Canadian TV, Computing and Home Theatre Forums banner

21 - 40 of 45 Posts

·
Registered
Joined
·
7,131 Posts
You could also use a coffee can waveguide antenna, as shown here. I've heard of these working reliably up to 5 miles (can to can.) I first heard of these being used over 10 years ago, so it's a well established design. If the coffee can doesn't work, you can always upgrade to a pizza sauce can or beef stew can design. :D

p.s. "Road Warriors" use similar designs to eavesdrop on private networks. If you see a car with a coffee can pointing out the window, it's a sure sign they are sniffing out unprotected networks.
 

·
Registered
Joined
·
84 Posts
Save the clutter and go with :
Two EOC-2610;s from EnGenius
There nice and compact 600mw 10dbi internal directional antennas
Power over Ethernet and no Signal loss,unless you want to add an external antenna to which is easily achieved as it has an external RP-SMA fitting under the hood.
http://www.engeniustech.com/datacom/products/details.aspx?id=246


Oh i just noticed you're in kitchener...not far from Belwood ont :)
This may be of interest these operate on the crowded 2.4Ghz and the less crowded 5Ghz band.
Available @ Tigerdirect not 2 far away.
 

·
Registered
Joined
·
598 Posts
Discussion Starter #23
I've read about the cantenna projects but I'm not getting any good information from the blog posts - ie really technical over-my-head graphs and such. And the db gain just doesn't seem to be much better than a bi-quad.

The quad antenna looks good, combined with a parabolic dish - I think this would give me great results.

Another contender is the helix antennas. Lots of good information, and within my abilities to build myself. This article has some good information -

http://www.safe-pc.net/helical.html

Just trying to wrap my head around the SWR and dielectric constants and velocity factors.

I've mapped out some good line of sight points of interest (open areas with no obstructions) from my place. There is a good one 1/2 mile away. I think the next step would be to build a few antennas and make some observations of signal strength and wifi speeds.

ps - the new egg coupon code doesn't seem to work here in Canada. I'm going to check out the surplus store just around the corner from me for USB wireless adapters with rubber ducky antennas.
 

·
Registered
Joined
·
598 Posts
Discussion Starter #24
Did some more reading and studying on the matter. Thought I would post this for anybody that is interested in long distance wifi.

First, I got my hands on a telescope to confirm a direct line of sight to my friends house. Surprisingly, even though I thought I had direct line of sight (looking a Google there were no big buildings in the way), there is a small 3 storey building in the way. Or more accurately my friends house is in a shallow valley.

No worries, with some careful measurements, I calculated a loss of about 25db for this lack of direct line of sight and I think I'm still within limits of being able to connect.

A great website I came across that can calculate all this stuff is here -

http://www.qsl.net/pa0hoo/helix_wifi/linkbudgetcalc/wlan_budgetcalc.html

------------

Finding the right radio was tricky. I looked at the USB wifi adapters however these things are all for Windows PCs. I wanted more diversity, something perhaps OS neutral. Most of my computers are Macs, then linux and I only have one Windows machine which is a laptop and I only use that to view video streams (when it's not updating Windows :p )

Power over ethernet radio was the logical choice. And I came across some nice options from a company called Ubiquiti Networks. In particular they have a nice radio called the Bullet.

http://ubnt.ca/home?page=shop.browse&category_id=9

Reasonably priced at $49 CAD for a 100mW radio. A little more for 800mW radio. The radio features a N type connector on one end and the ethernet connector on the other in a barrel shaped form factor. A segmented led display show the power levels so that comes in handy when aiming the antenna. This radio simply connects to any type antenna (biquad, helix, or dish)



Looking at the insides, the N connector solders directly to the board so there will be no losses in any cabling. So I should have very little loss.



Another thing that is nice about this radio is it is fully configurable, via a web browser interface, to be an access point, bridge, router, etc. It comes with a linux SDK for any custom work I want to do but I don't think I'll need to. It's also built from the ground up for long distance wifi, so you can easily adjust tricky items such as packet sizes and delays to wait for ACKs because of the distance with a simple 'distance to transmit slider' on the web interface.

I'm going to do a bit more reading a searching the web for as much info as possible, but I think I'll be ordering up a couple of these radios - unless I find something even better!

And I still can't make heads or tails of Industry Canada's limits on wifi transmissions, so I'll have to give them a call to find out what my legal unlicensed limit to transmit is.
 

·
Registered
Joined
·
8,188 Posts
Those "Bullets" look interesting. I see they also have higher power levels too. I wonder what they do about weatherproofing the ethernet side? You may also want to use shielded exterior grade ethernet cable.
 

·
Registered
Joined
·
7,131 Posts
And I still can't make heads or tails of Industry Canada's limits on wifi transmissions, so I'll have to give them a call to find out what my legal unlicensed limit to transmit is.
Right. Give them a heads up that you might be doing something illegal. ;) I think the guidelines and specs are published. Due to antenna gain, your installation will bend the rules somewhat. I have a feeling that unless you go over 100mw, nobody will notice. I would be most concerned about interference with/from other wireless devices in that building. Maybe a mast on each end will clear it.
 

·
Registered
Joined
·
598 Posts
Discussion Starter #27
Unless I'm mistaken Industry Canada is quite generous when it comes to output power. In fact, it seems to good to be true kind of output power!

Here is what I think I know-

main link on industry canada site -

http://www.ic.gc.ca/eic/site/smt-gst.nsf/eng/sf08655.html

The wifi radio has to comply with RSS-210

For systems employing digital modulation techniques operating in the 902-928 MHz, 2400-2483.5 MHz and 5725-5850 MHz bands, the maximum peak conducted power shall not exceed 1 W.

Systems in the 2400-2483.5 MHz and 5725-5850 MHz which have an e.i.r.p. above 4 W are permitted only for point-to-point systems (i.e. point-to-multipoint systems and multiple co-located transmitters transmitting the same information are prohibited from exceeding 4 W e.i.r.p.).

Point-to-point systems in these two bands may use higher e.i.r.p. as necessary for satisfactory operation provided that the higher e.i.r.p. is achieved by employing higher gain directional antennas and not higher transmitter output powers.


Max output is 1 W, max e.i.r.p. (the output of the radio + antenna gain) is as much as I need to go point to point, provided I use a better antenna and not a more powerful transmitter.

Handy link to make sure the wifi radio complies with RSS210

http://www.ic.gc.ca/app/sitt/reltel/srch/nwRdSrch.do?lang=eng

The Bullet radio @977mW has been certified by Industry Canada.

------

So, a quick call to industry canada will clear this up in my mind.
 

·
Registered
Joined
·
7,360 Posts
Surprisingly, even though I thought I had direct line of sight (looking a Google there were no big buildings in the way), there is a small 3 storey building in the way. Or more accurately my friends house is in a shallow valley.
If the building is made with a lot of metal, you may be effectively completely blocked. Any chance of mounting both antennas high enough so you do have line of sight ?
Max output is 1 W, max e.i.r.p. (the output of the radio + antenna gain) is as much as I need to go point to point, provided I use a better antenna and not a more powerful transmitter.
Yeah, 1000 mw limit is fine, get the 800mw unit.:p
 

·
Registered
Joined
·
598 Posts
Discussion Starter #29
"If the building is made with a lot of metal, you may be effectively completely blocked. Any chance of mounting both antennas high enough so you do have line of sight ?"

Looks like it might be metal - here is the building in question.



I have been using elevation maps to calculate how high the antenna must go to clear it. Quick calculations show an antenna 25ft from the ground (easily done with a chimney mount) will clear it. My next step is to borrow an accurate GPS unit to get super accurate elevations. I tried today with my iPhone's built in GPS but it's just not all that accurate for getting elevations.
 

·
Registered
Joined
·
7,131 Posts
A 100mW radio with a 10dB gain antenna is 1W e.i.r.p. 100mW with 16dB gain would be 4W e.i.r.p. That doesn't take things like line loss into account.
 

·
Registered
Joined
·
520 Posts
Quick calculations show an antenna 25ft from the ground (easily done with a chimney mount) will clear it
"Just Clearing it" may not be enough. If you haven't already found out about the "Fresnel" effect you might want to look it up. In a nutshell even with directional antennae there is a spread to the signal (like a cone radiating out) - the same with a receiving antenna's sensitivity. The two combined form a lens like pattern (hence "Fresnel") so if the building is in the middle of the path you have to take in to account the width of the fresnel zone at that point along the path. As you are relatively close it's probably not too much of a problem.

As one of my old profs used to say Radio Wave propagation is more of an art than a science. Sometimes things work for no apparent reason and sometimes they don't when logically they should.
 

·
Registered
Joined
·
8,188 Posts
One thing I noticed in the videos is that the Bullet has a shielded ethernet connector. So, if you're installing one of those, you'll be needing shielded, exterior grade ethernet cable. Working with that stuff is real "fun".
 

·
Registered
Joined
·
84 Posts
hmmm, no sense in using these boards if I have signal loss from the connector.

Here is a pic of single board computer with two wifi cards.



Each card has two ultra miniature connectors. The main antenna connector is in the corner if you are using is the one coax connection. I have mini cables that are about 11.5 inches long that break out the ultra miniature connector to an N-connector. If there is significant loss using this method then I'll have to look into a different wifi card - like the USB you mention that doesn't have these type of connectors.

Also, if I decide to go 802.11a at 5.8Ghz does the antenna scale linearly (can I change the dimensions to a fraction of 2.4/5.8 the lengths) or does the physical properties of the antenna change because of the higher frequency and at this point some sort of wave guide is better.
Fyi

Length: 8"
Thickness: 1.32mm cable.
Frequency Max rate: 6GHz
Cable Loss: 7.5dB/m
 

·
Registered
Joined
·
2,152 Posts
Over 2 years ago I setup a 5 km backhaul link from my home to my data center. My first attempt was to use two consumer grade linksys GL radios and 24 dBi grid antennas, believe it or not it worked… we had a good signal but we could not figure out at first why the link would run so slow at only 1 Mbit. As it turned out we discovered that once you go over the 2 km limit the acknowledgment timing gets confused and as a result keeps retransmitting the same packets over and over again slowing the whole thing down to a crawl. The DD_WRT firmware at that time had no working way to adjust this...

I then decided to test some professional WISP radios that would allow me to run a 5km backhaul link with lots of speed. I contacted Tranzeo and they were nice enough to ship me some TR-5a radios and 24 inch microwave dish antennas to test. These radios run on 5.2 GHz to 5.8 GHz. I found the 5.8 end of the band to have too much noise and use the 5.2 frequency. The link has been running now solid for 2 years in every type of weather… and not having to pay Rogers or Bell for Internet is the best part of all along with the awesome speed… 20 Mbit up and down and I’m on a private VLAN with the server room from my home.

 

·
Registered
Joined
·
598 Posts
Discussion Starter #38
The antennas look really nice! Over 5km and good speeds, this is giving me the confidence that my crazy plan just might work!

I've read about the distance being a factor in the wifi protocol. One of the reasons I was impressed with the Bullet products is they have a simple UI that can set the distance the antennas are apart and the software will figure out the rest for the timing.
 

·
Registered
Joined
·
84 Posts
That looks very expensive, how long was the payback from the initial investment there?
Ack timing has always been a feature in Tomato Firmware.
In addition to power levels it is also important to know how the 802.11 protocol uses acknowledge for each received frame. If acknowledge is not received the frame is re-transmitted. By default the maximum distance between transmitter and receiver is 1 mile (1.6 km). On longer distances the delay will force retransmissions.

thanks for the pics
 
21 - 40 of 45 Posts
Top