I thought some here my get some use out of this thread.

Hope it helps,
(509) 982-2181                                   Equipment sales
(408) 907-6910 (Vonage)                    Consulting services
42846865 (icq)                                    And I run my own wisp! (net meeting)

----- Original Message ----- From: "Marlon K. Schafer" <[EMAIL PROTECTED]>
To: <isp-wireless@isp-wireless.com>
Sent: Sunday, June 25, 2006 7:10 AM
Subject: Re: [isp-wireless] Sector Antenna's I want the Best!

David, as usual, great thoughts.

Let me add a bit here on the "but what should *I* do???" side of things.

I have both omni and sector solutions on the air. I have both non amped and amped of both. Let me show you how/where I get to each option.....

Very rural, few customers likely in the first 18 to 24 months. The AVERAGE ap out there will service from 75 to 100 subs jut fine as long as you don't have many hogs. So in a case where I'll likely have less than 50 users I'll go with an amplified omni unless the geography is such that there's just no way to get more than a few miles of service out to people.

Down town where there are a few companies in the shadows of my main wpops I may put in an omni. Low gain so that coverage only extends a few blocks etc.

I have some sites on hill sides that overlook many hundreds of houses. Very stiff competition so there number of customers isn't really an issue. Spectrum issues are though. There are at least 4 other wisps on the hill, one of whom is both stupid and mean. You know the type, never met an amp he didn't like..... He's got a sector that points to NO customers, only to my tower. The good news is that operators like that eventually run out of money and will go away. But in this location I've got SIX different systems. 4 of them are 2.4 and three of those are on the same channel. Remember you have not only the noise that your ap's see to deal with but also the noise that your cpe sees coming from his ap's. In my case I have sectors as narrow as 30* (panel antennas make nice sectors and are cheap!). I also rotate polarity ass needed to help bleed off as much noise as I can.

There are other spots that I have no customers behind a tower on a hill so sectors just make sense anyway. If long range is needed OR if I do have a customer or two out of the main lobe of the sector I may put an amp on the sector too. Usually there's no need though. It's surprising that a sector with an amp and running at the max legal 4 watts will often not give much more coverage than a sector with no amp running at 1 or 2 watts. The main reason for that is the noise that the amp injects into the system keeps your SNR about the same as often as not.

There's also a cost factor. A good Maxrad adjustable beam sector (the only ones I buy nowadays) is about $400 from EC. The H-Pol version is $250ish. I tried the cheap sectors from Maxrad and another company and all systems running them act just a bit strange. Drop in a good antenna and it's AMAZING how much better the system runs! A good omni antenna from Maxrad is about $100. Others that are worth having are in that ball park (remember, there's ALMOST NEVER a good reason to run an omni over 10dB!!!!!!!!!! The high gain ones are just more money for the sales geeks).

Sooo, 360* coverage with an omni and amp. $400 for a good ap, $250 for a good amp ($800? for a great one), $100 for an antenna, $25 to $150 for cables and we're at around $1000 for an omni based wpop. (plus backhaul but we'd need that anyhow) For a sectored solution you've got $400 for an ap, $400 for an antenna, $5 to $25 for cable (cat 5 this time). Times THREE or more puts us into the $1200 to $1500 range.

Now, there are some that say you should put in the sectors just because it's so close in price and you have better spectrum usage that way. I don't really buy into that theory. You'll usually use ALL spectrum when sector vs. one channel when using an omni. However, with sectors it's easier to avoid noise coming from different directions on different channels/polarities.

The FIRST thing you should do is figure out if there are enough customers in a given area to at least pay the bills. Then you need to run a spectrum sweep to make sure that the band you want to use is indeed usable (do NOT make the mistake I did and assume that because you are out in BFE you'll not have other wireless users in the area). NOTE: Netstubler or anything like it will NOT be of much use for this. You need a spectrum analyzer. Rent one for a week or two. It's well worth the cost as it'll often save you huge amounts of time and money down the road. Been there, done that, ain't doin' it again!!!!

Lastly, find a person or two who's advice you trust and don't deviate too much from what you read on these lists. Hang on to him and let him help design your system. He'll then also be able to help you fix the things that no one will know about till you've built it. There's always something that doesn't show up till that 20th customer is online.....

Hope that helps,

----- Original Message ----- From: "David Vrablic" <[EMAIL PROTECTED]>
To: <isp-wireless@isp-wireless.com>
Sent: Sunday, June 25, 2006 5:56 AM
Subject: Re: [isp-wireless] Sector Antenna's I want the Best!

Good morning Brad,
Thanks for a perfect explanation of where I was going with my opinions.
A sparsely populated rural market is a "whole nuther ball game" compared to a heavily populated urban market where Time Warner has lit up every household with a Linksys AP on Ch 6. (HEY! Time Warner, You're a Jerk!)
Now lets get back to answering the young mans question:
The truth is I have enough population that anywhere I would install a sector, Sales can get customers.
That fact changes the whole design picture.
This especially holds true if you are designing a meshed ring / star network.
You have to design it and plan for it, even if you don't build it.
You have to design your BH route / topology to cover the whole area you might wish to cover before you fill it in with sector or node coverage. You are right on the money with your consideration of the opposing beam width in use. If polarity of the beam "in either axis" is too wide, you are subject to waste and noise, if too narrow you loose coverage but increase reliability for where it does cover. I believe careful selection of antennas is where this thread started and is right where good sector coverage design starts.
The man asked for Best .
That can be Reliability, Ease of mounting, Predictable design plots, Polarity and overall bang for the buck. What I gleaned from the post was " What is the best first time out coverage approach to start a new WISP business with. My answer is: It all depends. (On a lot of things that only he can answer) I had a real eye opener when I started building my first Muni system and started reading up on the "Cylindrical Model" where you can envision coverage areas in 3D . The white paper treated the sectors like stacking cans of different sizes to represent the actual service areas in buildings. I realized that this was an omni approach and wasted a lot of coverage and had a lot of overlap that made for a lot of self generated spectral noise to overcome. I thought it over and would rather think of my directional sectors as wedges of cheese. I can even reuse the same frequency back to back if I am on the opposite side of a bell or clock tower. I believe you have to engineer everything not just put up an omni, that with any success, you will be pulling down and replacing with sectors anyway.
Changing subject here: Adaptive "Smart Antenna Systems" a look at the future?
Just for fun take a look at the Netgear MIMO AP.
(Fun to play with and has pretty glimmering lights dancing around in a circle.
Unfortunately it doesn't have very good coverage)
It has 7 internal antennas that scan in a circle and stop to transmit / receive for each transmission. Now take that concept "out of the box" with external antennas and a brave new world evolves.
It is something that is long overdue.
Also I would believe that "Point to point " rules would apply.
Oh well Back to the Cheese errr Drawing board I have a small village coverage to design this week. ;-)

Simple fact is the BG one channel amp & omni approach will only afford you
the capacity of that one radio.  In contrast sectors allow you multiple
radios giving you many times the scalability capacity and frequency agility.

     David L. Vrablic
 Wireless Network Engineer
   Logical Net Wireless
  Schenectady, NY, 12305
    Office 518-292-4519
      Cell 518-376-2940

WISPA Wireless List: wireless@wispa.org


Archives: http://lists.wispa.org/pipermail/wireless/

Reply via email to