You are here

Nanobridge M2 troubles LQ 100, NLQ 0

9 posts / 0 new
Last post
N1TEW
N1TEW's picture
Nanobridge M2 troubles LQ 100, NLQ 0

Hi guys, 

We have a new network here with lots of cheap Nanobridge M2s.

We've had lots of fun with this.  It has worked well as long as we are only a few miles apart.  Recently, however, we've tried longer links of 8, 11, and 20 miles. While they all have reasonable signal levels ranging from -69 to -80, I continue to suffer from LQ 90-100, NLQ 0-20.   The other end does agree and reports LQ 0, NLQ 100.

My troubles sound very much like this:  http://www.aredn.org/content/nanobridge-m5-issue

Because of the findings of the thread listed above, I've tried a few different power levels to no avail.  I've also tried distance levels equal to and greater than the actual distance.  I haven't been as methodical on trying different settings on the end that reports zero simply because I've always been at the end that reports 100.  

Any help for this newbie would be greatly appreciated.  I've linked support data below.  

Thanks!

-neal

Support data from zero end: https://drive.google.com/open?id=0B2bOti4a1NXlaURXUjBLcnRrbks0Zk1sOGgxUU9qTVBnT1ln

Support data from 100 end:https://drive.google.com/open?id=0B2bOti4a1NXlNVNwUTlsWjZxbWc







  

AE6XE
AE6XE's picture
N1TEW,
N1TEW,

The NSM2 devices, at times can achieve good links at 10+ miles, but they generally do not perform well at this distance in my experience.  RF rich environments, which is any metro area, they're not going  to be stable links out this far.   For the 10+ mile distances, take a look at the NanoBridge M5 devices.  These are also excellent price and can be sourced online for 5-packs at ~$50 each.    The nightly builds and next release will also support he PowerBeam M5 devices PBE-M5-300 and PBE-M5-400 to consider as well.  

why 5Ghz vs 2Ghz?  More channel options and less noise (better performance)...    It's also inefficient to run everything all on one channel in an area.  I know this is counter intuitive to the the benefits of a 'mesh'.  But because this is all built on top of 802.11 RF links, the performance can stay at a crawl if every devices is on the same channel.   I work to build in redundant paths, but with multiple RF channels to scale up data capacity.   3Ghz is good too, with many channel options, just more $$s.   It's not the band per se, it's having choices of clear and multiple channels to use.

Joe AE6XE 
N1TEW
N1TEW's picture
Thanks for the quick response

Thanks for the quick response, Joe.   

I completely understand what you're saying.  The reason I have trouble giving up on it is because the signal levels on these long perfect line of sight links are actually better than my shorter links that have more in the fresnel zone.  I have no problems getting 100/100 on links in the upper 70s, but the links I'm having trouble with are often -69 to -73.  One of the ones that reports zero packets is out in the country miles from anything.  

Based on the other thread, is it not possible that the AREDN GUI is trying to write some distance or TX power levels that the radio does not accept?  

I've got so many people bad mouthing mesh that I was really hoping this would work.  I just don't understand why my strongest RF paths can be 100 % in one direction and 0% in the other.

How hard is it to re-load the Ubiquiti firmware that I have on file?  I'd like to try the path with that and once it comes up I'll just use it for a dtd link. 

N1TEW
N1TEW's picture
Here's a picture to give you

Here's a picture to give you an idea of the local RF environment.  This radio hears me at -73 which is better than any of my 100/100 links, but decodes no packets while I decode 100% at the other end:

 

Image Attachments: 
AE6XE
AE6XE's picture
Wow, nice view.  I need a
Wow, nice view.  I need a vacation :) .    let me take a look at the support dumps.   Unless there's a tower site you share or right next door, yea, this shouldn't be a noise issue.  Something else is going on.    What's the line-of-sight distance?
N1TEW
N1TEW's picture
The pictured LOS is 20 miles.
The pictured LOS is 20 miles.  The other end is sparsely residential.  I'm also having trouble on links of 8 and 11 miles that are more into town.
AE6XE
AE6XE's picture
 N1TEW,   I read your
 N1TEW,   I read your original post too fast.   I was thinking NanoStation not NanoBridge.    You should be able to reach out past 10 miles with a NanoBridge M2 on both ends, and get out to 20+ miles.   

The device in the other thread is a Ubiquiti device designed and built for the EU.  It is an unknown entity.  The NBM2 certified for US part 15 are known devices.    We do known, e.g. 3Ghz bottom band edge, that the device will taper off the power output going to this band edge a little.  Maybe there is a bit of this happening with 2Ghz devices too.  However, the other thread was about not even getting power in the near field or local proximity within yards and miles--very different symptoms.

It is possible to tftp load the AirOS back in from AREDN (no sysupgrade path).   But this will not enable the use of ch -2 and generally move the link into more noise.  The environment (other signals, microwaves, home wifi, ...) in 2Ghz seems to be a bigger factor than in 3Ghz and 5Ghz.    Let' just say, I've started to shy away from 2Ghz more as I found it simply hasn't worked as well compared to the other bands for me in our local RF rich environment.    Your environment and mileage may vary...

What channel width are you using?   Going to 5Mhz bumps up the SNR 3dB from 10Mhz and may help.    I don't think anyone is able to use 20Mhz channels at these distances, at least I've not heard of anyone able to do this on 2Ghz.

I've not looked at the support dumps just yet...

Joe AE6XE
N1TEW
N1TEW's picture
Thanks for the advice.  The
Thanks for the advice.  The problem links are on 5 Mhz channels -1 and -2.  

My other links that have less signal strength are running 20 Mhz and have no trouble supporting our VOIP PBX.  
AE6XE
AE6XE's picture
Neil,  change the netmask for
Neil,  change the netmask for the wifi interface back to "255.0.0.0" on all the devices.   I'd have to think though the implication of using "255.255.255.0" netmask, but I suspect  OLSR, which determines LQ/NLQ and routing tables, might be broadcasting UDP packets to a missmatching subnet.  Thus, the mesh node is not 'hearing' these packets.   The design idea being the whole mesh is on the same subnet in this regard, so that any device can show up to the incident site and connect in. 

The SNR looks fine and you should be receiving packets, just the higher level IP layer may think the broadcast is going to someone else.     Basically, go back to out-of-box determination of the network settings to confirm everything works.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer