I have two XR2's in an rb100 with an atheros A wds backhaul. The XR2s will eventually power two sectoral H-Pole antennas but in testing using th existing 9dB omni, the performance does not even equal the prism pcmcia cards in the current soekris AP. One XR2 barely gets any association, I think it is defective. The other XR2 gets as many associations as the working AP but the signal levels are not as good... avg around -80-90dB The prism cards are getting -65-75dB on equivalent connections.

Has anyone else seen these performance problems? Any ideas? The power of the cards has been left at default - perhaps I should set it higher at 23dB (13dB showing because of the 10dB offset).

thank you,

How are you powering the board?

How are you powering the board?

48V .35A POE, about 25-30 meters

Try with one card, I am guessing it is a power issue, those cards draw too much power.

yea most def. a power issue i woudl think.

what is RB 100? is that RB 112? if so then the only thing i would think could POSSIBLY power 2 x XR2s is a RB 133.

even 133 may not beable to....i DO have a RB 133 powering: 1 SR9 1 8602 1 SR2 all works fine.

I have two XR2's in an rb100 with an atheros A wds backhaul. The XR2s will eventually power two sectoral H-Pole antennas but in testing using th existing 9dB omni, the performance does not even equal the prism pcmcia cards in the current soekris AP. One XR2 barely gets any association, I think it is defective. The other XR2 gets as many associations as the working AP but the signal levels are not as good... avg around -80-90dB The prism cards are getting -65-75dB on equivalent connections.

Has anyone else seen these performance problems? Any ideas? The power of the cards has been left at default - perhaps I should set it higher at 23dB (13dB showing because of the 10dB offset).

thank you,

Try setting the TX Power down to 7db on each card, these cards actually output 10db higher that you have set for, so 7 = 17db.
Ubiquiti XR2 SPecs: http://www.ubnt.com/downloads/xr2datasheet.pdf

I'm not sure but I don't think a 100 will do the job you are wanting either, I beleive you will at least need a 532...

It looks like a 133 but when we bought it I thought it was a 500 series, which I cannot find on routerboard.com - one with 3 mini PCI slots and 5 ethernet ports that is.

Should I buy a 12 V. 1A. POE supply then? Or could someone reccommend something better? So far I've only found 12V. 1A. from our usual vendors.

Thank you for your replies.

It looks like a 133 but when we bought it I thought it was a 500 series, which I cannot find on routerboard.com - one with 3 mini PCI slots and 5 ethernet ports that is.

Should I buy a 12 V. 1A. POE supply then? Or could someone reccommend something better? So far I've only found 12V. 1A. from our usual vendors.

Thank you for your replies.

It sounds as though you have an RB-153
Performance: MIPS32 4Kc based 175MHz CPU
Memory: 32 MB Memory / 64 MB Storage
Connectivity: 5 Ethernet, 3 miniPCI, 1 DB9

No the 12Vdc power supply @ 1 amp is sufficient.
Just make sure and re-pin the power jumper on the board for 12Vdc if the jumper option is on the 153.

With my experience you would really want to get a RB-532 or 532A.
I beleive that when people mention the board does not have enough power it seems they are referring to the processor primarily and then also the on-board transformer is slightly different from the 100 to 500 series RB's.

It sounds as though you have an RB-153
Performance: MIPS32 4Kc based 175MHz CPU
Memory: 32 MB Memory / 64 MB Storage
Connectivity: 5 Ethernet, 3 miniPCI, 1 DB9

Yes that is it.

No the 12Vdc power supply @ 1 amp is sufficient.
Just make sure and re-pin the power jumper on the board for 12Vdc if the jumper option is on the 153.

With my experience you would really want to get a RB-532 or 532A.
I beleive that when people mention the board does not have enough power it seems they are referring to the processor primarily and then also the on-board transformer is slightly different from the 100 to 500 series RB's.

[/quote]

Oh. I see. So the xr2's will require more processor as well as a higher current POE supply.

Ok, thanks very much!

Oh. I see. So the xr2's will require more processor as well as a higher current POE supply.

Ok, thanks very much!

Yes to run an AP you do require more CPU capacity.
However, your current POE rated at 1amp is more than ample to power the 532.


A note: Today I have managed to connect to one of my new RB-532/XR 2 at a client which is roughly 3 miles out from the AP.

Then I turned the TX power of the XR2 up to 15db which in reality is 25db to establish a better receive signal strength at the SU.

This weekend I will place the AP and sector at the top of this particular tower to enhance range and to view.

Oh. I see. So the xr2's will require more processor as well as a higher current POE supply.

Ok, thanks very much!

Yes to run an AP you do require more CPU capacity.
However, your current POE rated at 1amp is more than ample to power the 532.

No the current on is 48V 35mA. I will buy a 18V. 1A. for it.


One last question... buying a rb532 will leave me with only two minipci slots and I need a thrid for A backhaul. Should I buy an rb502 daughetrboard to get the extra slot and will the 532 power the 2XRs and an Atheros AR5212 doing A band with a good 18V. 1A. POE supply?

Otherwise if the rb520 doesn't have the ass I will route to another AP to do the backhaul.

thanks again.




A note: Today I have managed to connect to one of my new RB-532/XR 2 at a client which is roughly 3 miles out from the AP.

Then I turned the TX power of the XR2 up to 15db which in reality is 25db to establish a better receive signal strength at the SU.

This weekend I will place the AP and sector at the top of this particular tower to enhance range and to view.[/quote]

No the current on is 48V 35mA. I will buy a 18V. 1A. for it.


One last question... buying a rb532 will leave me with only two minipci slots and I need a thrid for A backhaul. Should I buy an rb502 daughetrboard to get the extra slot and will the 532 power the 2XRs and an Atheros AR5212 doing A band with a good 18V. 1A. POE supply?

Otherwise if the rb520 doesn't have the ass I will route to another AP to do the backhaul.

thanks again.

OK, the can be repinned to operate on 12 OR 24 volts I beleive it is.

Yes I beleive that the 532 in tandem with the 502 daughterboard would be adequate with the 18Vdc 1a POE injector.

I'm not sure if you are in US or other country, at either rate, I strongly suggest you speak with your supplier detailing your configuration requirement to make sure you are not making any mistakes in this purchase.
If you are in the US you could ask Eric Wiege at Titanwirelessonline.com.


Regards

kiira - I would suggest getting the new revisoin 5 532, much increased power(wattage) capacity. The older revision 4 boards had a maximun of 14watts including the 532 itself and extension cards which leaves about 10 watts for cards. According to Ubiquiti's datasheet the XR2 has a maximum draw of 1.3 amps at 3.3 volts (4.3 watts), two of these doesn't leave much for a third card. Plus I'm sure the 502 expansion uses some power. However the new revision 5 boards have a maximum of 24W with 16.5 available for extension cards. I haven't tested, but this should be ample for your needs. We have run 2 SR2's in a R4 board without issue (each card draws 1.1 amps) but have had issues with one SR2 and one SR5 (SR5 draws 1.3 amps) the extra .2 seemed to put it over the edge. We have run one XR5 (1.8 amps) and one XR2 (1.3 amps) in a R5 532 without issue. Hope this helps.

eburgash, giganet, everyone thanks for all this information. The owner bought these boards without consulting so I couldnt research them earlier, I have ordered a 532 and I believe it a revision 5, also a rb502 daughter card and an 18V. 1A. PS. We have two H-Pole arrays just waiting for them. Oh I am Finnish but living in the US.

thanks!

edited for spelling

We run the 48v/40w on all of our AP's (532a's) and clients(112's). We have had no issues with power. A couple of these are running 3 SR9's as AP's. We are about to add a sr5 in a couple of our boxes that have 502 boards fully populated. We will then be having 2 xr5 and 2 cm9 in those boxes. I think these 48/40 supplies in the 532 will be plenty.

We are currently testing the ap's in 24v to see if we can do a better job with solar power.

Also, we are testing the 112's at 18v so we can use the same power supplies we use on the 133 boards.

hope this helps
Kenny

We run the 48v/40w on all of our AP's (532a's) and clients(112's). We have had no issues with power. A couple of these are running 3 SR9's as AP's. We are about to add a sr5 in a couple of our boxes that have 502 boards fully populated. We will then be having 2 xr5 and 2 cm9 in those boxes. I think these 48/40 supplies in the 532 will be plenty.

We are currently testing the ap's in 24v to see if we can do a better job with solar power.

Also, we are testing the 112's at 18v so we can use the same power supplies we use on the 133 boards.

hope this helps
Kenny

Virtually of my towers are solar powered, I have both 12Vdc and 24Vdc systems.
I have 532's at one tower operating on 12Vdc without problem as I do have 532's at 24Vdc sites.

I have seen that some users claim that the 532 is less apt to emit un-wanteed noise opearting at 12Vdc.

If anyone needs input or suggestions how to budget your solar system output requiremenets I can help with that to save money and mistakes for those not able to calc true usage at a site.


Regards

I agree 100% with whoever said the 532r5 would help the situation out, the higher dedicated power to the mini-pci's has helped a lot of my customers out when coupled with the higher powered cards.

eric

Something is catestrophically wrong. I have the rb532 rev. 5 and 502 with 2 XR2 and an atheros 5G. backaul. Power is 18V. 1A. Associations are very low and weak. CCQ varies between 18% and 32% Noise floor is good at -101 though. I have the card I am testing at 9dB (19dB) which seemed about right. Using the old prism carded AP I saw all our customers associated from -65 to -75 and overall-tx-ccq=98%

I just don't know what to do. The owner thinks these cards are savior for our customer base and I do everything and they still suck!

signed,

upset, pissed and baffled

Sounds like your cpe are prism too,
if so, how many cpe per sector?

Have you selected the correct antenna connector?
If the board wasn't rebooting, it's unlikely to be power related.

Specs say XR2 consumes 2.97W @ 24dBm (nothing less in datasheet, so I'd assume it can't go lower?)
RB532 r5 has 20W spare for extension cards, so you've got loads left.
Your 18v/1A PSU supplies 18W, the 48v/0.35A supplies 16.8W
Your total is about 9.5W, so either PSU would be OK.

Be sure to put 2 of the cards on the RB532 and the third by itself on the daughter board.
The daughter board may not be able to power 2 high power cards.

I could be wrong, but I have more faith in the RB532 than I do the daughter board.

Have you selected the correct antenna connector?
If the board wasn't rebooting, it's unlikely to be power related.

Specs say XR2 consumes 2.97W @ 24dBm (nothing less in datasheet, so I'd assume it can't go lower?)
RB532 r5 has 20W spare for extension cards, so you've got loads left.
Your 18v/1A PSU supplies 18W, the 48v/0.35A supplies 16.8W
Your total is about 9.5W, so either PSU would be OK.

Antenna connector? I have mmcx and N-male


I have tried all different power supplies under more controlled conditions, every combination of cards and slots, including a single card in the RB532... everything. One card performs inferior to the other XR2. Neither shows connection strength as good as prism cards. I cannot get them to show any significant improvement - the best one seems to perform worse that prism cards, the defective seemingly XR2 performs very much worse.

Here is some info:
here are the registrations. This is inferior to the prism.

0 XR2a 00:02:6F:36:BE:A7 no -74dBm... 11Mbps
1 XR2a 00:11:7C:02:00:55 no -70dBm... 11Mbps
2 XR2a 00:02:6F:42:14:07 no -71dBm... 11Mbps
3 XR2a 00:02:6F:42:C4:ED no -80dBm... 11Mbps
4 XR2a 00:0A:52:01:2D:15 no -78dBm... 11Mbps
5 XR2a 00:C0:CA:17:FD:68 no -67dBm... 11Mbps
6 XR2a 00:60:B3:E2:90:D6 no -60dBm... 11Mbps
7 XR2a 00:60:B3:44:FE:CD no -77dBm... 11Mbps
8 XR2a 00:60:B3:E2:90:B8 no -72dBm... 11Mbps
9 XR2a 00:0B:6B:37:80:AE no -65dBm... 11Mbps
10 XR2a 00:0B:6B:4E:6C:CA no -53dBm... 11Mbps
11 XR2a 00:0B:6B:4E:6B:FF no -53dBm... 11Mbps

the prism card is typically like this:
1 prism1 00:13:10:1B:37:57 no -59dBm... 2Mbps
2 prism1 00:02:6F:42:C4:ED no -73dBm... 1Mbps
3 prism1 00:02:6F:42:14:07 no -59dBm... 11Mbps
4 prism1 00:11:7C:02:00:55 no -65dBm... 11Mbps
5 prism1 00:02:6F:36:BE:A7 no -74dBm... 11Mbps
6 prism1 00:0B:6B:4E:6B:FF no -44dBm... 11Mbps
7 prism1 00:60:B3:E2:90:B8 no -69dBm... 2Mbps
8 prism1 00:60:B3:E2:90:D6 no -53dBm... 11Mbps
9 prism1 00:0B:6B:4E:6C:CA no -37dBm... 11Mbps
10 prism1 00:60:B3:44:FE:CD no -68dBm... 11Mbps
11 prism1 00:12:17:75:C2:6A no -75dBm... 11Mbps
12 prism1 00:0B:6B:37:80:AE no -55dBm... 11Mbps


and here is the interface. I have tx-power set to 9 and tx-power-mode
set to default but have tried the all rates same as card rate as
well.

R name="XR2a" mtu=1500 mac-address=00:15:6D:63:00:F5 arp=enabled
disable-running-check=no
interface-type=Atheros AR5413 radio-name="00156D6300F5"
mode=ap-bridge
ssid="Hampden Wireless KeswickN" area=""
frequency-mode=manual-txpower country=no_country_set
antenna-gain=0 frequency=2412 band=2.4ghz-b scan-list=default
rate-set=default
supported-rates-b=1Mbps,2Mbps,5.5Mbps,11Mbps
supported-rates-a/g=6Mbps,9Mbps,12Mbps,18Mbps
basic-rates-b=1Mbps basic-rates-a/g=6Mbps max-station-count=2007
ack-timeout=dynamic
tx-power-mode=default noise-floor-threshold=default
periodic-calibration=enabled
periodic-calibration-interval=60 burst-time=disabled
dfs-mode=none antenna-mode=ant-a
wds-mode=disabled wds-default-bridge=none wds-default-cost=100
wds-cost-range=50-150
wds-ignore-ssid=no update-stats-interval=disabled
default-authentication=yes default-forwarding=no
default-ap-tx-limit=0 default-client-tx-limit=0
proprietary-extensions=post-2.9.25 hide-ssid=no
security-profile=default disconnect-timeout=3s
on-fail-retry-time=100ms preamle=both
current-tx-powers: 1Mbps:18,2Mbps:18,5.5Mbps:18,11Mbps:18



here is the monitor output

status: running-ap
band: 2.4ghz-b
frequency: 2412MHz
noise-floor: -93dBm
overall-tx-ccq: 84%
registered-clients: 14
authenticated-clients: 14
current-ack-timeout: 97
current-distance: 97
nstreme: no

status: running-ap
band: 2.4ghz-b
frequency: 2412MHz
noise-floor: -93dBm
overall-tx-ccq: 71%
registered-clients: 14
authenticated-clients: 14
current-ack-timeout: 97
current-distance: 97
nstreme: no

Have you selected the correct antenna connector?
If the board wasn't rebooting, it's unlikely to be power related.

Specs say XR2 consumes 2.97W @ 24dBm (nothing less in datasheet, so I'd assume it can't go lower?)
RB532 r5 has 20W spare for extension cards, so you've got loads left.
Your 18v/1A PSU supplies 18W, the 48v/0.35A supplies 16.8W
Your total is about 9.5W, so either PSU would be OK.

Antenna connector? I have mmcx and N-male

One thing I am curious of- Have you tried increasing the AP TX power from 9 to 11 and see if that improves your noise floor any?
(I have an XR5 running at 15 to achieve optimal perfomance with a noise floor of -103 and client connections at -65 or better anywhere from 2 to 11 miles from the AP)

Also try running a freq survey to see if there is any chance of offensive interference.
Re-channel the AP if it appears their is a high chance of interference, microwave ovens can be killers especially if they have a problem the user knows nothing about-

still cooks vittles what's wrong with that granny?

Are you actually using the XR2's in your client devices also?
That seems it could cause some form of problem especially concerning throughput.
If so have you tried using a Winstron CM9 a/b/g 65mw card in the client devices, I have good luck with this setup currently.

If you are using the XR's in client devices also try working with the TX power as that will have a direct affect on connection quality too.

I have been able to achieve far better throughput also by setting the data rates to configured only selecting 6MB for example rather than leaving more than one data rate selected, it simply relates to more time the client device is trying to re-establish it's communications if your data rates fluctuate when you watch Wireless => Registration TX/RX Rate. If that number isn't rock solid your clients will notice performance dips regularly.

Hi Kiira

If you use the MMCX connector the right setting for that is antenna-mode=ant-b
its like that on the SR5 anyway...the default is for the UFL conector...

//Rickard

Hi Kiira

If you use the MMCX connector the right setting for that is antenna-mode=ant-b
its like that on the SR5 anyway...the default is for the UFL conector...

//Rickard

I just double checked on an AP using an XR5 and it uses "antenna a", I would assume the XR2 to be the XR5 in this sense?

...

XR2 and XR5 have only one antenna connector

XR2 and XR5 have only one antenna connector

Uh-Huh! That's right

We have no or at least very few CPE with Atheros radios. Most are Tranzeo or prism or some other chip. Perhaps that is a problem?

I tuned one card to achieve an average 3dB improvement in signal at our central pop. It also reports %98-100 ccq and -100 noise floor. I don't know if I can expect better than that. There are only 20 or so clients registered against the AP usually.

Perhaps the advantages of the XR2 are more dramatic with longer distances as well? Almost all our clients are < 1k away. This is a city wisp.

I believe the second card is defective as it cannot do better than %72 ccq and 098 noise floor under the same conditions i.e. same slot, ssid, config etc.

thank you

Kiira

We have no or at least very few CPE with Atheros radios. Most are Tranzeo or prism or some other chip. Perhaps that is a problem?

I'm not sure but it seems there was another posting a couple months ago here that mentioned some form of incompatibility between the atheros and prizm chipsets??
Try running a MT forum search in "Wireless Networking" using keywords like: "atheros and prizm compatibility issues" those keywords should put you in the ballpark hopefully

I tuned one card to achieve an average 3dB improvement in signal at our central pop. It also reports %98-100 ccq and -100 noise floor. I don't know if I can expect better than that. There are only 20 or so clients registered against the AP usually.

-100 SNR is pretty good in my findings at least. The CCQ numbers too look good.
What is your average client RX signal quality on this AP?

Perhaps the advantages of the XR2 are more dramatic with longer distances as well? Almost all our clients are < 1k away. This is a city wisp.

Yeah, that is the primary benefit is the extended range and the ability to operate with less interference even in a noisy environment.

I believe the second card is defective as it cannot do better than %72 ccq and 098 noise floor under the same conditions i.e. same slot, ssid, config etc.

Before I ruled on the card I would absolutely verify the integrity of all pigtails, and LMR400/600 cable SWR.
Another thing to remember, when working with 2.4Ghz-- a small amount of moisture in any N connector or even moisture which wicked itself up into an LMR will really have a dramatic affect against all facets of performance.

Also, be positive the XR is grounded properly..
I had a week ground tie-in on a tower exposed by installing a XR equipped AP, the performance actually improved after I cleaned the corrosion from the #6 copper wire that runs up from the 8' ground rod the XR seemed to perform better.

One other question that comes to mind for me-- beings you are in a city envoironment.. have you got enough large scale buildings that could be creating a problem for this AP with signal bounce?

Have a good one

Oh yes I have verified and tested all cables and pigtails. I always check the obvious stuff first.

The average connection is around -57 to -62dB. for the XR2 that works well and that is certainly ok... at least at this location where I am trying to test it. It seems to be about 3dB. better than the prism cards. It cannot go through the splitter though... the splitter harms the signal badly.

thanks,

Kiira

What CPE equipment are you using?

What CPE equipment are you using?

It is a hodgepodge but we have standardized on Tranzeo over the last year. But there are Prism CPE as well and perhaps a few WRT54g type routers around and a few atheros, 8 of 50 or so customers. I am not so sure that the XR2 really performs all that better than the Prism 200mW cards. I wish I had some realistic idea of how much improvement I should be looking for. BTW - setting the ack-timeout to 80mS. helped a great deal... power to 9 (19 actual)dB. Our customers are < 1.5K. or so.

thanks for the reply,

Kiira