13

I have recently been contracted by a client of mine to facilitate the wireless communication of his "home" offices and a secondary site.

The primary site is the top two floors of a 5-story office building (15m height more or less) an the secondary is one of two open "lots" (which one is TBD by management). The ground distance from the secondary sites is a little more than 2km for the one closer and around 2.9km for the one furthest.

The link will be used to transmit the video feed of 1 (or even possibly two) IP cameras and some kind of Ethernet-enabled environmental or weather sensor. I have checked the necessary b/w for the cameras and both 900Mhz and 5.8Ghz are more than adequate for even 4 of them, much more for 2. I have also verified that there is clear line-of-sight to both possible installation points and that the 60% Fresnel Zone clearance is more than covered. Bear in mind that this is my first long distance link (long with or without quotes) and I hate to admit that wireless physics is far from my strong suit.

The ultimate point of my question is that although I have read a lot about frequency choice the last few days, I continue to find some ambiguity (I know it is just me that finds it ambiguous). Most sources, like this one, agree that although the lower frequencies have less losses over a given distance (free-space-loss I learned it is called) they need larger antennae for the same "strength" of trasmission (is "gain" really the same as "strength"?).

So, for the given distance of 2-3km and given also that all typical requirements are met, which is preferable (or do I dare say "better") frequency? Should I choose 900Mhz with a relatively "small" antenna on the basis that 3km is not really "long distance" and that it will provide a link with less attenuation ergo less retransmits ergo higher overall speed? Or should I choose the 5.8Ghz option for the superior b/w (I am still not very sure about this, please correct me if wrong) on the basis that at this distance there is no real difference so why not take the "better" one?

On a side note, should I stay to the beaten path of true WiFi or should I consider proprietary bridging solutions like the ones from Ubiquiti? I have a lot of experience with their Access Points and am really satisfied, so I would not mind integrating one more of their products in my client. In any case, I am looking for an optimal solution, choice of vendor is of very little concern at this point.

Forgive my ignorance and the possible mistaken use of language.

UPDATE: I arranged to have a spectrum analyzer on loan for a couple of days. I will make sure that the 900Mhz band is reasonably clear and proceed down that way.

UPDATE 2: I had the aforementioned equipment available to play with for a day and a half. The conclusive finding is that the 9Mhz band is almost "empty" in the area, as one suggested here, so that takes care for the frequency choice issue.

Concerning the equipment now, I am going with Ubiquiti AirMax Yagi antennae and matching RM900 2x2 radios. Preliminary testing on my part and from the client's employees shows that performance exceeds expectations.

On a side note, the chosen "lot" is the one that is 3km away.

dlyk1988
  • 1,644
  • 4
  • 24
  • 36
  • Over 300km has been done with 5GHz. I don't think you have much to worry about there. – Michael Hampton Sep 01 '13 at 01:19
  • @MichaelHampton 5.8GHz at 300km equates to 160dB freespace loss. The EIRP limit is 36dB, so your antenna and receiving radio sensitivity would have to add up to 134dB (including only a 10dB link budget). The best radio I could find quickly is rated at -96dBm@6Mb, so you'd need a 38dB antenna roughly 1 mile high to have a chance (assuming there's no hills/mountains between the antenna sites). – Chris S Sep 01 '13 at 01:32
  • 3
    @ChrisS http://www.gizmag.com/go/7878/ Oops, should have dropped the link to begin with. – Michael Hampton Sep 01 '13 at 01:40
  • @MichaelHampton Reliability of the connection is conspicuously missing from the article... – Chris S Sep 01 '13 at 02:10
  • @Chris S What the italians did with this link in the article is really astonishing. But I too have some reservations about its reliability. It is after all a not-for-profit project with an air of research and "let's do it" about it. I feel like I'm going with 900Mhz after all. – dlyk1988 Sep 01 '13 at 10:18
  • Bullshit. 3km can be VERY easily done even with WiFi, 2.4/5. In most situations with omni antennas. I've done tens of these with cheapest Ubiquity equipment, with Picostation HP and Nanostation pairs, just use directional antennas. 300km is not a problem for modern backhauls, bandwidth is a problem. – GioMac Sep 01 '13 at 10:37
  • 2
    @GioMac An omni antenna that'll do 3km? I need half a dozen of these mythical beasts... Where do you get them? – Michael Hampton Sep 01 '13 at 14:31
  • Yes, but same eq. from both sides. It's available everywhere. In Georgia we've installed Nanos in national parks as both backhauls and end user AP's. In this situation it works - no background noises, no interference, no sea, no reflections. There are many companies running them in the cities with directional antennas for 10+ km in unlicensed spectrum, PTP. I am certified Motorola Canopy technician, but my certifications become useless, mainly because of development of wifi mesh and Ubiquity & Mikrotik products, nobody buys Canopy for this distance anymore. – GioMac Sep 01 '13 at 23:26
  • It bears mentioning here that gain, in the context of an antenna, is more a measure of directivity than anything. A directional antenna with 12dB gain will radiate 16 times stronger than its input power in its direction, and correspondingly weaker in other directions (depending on its radiation pattern). A truly omnidirectional antenna will have no gain (and will radiate equally in all directions across all 3 planes). – Falcon Momot Sep 03 '13 at 19:21
  • 2
    I have had an 8KM link work flawlessly with line of site to a sector antenna with 500mW radios on either end. – hookenz Sep 03 '13 at 23:52
  • 1
    GioMac - Picostation and Nanostation use directional patch antennas not omni. However, you are right in saying that 3KM is not far. I've also installed several 2-5KM links with Nanostations. I actually had my 8KM link work on a nanostation M5. I switched to a bullet but only because there were some initial problems that turned out to be with the other end of the link. Over 8KM shared with other users I was able to pull data in both directions at ~11MBit/s. – hookenz Sep 04 '13 at 00:14
  • I was limited somewhat to 11MBit/s because the other end was set to a 20MHz wide spectrum although it supported 40MHz. – hookenz Sep 04 '13 at 00:15
  • 2
    Line of sight isn't always easy to achieve. Our access point was installed up at around 2000ft on top of a low mountain range overlooking the city. There was a back haul link back to another location of a similar distance which provided access to the data centre and internet. Oftentimes tall buildings (double story houses) or trees get in the way of providing line of site. – hookenz Sep 04 '13 at 21:41

2 Answers2

9

3km is approaching the limit of 5.8GHz equipment with reasonable size antennas and typical WiFi radios. With some of the best equipment you can only get ten times that distance with some caveats.

900MHz can easily go 3km, and much much farther. 900MHz is used for a wide variety of devices for this reason, so there's a good level of background noise. This might be a problem, especially if your neighbor has a 900MHz phone or similar (not a popular frequency for phones these days, but people hold on to technology for a long time too). I would avoid this frequency unless you're in the middle of nowhere (which you aren't). The 2.4GHz spectrum has worse problems with this, our microwave at work blasts 2.4GHz (I'm sure it violates FCC something or other, but nobody really cares as we have 5GHz WiFi).

Chris S
  • 77,337
  • 11
  • 120
  • 212
  • Which commercial radios operate on the 900Mhz band? Is it really that much common? – dlyk1988 Sep 01 '13 at 10:23
  • 2
    Commercial WiFI, I can only name Ubiquity off the top of my head, I'm sure I've seen others. It's usually a 802.11a radio, since the spec allows for several different frequencies (including 900MHz and 3650MHz). It's been a while since I did a survey, but 900MHz background noise was quite high at that time. YMMV. – Chris S Sep 01 '13 at 14:39
  • 1
    @Matt Please keep the commentary to yourself. If you think you know something I don't please feel free to write your own answer, with the correct information, and let the community assess each as they will. – Chris S Sep 04 '13 at 02:04
  • By the way, what is a "Reasonable sized antenna"? I've had 10KM on ones about twice the size of my hand with fairly inexpensive equipment. We're only talking a few hundred dollars. You have made it sound "specialized" when it's not. – hookenz Sep 04 '13 at 21:35
  • 3
    Keep in mind that in most places outside the USA, [900Mhz is a primary frequency for GSM](http://en.wikipedia.org/wiki/GSM_frequency_bands). – Andrew Sep 09 '13 at 00:47
4

Chris S's answer is incorrect in several areas.

3km is not approaching the limit of 5.8GHz. Similarly, that's not as big of a distance as for 2.4GHz wifi either. Chris says 10X, I say 100X the distance is possible. The longest range 5.8GHz wifi that I've heard of is 304KM with a 1.2M hand made antenna (See: Long Range Wifi). I believe it went over water so there were not any thing in the way of the signal. It used Ubiquiti radios. I don't know if it was reliable, but to get a connection and send data over that distance is nothing short of amazing.

I used to work for a wisp carrier and had wifi radios about twice the size of your hand easily going 10KM with good line of site. I personally had wifi on my house going 8KM without any issues. Amazingly, they were only 500mW.

We were using 5GHz radios and a mixture of Ubiquiti and Microtik hardware. They work almost flawlessly.

While it is true that with lower frequencies you need bigger antennas. You'll notice that your home wifi doesn't have large antennas. Neither do 5GHz ones.

In reality, 2.4GHz should perform better, but 5GHz in my experience works just as well. In theory, 5GHz could be affected by rain fade, but I didn't find that even over the 8KM link.

As for 2.4GHz and other devices like microwaves and cordless phones inside the office, it doesn't even factor in. The reason it doesn't factor in is that you'll be using directional antennas. Better still, if the antenna is on an iron roof you will be shielded from all that noise coming from inside the office/house.

Having said all of that, I will say that it's not perfect for every situation. You need:

  1. Line of site between the two buildings. Factoring in the freznel zone (good open clearance)
  2. A suitable place to mount the antenna to achieve point 1.
  3. Hopefully not much wifi traffic facing your directional antennas (no gaurantees there, but it should be workable) even in a reasonably crowded wifi space.

Freznel zone looks like this:enter image description here

By the way, I've had trees in the way of the freznel zone and in some cases right in the middle and it still worked, although it does affect the performance a lot. Install your antennas as high as practical/possible.

You can still get interference issues. Usually because of other devices transmitting in your direction at the same frequency. So it's not perfect. Directional antennas help a lot with that.

I can recommend the 5.8GHz radios simply because I know they work well. And yes, provide good bandwidth! From memory I used a Ubiquity Bullet M5 with a patch antenna (another brand). The other end was a less directional antenna feeding multiple clients but used a similar Microtik radio.

I don't know of anyone who uses 900MHz. But I do see you can buy them at around 3X the cost than 2.4 or 5.8. Stick away from it, it seems uncommon. In several countries the 900MHz spectrum is quite crowded anyway.

Update: You've been looking at the airmax antennas and the 900MHz radios. While I think performance should be great. I think you may be spending much more than you need to... perhaps a little overkill!? 3KM is not a long distance. I'd try some of these if you can.

http://www.gowifi.co.nz/antennas/5-ghz/directional/5.8-ghz-27-dbi-cast-reflector-grid-antenna.html combined with a Bullet M5.

Update2: You can also mitigate interference issues using polarized antennas. Here's a good write up that explains the concept. From memory, we used horizontal polarization. As I recall sometimes if I set the client radio to vertical polarization I could see maybe 12 AP's. But when setting it to horizontal I only saw 3 or 4 and half of them were ours.

hookenz
  • 14,132
  • 22
  • 86
  • 142
  • My choosing the 900Mhz option is based on my findings after using a spectrum analyzer. It seems that in my area both 2.4Ghz and 5.8Ghz bands are extremely crowded. That and the fact that the 900Mhz radios cost "only" about 2 times what the others cost is what drove me to my decision. – dlyk1988 Sep 03 '13 at 23:42
  • 2
    Are you connecting your spectrum analyzer to a directional antenna and aiming it very carefully at the remote office? Directional point to point antennas do cut out a lot of noise. – hookenz Sep 03 '13 at 23:46
  • 3
    Cherry picking words from my answer and comments and presenting them as what I said is both wrong and offensive. If you go re-read what I said I'm sure you'll find the meaning different than what you have presented. Also you present the possibility of a 304km link without disusing reliability , bandwidth, or other metrics critical to business applications. I should be surprised that someone working for an ISP completely disregards an Quality of Service discussion, but that seems pretty common on the whole. That link was done as a proof of concept by an amateur radio club. – Chris S Sep 04 '13 at 01:58
  • 2
    Further the antennas were only 5k feet ASL, so saying there was "not any thing in the way of the signal" is actually wrong. Go ahead, calculate the high required for those based on the curvature of the earth. And that link was mentioned in the comments above already. You make wild assumptions about unqualified WiFi gear, and spectrum availability/crowding; but then go on to say your experience and option is based on a single model of equipment. Feel free to make recommendations as you have, but don't be surprised when the OP tells you how wrong you are based on an actual radio survey. – Chris S Sep 04 '13 at 02:01
  • 1
    I didn't cherry pick your answer at all. No offence intended Chris. I actually didn't read the 304KM comment until after I posted about it. I'd heard about it around 3 years ago. – hookenz Sep 04 '13 at 09:08
  • 1
    I actually read the 304KM comment (as it was collapsed) well after I posted about it in my answer. I'd heard about the 304KM link around 3 years ago. Your comments about my wild assumptions are not wild at all. We used a range of gear. Mikrotik and Ubiquity worked well and better than some more expensive gear. We got good support from suppliers this gear. We had terrible support from other suppliers. We also tried trango but they performed badly. Along with some fancy stuff that we couldn't get working at all and had to sue the supplier. – hookenz Sep 04 '13 at 09:14