You are here

AirRouter HP oddity

6 posts / 0 new
Last post
N8NQH's picture
AirRouter HP oddity

UPDATE; it was not the AirRouter, read the responses for details !!!

I thought I had read an earlier discussion on this specific, but was unable to find it.

I was testing a known good AirRouter HP ... specifically using  the DTD port 4 for node connecting to a NanoStation NSM5; all nodes using the full 3.16 flash (not the beta).  My computer connected to ports 1,2,3 on the AirRouter

I  connected a powered known-good  NanoStation M5 into port 4 of this  AirRouter HP,  but the NS never appeared on the AirRouters Mesh Status page.  Thinking this AirRouter was misbehaving... I took out another AirRouter HP (also known good) and the same thing happened... the NanoStation NSM5 did not appear on the status page even though it was ethernet connected to the AirRouters port 4..

What is required for the AirRouter to DTD (ethernet link) to another node?


AE6XE's picture
This bites me every once in a
This bites me every once in a while, like just last week setting up for field day :) .    I'm guessing you have the NSM5 XW.    DtDlink is only on the secondary port, not the main, due to limitations of the internal switch.

NSM5 XW - main port:   LAN only
NSM5 XW - secondary port:  DtDlink and WAN only

Note, you can power NS's from the secondary port--a few forum posts on this. 

Highly recommend against the
Highly recommend against the feeding power via secondary port.

The understanding is your back feeding a mosfet and all the protection circuitry is on the other side. Fried mosfets are not unheard of as are rumors of full device failures.

Some of these scenarios include power always leaking from the primary port to the secondary port and when you go to plug in a non POE device it gets straight voltage into its pins.

In the end it's generally not worth back feeding from the secondary port.
k1ky's picture
2nd Port POE on XW Nanostation M5
Really? That means TWO runs up cable up the tower to run an XW Nanostation M5.  I have several installations with a single POE line going to the Nanostation XW secondary port on towers and mast pole mounts on structures.  These have been operational for over 4 months DTD to my switches so far and have survived some pretty nasty storms with no failures.  I thought the power was basically a "Pass-Through" on the two ports. Has anyone "traced" the circuitry, ohmed it out, seen a schematic?
AE6XE's picture
I found the first google
I found the first google postings on this power back-feeding by guys in the Ubiquiti forums when they turned on the primary->secondary power pass though in AirOS and the primary port failed and no longer powered up.  They figured out it still powered up on the secondary port and continued to serve the purpose.   I've yet to hear or read a post of anyone that has damaged a NS back feeding the power on the secondary port to date and I know people in my area that have been doing this for 2+ years with no issues.  I did this at field day last weekend.  

The experience so far says the risk doesn't jump up to damage these devices left and right.  The long term (years) usage of doing this is unknown, however.   
N8NQH's picture
yes, it is a XW         for

yes, it is a XW         for the Nanostation... I will give that secondary port a try.

FYI; on my NSM5 (XM) array (4 NSM5's - all on different 5.8 channels):

I am only using two POE's to power all four NanoStations.   I modified the NSM5's... jumping across the FET transistor (that switches the power from the Main to Secondary port) so that the Secondary port had power at all times now.  (this was done over a year ago, with XM NanoStations and BBHN; back then the secondary port had no data capability... and power was not activated to that secondary port either).  I forget which power line (either the plus or minus) that the FET interrupts, but I just simply soldered a jumper wire across the jacks on the inside.... on that one power "side" that was interrupted by the FET).

Here is how they were cabled in pairs (two pairs total):

 These have been running for 18 months now, no issues.

The only thing gained by doing this is that I only had to use two 1A POE's to power all 4 NSM5's.  Voltage is present on the two  CAT5e cables coming down from the NStations that do not go through a POE; ; fortunately the cheap 10/100 switch I combine them together with ignores pins 4-5, 7-8 (these "voltage" pins are missing on the jacks)... so this is where the voltage is stripped off.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer