mcm375

Hi I'm pretty much a wireless noob, but I did manage to figure out a little bit of info about link budgets. According to one resource, the requirement in the 2.4GHz range is 40 + 20log(N) where N = distance in meters.
Now, I want to set up a wireless link between two houses -- my own, and a neighbour. Without measuring I would say we are no more than 100 meters apart (in fact, probably more like 50-60, but I want to be on the safe side in my planning and we're talking about a reasonably small distance anyway..) So from this equation I figure I need 80dBm. According to the SeattleWireless wiki (http://www.seattlewireless.net/index.cgi/LinksysWrt54g), the WRT54G has 15dBm of xmit power and a receive sensitivity of -65db at 54Mbit.
I have read that there is a "fade margin" of 10dBm which should be taken into account (not quite sure how this figure is derived.) So that leaves me 10dBm that I need to make up with antennas, and I'm unsure which type will fit my requirements/preferences best:

1) Should be as discrete as possible (I'm going to be asking my neighbour if they would be interested in allowing me to set up this link on their house if I pay for their internet access.)
2) Preferably does not require external mounting, and if this can be satisfied, do double paned glass windows pose much of a problem for RF? How would this affect the link budget? The less cable routing and external house work, the better, at least as far as my neighbours place goes. I want to place as little burden on them as possible to make this an easy "sell", preferably I would like to setup a little unit on a window sill inside their house looking straight at mine, not sure if this is feasible though. I do not mind mounting some monstrosity of a device on the side of my house either, so perhaps that can make up for it (but no ghetto pringles or jelly meat cans please.. I do have some sense of aesthetics..)
3) Must penetrate snow and rain. I do have line of site to a reasonable portion of the target house from mine, there is one minor branch from a pine tree, but its nothing much to speak of (think of a branch 1" in diameter with a few pine needles on it..) How does this affect the link budget? Would a highly directional antenna like a yagi be required to punch through heavy snow fall reliably?

I'm guessing the answers to (1) and (3) are not compatible, so (3) is the higher priority requirement of the two. The neighbour side of (1) can probably be solved with a $50 bottle of wine...

Oh yes.. I'm planning on using two WRT54G's.. any reason this is a bad idea? My friend has one with the sveasoft firmware and he swears by it (though I don't believe I should need to up the power given decent antennas.)

G8tK33per

Lions/Tigers/Bears?

__________________
"Benjamin is nobody's friend. If Benjamin were an ice cream flavor, he'd be pralines and dick."

Sons of Confederate Veterans

mcm375

Well.. yeah.. actually.. there have been bear sightings around these parts before, and deer often run in the forests. I think the elevation of the houses involved over ground level should eliminate that as a problem though
I found this equation for the fresnel zone (unfortunately it uses imperial units.. eww): r = 43.3 * sqrt(d/4f) -- where d = distance in miles, f = frequency in Ghz, and r = fresnel clearance zone radius in feet. Using this formula I come up with a radius of 3.42 feet. I've read a little bit about 80% of this figure being acceptable, but I do not quite understand the implications of this figure: Does this indicate that I have to have at least 80% of 3.42 feet *completely clear* of anything but air at the midpoint for this to operate correctly? Or does it indicate that 20% of a 3.42 foot radius midpoint can be occluded by branches/trees/whatever and still xmit/recv a signal properly?

I don't think I have this kind of clear space at the midpoint, there are some sparse branches from a tall pine in the way (they are the lower branches which are kinda dead and easy to see through.. but they *could* potentially be removed..)

I think I'll take a photo from a few locations to give you an idea of the LOS I have.

Thorn

Yes. One dead branch may not be a problem. On (or more) branches, fuzzy with pine needles may be a problem. Pines by their nature are a problem for 2.4GHz. All those needles act like 1/4-wave, water-filled attenuators.

You may be close enough that you'll lose signal, but still get through.

__________________
Thorn
"I'm The Doctor. I'm a Time Lord. I am from the planet Gallifrey in the constellation Kasterborous. I'm 903 years old and I am the man who is going to save your lives and all 6 billion people on the planet below... You got a problem with that?"

mcm375

And what about the possibility of mounting behind a window? Does this pose a serious signal loss problem?

And how does the choice of antenna affect the ability to blast a signal through trees/windows at this short range? Could I potentially get away with a high gain omni antenna to avoid having to aim the things really well (as long as I can get the elevation about right, that is..)

According to this calculator: http://www.zeuswireless.com/support/som.php, I will have a 22.2 dB operating margin with only 6dBi antennas, a 15dBm xmit signal, -85dBm receive sensitivity (11mbit on the WRT54GS, which is more than enough for this application) and assuming 5dB losses on both Rx and Tx cables (assuming 50 foot @ 0.1dB loss/ft, which is a SUPER HIGH estimate.)

Is that kind of operating margin enough to handle windows, snow, rain and branches with a "light" amount of leaves (that you can see through) ??

King_Ice_Flash

Windows do not create anywhere near as much interference as trees, however, they do get covered with frost, rain, and dew. It is hard to say how much problem trees will cause without actually trying it. This is why people do site surveys. You never know what will happen.

__________________
"Yeah," said a voice from under the table, "you go to pieces so fast people get hit by the shrapnel."

mcm375

Mucking around with this calculator Very good LINK BUDGET calculator, it seems I should have a 20dB operating margin over this 100m distance using just 6dBi antennas and assuming pretty much the worst as far as cable losses go, while keeping the wireless base stations/routers at their normal xmit power and -80 dB sensitivity for 11mbit. Most figures I can find for trees suggest 0.35-0.5dB per meter of tree "material" (where leaves are worse than branches.) I probably have a cumulative total of 2-3 meters of tree "material" partially occluding my line of sight. Even if I figure for 10 meters of tree material at 0.5db/meter, I still have an operating margin of 15dB. From what I've read, this is pretty much bulletproof. Am I wrong here?

If not, what's the most user friendly antenna I can get away with at 6dBi? Is it possible to go with discrete rubber ducks mounted inside a window? If not, what do you think of this nice little antenna: http://www.rangeextender.com/mdmidian.html ?

It's directional, which poses an aiming problem I suppose, but its apparently nice and small. And it sure is cheap. I havn't got acrobat pdf reader installed so I didn't check out the full specs.

wrzwaldo

When the trees get wet it will add a bunch to that.

mcm375

Any idea of a number to wrap around that statement? Would it double the loss from trees to 1dB/meter? (I pulled that out of my ass..) Even given that, having a margin of 15dB is pretty good for handling such conditions is it not?

wrzwaldo

I would give it a go but I wouldn't bet my life on it.

mcm375

So much for "doing the math"...

mcm375

I'm almost at a point where hardware will be purchased for this wireless link to my neighbours across a green belt. I want to make this as low impact as possible for them, so the preference would be a 5-6 dBi gain omni "desk top" (something with a flat base) that would sit inside a window on a window sill or something like that. However I want to make sure this has enough grunt to blast through the light trees in between (not many trees, just a couple of branches getting in the way, still visible through them though) and any weather (we get heavy snow here.) So, I found this really neat mini panel antenna (http://www.rangeextender.com/mdmidian.html) which is 12dBi. So I was thinking about the possibility of mounting this externally on my house, pointing at a 5-6dBi omni in the other house. From what I can tell, this should be fine, the only possible problem being the RF radiating all over the place but from a functional stand point, this shouldn't be a problem right? I've also read that at higher gains (is 5-6 for an omni considered high?) the e plane flattens out (I assume e = elevation) making correct elevation important. Having said that, I believe I can get both antenna at the same height anyway, and we're talking about 100 meters MAX distance (probably more like 70) so perhaps this isn't such a big issue.

After running the numbers (including taking cables and connectors into account) I figure I should have a ~27dB operating margin at 100m. That should be enough to get through the trees within the fresnel zone even if they are wet (we're talking maybe 2-3 meters of 'tree material' here, tops) and any snow/rain that might be falling.

Is this do-able? Can you recommend any omni in particular for the neighbour's side (should have an RP-TNC connector for a WRT54GS.)

I'm also planning on using two WRT54GS's with 3rd party firmware to support WDS. I expect this to take care of the routing from a single internet connection to multiple hosts on both the neighbours and my side (via the wireless link between WRT's for my side.) Is this feasible?