that wifi is traditionally limited to 50Km @ 20Mhz width, you double the distance each time you halve the width so I had to go down to 5MHz which isn't that bad considering the fact that it's a dual N chain operating at 36Mbps... I can use a special "noack" configuration which will remove any distance limitation for any spectral width, but I rather have a stable fast link than an unstable ultra fast link (I could get it to work at 100Mbps!)
For the link from the chimney to our conference I used a pair of 4ft grids and a pair of Ubiquiti Bullet2HP operating at 2.484GHz, channel 14 is illegal in Israel but I was given a permit to use it. Basically I was allotted 20MHz between 2.48 and 2.5 but my equipment couldn't make it above 2.84 as a center frequency (instead of 2.49) so I have some interference from 2.479GHz to 2.482GHz (the top of channel 13), I am working at 10MHz spectral width (instead of 20) for this link. Total EIRP is ~57dBm (27dBi dish with 30dBm tx power)

I have attached a link analysis of the links.
Making this link work was one of the hardest things I ever did, it took me weeks to get all required permits and to do all the initial planning. I had only 2 days to install everything as I could not legally install the equipment myself on 2 out of the 3 locations (Haifa university and Ashkelon chimney. The conference end is a 4ft dish on a very large tripod which doesn't need special team to work on it), the law in Israel requires a certified climber to go on a mast or a pole to do such a work so I had to hire 2 teams to do the link from Haifa to Ashkelon and 1 team to be on the chimney to align the 4ft dish facing our conference site.

The entire story could make for a Hollywood film, really. We made it work in the last minute (literally) -
The plan was to install the link from our conference to the chimney on the first day (as well as the 3ft dish looking at haifa) and do install the 2nd link on the 2nd day.
We lost the first day as the placement for the 4ft dish looking at our conference was bad and it could not move to the required azimuth. Took us some time to realize that but it was too late, we could not get the required mounting materials to move it to a different place (which was annoying as I did a preinstallation tour with the climbing team manager so he could asses both sites.
We tried again on the 2nd day. After placing the dish on it's new location I fired up a very expensive rhode signal generator only to find out that it reports on an error and would not go above 1MHz... aligning narrow dishes using a signal generator and a spectrum analyzer is much easier than using the builtin antenna alignment tool that ubiquiti is providing with their equipment. We gave it a shot anyway but we didn't see anything.
I basically gave up at this point and told the climbers to start packing. I went to my laptop to shut it off and then I saw a pair of power indicator bars on my laptop screen, letting me know that the wifi units are actually seeing each other (very faintly).
I started jumping and shouting! we have a link! lets start to align the antennas.
Took us another 30 minutes and we were done aligning the dishes (-64dBm rx gain, not the maximum theoretical gain but good enough as we had limited time).
I then drove up to Haifa to work with the 2nd team (knowing that I have no way to use the spectrum/signal configuration and that I will have to rely on the builtin antenna alignment tool). Took me 3 hours to get there and it was getting dark. The guard at the enterence told me he was given instructions to kick us out once night is falling. This means I had only 40 minutes to hook up my equipment and align the dishes!!! (luckily the climbers team was already there and the dish was already installed).
To make a long story short - we managed to align the dishes to an acceptable alignment as the guard was nocking at the door, telling us to leave...!

[johnallg]

Sounds like it was a real adventure Gil. You should have assumed Murphy was going to assist. Kudos for pulling off a real coup! I'm really impressed with a 5ghz link of that distance at N speeds. Wow.

[gibsonlp]

The RX power fluctuates all the time between -57dBm (which is fantastic) to -86dBm which isn't that good...
I managed to deactivate the ack mechanism and raised the spectral bandwidth to 20MHz, I locked to a lower speed but was still able to push true 50mbps each way!

[billydlight]

[jps]

I managed to deactivate the ack mechanism...

No Way!!!