Long range (47-55 KM) point-to-point Wi-Fi link (depending on the chosen site)

[Halcyon]

I'm in the process of designing a point-to-point Wi-Fi link between two sites using Ubiquiti gear. My budget is limited so unfortunately I can shell out on the more expensive UBNT airFiber gear, so I'm limited to basically using the UBNT Rocket Prism AC radios. I'm also limited to using either the 2.4 GHz or 5 GHz unlicensed bands. Ideally I'd like to use the 5 GHz band, but I'm concerned about the trees and possibly rain fade over that distance. With 2.4 GHz, I'm concerned about congestion on the normal Wi-Fi channels.

The distance of the link to be established is 55 KM which is well within the specs and capability of the radio and dish. According to the UBNT wireless link simulator, the two sites are in line of sight of each other although there are some tree tops obscuring the view close to one of the stations but with a 30 dBi dish, the link should work.

Before buying this gear, I'd like to hear the opinions of those who have more experience with RF than I do or perhaps if you've set up a long-range link before.

The specs of the radio and the dish I was looking at are attached down below.

EDIT: Since this initial post, I've collected a lot of data and advice from various sources and have moved onto initial testing. I'll post results and actual data in blue so if you want to skip through the messages quickly for your own projects, meaningful posts should be easy to locate.

[grumpydoc]

How high are the antenna going to be?

i presume you have run the link end points into a propagation simulator to ensure you really do have line of sight - 55km is far enough to have to take curvature into account.

PS no direct experience of such links - just a few minutes thought experiment.

[Halcyon]

How high are the antenna going to be?

The figures I punched in the link simulator were 5 metres off the ground (which is conservative), they will be mounted on roof tops. Even with this, the fresnel zone clears the terrain no problems. The attached image are the results the link simulator spits out.

My concern with going 5 GHz is as I mentioned, the tree tops and to a lesser degree, rain fade. On the other hand going 2.4 GHz band would probably mean issues with congestion (would the beam be narrow enough to reject most of it?)

[Jeroen3]

On flat terrain you need to multiply the 5m by 10. But it looks like you're not around flat terrain.
Note that Mikrotik also sells long distance links, but 50km looks beyond their capabilities.

[HighVoltage]

I never designed a long distance antenna for WiFi but I was under the impression that both antennas have to be in sight with each other for best communication, right?
At 55km you have a drop in earth curvature by 138 m, if your antenna is 5 m above ground
https://dizzib.github.io/earth/curve-calc/?d0=50&h0=5&unit=metric

So, does this mean that you would need a 138m high antenna in a perfect world?

[Halcyon]

I never designed a long distance antenna for WiFi but I was under the impression that both antennas have to be in sight with each other for best communication, right?
At 55km you have a drop in earth curvature by 138 m, if your antenna is 5 m above ground
https://dizzib.github.io/earth/curve-calc/?d0=50&h0=5&unit=metric

So, does this mean that you would need a 138m high antenna in a perfect world?

Probably on flat land. But one station is on the top of a mountain range (some 500+ meters higher than the other) so it's not an issue in this case.

[Brumby]

I never designed a long distance antenna for WiFi but I was under the impression that both antennas have to be in sight with each other for best communication, right?
At 55km you have a drop in earth curvature by 138 m, if your antenna is 5 m above ground
https://dizzib.github.io/earth/curve-calc/?d0=50&h0=5&unit=metric

So, does this mean that you would need a 138m high antenna in a perfect world?

Probably on flat land. But one station is on the top of a mountain range (some 500+ meters higher than the other) so it's not an issue in this case.

Snap.

This was fairly apparent from the image supplied.

[Psi]

But will it still hear the 5.8ghz signal 55km away when you have Johnny flying his FPV plane in the park next door with a 1.5W 5.8Ghz video link?   

[Halcyon]

But will it still hear the 5.8ghz signal 55km away when you have Johnny flying his FPV plane in the park next door with a 1.5W 5.8Ghz video link?   

Johnny will get a rude shock since he'd be flying near a flight path for Police and Medical helicopters.

[Someone]

I'm also limited to using either the 2.4 GHz or 5 GHz unlicensed bands....

The distance of the link to be established is 55 KM which is well within the specs and capability of the radio and dish.

Given Australian power limits in the ISM band are EIRP and you've raised the concern about interference its a non-starter. You'll need a license for this sort of install just on the power requirement alone.
http://www.legislation.gov.au/Details/F2011C00543
https://www.acma.gov.au/theACMA/wireless-lans-in-the-24-ghz-band-faqs

[Halcyon]

Given Australian power limits in the ISM band are EIRP and you've raised the concern about interference its a non-starter. You'll need a license for this sort of install just on the power requirement alone.
http://www.legislation.gov.au/Details/F2011C00543
https://www.acma.gov.au/theACMA/wireless-lans-in-the-24-ghz-band-faqs

This type of install is covered under a class licence, no different to operating your own Wi-Fi access point at home. The Ubiquiti gear takes that into consideration when you set the country, antenna gain and cable loss values.

On some channels, you can legally go as high as 4 watts EIRP.

[borjam]

In order to know wether the link works or not I would recommend to use an application that takes into account an actual terrain profile.

On iOS I use one called RF Haversine. It's cheap and, as far as I know, quite effective although the user interface is a bit clumsy.

You input the coordinates, antenna heigth, frequency, TX EIRP, antenna gains, etc. and you tell the application to retrieve a terrain profile.

This is the result.

[Someone]

Given Australian power limits in the ISM band are EIRP and you've raised the concern about interference its a non-starter. You'll need a license for this sort of install just on the power requirement alone.
http://www.legislation.gov.au/Details/F2011C00543
https://www.acma.gov.au/theACMA/wireless-lans-in-the-24-ghz-band-faqs

This type of install is covered under a class licence, no different to operating your own Wi-Fi access point at home. The Ubiquiti gear takes that into consideration when you set the country, antenna gain and cable loss values.

On some channels, you can legally go as high as 4 watts EIRP.

Except you didn't put those requirements into your link model, and posted the image of it here showing well above the limits. The distance you want is right on the limits of whats possible within the power limits and would likely be unreliable. There are many other requirements beyond the power limits which you haven't mentioned any details into, you need to chat with ACMA and see what they think of the plan.

[Halcyon]

In order to know wether the link works or not I would recommend to use an application that takes into account an actual terrain profile.

As far as I know the Ubiquiti link simulator does all that, it takes into account topographical maps and lets you adjust antenna type, gain, height etc...

[borjam]

In order to know wether the link works or not I would recommend to use an application that takes into account an actual terrain profile.

As far as I know the Ubiquiti link simulator does all that, it takes into account topographical maps and lets you adjust antenna type, gain, height etc...

OOps! I stand corrected then!   

So much for being a Ubiquiti user (albeit using only Unifi stuff right now).

[Halcyon]

Except you didn't put those requirements into your link model, and posted the image of it here showing well above the limits. The distance you want is right on the limits of whats possible within the power limits and would likely be unreliable. There are many other requirements beyond the power limits which you haven't mentioned any details into, you need to chat with ACMA and see what they think of the plan.

You are correct. I've adjusted the figures based on a maximum EIRP allowed on 5 GHz. Similar result (slightly lower link speed). Still seems to work out.

The attachment below shows the link with a maximum EIRP of 36 dBm, before it was set to 46. No need to bother with ACMA (or as Dave would put it: Winner winner, chicken dinner!)

[BrianHG]

LOL, Get a larger dish with a focal beam so tight that it will be impossible to target each antenna reliably at 55km.  But for the instant you hit the target, you'll have a signal with enough gain which will ignore any surrounding interference.

[cdev]

There is a program called "splat" that might be helpful for figuring out coverage area and if a given height antenna is high enough. For any real distance the curvature of the earth requires that the antennas one or the other or both - get up high. For that far of a distance you likely will need some real height. Not a super lot but if you are hand held you're limited to only a very few miles because of the shape of the earth.

[The Soulman]

There is a program called "splat" that might be helpful for figuring out coverage area and if a given height antenna is high enough. For any real distance the curvature of the earth requires that the antennas one or the other or both - get up high. For that far of a distance you likely will need some real height. Not a super lot but if you are hand held you're limited to only a very few miles because of the shape of the earth.

I never designed a long distance antenna for WiFi but I was under the impression that both antennas have to be in sight with each other for best communication, right?
At 55km you have a drop in earth curvature by 138 m, if your antenna is 5 m above ground
https://dizzib.github.io/earth/curve-calc/?d0=50&h0=5&unit=metric

So, does this mean that you would need a 138m high antenna in a perfect world?

Probably on flat land. But one station is on the top of a mountain range (some 500+ meters higher than the other) so it's not an issue in this case.

Snap.

This was fairly apparent from the image supplied.

[stj]

get a couple of old C-band sat dishes - replace the lnb/polariser with your wifi antenna but keep the feedhorn.

[BrianHG]

get a couple of old C-band sat dishes - replace the lnb/polariser with your wifi antenna but keep the feedhorn.

It's what I was thinking.  It should work great for the 5Ghz band.  A quality 6 foot dish would make a nice tight beam.
I'm not sure about the 2.5Ghz, maybe you will want a 10 foot dish to get a good gain out of that one.
However, still, point to point, aligning the 2 dishes would be near impossible unless you deliberately adjust them slightly out of focus.
The feed-horn usually has a 'Z' of height alignment.
Some feed horns have a polarizer in them, I wonder how you would go about dealing with it...

[Brumby]

There is a program called "splat" that might be helpful for figuring out coverage area and if a given height antenna is high enough. For any real distance the curvature of the earth requires that the antennas one or the other or both - get up high. For that far of a distance you likely will need some real height. Not a super lot but if you are hand held you're limited to only a very few miles because of the shape of the earth.

500m elevation not enough?

[cdev]

That is definitely enough for you.

But aiming your antenna may be fairly difficult because the beam width you're going to need is narrow. You should definitely use the highest gain antenna you can find.

Properly "illuminating" a parabola, and issues relating to a dish "feed" are important.

Basically a feed is a small gain antenna at the focus of the dish that concentrates the signal at the surface you're reflecting off of.

One good example of a popular dish feed is a biquad antenna. But it needs to be just the right gain to work well, not too much gain or only the center of the reflecting surface will be illuminated and the rest of it will be wasted. A biquad with a square C shaped metal reflector small flat reflector with the top and bottom of the c part fairly short seems to work well..
You can extend the biquad in both directions. That is also popular as a feed.

Also, in Australia there used to be a satellite provider I think called Primestar? If you do a search on primestar wifi you'll likely find articles about converting the primstar antennas for wifi.

A number of years ago there was a web page about people in New Zealand (I think "orcon" was in their URL?) using woks and various informal paraboloid and similar cookware to set up medium range wifi links (11 km was typical) using USB wifi dongles.

The fact that it worked at all is a tribute to the effectiveness of the shape, but a USB wifi dongle usually has an omnidirectional chip antenna, not a good dish feed at all. But, nonetheless it worked over surprisingly long distances.

Search for "wok fi" and then plug the URL you get into archive.org's wayback machine.

If you do your research  and you'll likely succeed if you have 500 meter elevation and combine it with a good antenna.

[Halcyon]

But aiming your antenna may be fairly difficult because the beam width you're going to need is narrow. You should definitely use the highest gain antenna you can find.

I have the options of a 24dBi on the 2.4 GHz band or 30dBi and 34dBi dishes on the 5 GHz band. I'm just wondering if the beam would be narrow enough on the 2.4 GHz to reject a lot of the noise and interference? Anyone know how to calculate the beam width over that kind of distance?

Datasheets located here: https://www.ubnt.com/downloads/datasheets/rocketdish/rd_ds_web.pdf

[cdev]

Quote from: Halcyon on Today at 19:23:43

They should tell you in the specifications. the (typically specified as x db down) beam width is the attribute to look for. Also, gain is inversely related to thebeam width.

gain is typically referenced against a theoretical nonexistent "isotropic" antenna that sends energy equally in all directions, or a dipole.

there are very real and sometimes quite substantial differences in how an antenna receives signals from its sides or back. For example, a Yagi, log periodic and paraboloid will all have high gain but their pattersns can be quite different. All kinds of antennas are quite complex mathematically with few exceptions. The most predictable antennas are ones where the pattern is not dependent on frequency, but on shape.  Such antennas will present a similar gain profile and impednce at a wide range of frequencies. Also, polarization is important. Most antennas radiate a signal that is linearly polarized like w ave,  so in a anechoic chamber or in space, two antennas whose polarization is perpendicular to one another cancel out and there might be virtually no signal if you are trying to receive a vertically polarized signal with a horizontally olarized antenna. In practice, propagation conditions and reflections make signals a combination of polarizations once its been reflected off of some objects. But still the effect is quite marked with lne of sight communications. care should be taken that both antennas on both sides of the link radiate the same polarization. This also applies to circular polarization but there the vertical and horizontal compnents of the signal are 90 degrees out of phase and so can be described as left or right handed circular polarization.
The opposite polarization cancels out which with circular polarization can be exploited to cancel out even numbered reflections. "multipath".

For example, with GPS, which use right handed circular polarization.
>

[cdev]

Thats a "backfire" antenna.

https://en.wikipedia.org/wiki/Short_backfire_antenna

They are commonly used for backhaul links over long distances. The high gain figures cited sound like exaggerations for that kind of antenna, but then again it may be something else. What do you think?

I think STJ and BrianHG's suggestion to repurpose satellite dishes with a remade feed are likely to work better over a 50 km link which is much more than the usual backhaul link.

There are are some classic designs using old "primestar" dishes.

example:

http://www.trevormarshall.com/biquad.htm

http://trwn.gr/downloads/biquad_espec.pdf

http://martybugs.net/wireless/biquad/

Note that Brian HG is right about needing a larger dish (twice as large) for 2.4 GHz. Wind loading will be a big factor too. Wind loading with a dish twice the diameter will be much more than twice as much.

keep that in mind in areas where storms are an issue..

[Halcyon]

They should tell you in the specifications. the (typically specified as x db down) beam width is the attribute to look for. Also, gain is inversely related to thebeam width.

These are the specs for the 2.4 GHz antenna, which probably makes more sense to you than me. I'm not an RF guru.

Code: [Select]

Dimensions          650 x 650 x 295 mm
Frequency Range     2.3 - 2.7 GHz
Gain                24 dBi
HPOL Beamwidth      6.6° (3 dB)
VPOL Beamwidth      6.8° (3 dB)
F/B Ratio           28 dB
Max. VSWR           1.6:1
Polarisation        Dual-Linear
Cross-pol Isolation 35 dB Min.


[BrianHG]

A 6 foot C band dish will give you a gain of around 36db for 5Ghz.  (You have a chance of achieving full speed with this one with proper setup at 55km)
A 10 foot C Band dish will be in the 40db range.  (Expect full speed with this one at 55km, even with the 2.5ghz band)

Since the fall of C band dishes for consumers, you might find some cheap used ones as long as they are not warped, of designed by with a material which wont warp.  I suspect shoving a USB Wifi dongle in the cone area of the 10 foot dish, just because of the dish's mass size, would be functional, though not as fast or good as a WiFi transceiver designed to be fed though such an antenna.

Example 6 foot dish:
http://www.galaxy-marketing.com/WS1870P-6-foot-180cm-C-Band-Prime-Focus-Satellite-Dish-with-Polar-Mount.htm

[Halcyon]

A 6 foot C band dish will give you a gain of around 36db for 5Ghz.  (You have a chance of achieving full speed with this one with proper setup at 55km)
A 10 foot C Band dish will be in the 40db range.  (Expect full speed with this one at 55km, even with the 2.5ghz band)

Since the fall of C band dishes for consumers, you might find some cheap used ones as long as they are not warped, of designed by with a material which wont warp.  I suspect shoving a USB Wifi dongle in the cone area of the 10 foot dish, just because of the dish's mass size, would be functional, though not as fast or good as a WiFi transceiver designed to be fed though such an antenna.

You're quite right, however I can buy the UBNT 34 dBi (@ 5-5.8 GHz) dish for AUD$412.50 brand new. It's almost half the size.

My only concern with going 5 GHz is how well (if at all) will it penetrate through the tops of trees (just the top leafy parts, not the thick wood parts). There is about 40-50 metres worth of tree canopy within 20 metres of one side of the link. I just can't get quite high enough to clear them completely.

[BrianHG]

A 6 foot C band dish will give you a gain of around 36db for 5Ghz.  (You have a chance of achieving full speed with this one with proper setup at 55km)
A 10 foot C Band dish will be in the 40db range.  (Expect full speed with this one at 55km, even with the 2.5ghz band)

Since the fall of C band dishes for consumers, you might find some cheap used ones as long as they are not warped, of designed by with a material which wont warp.  I suspect shoving a USB Wifi dongle in the cone area of the 10 foot dish, just because of the dish's mass size, would be functional, though not as fast or good as a WiFi transceiver designed to be fed though such an antenna.

You're quite right, however I can buy the UBNT 34 dBi (@ 5-5.8 GHz) dish for AUD$412.50 brand new. It's almost half the size.

Yup, it's a whole lot smaller and at 3.6 feet and at 5GHz, it's rated closer to 31db.  The C band dish it rated 35.9db minimum at 4GHz., around 37-38db at 5.5Ghz.  I guess size is everything here.  Don't bother with the C band dish since you have the budget for the UBNT.  Note that that dish at 2.5Ghz is nowhere near 31db.  The only reason for using a C band dish would be if you are going through the hell of installing 2x 10 foot dishes to get that 35-40db gain down in the 2.5GHz to make it through those tree leaves.

I know from experience with a C band dish that in the summer, some satellite signals would weaken since my dish pointed through some trees very low to the ground pointing most westward to satellite F4, but, in the winter when the leaves fell off, the picture was clean.

One thing which might be out of your control, in the 5Ghz band, is microwave interference.  Once of my early installations in the city was near an airport and there was a zone where I pointed my dish which always had a a ton of unavoidable noise.  This is something which might be out of your control when it comes to using WiFi over such a long distance even with such focused antennas.

[Brumby]

I know from experience with a C band dish that in the summer, some satellite signals would weaken since my dish pointed through some trees very low to the ground pointing most westward to satellite F4, but, in the winter when the leaves fell off, the picture was clean.

I would pretty much say that the leaf problem would be consistent all year round.  There is a greater likelihood the vegetation will be evergreen, not deciduous.  There will, however, be an increased attenuation problem with rain, especially with raindrops on the leaves.

I can see this is very likely Halcyon's key concern.

One thing which might be out of your control, in the 5Ghz band, is microwave interference.  Once of my early installations in the city was near an airport and there was a zone where I pointed my dish which always had a a ton of unavoidable noise.  This is something which might be out of your control when it comes to using WiFi over such a long distance even with such focused antennas.

While I don't know where the link is to be situated, I am guessing that it doesn't have this as a significant problem.  There is a military air base to the north, but I would imagine it will be far enough off axis to not be a significant issue.  Sydney airport might be closer to being on-axis - but it is further away.  The new airport will be to the south - but it hasn't even had one sod of soil turned yet.

Then again - I could be completely off target.  Halcyon has been vague about location - probably for good reason.

[BrianHG]

I know from experience with a C band dish that in the summer, some satellite signals would weaken since my dish pointed through some trees very low to the ground pointing most westward to satellite F4, but, in the winter when the leaves fell off, the picture was clean.

I would pretty much say that the leaf problem would be consistent all year round.  There is a greater likelihood the vegetation will be evergreen, not deciduous.  There will, however, be an increased attenuation problem with rain, especially with raindrops on the leaves.

The trees were maple leafs.  Just as the leaves were gone by December, my signal on F4 vastly improved and I finally would be able to receive F1 (I think, it's been 20 years and I don't remember the satellite names from  back then...)  However, you are right in saying that Halcyon's situation may be different.

Weather-wise, rain won't be too bad at 5GHz, it's much worse for KU Band/15GHz, but, as you said, beaming 5GHz though water drenched leaves, basically much larger water drops clumped in a stationary location, will attenuate down 5GHz as well as 2.5GHz.

[stj]

if the dish is mesh, make sure it's intended for c or s band btw
the size of the holes is chosen so as not to let the correct waves pass through.
most big dishes are mesh to reduce wind effects.
also know that segmented dishes have less gain than single-piece "rolled" ones.

[Halcyon]

I would pretty much say that the leaf problem would be consistent all year round.  There is a greater likelihood the vegetation will be evergreen, not deciduous.  There will, however, be an increased attenuation problem with rain, especially with raindrops on the leaves.

I can see this is very likely Halcyon's key concern.

Correct, they are mostly gum trees so leaves are present all year. I can tolerate some unreliability in significant rain (it's not going to be a "mission critical" link) but for the most part when the weather is fine, it needs to be reliable. I mean we are talking a 55 KM link using well under $1000 worth of equipment at each end, I'm not expecting carrier-grade back haul type stuff, but I at least want to increase my chances of it working (at all) before investing money either with 2.4 GHz or 5 GHz. As I said, my only real issue with 5 GHz is physical obstruction because of the tree tops, maybe it'll be a problem, maybe not? Other than that, I have direct line of site to the other station (and I can see the lights at night in between the trees).

While I don't know where the link is to be situated, I am guessing that it doesn't have this as a significant problem.  There is a military air base to the north, but I would imagine it will be far enough off axis to not be a significant issue.  Sydney airport might be closer to being on-axis - but it is further away.  The new airport will be to the south - but it hasn't even had one sod of soil turned yet.

Richmond and Sydney airports pose no problems. Even BOMs weather radars are far enough out of band not to cause issues. There is a radar at the Air Services Site at Cecil Hills which might pose a problem (you're saying that they use 5 GHz? I don't know much about air traffic radar systems). Badgery's Creek (when it's operational) might pose an issue but it's far enough a way not to be a problem in the foreseeable future and by then, the gear will probably be upgraded anyway.

[cdev]

What would help would be a diversity link which means using several antennas at the receiver and circuitry to select among the signal at several antennas- for the best one- in real time. Many access points - of course have multiple antennas. So, you're all set there, however to maximize the value of multiple antennas you should try to separate them by some distance, without too much attenuation. (Use a top quality coax like LMR400)

that way any trees or whatever in between, even if one path was blocked (say from tree branches swaying in the wind) another one might be open.

But it would be best if there weren't any trees in between the two. The presence of trees really changes everything. From a good chance of success to a good chance of failure.

[BrianHG]

On the subject of trees, you could always do my final solution.  I little logging, or at least trimming on the worst case culprits.
If your tree problem exceeds the number of which you can manicure, that's way too much foliage and forget about getting the 2.5Ghz working let alone the 5GHz band....

[Brumby]

With one end of the link being up in the Blue Mountains, I highly doubt logging or "arboreal manicuring" are going to be possible.

Even in residential areas, trees are not allowed to be chopped down or even heavily trimmed without good reason and council approval.  Once you are outside a property boundary, you are likely to be facing Crown land which will be natural bush - and you aren't going to get anyone giving you an OK for any sort of trimming.  Do something 'on the sly' and expect there to be someone official checking it out - and if there is a communications dish with a swathe cut out in front of it ... you're pretty much sunk.

While the eastern side of the Blue Mountains does fall away fairly quickly, there is still going to be some amount of tree canopy to deal with, unless you have a site on the edge of a cliff - which obviously isn't the case.  I will state - with little risk of being corrected - that this canopy is not going to be accessible for "trimming" because of legal impediments and, quite likely, because of physical access issues.

The project is going to have to be handled with the canopy "as is".

[stj]

then the dish on the high point will need to be on a tower.

[Brumby]

Again, council regulations may come into play.  Height restrictions and approvals for building a structure are just two immediate thoughts.  There is also the cost factor.  That alone might put such a solution well outside the budget.

[Halcyon]

Again, council regulations may come into play.  Height restrictions and approvals for building a structure are just two immediate thoughts.  There is also the cost factor.  That alone might put such a solution well outside the budget.

Correct. This is more for fun and games than anything serious. (Unless Dave wants to do a short episode about it?)

I know it's a really hard question to answer because of so many variables at play, but I guess what I'm after is a go/no-go on 2.4 GHz or 5 GHz options. Worst case, I can sell the equipment with minimal loss and chop it up as another experiment that didn't quite work.

What would you guys do in my situation?

[Brumby]

Ask the manufacturer/supplier if they have any further material on the vegetation subject - or if they know of any material produced by third parties - such as Youtubers.

Ask them if they have trial equipment.  (I know - a long shot, but it can't hurt to ask, can it?)

Get both and try them out.  Keep the one that performs best and sell the other.

Bit the bullet and just choose one.  If you find it difficult to choose which one you should buy, try looking at the reasons for working out which one you shouldn't buy.

Win Lotto and do whatever you like.

[Halcyon]

I think I'll just go for it. Will update this thread once I do.

[radar_macgyver]

55 km might be ambitious to do in a single hop; consider multiple hops.

My work's field site uses a C-band radio link, but only for a short hop (about 6 km) with 18 dBi dishes. At one point, I noticed the radio's SNR plots slowly drop over the course of the year. Turned out in the springtime, some new trees in the way had started putting out foliage, and that was enough to reduce the SNR.

Check what band your nearby weather radar operates at. If it's 5.6 GHz, your radio will periodically enter 'DFS' mode and try to retune itself to avoid interfering with the radar. You generally won't see interference from S-band radar at 2.4 GHz.

[Brumby]

55 km might be ambitious to do in a single hop; consider multiple hops.

Ambitious is not as big of a problem as multiple hops would be.

I am seeing this as a personal project where the two endpoints are at locations which allow for an antenna installation and where Halcyon has some rights or permission to use.  Adding one mid point will double the cost of the basic equipment, plus it will need a third site where affordable access and the appropriate permissions can be obtained.

But more important is the geography involved.  With one site in the Blue Mountains "near the edge" and a second down in the Sydney basin, any third point would have to be in the Sydney basin as well.

Assuming the mid point can cater for the foliage attenuation, there are no significant locations in the Sydney basin where you are going to get any elevation advantage between two points - and those that do represent more favourable locations are likely to have commercial communications towers in place anyway.

While a great idea to address this sort of issue, I just cannot see multi-hop being practical nor affordable.  I can see bigger dishes being a more appealing option.


I'm no RF nut - but here's a thought.....

What if you were to go for the 5GHz solution, but using a dish twice the size?  You would still end up with a dish that would be much the same size as for 2.4GHz, but with a wider beam, it may have more opportunities to find 'gaps' through a leafy canopy.

Opinions anyone?

[BrianHG]

The vast signal gain of the larger dish would be the overriding improvement factor.  I guess beam diameter would increase as well.
If the OP gets a functional link with the UBNT dish, but not a good enough sustained bit rate, he can always cheat a little by cranking up the output transmit power.  By the look of his location and beam direction, I don't anyone would notice.
Just make sure he is using a UBNT P-P dishes (Point to Point).

Also, if he is crafty enough, he can also enlarge the dishes himself with any sheet metal if he can create the proper curve.
He might only need to increase the dish size on one side.  There is no law that says he would need to improve both dishes or that both dishes need to be identical in size and gain...

[Brumby]

He might only need to increase the dish size on one side.  There is no law that says he would need to improve both dishes or that both dishes need to be identical in size and gain...

That sounds like a plan.  I'd suggest this for the mountain top dish where the problem foliage could be helpful in helping it blend in - or at least not be as obvious as one might be down in suburban Sydney.

[Halcyon]

55 km might be ambitious to do in a single hop; consider multiple hops.

55 KM is well within the specs for the RP-5AC-Gen2 radio (Ubiquiti claim up to 200+ KM but that's obviously with the highest gain antenna, perfect height/terrain, low noise levels, correct tongue angle etc...). Multiple hops isn't possible.

I'll be doing some initial tests over the weekend with telescopes and light sources to see if we do indeed have line of sight. There is an opening of a few degrees where trees are sparse and we might just be able to get over the top of them (assuming that's the right direction). I might even do a really crude test tonight with a high power LED torch and some binoculars. It's quite a clear night.

What if you were to go for the 5GHz solution, but using a dish twice the size?  You would still end up with a dish that would be much the same size as for 2.4GHz, but with a wider beam, it may have more opportunities to find 'gaps' through a leafy canopy.

Opinions anyone?

That's kind of what I'm leaning towards, but not having enough RF-voodoo experience, I'm finding it very hard to make a call one way or another.

he can always cheat a little by cranking up the output transmit power. 

That would be highly illegal. I mean, I wouldn't dare to input lower antenna gain figures so that the radios increase their output in-turn making the EIRP higher than what's permissible by ACMA. No way in the world. What kind of despicable human being would do that?   

[Brumby]

... but not having enough RF-voodoo experience, I'm finding it very hard to make a call one way or another.

Please understand - my RF voodoo is limited to SWRing in 27MHz mobile antennas.  Everything else is just ideas....

[Halcyon]

... but not having enough RF-voodoo experience, I'm finding it very hard to make a call one way or another.

Please understand - my RF voodoo is limited to SWRing in 27MHz mobile antennas.  Everything else is just ideas....

Much appreciated though, you're one step ahead of me.

[stj]

where specifically would these be, could you put them on the wall or roof of a building, or even bolted to a tree??

another option is to use a tight beam and ram the power up to full 400-600mW
because if it's a well defined beam - it's not going to effect anybody any way.

[Halcyon]

where specifically would these be, could you put them on the wall or roof of a building, or even bolted to a tree??

They will be attached to buildings. Trees are no good as the slightest bit of wind would throw the beam out of alignment.

[cdev]

This is the New Zealand man's web page about using woks for long distance wifi links. This is from a decade ago, his web page is no longer online.

https://web.archive.org/web/20081002032952/http://www.usbwifi.orconhosting.net.nz/

An example of the surprising successes people had.

https://archive.li/G9Vm

$10 wok keeps TV station on air

[vk2amv]

It's a pity you are doing this project now and now in 3 months time.

I am decomissioning a 21KM Point-To-Point Wi-Fi link in January that has been up and running since October 2012.
It has been up an running from the edge of Parkes, out to the in-law's place in Gunningbland providing them with internet with pretty much 100% reliability aside from one hardware failure in that time.
The in-laws are selling up and moving out of their place in January so the link will be pulled down then.

It is using Ubiquiti Rocket M5's (802.11N Gear) with 30dBi Ubiquiti dish's (They are 65CM in diameter).

I normally run it with conservative settings, choosing pure link stability over speeds because it is only terminating to an ADSL connection in town, but on testing it has done a reliable 240Megabits link speed.
Was quite cool sending data at that speed over the 21KM wireless link when I set it up back in 2012.

If you were doing your project after the time I am decomissioning this link I would have been happy to loan you the Ubiquiti gear to have a play with so you don't have to be out of pocket to do testing.

Hell I would have come down your way and given you a hand as it would be a fun project doing a long haul link from the Blue Mountains. (I think that is where you said you were doing one end of the link from)

It can be a hell of a lot of fun.
Or annoying at times as in my case for the 21KM link I had a 50KM drive by road from end to end, so when I am trying to diagnose a problem and need to drive between sites back and forth, it sucks.

Oh well.
I have not had a look at your path maps yet, I miscalculated and my internet is shaped until Sunday night, but I will have a look on Monday or Tuesday and post back some thoughts.

Anyway best of luck with the project.

[Halcyon]

It's a pity you are doing this project now and now in 3 months time.

Nothing's really preventing us from waiting other than the desire to have a play sooner rather than later.

I am decomissioning a 21KM Point-To-Point Wi-Fi link in January that has been up and running since October 2012.
It has been up an running from the edge of Parkes, out to the in-law's place in Gunningbland providing them with internet with pretty much 100% reliability aside from one hardware failure in that time.
The in-laws are selling up and moving out of their place in January so the link will be pulled down then.

It is using Ubiquiti Rocket M5's (802.11N Gear) with 30dBi Ubiquiti dish's (They are 65CM in diameter).

I normally run it with conservative settings, choosing pure link stability over speeds because it is only terminating to an ADSL connection in town, but on testing it has done a reliable 240Megabits link speed.
Was quite cool sending data at that speed over the 21KM wireless link when I set it up back in 2012.

Yeah I bet! It's still cool in 2017. For consumers to have affordable access to this kind of gear was unheard of 10 years ago.

How did you go with aligning each end? I'd imagine where it was is quite flat so no obstacles to overcome?

Hell I would have come down your way and given you a hand as it would be a fun project doing a long haul link from the Blue Mountains. (I think that is where you said you were doing one end of the link from)

It can be a hell of a lot of fun.

Yep. Without giving too much detail away, one station will be up in the Blue Mountains (more or less along the ridge line facing east) and the other station is down in the South/South-West of Sydney.

If we can get this to work, the idea later down the track would be to extend the network into more of a mesh between friends and colleagues. Why? Because we're a bunch of nerds and we can. :-)

[IanB]

One thought I have about trees, and the impossibility of trimming them, is that trees grow. Treetops that are only slightly impinging this year could grow much further into the sight line with each following year.

[Halcyon]

One thought I have about trees, and the impossibility of trimming them, is that trees grow. Treetops that are only slightly impinging this year could grow much further into the sight line with each following year.

I did think about that. Most of the trees here are fully grown so they shouldn't get much larger. The occasional storm and strong winds generally take care of branches that tend to get too big.

[vk2amv]

Nothing's really preventing us from waiting other than the desire to have a play sooner rather than later.

Well if you can wait until January when I decomission this 21KM link and pull the gear down, the offer of a free loan of this gear is there and genuine.
I honestly don't have any future plans for the gear at this stage, so I am happy to let you have a play with it.

I am also happy to give you help for a weekend to try and get a link up and running and offer my experiance. (My 21KM link is not the only point-to-point link I have done, just the longest in a large number of links I have setup)
I would enjoy the project of trying for a 55KM link for a bit of fun.

Feel free to send me a private message and I will give you my direct email if you want.

If we can get this to work, the idea later down the track would be to extend the network into more of a mesh between friends and colleagues. Why? Because we're a bunch of nerds and we can. :-)

And you need no other reason than that.
It can be a hell of a lot of fun, and is well worth playing with.

[cdev]

In my experience trees need constant attention. They're always growing bigger,so if you don't want them where they're growing you have to trim them back. When they die, you have to remove them or they may fall down on somebody. Trees require maintenance. There is no way around it. So budget for it, or face the fact that you'll have to cut them down and be done with it. (which sucks because lets face it, trees are often beautiful and being alive they arguably have life and interests on some level. Cutting down a tree is killing a living thing. It never feels good. Animals live in trees. And so on.

If you can do it, figure out if your project might not need to cut down trees. Maybe find a path that does not impinge on trees using a very bright light like a laser.. Perhaps - get a much much better antenna or something, if you can push your signal through, go for it, but if trees are in the line of sight, you may want to try to find a better pair of locations. Is there any way you could set up a laser or very very bright LED and visually evaluate the path to see if the trees are blocking the line of sight. Surveying done any time ago, before GPSs got so good, is quite often wrong in the vertical direction, often by several meters or more. . Another alternative is RTK GPS. Even cheap GPS now are more accurate than many maps. You should go to both locations and get a GPS fix on both of them - after letting the GPS run for some time (the fix gets more accurate as you stay in one place for awhile. Elevation is particularly problematic for some GPSs. Its often wrong, unless you take care to apply all the necessary corrections. GPS fixes on altitude are often quite a bit less accurate than horizontally. So you have tobe willing to get more accuratye elevation data and teh way to do that is via GPS "precise polt positioning or real time kinematics.. RTK. See rtklib.com for an open source GPS technology that can give you very accurate data and not cost an arm and a leg.

Scan the academic literature. Maybe there is soome way you can push your signal through the trees using diversity or beam forming or mimo or something.

Math!


Quote from: Halcyon on Today at 21:15:27>Quote from: IanB on Today at 21:06:43

One thought I have about trees, and the impossibility of trimming them, is that trees grow. Treetops that are only slightly impinging this year could grow much further into the sight line with each following year.

I did think about that. Most of the trees here are fully grown so they shouldn't get much larger. The occasional storm and strong winds generally take care of branches that tend to get too big.

[Brumby]

In my experience trees need constant attention. They're always growing bigger,so if you don't want them where they're growing you have to trim them back. When they die, you have to remove them or they may fall down on somebody. Trees require maintenance. There is no way around it. So budget for it, or face the fact that you'll have to cut them down and be done with it. (which sucks because lets face it, trees are often beautiful and being alive they arguably have life and interests on some level. Cutting down a tree is killing a living thing. It never feels good. Animals live in trees. And so on.

Again, not knowing exactly where the Blue Mountains link will be placed, I cannot say with absolute certainty - but I can say that there is a very high likelihood that these trees will be in protected reserves.  No cutting or trimming is allowed - and often you are not even permitted to collect fallen branches as they, too, provide homes for creatures and eventually fertilise the forest floor.  If a branch happens to fall on someone, then bad luck.  The exception to this would be along recognised trails, where there is active maintenance to ensure the safety of those who walk them.

I feel relatively confident in saying this is very much a "hands off and deal with it" scenario.

[ivan747]

Rain fade vs frequency at several different types of rain:


http://www.e-band.com/get.php?i.119:w.698:h.478

I remember there are figures on fading by leaves and vegetation. I believe you just add a fixed figure to your margin, something like x dB per km. Try to google it.


There's also extra fading you have to consider in your link budget if you're going over a body of water, I believe it's actually scattering but I can't remember.

There are books on RF design, maybe you can pick up an old edition for cheap on eBay after sampling a new edition on Amazon.


Sent from my iPhone using Tapatalk

[Brumby]

I don't think bodies of water are going to be an issue.

The closest are the Nepean River and the artificial Penrith Lakes, but they are so close to the mountains that they are going to be well below any vertical beam spread.  The only other significant body is Prospect Reservoir - but it isn't that massive and it's a fair way from the Blue Mountains.

[cdev]

If you can find a place between the two terminuses of this link which has good visibility of both, you could perhaps (if you could get permission) mount a reflector there to complete the circuit so to speak.

Similar to how this is done on microwave towers. It could be something as simple as an old highway sign or similar. (a strong, flat piece of metal sufficiently large.) just mount it so securely its not going to move significantly in even the highest winds. Adjusting it properly might be difficult - thats where having he best possible elevation data comes in. (see below)

One thing that must be truly awesome is that Australia being south of Japan, also has coverage from the Japanese QZSS system and Beidou, the Chinese GPS system. Its my understanding that far more sats are visible in Asia for this reason. Just use a newer GPS that supports them.

I wrote a long post about RTKlib and hardware that supports it and setting up an RTK system but it may not be necessary IF you use a system that supports the newer sats available in your area. Thats important because the *vertical* accuracy of older GPSs often leaves much to be desired, that is to say its often way off. (typically the offset stays the same in a given area)  For some of my old GPS's the vertical direction is consistently overestimated. My real elevation as measured properly is substantially less than they tell me. This has to do with a bunch of factors, which you'll learn about if you explore RTKlib. (which is also a lot of fun learning)

Anyway, with that caveat getting an accurate position when you have a clear sky view is fairly easy and with the multi-constellation units is fairly rapid, even without RTK. especially if the unit has been running for awhile with its antenna having a good view of the sky (on top of the car that gets you there?) Make sure your fix quiets down after you turn the unit on have it still and running for awhile with a good sky view before taking your measurements.

So anyway, one of my GPS's that seems extremely accurate and supports GPS, QZSS and Beidou - my navspark mini, cost me less than $10 and gets Beidou sats (only one or two are ever in view here in he US) But those sats add to the quality of the fix significantly. Enough for the indoor movements of the GPS's antenna to seem accurate. (for example, moving the antenna from next to the window to computer desk, a distance of around two meters, was recorded with what appeared to be fidelity) This is without DGPS of any kind, inside of a two story wooden house with the usual levels of wiring and stuff in it. That's very impressive to me. So, you likely would get even more accuracy, being in Australia because of the way these newer GNSS satellites orbits are designed to keep them over one part of the world, as the earth turns beneath it. The QZSS system was designed to augment GPS for very densely populated Japan, and he fact hat Australia is directly south of Japan means it spends as much time over Australia as it does Japan. So you get that extra accuracy for free. Similarly with Beidou and China.

Record a log and record the time so later you can correlate the two and when you get home figure out what the exact heading and elevation the links between them would need to be.

Do this by making a coverage map in one of the aforementioned programs using the exact same scale (and if you ca perhaps also the same coordinates to define the image borders) and then overlay the two in an imaging program and "diff" them, start in the "difference" layer mode perhaps and the areas that are covered by both signals at the highest level should be visible.

Those places likely would be good spots for a reflector!

[Brumby]

Unless there is a clear view of the Sydney basin in some other direction, that's a lot of work ... to achieve nothing.  Even if there was (which, knowing the typical situation of residences in the Blue Mountains, is not likely) the additional distance the signal would need to traverse could possibly (likely) result in more losses than the wet tree leaves.

There is about 40-50 metres worth of tree canopy within 20 metres of one side of the link. I just can't get quite high enough to clear them completely.

So, about 70 metres from the mountain top location, there's clear line of sight.  The only place to put a reflector would be somewhere in that 70 metres, hovering above the tree tops.

To me, other than assessing the options, the best solution is going to get as high as possible at the mountain end.  You face issues with council regulations and cost - to get something high enough and rigid enough to do the job.

[Halcyon]

To me, other than assessing the options, the best solution is going to get as high as possible at the mountain end.  You face issues with council regulations and cost - to get something high enough and rigid enough to do the job.

I tend to agree. Adding reflectors, bouncing the beam off objects etc... is out of the question. If I had access to an intermediate site, I'd just throw a repeater there.

I think the answer here is to get slightly higher (which I can at one end). If the trees weren't there, I'd buy the gear tomorrow as there is direct line of sight.
This week, I'm going to try and borrow a telescope to visually confirm the path. There is a narrow path between trees which might just be enough.

[cdev]

A repeater would be much better than a reflector if its possible.



Quote from: Brumby on Today at 02:03:36

Unless there is a clear view of the Sydney basin in some other direction, that's a lot of work ... to achieve nothing.  Even if there was (which, knowing the typical situation of residences in the Blue Mountains, is not likely) the additional distance the signal would need to traverse could possibly (likely) result in more losses than the wet tree leaves.
>Quote from: Halcyon on 2017-09-14, 03:12:36

There is about 40-50 metres worth of tree canopy within 20 metres of one side of the link. I just can't get quite high enough to clear them completely.

So, about 70 metres from the mountain top location, there's clear line of sight.  The only place to put a reflector would be somewhere in that 70 metres, hovering above the tree tops.

To me, other than assessing the options, the best solution is going to get as high as possible at the mountain end.  You face issues with council regulations and cost - to get something high enough and rigid enough to do the job.

[Halcyon]

I had the pleasure of meeting forum member vk2amv this afternoon who dropped off some decommissioned Ubiquiti radios and dishes for me to play around with.

For anyone still following this thread, I'll be doing some testing with these over the next few weekends across the 55 km proposed link path and will post results, experiences and photos in this thread.

@vk2amv -- Drinks are on me when we next meet!

 

[Brumby]

Kudos to vk2amv !!      

This is going to be a big help!

[cdev]

How very very cool!


****GOOD LUCK!!!****

Will be following this-

Please take some pictures and screenshots and post them so we know what the equipment, terrain and link signal parameters in the real world testing look like!

[Halcyon]

I've been thinking about how to mount the dish (as I don't have a TV mast on my roof). I have a brick wall where I can attach a mast directly to several points up the wall and have it protrude past the roof line.

I'm not paying ridiculous amounts for a "proper" antenna mast, which is essentially just a metal pipe. I thought about using 900mm lengths of 20mm diameter galvanised steel pipe such as these: https://www.bunnings.com.au/kinetic-20-x-900mm-galvanised-pipe_p4700271 or alternatively https://www.bunnings.com.au/arc-2-4m-32nb-galvanised-steel-pipe_p1100586

It's cheap, I can built it as high as I need with screw-on joints and it should be sturdy enough to withstand strong with without flexing too much.

Does anyone have any thoughts?

[Brumby]

I'd be inclined to minimise the number of screw on joints.  If it can be done, a single length is going to be more rigid.  Mounting points will need to be rock solid, obviously - and I wonder about how much movement there will be on the dish under wind load.  The thought of guy wires to help stabilise the top of the mast did cross my mind - but I have no practical experience with these, so I don't know if they are a good idea or not.  (Points of attachment are the first question.)   ... but, then, this is going to depend on just how high the mast will reach.

[cdev]

You should figure out the wind loading with pencil and paper. Any kind of solid dish is likely to cause an extreme amout of wind loading in storms. Get a range of opinion. You want your antenna staying put in high winds and not becoming a projectile. Winds are getting stronger. Keep that in mind. Winds in 10 or 20 years time may be significantly stronger than today. Think ahead.

I've been thinking about how to mount the dish (as I don't have a TV mast on my roof). I have a brick wall where I can attach a mast directly to several points up the wall and have it protrude past the roof line.

I'm not paying ridiculous amounts for a "proper" antenna mast, which is essentially just a metal pipe. I thought about using 900mm lengths of 20mm diameter galvanised steel pipe such as these: https://www.bunnings.com.au/kinetic-20-x-900mm-galvanised-pipe_p4700271 or alternatively https://www.bunnings.com.au/arc-2-4m-32nb-galvanised-steel-pipe_p1100586

It's cheap, I can built it as high as I need with screw-on joints and it should be sturdy enough to withstand strong with without flexing too much.

Does anyone have any thoughts?

[Halcyon]

I'd be inclined to minimise the number of screw on joints.  If it can be done, a single length is going to be more rigid.  Mounting points will need to be rock solid, obviously - and I wonder about how much movement there will be on the dish under wind load.  The thought of guy wires to help stabilise the top of the mast did cross my mind - but I have no practical experience with these, so I don't know if they are a good idea or not.  (Points of attachment are the first question.)   ... but, then, this is going to depend on just how high the mast will reach.

I'm thinking of mounting about 2 metres of the pole at several points along the brick wall (with clamps similar to saddle clips, but stronger) then have another ~2 metres protruding up past the roof line. I don't think I'll need anything much higher than that.

You should figure out the wind loading with pencil and paper. Any kind of solid dish is likely to cause an extreme amout of wind loading in storms. Get a range of opinion. You want your antenna staying put in high winds and not becoming a projectile.

The dish, radio and supplied brackets are good for up to 200km/hr winds. Obviously the mast will need to withstand that as well. I'm looking at about 8.5 KG of weight on the mast.

[cdev]

Even a relatively short two or three meter mast is a LEVER with all that implies for strength of force. So, if you go up any distance, guying is absolutely essential unless you want to use a trussed steel tower.

With guying you can reach for the skies and have an antenna system thats unlikely to fall down in 95% of all storms. If you're lucky that other 5% misses you!

Don't even think about using a tall mast without it, it wont work, even with the strongest steel pipe.

I'd be inclined to minimise the number of screw on joints.  If it can be done, a single length is going to be more rigid.  Mounting points will need to be rock solid, obviously - and I wonder about how much movement there will be on the dish under wind load.  The thought of guy wires to help stabilise the top of the mast did cross my mind - but I have no practical experience with these, so I don't know if they are a good idea or not.  (Points of attachment are the first question.)   ... but, then, this is going to depend on just how high the mast will reach.

[cdev]

A mast has pieces that fit into one another. A guying system is just a ring and wires or very strong ropes that prevents lateral motion. The mast is not going up, it doesn't need to be attached strongly to the piece below it its not going to fly away, what it needs is to be prevented from going sideways even a small amount. If you have a tall mast you should use two rings with a minimum of wour guide ropes going to each one. You can use the same anchor points. Look at how they brace utility poles. Like that. You can use scrap wood to make enclosures for concrete pours and put anchors in the concrete so when it sets you'll have an extremely strong anchor post. It will last longer than you will. Concrete is very easy to work with. Even I can do it. They bring it to you and pour it in.

The most important rule is stay away from utility wires. Always assume your antenna is going to fall down someday. Make it impossible for it to ever fall on a utility wire and start a fire by physical distance. You'll notice you never see antennas anywhere near power lines. Thats because laws likely prevent their erection anywhere near them. .

[Red Squirrel]

Yep guyed towers hardly have any ground support, this surprised me the first time I saw it.





This is one tower I went at, it's probably about 150 foot tall if I had to guess.

That said if building a tower from scratch I would do self standing personally, less maintenance and simpler.  More expensive though.

[tautech]

65cm dish is not small and will need a reasonable mount if it is withstand any wind event. The 32mm galv pipe would be my choice. Think also how much time/km's you'll spend on alignment....do it once and right.

[Brumby]

I wasn't really thinking of guying for support against failure - but that is something to consider.  I was thinking more along the lines of rigidity - preventing the tip of the mast from bending over under wind load.

With the wind blowing directly from behind the dish, the dish will point down a bit.  A 1cm deviation at the end of a 2m section of mast, will change the beam centreline by at least 275m at the distance involved.

What range of geometry is available from the dishes?

[cdev]

Have you ever heard of a tensegrity mast?

[cdev]

Are you sure the gain is that high?

[Halcyon]

Think also how much time/km's you'll spend on alignment....do it once and right.

Very good advice!

Are you sure the gain is that high?

The antenna I'm looking at is the AF-5G30-S45 65cm, 30 dBi dish, but if I can get away with the AF-5G23-S45 (37cm, 23 dBi) then I'll use that.

The ones I have for testing are around the 37cm mark I think.

[cdev]

Did the manufacturer give you a figure for the beam width - how critical the alignment is?

You should consider using a mesh dish to reduce wind loading.

Any sold flat or concave antenna is going to have to be mounted extremely well or it will move with the wind.

There are very high gain panel antennas too. Typically PCB elements and a reflector. 17db is quite achievable at 2400 MHz in a compact design.

At 5 GHz the loss in the PCB material might be too high.

You should consult your area's historical weather data to find out what kind of winds are seen there during storms.

[Halcyon]

Did the manufacturer give you a figure for the beam width - how critical the alignment is?

You should consider using a mesh dish to reduce wind loading.

Any sold flat or concave antenna is going to have to be mounted extremely well or it will move with the wind.

There are very high gain panel antennas too. Typically PCB elements and a reflector. 17db is quite achievable at 2400 MHz in a compact design.

At 5 GHz the loss in the PCB material might be too high.

You should consult your area's historical weather data to find out what kind of winds are seen there during storms.

I'll be using 5 GHz so a mesh dish isn't really that suitable (or available from UBNT). I need the higher gain to work over such a long distance. Remember, these radios don't output a signal any stronger than good quality Wi-Fi gear and the EIRP must remain within the ACMA limits.

Alignment will be fairly narrow being 5 GHz. There are 3 options for antenna that I'm looking at:

AF-5G23-S45 - 37.8cm, 23 dBi dish (10° beamwidth)
AF-5G30-S45 - 65cm, 30 dBi dish (5.8° beamwidth)
AF-5634-S45 - 105cm, 34 dBi dish (3° beamwidth)

If I can get away with using the smaller dishes, that's great, but I have trees to contend with so it will depend on how much of the 1st Fresnel zone is obstructed.

All three antennas have a 45° slant which should assist with noise rejection.

Ubiquiti also make a "Precision Alignment Kit" for fine tuning alignment over long distances. Don't forget though, I'm "only" needing 55 KM range, the Airfibre radio that I'm considering will do 200+ KM with the right antenna and conditions.

[Old Don]

https://www.wired.com/2007/06/w_wifi_record_2/

[ez24]

Watching because this might help me get 55 meters

[cdev]

This is a good resource to get the wind velocity in realtime .

https://www.windy.com/?-34.253,142.559,6

[cdev]

Here are the new record holders and their antenna.

[Halcyon]

Hi all, apologies for the lack of posts/updates over the last few weeks. Work has been crazy busy so I haven't had a lot of time to test.

I did visit and visually check out the 'Station B' end and it looks like we might have issues getting over the top of a few rooftops.

However not all is lost, I have alternative plans to set up a link to another site (47 kilometers away) which is looking much more promising. I'll still conduct some tests in the coming weeks to the original site but I don't like the chances of it working.

Since I'm on Australia's new National Broadband Network (NBN) for my internet connection (with no real alternatives) and considering they refuse to replace any of the faulty and ageing copper in the street (which is causing me drop-outs and less than expected line sync speeds), I'm looking at moving my service to the alternate address (which is much closer to an NBN/VDSL2 node than I am) and access the internet over the Ubiquiti link. It should (in theory) be much, much faster and more reliable than what I'm experiencing now.

Updates will be forthcoming soon! I promise!