Rain Fade, Smart Antenna

Hello all,

 

I searched the forums for any info on Rain Fade with a 2000 AP with Cambium Sector and Smart Antenna and came up empty so I am going to post my latest experiences and hope some people reply with their experience.  First a bit of background on our AP's and how they are set up.

 

1.  ePMP2000 with Cambium Sector and Smart Antenna mounted on the bottom of the Sector.

2.  Running Front and Back on 4 AP's on a 100-ft. tower.

3.  15MGz separation on 40MGz.

4.  Excellent results with anywhere from 50-140 Ag. thru to customer's SM's.  NOTE:  I know this for a fact because I am one of the users.  

5.  Most SM's so far on 1 AP:  44

6.  AP was set to "FORCE SMART ANTENNA"

 

Here is what happened to my set up 2 days ago.  

 

1.  It started Raining... HARD... LIKE REALLY HARD!  Most Tower side signals stayed good and useable.  But I noticed the SM side back to tower started getting higher and higher and soon not letting me get into the SM and then customers started calling.   So of course, I started looking things over.  Long story short... I put the AP on "Auto" for the Smart Antenna.  When people got on their signals were GOOD again and customers could function.

 

2.  But after a while I noticed about 3-5 customers were STILL connected to the Smart Antenna and their signals were bad.  So, I when back in the AP and set it to "Forced Sector".  Everyone got back on and signal were good including the 3-5 that were still trying to use the Beam Forming before I changed to Force Sector.

 

3.  So since I live in a part of the U.S. that averages me 50 Inches of rain a year I need my equipment set to "Forced Sector" or I will be doing this every time it rains hard which is often.  Good news is the customers did not experience any change in their service on this setting.  But I do wonder when I get 100 customers on the APs if that will change?

 

CONCLUSION:  Rain Fade occurred because the Beam Forming produces a 'small beam' there for Hard Rain will block it just like say the 60-80 Wireless spectrum.  I know the higher the weaker but.... the small beam did get blocked by the Rain.

 

SORRY:  I know the thread is getting LONG!

 

Questions to Cambium Engineers and OR Forum users that have experienced this and have answers.

 

1.  My Engineer thinks that even though we have it set to "Forced Sector" that the upload is still going to the Smart Antenna just NOT Beam Forming.  I think it is going ONLY through the Main Bigger Sector?  So, Cambium Engineers which antenna is the upload going to.  If it is going to the Main then I assume you no longer have the 4 chain but only the 2 chains?

 

2.  I would like to leave things on the "Auto" setting for obvious reasons but since I found a few customers still on the Smart with bad signal it makes me not want to???  Should it end up putting them on the Sector if I would have waited long enough?

 

3.  Has Cambium TESTED these conditions?  If so pls. reply with findings.

 

Thanks to ALL for putting up with the LOOOOOOOOG Post!

1 Like

Hi,

Thank you for sharing your observation and experience.

First of all I'd like to provide answers to your questions.
1. UL Traffic is going to Sector Antenna in case Forced Sector Antenna option is configured.
Smart Antenna does not operate at all in this case.

2. In Auto mode, each SM estimate possible UL throughput it can get through Smart Antenna and through Sector Antenna and select better one.
Each SM perform re-alignment each 10 minutes, to keep track whether environment's conditions has been changed.

3. Hundreds of field tests were performed before ePMP2000 with Smart Antenna was introduced on market.
We didn't observe described behavior. Both Sector Antenna and Smart Antenna get lower RSSI when it's raining and it is expected because of Rain Fade.

Could you please verify Smart Antenna operates well and issue is not related hardware or deployment:
1. Please check Smart Antenna Disconnects counter on Monitor -> Performance page.
2. Please check Device initialization status on Monitor -> Wireless page.

We will investigate your case and schedule additional tests on this.
Also we can try to debug isssue on your link if it is accepatble for you.

Thank you.

1 Like

Thanks for the fast reply.

3 files are attached to answer your questions.

1.  “0” for smart antenna disconnects
2.  Initialization page status
3.  Page to show how good the signals are and the SM distances to show how ‘well’ the deployments of our clients are.  We walk away from a potential customer if there will not be excellent QOS.

As far as ‘debugging’?  I would say from the files I sent you there would be no debugging needed, but I did not Engineer the radio so I could obviously be wrong. 

Our tower wiring harnesses are less than a year old and are Cat6 double shielded cable with a ground wire, 9 wires total, with metal connectors. There is an 18X18 inch steel box at the top of the tower with Tycon Gig Ethernet surge protectors in the box at the top of the tower and in the Head End building at the bottom.  The Cat6 double shielded cable is incased in Metallic PVC all the way from inside the Head end building (Just like Carrier class Cellular) until it goes into the bottom of the ePMP2000 radio.  (would be nice if your radio had a metal case then NO lightening or surge would have a path to the Ethernet).

So, whatever you think is the next move just let me know.

Thanks for the forum and the attention to this issue.

Randall Fowler
Owner/CEO
HBE Internet
870-423-6223 – Office
iamrbf@gmail.com

Hi,

Thank you for update.

Yes, you are right Smart Antenna operates in proper way.

So for now the main challenge is to to find reason why SMs were not able to communicate with AP through Smart Antenna but were not switched to Sector Antenna during it was raining.

Probably rain was to heavy and SMs had connection issues even with Sector Antenna, but after the time you had spent on AP re-configuration rain had died down and everithing started to operate as usual.

SMs were not switched to Sector Antenna because they were not dropped by AP and SM were waiting for 10 minutes to start re-alignment.

So I'd like to propose you to leave AP operates in Auto mode and monitor whether it will be affected by rain, fog or something else. We will continue to test this on our side as well.

Then if you will see reproduce on your side please let us know asap and we'll troubleshoot it together.

You can reach me at fedor.trutsko(at)cambiumnetworks.(dot)com

Thank you.

Latest testing shortly after the last major rain storm.  This morning for about 3 hrs. I tested 4 AP’s.  Here are the numbers of SM units and the results.

  1. 44 Clients. 2. 39 Clients. 3. 20 Clients. 4. 10 Clients. (SM’s)

I was in each AP looking over customer’s signals.  I concentrated on the worst signals.  As I have stated our customers are in a heavily wooded National Forest Area so we fight with tree issues all the time.

What I started noticing was most ALL the higher signals (customer side always) were showing they were connected to the Smart Antenna.  When I would try, and get into them my attempt would stall out or take an extremely long time to connect to the SM.  Which we all know is a result of a high customer side signal.  Once I connected I did some ‘link testing’ from the SM to the AP so I forced the signal to transmit to ensure a “data driven’ signal. When I did this the signals stayed high and did not come down on the customer side.  If the signal did come down I would do several tests to make sure it stayed down.  I think we have all experienced this situation.  I call it the ‘SM sleep mode’.  My experience is it often the case that a ‘idle’ SM will elevate to a high customer side signal but when you force traffic it will generate the proper power level and lower its signal. 

But on the SM’s that the signal did not come down I changed the SM setting to ‘Forced Sector’ and applied.  After connecting to the AP I once again did the ‘data transfer’ testing between the AP and SM from the SM side.  My results were that the SM side signal stayed consistently lower and the “Upload” side often ‘Hit’ the through put setting max. level. Results were a better overall QOS.

Since I personally know, mostly, where the SM is located and what it is shooting through back to the AP I have come up with the following theory.  Believe me when I say this is all on the fly results in real world testing.  I am by NO means saying this is THE WAY IT IS.  The whole idea of the forums, as I understand, is to share and hopefully give feedback to the Cambium Engineers and the other WISP members so we all learn and benefit.

A couple of things started to surface as I went through looking at over 100 real worlds connected SM units.  The customers (SMs) that were basically JUST shooting through trees, meaning NO ground or Fresnel issue seemed to be easier to determine if they needed the ‘Forced Sector’.  It was a straight forward data through put testing result.  Better signal, more consistent upload number and that gave me the info to set the SM on Auto or Forced.

Now when it came to Fresnel this is where it really seemed to help to know the location and what the path was back to the AP.  If it was ‘Fresnel Only’ meaning NO trees the Smart Antenna either cleared the earth or it did not.  When it did not is when the ‘Forced Sector’ worked best.  If it was with trees and ‘somewhat’ an earth Fresnel issue knowing the path seemed to give me a better feel for what to set the SM to.  I just used my common knowledge on ALL factors and made a judgement call and felt good about that.

Now for the number of Clients (SMs) that I forced:

AP with 44 – 5 were changed to “Forced Sector” with obvious better overall results.

AP with 39 – 5 were changed to “Forced Sector” with obvious better overall results.

AP with 20 – 2 were changed to “Forced Sector” with obvious better overall results.

AP with 10 – 2 were changed to “Forced Sector” with obvious better overall results.

So even though the changes, 14 total, seem small, it equals right at 10% of the customers.  When I can improve 10% overall on my customers I consider that enough to warrant taking the time to get the signals better and overall better QOS.

Another note that I believe really helps us tweak or fine tune customers is the fact that our Installation Techs take a picture from behind the SM back towards the AP (tower) so that when we are trying to figure out signal issues we have the real-world picture.

Thanks for letting me post such a loooooooooooooong string.  I do believe the ePMP2000 software does an EXCELLENT job of adjusting the antenna it believes it should be using.  As always with Wireless the signal path on PTMP is always changing and always seasonal.  Understanding that different day to day weather changes can promote me to ‘manual’ override some of the auto settings of a radio I believe is a reality in my location.  I will try and keep updating this string as I add SMs and continue testing.

Hi.  I'm curious if you tried waiting more than 10 minutes for the AP to do it's re-evaluation of what's the best?  Perhaps you already tried that and I missed it...    but I'm wondering if maybe after 10 minutes of heavy rain soaked trees, maybe the AP would have done it's re-evaluation and moved those 14 clients over to the other antenna.

I have.... it did not move them on its own.  NOTE:  I also have the MCS on “O”.  What my research says is that it takes very little transmit on “O” so it might be that there is enough that makes it thru that the SA does not think it needs to change??

 As stated in the above thread these were customers that have tree issues as well as Fresnel issues.  Knowing the path of most all those customers I put on 'Forced' I am, now, convinced that is what they need.  They also 're-connect' to the tower faster because of the nature of the Engineering behind the Smart Antenna VS a Standard Sector.

 By NO means am I stating, as also mentioned above, that is the way it will be for another WISP.  I spend a bucket load of time watching my customer’s connections whenever we make equipment changes.  Just seems like these changes have helped those customers. 

I also looked at the performance tab under the AP ‘Monitor’ tab.  They have LESS packet loss and re-transmission with the Forced Sector on these customers. 

I wouldnt be suprised if part of the problem is due to multipath scattering from the wet leaves of the trees,

I have seen signals change significantly if the path clips or passes  through trees between wet and dry leaves.

1 Like