Page 1 of 1

sr9 and client distance

Posted: Tue Jan 23, 2007 9:51 am
by firebat
What are folks seeing with regard to distance and throughput with the SR9 900MHZ? Will these go 5-6 miles through a few trees (a few trees about mid way and a few more around the clients house)?

Posted: Wed Jan 24, 2007 12:52 am
by slipstream1
AP:SR9 in RB532 w/ 13 dbi 90 degree sector @ 110' EIRP 36 dbm
CPE: SR9 in RB112 w/ 13 dbi yagi @ 8' EIRP 36 dbm
Distance: 11.3 miles NLOS
Download to site on internet: 1.34 Mbps
Upload to site on internet: 800 Kbps

This is over a single T-1, so the SR9 link almost maxed out my T-1 @ 11.3 miles sitting right next to an operational Motorola Canopy 900 Mhz system. I was checking to see if the operational Canopy 900 Mhz system would cause interference on my 900 link.

At just about any location inside 8 miles, I could max out the T-1 on uploads and downloads. I was really impressed with the performance. :twisted:

Posted: Wed Jan 24, 2007 2:28 am
by firebat
Thanks for the info. How many trees, obstructions, etc did you have between the AP and the client? That sounds pretty promising. :)

Posted: Wed Jan 24, 2007 5:05 am
by slipstream1
Northeast Texas. The pineywoods region. Lots of pine trees and live oaks.

Posted: Wed Jan 24, 2007 8:15 am
by 0ldman
Pretty flat elevations, right?

Posted: Wed Jan 24, 2007 2:40 pm
by slipstream1
We have alot of rolling hills and some 200' to 300' elevation changes. The direction that I did this test in was across a river bottom so the elevation was fairly flat, but still impressive.

Posted: Wed Jan 24, 2007 5:53 pm
by ejansson
What was your signal strength and what channel size where you using 5/10/20 ?

Erik

Posted: Wed Jan 24, 2007 6:53 pm
by slipstream1
Signal strength was -81 on 5 mhz channel spacing.

Posted: Wed Jan 24, 2007 7:18 pm
by dbostrom
We're seeing results very similar to slipstream1's. However that news about coexisting w/Canopy is very good to hear; we don't have the problem in our area.

Posted: Wed Jan 24, 2007 9:29 pm
by ejansson
Can any of you guys tell me what through put you are getting at the different modulations, and have you tried the 10mhz channel at all? and if so what did you see for through put.

Erik

Posted: Wed Jan 24, 2007 10:03 pm
by dbostrom
For right now the best information on the effect of rates, bandwidth are probably the following writeups by Ubiquiti.

Point-to-point:
http://www.ubnt.com/downloads/SR9_Mtik_PtP.pdf

and point to multipoint:
http://www.ubnt.com/downloads/SR9_Mtik_PMP.pdf

Our own experience closely matches the data therein.

Posted: Wed Jan 24, 2007 11:35 pm
by slipstream1
I am just going to use the 5 mhz channels for now. If I expand later on, I may include some 10 mhz channels. I noticed that with one card on 5 mhz channels and 1 card on 10 mhz channels, the cards don't see each other. I guess they will add a little more noise but maybe not add significant interference to wach other.

Posted: Thu Jan 25, 2007 3:29 am
by jwcn
They refer to antenna/cable mismatch keeping within 1.5:1

How does one calculate this?

Posted: Thu Jan 25, 2007 3:35 am
by ejansson
You need a meter. For 900mhz only M2 makes a great little unit that will test your antenna and cable system any where on the 900mhz band. They also make great 14 and 17dbi yagi antennas. http://www.m2inc.com/main%20html/900acctemp.html

Erik

Posted: Thu Jan 25, 2007 5:23 am
by jwcn
There isn't a way to calculate based on what specs the Antenna and cable have?

Posted: Thu Jan 25, 2007 5:00 pm
by jwcn
Slipstream,

What antennas were you using? Can you share your hardware configuration as I'm not saying anywhere near that kind of performance.

Thanks!

Posted: Thu Jan 25, 2007 8:14 pm
by dbostrom
There isn't a way to calculate based on what specs the Antenna and cable have?
No, its a factor of the quality of the antenna design (decent design has low reflected power even over the relatively wide bandwidths required for this application), the quality of your connector workmanship (properly assembled connectors do not cause reflected power), and the assumption that the impedance of your antenna and cable are correct for the transceiver (a given if you're buying from Wifi vendors).

Posted: Fri Jan 26, 2007 2:21 am
by jober
AP:SR9 in RB532 w/ 13 dbi 90 degree sector @ 110' EIRP 36 dbm
CPE: SR9 in RB112 w/ 13 dbi yagi @ 8' EIRP 36 dbm
Distance: 11.3 miles NLOS
Download to site on internet: 1.34 Mbps
Upload to site on internet: 800 Kbps

This is over a single T-1, so the SR9 link almost maxed out my T-1 @ 11.3 miles sitting right next to an operational Motorola Canopy 900 Mhz system. I was checking to see if the operational Canopy 900 Mhz system would cause interference on my 900 link.

At just about any location inside 8 miles, I could max out the T-1 on uploads and downloads. I was really impressed with the performance. :twisted:
Humm, Is there something different about the way the Moto 900 APs work then the 2.4 Moto. I only ask because a friend of mine was trying to add Moto 2.4 equipment to his Mikrotik network and as soon as he powers up the Canopy AP his MT APs just stop working. Even if they are on ch1(MT) and ch11(Moto). The Moto AP seems to kill the whole 2.4Ghz band.

Posted: Fri Jan 26, 2007 3:07 pm
by slipstream1
jcwn:
I am using some DB Products 90 degree 13 dbi sector antennas that are hand-me-downs from a cellular carrier. They have a 40 db F/B ratio and just perform awesome. The CPE side is a 13 dbi Hyperlink yagi. This is a little larger antenna than I would like to use, almost 5' long, I am planning to use a 9 or 10 dbi yagi, less than 3' long, in future installations. The configuration on the RB's are plain jane most config set to default.

jober:
I was testing around a competitors Moto 900 mhz system to determine if there would be any interference on the Mikrotik/Ubiquiti setup that I plan on using, before i spent the moolah for a large scale deployment. I could see no substantial interference.

Posted: Fri Jan 26, 2007 3:15 pm
by slipstream1
Let me also state that I turned the power down to stay within FCC EIRP guidelines for 900 mhz, 36 dbm AP and Client side.

I do not see why we cannot get the ranges that the cellular carriers are getting, maybe even more. Yes, they can use higher power at the base station, but those little phones can only muster 100 mW or so max.

Posted: Fri Jan 26, 2007 5:04 pm
by MyThoughts
Cell phones and cell phone carrier class equipment have RX sensitivities of around -114 to -120 dB.

Atheros extended range enables rx sensitivities of around -120 dB on supported cards however RouterOS does not support this feature YET :)

Posted: Fri Jan 26, 2007 6:01 pm
by slipstream1
I am aware of that, but we can generate more power at the client end. I think it will even out and improve the client end distance.

Posted: Fri Jan 26, 2007 7:34 pm
by ghmorris
The issue is the noise floor. When the noise floor is at -75 you're going to have a problem.

Unlicensed spectrum and all that compared to a cell carrier.

Always design for a very high noise floor, you know its just a matter of time...

George

Posted: Sat Jan 27, 2007 9:22 am
by 0ldman
I am aware of that, but we can generate more power at the client end. I think it will even out and improve the client end distance.
A bud of mine used to work for a cell company. They have amps on the tower as well to take those dinky phone signals and amplify them before they get to the receiver.

I agree, we should be able to get similar results as a handheld phone is like 500mw max, I don't even know if they hit that anymore.

Another benefit we have is fixed point to multipoint. The phones don't have much room for an antenna. Quite a few of the aftermarket cell phone antennas are only 3dBi and show signifigant improvement over stock.

Posted: Sat Jan 27, 2007 1:48 pm
by slipstream1
I work in the cell phone industry also. Not all of the carriers use TTA's (tower top amps). Cingular is a big TTA user, but companies like Alltel do not.

I do not think that the modern cell phones reach 500 mW either. In hte search for more battery life, they are using smaller, less powerful transmitters. The digital technologies, CDMA, PCS, TDMA, and GSM use a lot less power than the old analog bay stations that Alltel and the like used. If I am not mistaken, in some areas, Alltel analog bay stations were cranking out close to 100 Watts, not mW but Watts. That would set the trees on fire.

Anyway, I am very optimistic about the SR9 and routerboard combos and how they will perform in the pineywoods region of Texas. I believe it will be another good tool to be used like alcohol, always best when used in moderation.

Posted: Thu Feb 08, 2007 5:42 am
by gregdhayes
RB532/SR9 - Pac Wireless 120 deg Horiz Ant - 190' on tower.

Customer site: 20' off ground. RB112/SR9, pac wireless 15 Dbi Grid - 7 miles out, pretty heavy trees - Signal -70, noise -98.

Can do a Torch and get pretty consistant 11MB.

Another customer, Very heavy trees, rolling hills, 5.5 miles out, Ant 8' off the ground. Signal -80, same Torch speeds.

Using 2Ghz-10Mhz channel. Can't seem to connect anybody using 5Mhz channel???

Posted: Thu Feb 08, 2007 4:17 pm
by dbostrom
We're fairly busy right now circling back and installing customers who did not want the expense or eyesore of a telescoping pole or truss for 2400mhz installations. Vegetation is a major problem for us but the 900 gear seems to be doing the trick so far.

If there's a challenge here, it's confining radiation from 900mhz APs, so as not to get too much overlap, a good problem to have on balance. That and Fresnel interference, which we've found can usually be overcome by tweaking CPE placement.

So far we're using RB112 CPE on 5mhz with good results. However we're already seeing hidden node issues crop up. MT, still waiting for that RTS control.... :wink: (and how about 802.11 MAC-level fragmentation too, while you're at it?)