900mhz AP recieve is down 10 to 15db on recieve only

We just installed a 532 with Ub 900mhz card on it, we are using the 5mhz channel size at 907mhz. 4 miles out we are receiving a great -64db in the wide open areas. However the ap sees on only about -75-80db down about 10-15db. It is a couple of miles from the first trees and the ap is 120ft up H pol, cpe is 112 in 12dbi rootenna. AP antenna is Dual pol Tiltek 120 sector using h pol only

We have completely swapped the ap (antenna and lmr are unchanged). We get the same results with a yagi at the cpe.

Any one seen any issues like this? It is something we need to get sorted out quickly

thanks

Erik

tx power set to auto?

You should still see more signal.

Swapped the SR9 on the AP yet? I think the diversity chip is worth around 13dB if its blown.

Have you tried changing the antenna port on the AP?

We’ve seen the same thing. We’ll try swapping the SR9 as well.

Hello,

It sounds as though the diversity switch took an ESD hit, and cause the card to be partially deaf. Please swap it with another card and see if it solves the problem.

Let me know!

Thanks,

Mike

We swapped the ap no change. I’m going to bench it here tonite or in the morning and also swap cards around… I think I have a sealed one somewhere still.

Mike, If it is a hit on the diversity, would this be a handling issues or a possible ground on the case/tower issue. We are about to roll out a lot of these and want to get any issues resolved before they are in the field :wink:

Erik

We’ve seen this issue when the AP (in relatively advantageous position w/wide aperture antenna) can see interference but CPE cannot. Generally CPE is looking through a relatively narrow aperture antenna and is not as exposed to whatever signals may be causing interference, thus less affected by interference. Interference on the SR9 cards seems to cause low RSSI readings.

We did move around and try different channels, the noise floor was in the 90’s, we go the same results on higher channels where the noise floor was in the -80 to -70. Besides with a signal of -64 you should be able to get through the noise i would think.

Erik

Check to see if noise floor is similar at both AP and CPE. Another helpful diagnostic is RX/TX CCQ. In the interference case I mentioned, we saw decent RX CCQ at CPE but poor RX CCQ --from-- CPE at AP.

If you’re really desperate you might try the 3.06 beta which actually does seem to sport wireless drivers that improve performance in noise and other challenging situations.

looks like the card, here on the bench, I’m down the same amount and I have set all tx power to 17db. I will try a fresh card and see if things improve. I would be strange that we would see 2 in a row with this problem.

Mike: does it make any sense to ground the shielding on the card or is this likely a charged person handling problem.


Erik

Hey ejannson,

From my experience adding additional ground to the shielding has never helped (in lab experiments). The shield is already grounded through the ground plane of the PCB. It is very likely that a charged person could have zapped the card, or that static charge in the air (lightning storm or something ) could have done it.

Thanks,

Mike

Why is it every card produced by UBT has problems with noise, esd, and excuses for it, I’ve been through many SR2’s all with the system grounded and they go deaf. we hear the excuse of the dervisity switch. well then fix it, cm9’s perform better and do not have this problem. SR9’s work fine as long as you reboot the system once or twice an hour. they cannot handle thier own noise(this was tested in lab). I have not tried the XR series, I,ve got a business to run and can’t keep loosing money on this. Who know’s how many promises Robert at UBT broke with me.
O.K. I feel better now but will UBT ever own up to this Robert?
Boo

Something wrong w/your lab, maybe. We’ve got oodles of SR9 showing multi-day or multi-week uptime.

No problems with the lab just did everything like they said to do, theres a little more to it though and I’m sorry if I led the forum in the wrong direction.

Boo

I have SR2 and SR5 cards in a production ISP - that have been up and running for nearly 2 years. The only reboots are when I upgrade. They are sure not deaf. SR9’s do not require reboots. The only problem we encountered were very high noise environments. Very common here in 900Mhz. I would have to look at your setup, but if you are having that many problems - you are definitely doing something wrong. There are too many people that use these cards, with 0 problems. I normally see 3 people on this forum that whine about UBNT - and they are always the same people. I just have to wonder how they are deploying these cards. I personally use them, as does my neighboring WISP - 0 problems. I can’t imagine one of us would not see issues if there were any. There are many WISP’s that post here - using them successfully. UBNT has always been very forthcoming if there were problems - see the cold weather post about SR9’s.

Hello Boo,

I am sorry you are experiencing problems. I do not know who/what you have been dealing with Robert about, but you can email me @ support@ubnt.com if you are having problems.

Now about Sr9. When you say you you have to reboot every hour or 2 that seems that the Sr9’s are not the cause, unless those specific 2 units are bad. We have thousands of cards in the field working properly, with no issues. It sounds as though it may be and environment (every environment is different and it is hard to predict if the interference levels at every sight.), or implementation issue. To be honest thought the 900Mhz solution does not work for 100% of the people, but this goes back to either 95% environment and 5% bad implementation and planning.


So from here, please email me and deal with me directly and I will help resolve your issue. Or if you are already dealing with us please respond to those emails and we will see where we can go. Also if you can give me complete specs to your setup to try and duplicate that would be great.

Thank you,

Mike

I’ve had no problems with SR9’s. I’ve had a test link up for weeks solid.

No reboots, lots of changes to the config made in testing.

I haven’t gotten my real world connection working yet, even too many trees for 900MHz, but I’m working on that tomorrow.

I’m in a low noise area, for certain…

I had a SR9 ap up with one client for six months - no problems.

I added additional clients (around 10) and within a week it died. Replaced it with a brand new SR9 RB532a and 48volt poe and 5dBi omni antenna. Within a week it also died.


I have SR2’s - low clients they work wonders. For backhauls they work wonders. I have an AP with 20 clients and ever 1-2 months the card goes deaf. I have other AP’s with 1-2 clients and I never have a problem.

These cards go deaf like there is no tomorrow. Lab enviroments are useless unless you are actually putting the cards under load with multiple clients.

I plan on deploying 3 more SR9 AP’s at tower sites this week. If there is a better SBC to use I’d love to know.

And less someone say I have tower issues… I replaced the defective SR9 AP on the tower with a Moto Canopy AP same antenna and have had no issues.

ejansson;

at 900MHz you are competing with highspeedcrow, Canopy.
That is the noise which is dumbying your connections.

I’m having the same problems with the SR9s also the SR2. They (SR2) go deaf after a while.
I sent them in to UB but by then the “warrantee” was out. 25 of them and they sent them back to me.
SR9 is very sensitive. I have a 9+dbi omni in the field and a 13dbi yagi. We are trying to roll these out on all our edge towers but need to figure out a plan so that it is a constant. MT 2.9.46 was an awesome upgrade for the AP (TY MT). I no longer have to buy Deliberant APs. But its hard to know how the equipment is working when MT doesn’t tell you specifics. Any idea about when that will be?

BTW - the whiners using UB SR2’s are right on track - It’s cost me a lot of customers to figure that one out - on top of the MT AP problem we had.

We have 70+ towers and over 150 hotspots in the field and this gets expensive when it doesn’t work as stated.
Thats why we are concerned with the SR9 and the problems we are having here.
We don’t know whose fault it is!
Thank you,
Rod

OK - I worked on the problem more today and found that the modulation was acting more like a “G” connection then a “B” connection. Now I know you RF Gods out there are gonna chew on this but I freely admit that I am NOT a RF God. Just an observation.

The 2ghz-5mhz and 10mhz where trying to connect at 6-9-12-18+ rates and we all know that “G” is notorious for behaving more like “A” then “B” in modulation.
When I turned off the data rates for the “G” and tried to make them “B” only the 2ghz-5mhz-10mhz complained and behaved badly.
I then set the interfaces to “B” and the data rates to 1-11 only and everything behaved much better. I’m still tweaking the connections and signal testing the antenna angles but I think we have a beginning solution that we can implement and then build on.
The only real aspect of the winbox data was the ack timeout - it really gets bad(408+) when the connections are not set right. Good (30-) 3mb thru-put down; 1mb up
Would love feedback to enhance my knowledgebase.
BTW - my connection is 1 mile thru heavy wooded area and a hill.
Omni(SR9) is 90 feet on tower and elevation is 20 lower then the CPE(18dbi yagi w/SR9) which is on a 30 foot roof - which means they are pretty darn close to equal height.
They are both vertical polarization.
Both are using “default” power settings.
Thank you,
Rod