Slashdot Mirror
Wide Area Wireless on a Shoestring Budget?
wkytechhead asks: "My father owns a greenhouse operation that covers a few dozen acres. He has a number of greenhouses some up to 1000' plus apart that he would like to network. Currently he is using a HomePNA based network via regular RJ12. He has decided that he would like to go at least partially wireless. Some consideration has been given to using the backbone with fiber convertors, but he would really like to do full wireless. I have checked into homemade and commercial 'Cantennas' but I am not sure if they are worth the money. How would my fellow geeks go about wirelessly networking a large outdoor area for as cheap as possible?"
You should check out these antennas from HyperLink Technologies. For outdoor applications these should work very well for you.
Some golf courses and other large areas use remote-DSL for such links. Maybe that would apply to him? Many cisco DSL modems can be operated in server mode, only downside is you must run RJ12 separately to each location.
Otherwise, run ethernet?
if you are going to go wireless, get some good APs and sector antennas, or alternately setup a bunch of repeater stations that use different channels to avoid interference.
--IronHelix
minidishes (commonly used in the UK for digital satellite television) can be easily adapted to use a WiFi signal boosters. They're highly directiuonal of course, but ranges of 2 or more kilometers LOS are not unknown. Plus, no boosting equipment is required, just a modification to the antenna.
For outdoors, I think your SOL, but for inside the greenhouses, investigate something called "leaxy coax." It's basically a coaxial cable with little to no shielding, and a couple of companies have recently made it usuable with 802.11b/g.
Wire cheap antennas to cows.
:)
Problem solved
I would find the central point and place broadcast point there and at the end of that one range place 2 at the other points to max. range and give full coverage. Nice idea. I am working on something like that for my wine cellar.
http://www.seattlewireless.net/index.cgi/AntennaHo wTo
Is a WiFi antenna howto and I build the can-waveguide out of two tomato sauce cans with an N connector. It totaled to about 5 bucks worth of parts and works very well. The tuning for it though might be off and I hand redid with a sheet metal snipper.
Unless the sites are spaced something like along a road, a mesh network would probably work and be much easier to setup as well as much more robust (I presume the network carries vital info on temperatures, greenhouse equipment status, etc).
2 2/wirelessmesh.html for an overview
See http://www.oreillynet.com/pub/a/wireless/2004/01/
Eat Lamb, 1 million coyotes can't be wrong
I've had great luck with a 24dBi Parabolic Antenna at over 1.5 miles away.
. htm
If all of these sites are within a mile with reasonable line of sight and not many trees in the way, I'd suggest putting the parabolic antennaes at each location directed toward your main location.
At the main location go with a nice sector antenna array.
Heres a link:
http://squitter.com/products/antennas/parabolic24
Homebrew your AP and end nodes with a workstation or soekris board and a netgate radio...or connect the antennas to commercial AP/Bridge units.
Depending on how many sites you have to link, you should get out at a reasonable cost.
This is our amateur radio repeater that uses VoIP to link to other repeaters around the world. The link is 802.11b
Here is the AP that provides access.
Nothing special to do a 1000' link. Just a parabolic grid antenna on the client side.
Here are some parabolic grid antennas.
--fatboy
Cantennas are the wrong way to go, as their propagation pattern approximates linear, like the yagi designs. What you want is an omni, sometimes called marine, antenna that will spread signal in a plane. If you're in a greenhouse, I'm assuming you don't want strong signal going up or down, but horizontal in the plane of people walking around. Here's an example of one I grabbed from Google: radiolabs omni antenna For about another $30 you can pick up pigtails on eBay that let you attach these to the usual netgear/buffalo/d-link/linksys/etc. accesspoints. You can place them for effective 10Mbit coverage about one for every 2 acres assuming clean line of sight to the antennas and no major obstacles. Note that vegetation would definitely impact signal propagation in the 5.4ghz band.
http://tinyurl.com/4ny52
I would have workers every 100' and they would throw packets back and forth. The workers would not need to know TCP/IP since they would not even need to operate at Level 1, they would just act as repeaters/fowarders, only needing to know whether to throw a packet to the next or to the previous worker and that would be easy because they would just have to make sure not to return the packet to the worker they got it from. This repeater/forwarding mechanism works well and requires no additional expenses since the workers would be hired to do actual work in the garden anyway. Oh and did I mention the workers are wireless as well?
Yeah. The FCC typically just responds to complaints. As long as you don't paint a picture of Janet Jackson's nipple on the antenna you should be fine.
blog
Can't get much cheaper. http://www.faqs.org/rfcs/rfc1149.html
Well, at least you'd know which way the thing was pointed...
The Spoon
Updated 6/28/2011
The first question you have to ask yourself is what you want to get out of the network. Do you want full roaming capabilities, or do you want localized points of presence in every greenhouse? Are you wanting to implement VoIP? Network aware control and monitoring systems?
If you are looking for a setup like what I had (points of presence in every greenhouse), all you need are some cheap(ish) base stations for the endpoints, and a kick ass 24dB omnidirectional antenna for the master base station at the head office. Build small networks at each pop and bango. If you want roaming in a few areas, throw in a small consumer wireless base stations and hook it up to the pop.
If, however, you are looking at using the network for business critical voice services or control and monitoring systems, then you had better look seriously at fiber (fiber ring preferably). You never know when some bird lays a massive shit on the antenna. Wireless is simply not as reliable as wired.
Feed the need: Digitaladdiction.net
RadioLabs.com has an excellent selection of wireless antennas, coax, and connectors that will work for you.s s-2.html
http://www.radiolabs.com/products/wireless/wirele
Also, I agree with an earlier post suggesting the Linksys WRT54G or GS running the SveaSoft firmware.
http://www.sveasoft.com/
With a slightly better antenna and a few WRT54G's in a WDS configuration, you could easily provide both wireless and wired connections in the remote buildings and have some wireless coverage between the buildings as a bonus.
With enough of the WRT54G's in the WDS configuration and using OSPF, you could create a "self healing" component to the network.
G'Luck
You'd be suprised the kind complaint they respond to. My house used to have a very large antenna for TV(35 foot I think) because we were on the dark side of a mountain and couldn't recieve a signal from NYC. We also had a special motorized/pointable CB antenna. The guy next door, instead looking at the mountain and thinking, "hey the only reason they can a signal is because they have a giant antenna" thinks "hey they are plotting against me, and blocking my signal"
The FCC actually came and did tests, said that was silly and went home.
=1000101
Or you can search froogle for them
"2.4 ghz parabolic antenna"
Lowest price is around 33$ for a 15dbi gain. 50$ is about right for a 24dbi gain. Though a parabolic is only going to be truly handy for sight to sight. They do have low cost omni directional antennas. This is difficult to advise without knowing the layout.
ie, possibly sight to sight parabolic for the backbone link and then omni's distributed throughout.
If you are really worried about distance, just dig up a router that can be tweaked for a bit more output power.
Now, if completely unsure of what your needs are, here is a handy calculator to help you figure your distance needs out.
http://www.signull.com/fsc.php
For the uninformed, cable loss is going to be calculated by the amount of signal loss expected when using some portion of cable from the transciever to the antenna. It should be expressed quite clearly on type of cable you are going to be using.
There are also some common models output numbers listed there as well.
That really sums up everything. It should be quite cost effective to implement a few antennas and a cheap 802.11b setup.
"You should always go to other people's funerals; otherwise, they won't come to yours." -- Yogi Berra
Just convince the city to run two perpendicular streets through your property. The resulting intersection will produce no fewer than 4 Starbucks coffee shops, each with their own wi-fi transmitter. Problem solved.
So, the limit is 30 dbm (1000 milliwatts), most wireless cards are about 15dbm, which gives us 15 dbm of slack. 24dbi exceeds 6 dbi by 18, so this would be illegal for a point-to-multipoint topology. However, if it was a point-to-point link, that 18dbi of excess gain results in a reduction of only 6dbm of permitted power, so you could use a 24dbm radio (or a little over 200mw) legally. (Though ianal or an rf engineer, so take this with a grain of salt.)
At that distance, regular old 802.11 will be fine. Two $50 directionals, use a of the shelf AP at one end and an off the shelf network card at the other. Keep the cable runs short. Long cable runs mean you need rather expensive cable. If you're unable to place the PC with the wireless card close to the antenna, use a Pentium class throwaway stuffed in the attic to bridge it to Ethernet, or a second AP. (Make sure the two AP will interoperate without too much firmware headache.)
You should be able to deal with minor tree obstruction.
.sig: Now legally binding!
www.smartbridges.com
the all in one units.
simple. cheap.
Use a link budget calculator. (The link is one I just found with google).
Basically, you need a certain signal to noise ratio for a digital radio connection to work at all at its lowest speed. Increased signal to noise ratios get you more speed and some margin of reliability.
Signal decreases with the square of the distance. If you double the distance you'll have one quarter the signal, or 6db less (decibels are logarithmic - 3db is a ratio of 2, 10db is a ratio of 10). So, everything else being equal, you'd need 6db more gain on the antenna at either end to get the same results.
For your particular scenario you'd probably be fine with just a couple cantennas or other moderate-gain antennas.
One thing to watch out for when shooting through trees is that they may not have leaves now but they will in a few months!
A lot of those wireless cable antenna are actually in the 2.5ghz range.
Still, they will work just fine, with some gain loss...
Here is a smaller article on the matter... to adapt one. (kinda helpful if you really really don't know what you are doing)
http://www.qsl.net/n9zia/wireless/page04.html
"You should always go to other people's funerals; otherwise, they won't come to yours." -- Yogi Berra
A friend pointed the site out to me about a week before it hit slashdot, so by the time the original story broke here, I had built one.
Long story short. They work really well. I've pwned every wireless access point within a 3Km radius of my house. Free Internet anyone? :-)
I find your ideas intriguing and I wish to subscribe to your newsletter.
I would recommend that you purchase a BOOK.
Wireless LAN's End to End
and
Wireless Networks The Definitive Guide
are both excellent books that clearly explain the issues involved with DESIGNING a wireless network.
A +24dbi parabolic dish is more than innappropriate for what you are doing, so is running firmware hacks to "Pump Up Your Power".
It's not about how much signal you can irradiate the land with, it's about placement and antennae with the correct signal pattern.