1.3 Miles PtP Link - Seeking suggestions.

I've got a 1.3mi distance I'd like to do a PtP link with 2 WRT54GL's. Its not exactly a pure LOS between the 2 locations, but the obstacles are minimal.

Using a 2400MHz Tx frequency, the 1.31mi distance, along with a 200mW Tx power, -80dBm Rx sensitivity, 1dB loss for connectors and cable, and 14dB antenna gain on the Tx and Rx, I get a Theoretical System Operating Margin of 23.5 dB.

This is using a helical antenna, chosen because it is of a circular polarization and according to my research, the best for getting through and around obstructions.

Does this configuration sound about right?

Reply to
SM
Loading thread data ...

I travel and connect over those distances routinely with nothing special, except effort and experience. Easily doable..With higher gain antenna.."piece-of- cake." Luck

Reply to
Jack Daniels

"SM" hath wroth:

Assuming your numbers are correct, a 23dB fade margin should work. You didn't specify the connection speed that you were getting the -80dBm Rx sensitivity so I'll guess you're going for about a 24Mbits/sec connection. That might be a bit optimistic in the presense of any interference. 1dB for connectors and cables seems a bit low unless the WRT54GL is mounted directly on the back of the antenna. 200mw will cook the WRT54GL eventually. I wouldn't go much over 100mw (+20dBm) tx power. Your "minimal obstacles" are still obstacles and undoubtably infringe on the Fresnel Zone. However, without knowing their composition, I can't assign a loss value estimate.

I've had some experience with helicals and do not agree with your assessment. They're great for eliminating odd number reflections and multipath. They're also good for dealing with odd mounting and pointing arrangements. However, they are not useful for "drilling" through obstructions. Details please?

I'm too busy right now to run the numbers. See:

formatting link
'll run them later...

Reply to
Jeff Liebermann

Jeff Liebermann hath wroth:

TX power = +20dBm (100mw) TX coax loss = -2dB TX ant gain = +14dBi Distance = 1.3 miles RX ant gain = +14dBi RX coax loss = -2dB RX sens = -82dBm (at 18Mbits/sec) Fade margin = unknown Plugging into:

yields a fade margin of 19.5dB. That's close enough to 20dB to be workable.

The catch is that this is the *BEST* case estimate. It just doesn't get any better than the above numbers, but always ends up worse. Antennas don't quite have the advertised gain. Radios don't quite put out the rated power or have the rated sensitivity. Interference always causes problems. Of course, I added no consideration for whatever obstructions you have in the line of sight.

Who's helical antennas were you considering?

Reply to
Jeff Liebermann

The intent is to setup for 802.11b, shooting for 11 Mbps. I couldn't find a manufacturer's published spec for Rx sensitivity, but research at

formatting link
found that the WRT54GL had an Rx sensitivity of -80db in 802.11b mode (-65db in 802.11g)

Next intent was to rig the WRT54GL for outdoor habitation - use POE - and have it live at the antenna height thereby reducing many feet of cable line-loss.

All obstacles from point "A" to point "B" are trees and residential houses of roughly the same size and shape. (See attached photo.) And this is why I 'm asking for input. As point "B" is about 30' higher than point "A" with no rise in-between, I hoped using a 14db helical antenna from 50' mast at Point "A" to roof-top at point "B" would get me past all the odd shapes.

Or should I just use a 23db parabolic antennas with vertical polarization (this would give me a 37.5db fade margin)? Would this amount of fade margin get me past the various trees and roofs, or must I install masts at both ends?

Will an increased fade margin help me with my obstacles, or will the only thing that helps me is height?

If an increased fade margin does help, I'm thinking of going with parabolics of 23db gain.

The antennas are made by Wade here in Canada. See below to a link to the specs.

formatting link
formatting link
Thanks.

Reply to
SM

Forgot the link to the picture between location A and B.

Its here:

formatting link
Thanks.

Reply to
SM

That depends how you phrase it.

Increasing you acceptable fade margin means you are setting your expectations for reliability higher.

If you design you system with a higher allowable fade margin, it will work under worsening conditions. If you allow for only a slight fade margin, it may fall under adverse conditions.

Increasing the antenna heights will help you clear the obstacles better, but the actually hight may have to be varied at one end for optimum signal as you will find peaks and nulls up and down the mast.

[snipped from a previous post of mine]

decaturtxcowboy wrote: > Jeff Liebermann wrote: >> On Sat, 18 Nov 2006 01:01:17 GMT, decaturtxcowboy >> wrote: >>

Reply to
decaturtxcowboy

"SM" hath wroth:

Those numbers are garbage. Sensitivity varies with connection speed. See table at:

These are for a DLink DI-624, but are close enough for ballpark calculations. At 11Mbits/sec, use -82dBm.

However, I wouldn't use 802.11b modes. 802.11g is more effective in the presense of reflections and interference. It's also more sensitive for the same speeds. Try a 12Mbits/sec OFDM connection for a sensitivity of -84dBm.

Good idea, but there are some serious environmental concerns with outdoor repackaging. Keeping the moisture off the circuit board is always a problem. Lots of RF connectors. However, it's been done hundreds of times, so it will work.

Ummmm.... I can see you've never tried to install a 50ft mast. It's not easy. The common Radio Shack 30ft push up mast is also a challenge. Finding points to anchor the guy wires is the major headache. Aligning the antenna 50ft in the air is VERY difficult. A

24dBi dish has a -3dB beamwidth of about 10 degrees. That means you have to align the antenna to within about +/- 4 degrees. Mark a 10 degree angle on a piece of paper and visually see how well you can align it. The pole has to be adjusted and aligned vertically to the same accuracy as the antenna.

Basically, you're asking if it's better to use high gain at the rooftop level and try to drill through the obstructions, or if it's better to use a 50ft pole and a lower gain antenna. The 50ft pole with a lower gain (10 to 12dBi) antenna will probably work best. However, I question your ability to actually install a 50ft antenna mast.

Unfortunately, the picture you posted is useless:

It doesn't show any of the obstructions or Fresnel Zone clearances. What's needed is a side view picture, from ground level, showing the endpoints of the link. That's often difficult to do if there's no convenient viewing point. If you can't do that, two photos, along the line of sight, with the target marked, will suffice.

Looking at the photo and from your description, it appears that you're barely skimming across the rooftops with several large trees in the way. First, the Fresnel Zone needs to be considered. At 1.3 miles, that's about 22ft at 80% clearance at midpoint. I don't think you have 22ft of ground clearance, much less building clearance. This is not going to work too well without clearance.

Obstacles create both attenuation and reflections. An increase in fade margin will certainly help with attenuation. However, reflections are a bigger problem. Position the antennas in a null, and the signal will completely disappear. Worse, it will not stay put. If something moves along the path, then the reflections will change and the null might get better, but more probably worse. It's really a difficult call. For links without spacial diversity, I move the antenna up and down the tower until I find a location that doesn't seem to be anywhere near a null and hope for the best. That's not a great way to do it, but I usually don't have much of a choice. In your case, I don't know what to suggest.

I wonder how well it works with a bird sitting on the helical? 14dBi gain with 36 degree beamwidth. That should be fairly easy to aim.

Dish 23dBi with 11 degree beamwidth. 3 times as difficult to aim but not impossible. I don't think you can aim it when mounted on top of a

50ft pole. A climbable tower would be no problem, but not a pole. Incidentally, we have a local CATV cooperative that has a 2.4GHz 24dBi dish on top of what I guess is a 60ft pipe. It's fun to watch it swing around in the wind. It works only because they have 40dB of fade margin.

I'm not sure what to suggest. If you can borrow some directional antennas and actually try the path, it might give you a clue if you're close. I use a 25ft fiberglass window washing pole for "site surveys" which really is the same "try it first" type of test. Throw something together temporarily and see if there's hope. If it looks good, build something permanent. If not, try something different.

Incidentally, a big problem with high gain antennas is picking up interference from other systems along the line of sight. I don't mean just along the 1.3 mile path, but also well beyond the intended target. The high gain antennas will make any interference much worse. They do eliminate interference sources to the sides of the antenna pattern, but greatly increase interference inside the gain pattern. When you're testing for a useable path, also use a sniffer (Kismet) to test for other Wi-Fi systems along the path.

Reply to
Jeff Liebermann

Jeff,

Thanks for the info.

Would setting up for horizontal polarization with the big 23db parabolic dish (presuming we can aim them properly) help at all with the long-distance interference?

Reply to
SM

That's been the topic of many a technical arguement. In theory, most of the access points are vertically polarized. So, running your link with horizontal polarization should see a considerable reduction in interference. Unfortunately, reality doesn't follow theory. The typical laptop is actually horizontally polarized. One a signal bounces around a room, it can almost any polarization. Long ago, I did a study which measure the polarization of 2.4GHz signals in varying conditions. Close in, the source polarization was maintained. Everything else was almost random. At this point, I would say that it doesn't really matter, but it's probably a good idea to do it anyway.

Incidentally, one of the benifits of a helical antenna is that it is insensitive to linear polarization (vert or horiz) with only a 3dB loss in gain. That's handy when you're trying to track a moving vehicle or deal with a reflection infested environment. However, it's useless when you want to use polarization to minimize interference pickup.

Incidentally, if you need some hint as to what it's like to aim a dish, just grab a DBS satellite pizza dish and try to aim it at the satellite. It doesn't need to be activated for this exercise. The birds are 9 degrees apart in the sky and you have to be within about

+/- 2 degrees in order for it to work. It's fairly easy with an inline signal strength meter ($10 almost anywhere). Once you get a feel for what I mean as critical, then put the antenna on the end of a 10ft pole and try again to aim it. You'll find that it's not very easy. If you have a 25ft fiberglass collapsable window washer pole, try it again with a longer pole. The 24dBi dish is considerably heavier and more cantilevered, but about half as critical as the pizza dish. Still, it's quite a challenge to aim.
Reply to
Jeff Liebermann

Jeff;

Thank you for your input.

As a result of all the great info being received, I now have the following as a design spec:

TX power = +20dBm TX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) TX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) Distance = 1.31 miles RX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) RX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) RX sens = -82dBm (at 11Mbits/sec CCK) Theoretical Fade margin = 31.5dBm

This weekend I am undertaking to do a detailed site-survey which will set out in cross-section the elevation, buildings and trees along the path between point "A" and point "B". The results could change the above setup.

for the two points. Please note I am now using the term "tower" - not "mast" - as I have taken your admonitions about aiming parabolic dishes at the top of masts to heart.

I'm guessing I need 50' towers. We shall see.

Reply to
SM

Jeff;

Thank you for your input.

As a result of all the great info being received, I now have the following as a design spec:

TX power = +20dBm TX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) TX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) Distance = 1.31 miles RX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) RX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) RX sens = -82dBm (at 11Mbits/sec CCK) Theoretical Fade margin = 31.5dBm

This weekend I am undertaking to do a detailed site-survey which will set out in cross-section the elevation, buildings and trees along the path between point "A" and point "B". The results could change the above setup.

for the two points. Please note I am now using the term "tower" - not "mast" - as I have taken your admonitions about aiming parabolic dishes at the top of masts to heart.

I'm guessing I need 50' towers. We shall see.

Reply to
SM

"SM" hath wroth:

I suggest you not use 802.11b and use 802.11g 12Mbits/sec mode at about -84dBm sensitivity.

Reply to
Jeff Liebermann

Jeff;

Thank you for your input.

As a result of all the great info being received, I now have the following as a design spec:

TX power = +20dBm TX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) TX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) Distance = 1.31 miles RX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) RX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) RX sens = -82dBm (at 11Mbits/sec CCK) Theoretical Fade margin = 31.5dBm

This weekend I am undertaking to do a detailed site-survey which will set out in cross-section the elevation, buildings and trees along the path between point "A" and point "B". The results could change the above setup.

for the two points. Please note I am now using the term "tower" - not "mast" - as I have taken your admonitions about aiming parabolic dishes at the top of masts to heart.

I'm guessing I need 50' towers. We shall see.

Reply to
SM

Jeff;

Thank you for your input.

As a result of all the great info being received, I now have the following as a design spec:

TX power = +20dBm TX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) TX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) Distance = 1.31 miles RX ant gain = +23dBi (Wade D2423 grid parabolic - horizontally polarized) RX coax loss = -5dB (70' LMR400 plus 4 connectors and 1 lightning arrestor) RX sens = -82dBm (at 11Mbits/sec CCK) Theoretical Fade margin = 31.5dBm

This weekend I am undertaking to do a detailed site-survey which will set out in cross-section the elevation, buildings and trees along the path between point "A" and point "B". The results could change the above setup.

for the two points. Please note I am now using the term "tower" - not "mast" - as I have taken your admonitions about aiming parabolic dishes at the top of masts to heart.

I'm guessing I need 50' towers. We shall see.

Reply to
SM

this thread has been very interesting reading... what is a Fresnel Zone relating to RF - I've only become aware of it with optics & lighthouses

tnx - WA9TKA -

Reply to
P.Schuman

Fresnel lens...same guy.

Reply to
decaturtxcowboy

"SM" wrote in news:1166196482.483531.196020 @f1g2000cwa.googlegroups.com:

SM,

It seems as though you are really into this, and like learning new things. If so, there is a free RF propagation tool available for d/l, it's called RadioMobile. It's not the greatest tool, as it's free, but gives out OK data.

formatting link
It's a little hard to learn how to use, because it's written by a Ham :), but there are tutorials at that page also. Even if it isn't totally accurate, it will give you a general idea of the numbers. It will also show you a profile view of your link as well, showing the fresnel zone clearance and requirements of said link.

Or, if you don't want to learn that, you can just supply the lat/longs for both locations and I can prop it for you. Just post here or you can e-mail me, removing all .'s except for one of the double's at the end.

Regards,

DanS

Reply to
DanS

It's just edge diffraction. In optics, it's usually an aperature calculation. At lower frequencies than light, the aperature is between the ground and buildings on 3 sides. Same effect.

The edge diffracted part of the signal mixes with the signal along the line of sight and cancels at specific radii forming "rings" of varying signal strength. The diffracted part of the signal has a longer path length than the line of sight part. If the difference in phase between the line of sight part, and the diffracted part is 180 degrees, you get a null, thus forming the "rings".

formatting link

Reply to
Jeff Liebermann

tnx for the info & links.... BTW - here's a pretty simple image from a link on wiki -

formatting link

Reply to
Phil Schuman

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.