Community discussions

MikroTik App
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

SR9 distance 5MHz vs 10MHz

Thu Apr 19, 2007 6:58 am

My first day of testing with two rb133c + SR9 set as a bridge. With one side set 9ft above the ground (my porch) and only 4 ft above the ground (next obstical in line), I got 1 mile distance through a half mile of trees with the yagi pointed the wrong way.

I did notice I gained signal by switching to 5MHz channels. What difference have you guys noticed with 5 vs 10MHz channels on range with the SR9, or any card for that matter.

I'm weighing the options, coverage vs bandwidth.
 
User avatar
janisk
MikroTik Support
MikroTik Support
Posts: 6263
Joined: Tue Feb 14, 2006 9:46 am
Location: Riga, Latvia

Thu Apr 19, 2007 3:29 pm

IMO you have less noise incoming, and therefore you can shoot further, but you have less bandwidth
 
User avatar
jp
Long time Member
Long time Member
Posts: 609
Joined: Wed Mar 02, 2005 5:06 am
Location: Maine
Contact:

Thu Apr 19, 2007 4:40 pm

You are putting the same power into a smaller channel width (1/2 the width) - that will increase the receive level by 3db.
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Thu Apr 19, 2007 5:34 pm

I noticed both of these, glad to have it confirmed. Keep'em coming.

Real world experiences as well.
 
hytanium
Member Candidate
Member Candidate
Posts: 201
Joined: Thu Jan 18, 2007 9:10 pm

Fri Apr 20, 2007 4:22 am

Oldman,

How many customers can you support at 5mhz with an SLA of 500Kbps? How about 10MHz.

I also want to know, can you run 10Mhz and support 3 sectors or does the overlapping create self interferance??
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Fri Apr 20, 2007 5:35 am

I haven't gotten that far yet. Still have to get the equipment on the tower for a point to point shot for stability testing. I'm shooting past a cell tower (high 800MHz range). Noise is -96 at worst, -102 at best.

A few other WISP's use 9dBi yagi in place of sectors, 2 yagi, back to back, vertically polarized seems to come close to 360* coverage. It does seem that 10MHz channels will be required for a large customer base in 900MHz, but I haven't thoroughly tested 10MHz channel tree penetration.

I've gotten really good range off of sidelobes so far. The biggest problem I have is height. Being only 9ft up, the ground is blocking quite a bit of signal. I am on top of one of the high points in this area, but there are quite a few hills.

Putting one AP in the shop (metal building) and dropping tx to 17, I get -85 signal, still 18Mb on a 5MHz channel, I get close to 4Mb on the bandwidth test, 1700Kbps on send and recieve. This is with one RB133c running btest server. I don't have a PC on the other side right now.
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Fri Apr 20, 2007 5:53 am

768 up and 2Mb down seems to be the limit, or any combination in between. Go over 2Mb down, upload suffers, go over 768k up, download suffers. 1.544Mb synchronious seems to max out the bandwidth. Anything over, the ping starts going crazy and the bandwidth test gets erratic as well.

Looks to me like 5MHz channel connected at 18Mb (which isn't that tough) is equal to one T1 line.

edit: looking at it, the client side rb133c was pegged at 100% CPU during the test. The link may do better that initial testing shows.
 
hytanium
Member Candidate
Member Candidate
Posts: 201
Joined: Thu Jan 18, 2007 9:10 pm

Fri Apr 20, 2007 6:26 am

I have a 9 db Yagi at a little over 200 ft right now. Had to buy a cavity filter to get above -70 Noise floor from paging system, now sitting pretty steady between -93 to -97 noise floor.

The tower is on the city, when I travel out to rural areas with rolling hills and stands of trees and point a 13dBi Yagi ( 8miles out) I get a -68 to - 78 signal (depends on location) . I am runnig at 10Mhz and I can move 9-12Mbps consistantly half duplex.

My feeling is 900Mhz is the answer for NLOS but use it only if you can't reach customers with 2.4 or 5.8. I was origionally hoping to use 900Mhz as the only access for my wisp, but given the limited number of non-overlapping channels, I feel that capacity will be capped at 20-30 users per AP and cell planning would be very difficult (Self interferance). Cavity filters are a must if you want to minimize out of band interferance. Nstreme increases throughput, but I find that overall distance is reduced on the system. I have been playing with the SR9 for about 6 months now, and have done simular tests at about 8ft above ground and through 2 miles of tree trunks and have been able to achieve reliable connections at aver 4 miles out. I am now waiting to see what happens once vegetation starts growing in the forest and measure impact on the connectivity.
 
jober
Long time Member
Long time Member
Posts: 690
Joined: Fri May 28, 2004 12:16 pm
Location: Louisiana,USA

Fri Apr 20, 2007 6:35 am

 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Fri Apr 20, 2007 6:52 am

I am now waiting to see what happens once vegetation starts growing in the forest and measure impact on the connectivity.
For the record I'm in central Alabama, its very green right now. A solid half mile of pine on a 1 mile shot worked. A fairly thick run of hardwood scattered over a 1 mile shot put me down, but the ground level in between was close enough it may have been the problem. The elevation maps show the shot should have worked, but there are some monster oak's scattered throughout the shot.
 
jober
Long time Member
Long time Member
Posts: 690
Joined: Fri May 28, 2004 12:16 pm
Location: Louisiana,USA

Fri Apr 20, 2007 5:09 pm

Man I wish I had your luck with the sr9's. I tryed to setup a sr9 link but didn't get a usable connection. I used rb532s on both sides. They had sr9 cards with 13dbi antennas. The tower side antenna was 270 feet up and it has a noise floor of -81 to -75.
The way it's working now is with SR2 cards on both sides and 24dbi grids. Also running 5mhz bandwidth.
 
User avatar
warwick09
Member Candidate
Member Candidate
Posts: 190
Joined: Mon Aug 07, 2006 1:34 pm
Location: The Bahamas / Florida

Fri Apr 20, 2007 5:22 pm

Well Jober your noise floor is terrible... Have you tried Hori. Pol, Cavity Filter?

You cannot essentially blame it on the card as you did in one post :?:

Regards
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Fri Apr 20, 2007 5:37 pm

Noise floor is the ticket. 900MHz carries, and the SR9 doesn't like out of band noise. I've noticed, consistantly, just before I get a signal lock while lining up the antenna, my noise drops from -98ish to about -92. 900MHz is definately different than 2.4GHz.
 
jober
Long time Member
Long time Member
Posts: 690
Joined: Fri May 28, 2004 12:16 pm
Location: Louisiana,USA

Sat Apr 21, 2007 1:41 am

I don't really think that it's a bad card, just a bad area I'm in. Maybe the next time I'm up the tower I will move the antenna down to the 200 foot mount to see if it helps with the noise.
Also I don't want to piss away money on the filters after hearing a few people say that they don't help.
I guess I just need more time to play with them, but time is some thing I don't seem to have these days.
Oh, and the antennas are in horizontal polarity.
 
hytanium
Member Candidate
Member Candidate
Posts: 201
Joined: Thu Jan 18, 2007 9:10 pm

Sat Apr 21, 2007 2:14 am

Jober,

The cavity filters definately do the trick, especially in urban areas with AP's at high elevations. The filters remove all out of band noise.

Has anyone tested 900 full band filters? They are about the same price as Ubiquiti's. I bought the Ubiquiti filter, but you have to choose a center freq of 912 or 917... (2 Models) I think this only matters if you are using the 20Mhz setting.
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Sat Apr 21, 2007 8:16 am

I've read countless posts on several forums and google searches where the filters help. I've also read (but not confirmed) that 900MHz is best when only about 30ft above the tree line. 200ft may give far more noise as 900MHz carries very far. With Louisiana's elevation, you're likely to get 900MHz noise from up in MS, possibly as far as Jackson. I haven't looked at the elevation maps in detail for MS, but it wouldn't suprise me.
 
jo2jo
Forum Guru
Forum Guru
Posts: 1003
Joined: Fri May 26, 2006 1:25 am

Sat Apr 21, 2007 11:46 am

jober, the cavity filters HELP. all other stuff (amps etc..) are crap.

2 things:

1) noise floor is very important for 900mhz
2) even MT has pointed out that the noise floor reading from the sr9 may be off...but ive found them to be very good and consistant.. just not very precise.



the stuff works though..i have MT + sr9 deployed in a few locations around a major US city...so i see a *little* noise :) at 2 or 3 sites, (i suspect heavy 480v electrical interference at one however) ONLY 802.11b will work...everything else is less than 100kB bandwith and keeps droping out....remember all the sr9 is, is a SR2 with a special 2.4 to 900mhz converter built in..

EDIT: that is 802.11b on the sr9

if you need a 100% up time, mission critical link, do not use 900 generally..
Last edited by jo2jo on Mon May 14, 2007 6:54 pm, edited 1 time in total.
 
TomKolb
newbie
Posts: 32
Joined: Thu Jul 14, 2005 3:51 am

Sat Apr 21, 2007 4:51 pm

Hi Oldman and all, Have you tested Pine tree penetration when you have a lite mist or heavy morning due. We have when you have the needles covered by very small water droplets Pine trees become very difficult to penetrate. In many cases even 1 big tree will kill the connection.
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Sat Apr 21, 2007 6:20 pm

It hasn't rained here yet, and we haven't had much dew in the mornings either.

Thanks for the info.
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Sat Apr 21, 2007 8:05 pm

Another interesting tidbit.

I set both radios at 13dB tx, got the signal down around -85, 6Mb on 10MHz channel, stable.

I set it to 5MHz channel and *lost* about 4dB signal and had an inconsistant connection. Maybe the channel wasn't wide enough to take advantage of OFDM's error correcting capability?
 
jober
Long time Member
Long time Member
Posts: 690
Joined: Fri May 28, 2004 12:16 pm
Location: Louisiana,USA

Sat Apr 21, 2007 8:13 pm

Pine tree you say. I think I may be the poster child for Where not to start a WISP. You can't see 90% on the town from the tower in the day time but at night I see all the lights through the trees.
Lots and lots of pine trees.
 
jober
Long time Member
Long time Member
Posts: 690
Joined: Fri May 28, 2004 12:16 pm
Location: Louisiana,USA

Sat Apr 21, 2007 8:32 pm

Another interesting tidbit.

I set both radios at 13dB tx, got the signal down around -85, 6Mb on 10MHz channel, stable.

I set it to 5MHz channel and *lost* about 4dB signal and had an inconsistant connection. Maybe the channel wasn't wide enough to take advantage of OFDM's error correcting capability?
I saw the same thing when I first started testing the SR9 cards. It seemed that 10mhz worked a lot better than 5mhz.
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Sun Apr 22, 2007 12:48 am

With signal in the -70's, seems I gained signal when going to 5MHz, I'll have to do some more testing in open area. Being in a metal building is going to do some odd stuff as well.
 
rcm
just joined
Posts: 10
Joined: Wed Oct 11, 2006 12:31 am

sr9

Wed Apr 25, 2007 7:40 pm

We have 2 Mikrotik AP 532's with SR9 in middle TN close to AL ... So far the AP is running pretty good with several customers on it. The best link so far is with a rb112 and a 6db diversity antenna about a mile away from the Tower @ a -59 rssi inside the house. We've had customers up to 5 miles NLOS, so LOS may be possible further out of course. LOS with sr9 is great of course, NLOS depends on what obstacles are in the way and how far out the customer is. The water tower we are on is about 100 ft tall. Of course you are going to have to deal with interference with sr9, but you are also going to have to deal with interference even with 2.4 and 5.8 .

We are using 5mhz channel spacing to get away from interference from other applications in the area. I haven't tried 10Mhz or 20Mhz, and I don't plan on trying that channel spacing because of interference. Cell towers and Paging towers are below and above the spectrum we use. I don't like that Ubiquiti only has 20 Mhz cavity filters because all I need is 5.5mhz filter to filter out everything else other than my center frequency. That would help out from self interfering as well. We've noticed that in the 900 mhz spectrum, most of the equipment put up by other businesses are on Vertical Polarization. We chose to go with HPOL on our new capsite to get away from that. I believe somebody was telling me you'll get about 10db worth of separation which really helps. Using HPOL sectors with a good front to back ratio is also a plus.

Sr9 is a good solution for what we deal with when it comes to terrain in TN. Rolling hills and a bunch of tress makes it difficult to use 2.4 or 5.8 without setting up a lot of cap sites ( mini pops )

Some of you have already mentioned about Pine trees and you are right. Pine tress obstruct the signal and we have noticed that. Everytime we go to a house to set up a link we always hate to see pine trees. They are the only ones that we have noticed that will really put a hamper on our 900mhz signal. We are greatful that they are not everywhere in this part of the country.
 
User avatar
tgrand
Long time Member
Long time Member
Posts: 667
Joined: Mon Aug 21, 2006 2:57 am
Location: Winnipeg, Manitoba, Canada

Thu Apr 26, 2007 1:46 pm

Yes a 5.5MHz cavity filter would be a nicety.
If anyone knows where one can find a cavity filter of this nature, it would be greatly appreciated.

If you notice on the 5MHz channels there are 2 side bands which are essentially wasted energy and also noise to others.
It would be great to filter thos bands so that the amps do not over work and create noise.
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Mon May 14, 2007 4:27 pm

Did some more testing this weekend. Got one half of the bridge on up about 125ft 3.25 miles away.

At ground level I had around -88 in one location a hundred feet behind my home off in the woods. I changed from 10MHz channel to 5MHz and got -84, inconsistant pings and could barely get into the other rb133c to make changes.

I mounted my yagi on the basket of a boom truck, got it up around 50ft, got a -79 on my side, -89 on the tower. Changed channels, got it to -79 and -84. I went about 50ft up in the basket, I can see 6+ miles north, probably around 5 miles west. My shot to the north east is blocked by pine that is within 200ft of my home. Pings are decent, 6-9ms for about 10 pings, then a jump to 200+ms, but the link is there.

If everything goes as planned, I'll have a temporary shot up this afternoon with my yagi about 45-50ft up a 65+ft tree.
 
palmczak
Frequent Visitor
Frequent Visitor
Posts: 78
Joined: Sat May 29, 2004 7:53 pm

Wed May 23, 2007 12:11 am

These guys will build a cavity to your specification. They build the cavity filers for a proprietary 900MHZ system we use. This system uses a 5.5mhz wide channel. They work extrememly well.

http://www.dci.ca/

Yes a 5.5MHz cavity filter would be a nicety.
If anyone knows where one can find a cavity filter of this nature, it would be greatly appreciated.

If you notice on the 5MHz channels there are 2 side bands which are essentially wasted energy and also noise to others.
It would be great to filter thos bands so that the amps do not over work and create noise.
 
User avatar
tgrand
Long time Member
Long time Member
Posts: 667
Joined: Mon Aug 21, 2006 2:57 am
Location: Winnipeg, Manitoba, Canada

Re: SR9 distance 5MHz vs 10MHz

Mon May 28, 2007 2:32 pm

Just for the record:

Using 802.11g as a reference of 0db

20MHz wide channel = 0 db (802.11g)
+1 db (802.11b)
10MHz wide channel = +3 db
5MHz wide channel = +6 db

However the strength of "B" is not significantly stronger than "G".
"B" is however more susceptable to signal distortion, due to interferers such as obstacles, and competing signals.

CORRECTION....
"B" is however less susceptable to signal distortion
 
User avatar
Adam McLaughlin
Member Candidate
Member Candidate
Posts: 149
Joined: Sat Mar 31, 2007 5:36 am
Location: Santa Rosa, California
Contact:

Re: SR9 distance 5MHz vs 10MHz

Sat Jun 23, 2007 8:00 am

When you guys tried to configure this card with your RB133C, did the Wlan1 interface look just like a R52?

I put a SR9 into my RB133C, and I see nothing that would indicate 900 Mhz operation...

Adam
 
User avatar
ghmorris
Member Candidate
Member Candidate
Posts: 284
Joined: Fri May 28, 2004 12:14 pm
Location: Minden, Ontario, Canada
Contact:

Re: SR9 distance 5MHz vs 10MHz

Sat Jun 23, 2007 11:21 pm

Correct. You just have to pick the correct 2gig - 5Mhz operation, then one of the four frequencies that an SR9 can operate on.

Works great! We have well over 250 in the field.

George
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Re: SR9 distance 5MHz vs 10MHz

Sun Jun 24, 2007 1:43 am

Just to follow up, I'm running in B mode with the link up for a few weeks now. The interference doesn't knock out the link now, just makes it slow down a bit.

The connection has run with a SNR of 1dB on occasion, average signal is -86. Yes, its marginal, but 5MHz G mode dropped out completely when the noise showed up, even with -79 signal. B mode has only dropped for about 2 minutes, never completely, just lost connection and reassociated. Rain didn't seem to cause any problems, in fact, my CCQ went from 60's to 90's during and imediately after a nice rain came through.
 
jarosoup
Long time Member
Long time Member
Posts: 596
Joined: Sun Aug 22, 2004 9:02 am

Re: SR9 distance 5MHz vs 10MHz

Sun Jun 24, 2007 2:46 am

I'm a little confused with b/g mode and b-only mode when using a width of 5 or 10MHz. There are only 2 options for 5 and 10MHz widths, 2ghz-5mhz and 2ghz-10mhz. Using either of these appears to be b/g mode as they use the a/g rates either default or defined. How can one use the SR9 in b-only mode with 5 or 10mhz widths?

As for our personal experiences with the SR9, it is a good solution if there isn't a lot of interference around. We're only able to get about 5Mbps at 10MHz in the field, but on the bench they will do 10Mbps. We had a hard time getting 5MHz to work right, but I think this was due to existing 900MHz interference in the air.
 
User avatar
Adam McLaughlin
Member Candidate
Member Candidate
Posts: 149
Joined: Sat Mar 31, 2007 5:36 am
Location: Santa Rosa, California
Contact:

Re: SR9 distance 5MHz vs 10MHz

Sun Jun 24, 2007 3:45 am

I have a unique situation myself.

I setup a client with an EOC-3220 a few months back when the A.P. was a Prism card in a BriLan board. Everything was fine and dandy, all rosy until I went and swapped out that A.P. with a RB532 / 564 with a XR2.

Now, the client's CCQ is about 15 to 18, and he can barely surf.

Everything was great with the EOC-3220 and the prism unit. Signal levels were in the low 70s on both sides. The signal levels did not change when we swapped over to the RB532 / 564 / XR2 at the A.P.

The client is separated from the A.P. by less than a mile...

Well, no matter what I could do, I could not bring the CCQ up to something respectable. Tried moving the antenna, the radio, sniffing all around the property, etc.

So, I am looking to swap the guy to 900 Mhz to see what I can do to bring his life up to par. I bought a SR9 for each side. Going to put the omni at the A.P., and a grid at the client side. Going to use a RB133C at the client side as well.

Hoping to see what I can do for the guy - there are many trees in the way between him and the A.P. this is the forest, after all.

Thanks for the advice to get me started on this guys. I will pick a 10 Mhz wide channel for starters. There is not any interference out there in the wild woods of this Forest.

Adam
 
0ldman
Forum Guru
Forum Guru
Topic Author
Posts: 1465
Joined: Thu Jul 27, 2006 5:01 am

Re: SR9 distance 5MHz vs 10MHz

Mon Jun 25, 2007 5:32 pm

I'm a little confused with b/g mode and b-only mode when using a width of 5 or 10MHz. There are only 2 options for 5 and 10MHz widths, 2ghz-5mhz and 2ghz-10mhz. Using either of these appears to be b/g mode as they use the a/g rates either default or defined. How can one use the SR9 in b-only mode with 5 or 10mhz widths?

As for our personal experiences with the SR9, it is a good solution if there isn't a lot of interference around. We're only able to get about 5Mbps at 10MHz in the field, but on the bench they will do 10Mbps. We had a hard time getting 5MHz to work right, but I think this was due to existing 900MHz interference in the air.
I have not been able to use B speeds in 5 or 10MHz mode.

Who is online

Users browsing this forum: nickcarr and 25 guests