Wireless Bridge Help

Answered Question
Apr 21st, 2012

Hi all,

I have a Wireless P2P bridge set up between two buildings.  They are about 450 feet apart, Root bridge has a 16dbi Yagi antenna (ebay) and nonroot has a 12dbi Flat Panel antenna.  Both bridges are 1231G APs set to Root and Nonroot mode, respectively.  I have a few questions.  Right now the link is working and is reliably connected.  My data rates right now are 1.0  ,  2.0  ,  5.5  ,  6.0  ,  9.0  ,  11.0  ,  12.0  ,  18.0  ,  24.0  ,  36.0  ,  48.0  ,  54.0                                   Mb/sec on both sides.  I consistenly get on the lower end of the scale, usually hovering around 5.5-12mb/sec range on the nonroot side.  Root is typically transmitting at 48 or 54 mb. 

My signal should be pretty good at -61 and 34-37 SNR, right?

At work we have a wireless bridge (1310 hardware) and it is set to :::speed  basic-9.0 basic-12.0 basic-18.0 basic-24.0 basic-36.0 basic-48.0 basic-54.0::: so I went into mine (1231G hardware) and set it to the same.  Voila!  I connected at 24-54mb.  All seemed well - until I noticed that now that I wasnt using the B rates, I had a spotty connection that was going up and down.  Changed it back to all rates, and I again have a stable connection. 

What can I do to stabilize and set the connection to OFDM rates so it will be fast and reliable?  The distance really shouldnt be an issue, and I can see from one antenna to the other.  Would 2 panel antennas be better?  Should I find higher gain antennas?  Should I switch to the 1310 hardware because it is designed as a bridge?  HELP!!! 

Here is the output from the Root side:

Root#sh dot11 associations all-client

Address           : 0014.a86f.2b50     Name             : NonRoot

IP Address        : 10.10.81.3         Interface        : Dot11Radio 0

Device            : 11g-bridge         Software Version : 12.3

CCX Version       : NONE

State             : EAP-Assoc          Parent           : self              

SSID              : WT3342NET

VLAN              : 0

Hops to Infra     : 1                  Association Id   : 160

Clients Associated: 1                  Repeaters associated: 0

Tunnel Address    : 0.0.0.0

Key Mgmt type     : NONE               Encryption       : WEP

Current Rate      : 54.0               Capability       : WMM ShortHdr ShortSlot

Supported Rates   : 1.0 2.0 5.5 6.0 9.0 11.0 12.0 18.0 24.0 36.0 48.0 54.0

Voice Rates       : disabled

Signal Strength   : -61  dBm           Connected for    : 925 seconds

Signal to Noise   : 34  dB            Activity Timeout : 28 seconds

Power-save        : Off                Last Activity    : 2 seconds ago

Apsd DE AC(s)     : NONE

Packets Input     : 7521               Packets Output   : 4241     

Bytes Input       : 2893219            Bytes Output     : 650362   

Duplicates Rcvd   : 0                  Data Retries     : 1375     

Decrypt Failed    : 0                  RTS Retries      : 0        

MIC Failed        : 0                  MIC Missing      : 0        

Packets Redirected: 0                  Redirect Filtered: 0        

Session timeout   : 0 seconds

Reauthenticate in : never

And here is the output from the NonRoot side:

NonRoot#sh dot11 associations all-client

Address           : 0013.c365.6190     Name             : Root

IP Address        : 10.10.81.2         Interface        : Dot11Radio 0

Device            : 11g-bridge         Software Version : 12.3

CCX Version       : NONE

State             : EAP-Assoc          Parent           : Our Parent        

SSID              : WT3342NET

VLAN              : 0

Hops to Infra     : 0                  Association Id   : 143

Tunnel Address    : 0.0.0.0

Key Mgmt type     : NONE               Encryption       : WEP

Current Rate      : 5.5                Capability       : WMM ShortHdr ShortSlot

Supported Rates   : 1.0 2.0 5.5 6.0 9.0 11.0 12.0 18.0 24.0 36.0 48.0 54.0

Voice Rates       : disabled

Signal Strength   : -60  dBm           Connected for    : 1012 seconds

Signal to Noise   : 37  dB            Activity Timeout : 15 seconds

Power-save        : Off                Last Activity    : 0 seconds ago

Apsd DE AC(s)     : NONE

Packets Input     : 14106              Packets Output   : 7646     

Bytes Input       : 2176272            Bytes Output     : 2908380  

Duplicates Rcvd   : 0                  Data Retries     : 3136     

Decrypt Failed    : 0                  RTS Retries      : 661      

MIC Failed        : 0                  MIC Missing      : 0        

Packets Redirected: 0                  Redirect Filtered: 0 

I have this problem too.
0 votes
Correct Answer by George Stefanick about 1 year 11 months ago

Jason,

The wifi on 2.4ghz is pretty loud there. Were you up near the bridges when you took these readings. We should always be 20 dbm difference between us and other competing wifi. If the other wifi is at 60 already this means we need to be at 40 which is a lot of effort. Remember wifi is half duplex. Only 1 device can talk at a time.

Lets try this before you do anything else. After seeing the 2.4 GHz from your pics. Lets send smaller frames over the bridges. See the attached link to show how.

In theory, when you send traffic, lots of traffic. These larger frames get crushed by interference more often then smaller frames. If we make smaller frames and your experience improves we can then prove this of a interference issue and gives a sound reason for 5 GHz.

http://www.cisco.com/en/US/docs/wireless/access_point/1300/12.3_7_JA/configuration/guide/b37rf.html#wp1062530

Try the smalls frame and then start your way up and see if your performance improves.

I haven't used the ubiquiti. Although I have used http://www.tranzeo.com/ 5GHz bridges.

  • 1
  • 2
  • 3
  • 4
  • 5
Average Rating: 5 (23 ratings)
George Stefanick Sat, 04/21/2012 - 08:59

Hey JHoffmann,

Im a get back to basic kinda guy. Lets talk through this shall well. Im sure Scott, Steve and Leo may chime in as well for added input.

1. 450 ft is a very short distance indeed. These bridges can shoot miles.

2. It is always best pratice to use identical antennas on both side. You have a significant mismatch. 12dBi vs 16 dBi. What degree of antennas are these? You should get a nice panel with a 90 degree or better yet a yagi.

3. It looks like you are using dot 0.. Which is the 2.4GHZ side. Is their interference in the area? Does your aps have 5 ghz radios, have you tried doing your bridge on 5 GHz?

4. What is your TX power set at on both bridges?

jhoffmann015 Sat, 04/21/2012 - 09:45

George - thanks for getting back to me!  I really want to figure this out, and having a similar setup at work (albeit with 1310 hardware which I believe has built-in flat panel antennas but not sure of the db rating), and having that bridge stable around 36-54mb, I want to try and get this as stable as possible.

I have a yagi on the root side, and the panel on the nonroot side.  The yagi is a cheap one from ebay which I can switch out, and the flat panel patch antenna is from L-com,2.4ghz 14dBi, not 12 as I had thought.  The pattern is 30x30 degrees.  ( http://www.l-com.com/item.aspx?id=22114 ).  The yagi I have no idea what the degree is on it, I sort of just got excited that it worked all the way at the other house.

3. I only have one radio installed in the AP, the 2.4 G radio.  No 5 here.

4. Transmitter Power CCK / OFDM - 100 mW / 30 mW

For now, I'd like to see what I can do with the hardware and antennas I have, and I can put in an order for another panel antenna so they both match.  It is my understanding that it's easier to aim 2 panel antennas than 2 yagis, correct?

Thanks for all your help - if you need any other info or sh run output let me know.

Thanks!

Jon

George Stefanick Sat, 04/21/2012 - 09:59

Jon,

First lets do a little RF math, shall we...

If your radio is TX at 100mW and you add the following antennas lets look at the ERP.

100mW + 12 dBi = 1600mW !

100mW + 16 dBi = 3200mW+ !

Your laptop, just as an example, TX at 40mW.

So you have one side yelling 2x as loud as the other side. Not to mention, these are VERY VERY VERY loud.

1) I would get these both on the same link budget.

a. Power the 12 dBi guy down to 12.5 dBm lets look at that math after this change:

     - 12.5mW + 12 dBi = 200mW

b. Power the 16 dBi guy down to 6.25 dBm lets look at that math after this change:

     - 6.25 + 16 dBi = 200mW (there around you have an extra dBi there, but thats fine)

You will see now you have a closley configured link budget.

2) At 450 feet either a yagi or panel will work. Do you have any obstructions between these 2 bridges, like trees, houses etc ?

jhoffmann015 Sat, 04/21/2012 - 11:11

I have very clear line of sight but there is a small gangly tree close to one house.  I can see through it, and judging by what I read and looked at regarding Fresnel zones, it seems like it is close enough not to really make too much of a difference.  I will power both down and see what I can get - and report back in a few.

thanks

jhoffmann015 Sat, 04/21/2012 - 11:52

Configs are attached - changing the TX power I get similar results - sometimes it likes to connect at 54mb, but usually hovers around 6-12.  Download speeds at the root side (on my actual network) are 25mb down, 4.5 up.  Download speeds on the Nonroot (remote) side are 5.5 down /3.5 up.

What is the difference between CCK and OFDM?  in our 1310 setup at work we only use

speed  basic-9.0 basic-12.0 basic-18.0 basic-24.0 basic-36.0 basic-48.0 basic-54.0

when I try to set that in mine, that's when the connection drops and becomes unstable.  Am I missing something else? 

George Stefanick Sat, 04/21/2012 - 12:51

CCKM - OFDM is the modualtion.

CCKM modulates 1,2,5,11 and OFDM does 9,12,18,24,36,48,53

Well you are not allowing the lower rates. The problem with that, if you do that to your setup your bridges will break. Becuase you wont support the 5,5 that your bridge is at now. Make sense?

Are you sure the antennas are connected properly? Are they attached to the "primary" buld head?

jhoffmann015 Sat, 04/21/2012 - 14:21

Is there no way even if I set it at both ends to just support 9 12 18 24 36 48 and 54?   This link is for file transfer and internet access, so Id like it to be as snappy as possible.  I will order a matching panel antenna for my side - they seem to be more forgiving to alignment too, correct?  I do have the antennas connected to the Primary connector on the APs.  From there I lose about 3-4db through the run of cable (LMR400 @ 50'), is that of much concern at this distance?

I am glad it connects and is stable at lower speeds, but I wanted to get it running at full throttle so to speak...should I be doing anything different on the "core" switch it's connected to?  (3550)  I just have the port for the root bridge in th 81 VLAN as an access port.

jhoffmann015 Sat, 04/21/2012 - 14:26

should I let it scan all available frequencies?  I have it set to 1, 6, and 11, but when I set it to all, it picked 8.  Does that make a difference?  Sorry for all the questions - this depth of learning for RF is new to me.

Scott Fella Sat, 04/21/2012 - 14:32

Don't let it scan.. Use 1, 6, or 11. The 50' of LMR is fine. Is this a new setup? The antenna is not filled with moisture or the cable not in coax seal? Sometimes you can lower the TX little by little until you loose connection. This gives you an idea of you min TX power. So if you hard set the mandatory rate to 54mbps, you drop the link.

Thanks,

Scott Fella

Sent from my iPhone

George Stefanick Sat, 04/21/2012 - 14:32

Well when you bridge they both need to be on the same channel.

you could download like ssid inssider to see what your wifi looks like and pick a channel that is less occupied.

George Stefanick Sat, 04/21/2012 - 14:31

Yea the panels are more forgiving. But the distance is so close you could probably point it in the opposite direction and the back lobe would hit it.

These are things I would look at.

- Change the channel and see if that helps. Maybe there is interference on one end of the bridge. So if you are on channel 1 go to channel 11. Are you close to other neighbors with wifi networks?

- Check the antenna connectors

- Make sure the antennas are conncted to the primary bulk head

- Get matching antennas

jhoffmann015 Sat, 04/21/2012 - 14:36

When I get home tonight I will change the channels and see - and yes, we are in the city so there are a handful of closeby networks.  I will also check the connectors and put an order in for a matching panel.  In a moment of frustration I also picked up 2 1310 bridges off fleabay for under 200.00 for the pair.  If this works out I may just sell them back though. 

Should I still allow all data rates, or just not set lower ones to "disable"?

George Stefanick Sat, 04/21/2012 - 15:22

If you disable the lower rates, your connection breaks, right?

So what these rates are, they are called PHY rates. If they drop from 54 to 5 for exmaple, a number of things like interfrence, link budget issues, antennas problems etc ..

jhoffmann015 Sat, 04/21/2012 - 15:31

right, if I only use OFDM rates, the connection gets very unstable.  With the CCKM also enabled it is great- never loses association.  I will adjust tonight and look into the channel availability as well.

Scott - no water in the antennas, but no coax seal either.  I ordered some today so I will put that on the connector end when it comes in.  They are, however, finished with shrink tubing and silicone to keep the water out.

Again, thanks for all the help guys - I really appreciate it!

Scott Fella Sat, 04/21/2012 - 16:25

One thing with antennas... At least outdoor antennas, there is a moisture hole built in and it requires the antenna to be mounted the correct way. Also make sure you have a drip loop on the antenna cable to prevent water from traveling toward the connector. You can also verify the cable is crimped correctly by taking a multimeter an touching the inner wire and outer shield. There should be no connectivity between the two. Of course the antennas need to be disconnected.

Thanks,

Scott Fella

Sent from my iPhone

jhoffmann015 Sun, 04/22/2012 - 09:24

Both antennas are in Right / Primary jack, the cables are all brand new and test open from the core to the shield.  I sealed up the connectors for now with silicone and will re-do it with the coax seal when I get it. 

Scanning with KisMAC, I found that of 1, 6, and 11, there were only 2 APs broadcasting on ch. 1, so I selected that on both sides.  I also enabled all data rates on both sides.  It still fluctuates wildly from 5.5 - 54mb so I can only assume theres something funky with the crappy ebay antenna on my side.  Panel antenna ordered, will see if that changes anything.  In the meantime I also ordered 2 1310 bridges and will try those in the case that the panel antenna does not change things (which I imagine it will).

Anyone know why I cant run dot11 dot11radio 0 antenna-alignment ?  I thought the command was able to run on any non-root bridge?  When I do a sh antenna-alignment, I get the message that it hasnt run yet.

Any ideas on that one?

jhoffmann015 Tue, 04/24/2012 - 10:04

OK, I can say comfortably that I see 54mb link between the bridges, with the occasional drop to 36 or 48.  My internet connection at the root side (on the core switch) is 25mb down, 5 up.  On the non root connected network I'm only getting 7-8 mb down, and 3-4mb up.  Is this just because of the nature of the bridge, or should I see the full 25 since I'm connected at 54?  Is the true throughput of the 54mb link actually 7 megs?  I know the authentication has some overhead but I didnt think it was that much? 

George Stefanick Tue, 04/24/2012 - 10:12

So what did you do to improve the performance ?

Are you indeed getting 25/6 on the root side, you tested this ?

You should see close to that on the non-root time.

jhoffmann015 Tue, 04/24/2012 - 14:00

I changed the channel on both sides to 1 (the less congested of all wifi networks in our area), and the link seems pretty stable around 36-54mb. 

As for throughput, on the nonroot side off a wire or wireless connection, I am seeing 22.95/4.20 (see attached).  On the nonroot side I am seeing 6.35/4.13 (see attached)

Could it be the antenna mismatch?

George Stefanick Tue, 04/24/2012 - 14:32

Good so my idea of checking the channels was helpful.

Well anything is possible, but I wouldnt think so much. If you are keeping stable on both ends at 36-54. I suspect there could be congestion, maybe. Other interferece ?

Have you tried perhaps channel 11 or 6 ?

jhoffmann015 Tue, 05/01/2012 - 23:35

Hi George,

Tonight I installed a matching panel antenna (l-com HP2414) on my side, and have similar results (actually worse results at first).  I upped my side transmit to CCK 30 and OFDM 30 from 20/20 and now have about the same performance (see red below).  I have a question - how important is it to have the wifi antenna clear?  I have it mounted on the same mast as our TV antenna- see pic attached.  Should I move it to it's own mast?  I can mount another one on a second chimney if necessary, or raise the TV antenna up a few feet. 

I am also not past sticking a yagi at both sides if that will improve my signal to noise ratio...connection still is around 36-54mb, but download speeds on the other side are still only 6mb, not the full 25 we get on the wire off the core, or even the 16 I got on the test nonroot bridge I set up at the same side as the root.

NonRoot#sh dot11 associations all

Address           : 0013.c365.6190     Name             : Root

IP Address        : 10.10.81.2         Interface        : Dot11Radio 0

Device            : 11g-bridge         Software Version : 12.3

CCX Version       : NONE

State             : EAP-Assoc          Parent           : Our Parent        

SSID              : WT3342NET

VLAN              : 0

Hops to Infra     : 0                  Association Id   : 199

Tunnel Address    : 0.0.0.0

Key Mgmt type     : NONE               Encryption       : WEP

Current Rate      : 54.0               Capability       : WMM ShortHdr ShortSlot

Supported Rates   : 1.0 2.0 5.5 6.0 9.0 11.0 12.0 18.0 24.0 36.0 48.0 54.0

Voice Rates       : disabled

Signal Strength   : -65  dBm           Connected for    : 9517 seconds

Signal to Noise   : 32  dB            Activity Timeout : 15 seconds

Power-save        : Off                Last Activity    : 0 seconds ago

Apsd DE AC(s)     : NONE

Packets Input     : 95379              Packets Output   : 10182    

Bytes Input       : 15774260           Bytes Output     : 2163933  

Duplicates Rcvd   : 0                  Data Retries     : 7927     

Decrypt Failed    : 0                  RTS Retries      : 168      

MIC Failed        : 0                  MIC Missing      : 0        

Packets Redirected: 0                  Redirect Filtered: 0        

jhoffmann015 Tue, 05/01/2012 - 23:36

tomorrow I will check channels 11 and 6 and see what that test yields...

George Stefanick Thu, 05/03/2012 - 08:44

Try this swap your bridges around. And see if the problem follows.

Sent from Cisco Technical Support iPhone App

jhoffmann015 Fri, 05/04/2012 - 21:11

George,

So far these are the steps I've followed:

1. Purchase and install matching 14dbi panel antenna - check.

2. Seal connections with CoaxSeal - check.

3. Set channel to 1 - check.

4. replace 1200 APs with 1310 bridge hardware - learned the hard (read:long) way that the -R after the model meant there was no built-in antenna in this model.....check.

5. Dumped same config on the 1310 bridges - check.

6. New 1310s associate at a little better s/n ratio, but still same performance.

Here's what I will do tomorrow:

1. Check performance on Ch. 11 and 6

2. Back power down to a lower transmit

3. check antenna alignment (will have to wait till Monday morning for this one).

Any other suggestions?  What are your thoughts to being so close to another (very large) TV antenna (see pics in post above).

jhoffmann015 Sat, 05/05/2012 - 23:15

Performance on ch 6 and 11 was worse than 1 - so back to 1 we went.  Power was dropped from 100/100 to 50/50mw, with a 14dbi antenna on either side.   Watching the signal level, this seems to get me the best signal without dropping.

Same results.

I ran a wifi scan at the nonroot side, and boy is there a ton of wireless over there - more than on my side.  While I think this may have something to do with it, there is not much using ch. 1, most is on 6, and most signal levels were -88 to -93db.  At this point, I am assuming it has something to do with the antenna alignment, which hopefully I can check on Monday morning.

Will report back. 

On our bridge at work we have hold-queue 80 in in the config - is this important and should I add it?  Reading up it seems as though I may want to add this line in, but I wanted to get your opinion first if it was important enough.

Thanks - hopefully I wont have to settle for 1/4 the down speed!  If I align the antenna better and still get the same results, then I will leave well enough alone - after all it is good enough for web browsing, but I would like to speed up file transfers between the two sites...

George Stefanick Sun, 05/06/2012 - 12:04

So did you swap the bridges by chance? Did this issue follow or did it still stay the same. I think based on everything you did, good job btw, this may be a medium congestion problem.

How many wireless networks do you see. Can you use a tool like http://www.metageek.net/products/inssider/ and take a peek at how many wifi networks you see on each side?

If this is a medium problem the only other choice is to go 5 GHz which is the 1400 bridges.

jhoffmann015 Sun, 05/06/2012 - 15:12

I did not swap the bridges because I changed out the hardware altogether. To be clear, your thinking may be if I switch the station roles on either side, it may help? Does it matter which end is root and non-root?

I did do a wifi scan with INSSIDER but need to go back and do two seperate ones at each location.

If antenna alignment is the issue, I am not going to go with the 1410 bridge unless I come into 6,000.00 for the pair of them, so if I can indeed not get this setup to work more reliably, I will be looking into some of the Ubiquiti products - they have some very promising looking products that are at an awesome price point; I just hope that I dont get what I pay for should I go this route, and I am not clear on whether or not they support multiple vlan trunking from a cisco switch. More research I suppose...

Here's to hoping alignment solves enough of the issue to become more reliable!

More tomorrow night!

Thanks again

George Stefanick Sun, 05/06/2012 - 20:54

Ok, yea if you swapped out the bridges, then yes ignore that 'swapping' comment. Lets see what the inssider data looks like when you collect it.

jhoffmann015 Mon, 05/07/2012 - 14:29

Here is a link to the pics I took from speed tests at different channel settings (see filename for channel). 

https://plus.google.com/photos/113835363852808697523/albums/5739905011320472641?authkey=CNKx2JW7kc2Pcw

INSSIDer is in there as well - 2647 is the root side, 3342 is the nonroot side.  I actually got the best (8+mb down) results with ch. 11.

Would a yagi at either end be better?  It is a pretty clear straight shot from one end to the other. 

I may try my 1200 APs with the 802.11A add-on card, and just swap out the panels for 5ghz ones.  Thats the cheaper way, if that still doesnt yield results I may give the ubiquiti stuff a shot - do you have any experience with them?  From all accounts, its a really nice product, and lots of bang for the buck.

Take a look at the pics/results, and let me know your thoughts.  File transfer from a server on the root side to a client on the nonroot side was ~1mb/sec...

All better speeds, though...

      

EDIT:  To see filenames of the pics, select Details on the right hand side while viewing the picture.

Correct Answer
George Stefanick Mon, 05/07/2012 - 20:35

Jason,

The wifi on 2.4ghz is pretty loud there. Were you up near the bridges when you took these readings. We should always be 20 dbm difference between us and other competing wifi. If the other wifi is at 60 already this means we need to be at 40 which is a lot of effort. Remember wifi is half duplex. Only 1 device can talk at a time.

Lets try this before you do anything else. After seeing the 2.4 GHz from your pics. Lets send smaller frames over the bridges. See the attached link to show how.

In theory, when you send traffic, lots of traffic. These larger frames get crushed by interference more often then smaller frames. If we make smaller frames and your experience improves we can then prove this of a interference issue and gives a sound reason for 5 GHz.

http://www.cisco.com/en/US/docs/wireless/access_point/1300/12.3_7_JA/configuration/guide/b37rf.html#wp1062530

Try the smalls frame and then start your way up and see if your performance improves.

I haven't used the ubiquiti. Although I have used http://www.tranzeo.com/ 5GHz bridges.

jhoffmann015 Mon, 05/07/2012 - 20:27

Well.........this tells a very congested story as well...

Root BridgeNonroot Bridge

Root#dot11 dot11Radio 0 carrier busy

Root#sh dot11 carrier busy

Frequency  Carrier Busy %

---------  --------------

   2412         50

   2417         32

   2422         24

   2427         12

   2432         22

   2437         42

   2442         15

   2447         12

   2452         12

   2457         17

   2462         16

NonRoot#dot11 dot11Radio 0 carrier busy

NonRoot#sh dot11 carrier busy

Frequency  Carrier Busy %

---------  --------------

   2412         40

   2417         17

   2422         16

   2427         14

   2432          9

   2437         32

   2442         16

   2447         13

   2452          7

   2457         12

   2462         20



Maybe time for 5ghz?  I'm the only one in both areas (1140 AP) broadcasting ANY SSID on 5ghz band...I ordered a pair of 1242AG APs to test this theory out.  I understand the 5.8 antennas are bigger and need to be stronger?

George Stefanick Mon, 05/07/2012 - 20:36

5ghz doesnt shoot as far. But in your case you are so close it shouldnt be an issue. Try sizing down the frames and see what happens.

jhoffmann015 Mon, 05/07/2012 - 20:39

not to sound ignorant but how do I size down the frames? 

edit:  conf t / int dot1 0 / mtu ?

what size should I use?  Match the Ethernet frame size ~1500mtu?

I am looking at a 21-23dbi antenna for the 5.8 radios to test with (panels again)

George Stefanick Mon, 05/07/2012 - 20:41

Click on that link. Its call Configuring the Fragmentation Threshold.

Thats more than enough dbi ..

jhoffmann015 Mon, 05/07/2012 - 20:44

do I need to do this on both radios (root and nonroot bridges) on the dot1 0 interface so that they match?

George Stefanick Mon, 05/07/2012 - 20:49

No actually do this on the side having issues. when the bridge sends frames, it will chop them down smaller.

jhoffmann015 Mon, 05/07/2012 - 21:32

ok - on the root side, I dropped the fragmentation Threshold down to 1500 and I now get almost 10m down/ 4m up.  Much better than yesterdays 5-6m down and 2-3 up.

Between re-aligning the antennas, playing with the power, and changing the Fragmentation Theshold, we've gone up quite a bit from the initial 5.5 down and 3.5 up.

Can you point me in the direction of a good tool to calculate link power (ERP I think you used in the second post)?  I'd like to re-calculate now that I'm using matching 14dbi antennas... 

George Stefanick Mon, 05/07/2012 - 21:39

Look at you big guy! NICE!

These are the times I wish we had RF glasses and actually see whats going on in the air.

You are so close, personally, so long as your links are within 3 - 6 db i would leave it alone. In fact, at 14 dBi and the power your bridges make you are WAY WAY powerful! You could shoot those things for a few miles.

If you want to get nerdy start to power the bridges down and test the links and see where your sweet spot is. I dont think you will see much of a change.

if you find any of this helpful rate a few of the post. Im work on my new color star  and trying to catch up to steve on the ratings board! LOL

jhoffmann015 Mon, 05/07/2012 - 21:52

I may get nerdy with it but for now this will work.  RF Glasses would be extremely helpful now.

Here's what I did (and I did actually hit 10m mark!)

fragment-threshold 1500

rts threshold 4000

beacon period 50

Tomorrow I will test file transfer from our fileserver; hoping for more than 1mb there....

So I'm at about 9.3-9.8mb down right now; changing the Frag helped a lot which points to interference correct?  Sounds like a good reason to get even nerdier, add to my pile of Cisco APs, and go for the 5.8 ghz 1142AGs with new panel antennas. 

If you couldnt tell already I just can't leave well enough alone...lol

This has been the most insightful and helpful experience - and thank you for not treating me like a moron when I didnt understand something like is so common in forums now.  Once I get the 5ghz equipment I will post back here and update with the new results (probably not for another week or so).

Thanks again!

George Stefanick Mon, 05/07/2012 - 22:02

Ive been on the CSC for a few years, which is unpaid btw, we all volunteer our time to help others. So no worries you wont get that treatment here. We all know the pain, because we have all been there ...

A few things ..

Correct, if you frag the frames and it improves performance that is indeed indicative of interference. Again, if you play with the sizing, you may see even better performance. Do some test. Send a ftp at different thresholds and compare performance.

Before you run the 5 GHz, run INSSIDER on 5 GHz to make sure its clean, which it will be. There are a ton of channels on the 5 GHz side. But check anyway.

For testing you could use the same 2.4 panels on the 5ghz .. Its not optimal but you are so close it may not really make a difference for testing

jhoffmann015 Mon, 05/07/2012 - 22:15

When I ran INSSIDER I did look at the 5.8 scans, and there was zip zero nothing there - just my own AP on the root side (a 1140).  So I think that will be much better.  I will have the APs before the antennas, so I will test with the existing panels.  We'll see what happens....

When I can get over to the other (nonroot) side tomorrow I will run some FTP tests, for now I have been just running OOKLA speedtests each time I make a change and noting the difference.  I did play with a few different rates for frag, and seems that this is the sweet spot.

blakekrone Mon, 05/07/2012 - 22:17

Good stuff in this thread! Big kudos to George for keeping at it.

I still think though you are using too high powered of antennas for such a short link. Remember that turning up the power doesn't fix the issue, it hurts. If I start shouting in a crowded room that doesn't mean George can hear me any better (he is a lilttle hard of hearing, jk!!), he still needs to process the background noise.

If you were going to try 5GHz then I would try 1242's with the AIR-ANT5170P-R antennas or 1262's and AIR-ANT5160NP-R= antennas.

jhoffmann015 Mon, 05/07/2012 - 22:25

I actually am on 30mw for both sides and am losing ~3-4db between the cable and connectors (LMR400 but at the antenna end there's a 1.5ft RG58 adapter pigtail.  So from what I am understanding, I basically have a 9-10dbi antenna after cable loss.

Powering up from 30mw does not help so I left it alone at that.  Funny that Cisco docs suggest maxing it out...

Huge kudos to George for not abandoning me!  I've learned a lot about RF tuning from this and will always remember it!

I will look into the AIR-ANT5170P-R  antennas, for the most part I was going L-com because Cisco antennas are so friggin expensive...

I was looking at http://www.l-com.com/item.aspx?id=22369 but they also make an 11dbi antenna, maybe I'll go that route instead.

blakekrone Mon, 05/07/2012 - 22:28

While I agree they are expensive you do get what you pay for. I've been to the labs Cisco uses for their antenna development, they take this very seriously and their specifications are very precise.

Just as an FYI I have seen issues with using non-Cisco branded antennas when using WCS to manage the controllers.

jhoffmann015 Fri, 05/18/2012 - 05:57

Alright guys, I have received the new AP1242AG Access points, and configured them.  They work with little rubber ducky 2.4ghz antennas but not with the 2.4 panels.  I have ordered 5ghz panels@19dbi each, and will post back when they come in.

Thanks for everyone's help with this - I feel like I'm closer now to full speed ahead!

Actions

Login or Register to take actions

This Discussion

Posted April 21, 2012 at 1:41 AM
Stats:
Replies:56 Avg. Rating:5
Views:3126 Votes:0
Shares:0

Related Content

Discussions Leaderboard