PoE is enabled and port has link, power should not be 0

DOWNLOAD THE LATEST FIRMWARE HERE
User avatar
sirhc
Employee
Employee
 
Posts: 7347
Joined: Tue Apr 08, 2014 3:48 pm
Location: Lancaster, PA
Has thanked: 1597 times
Been thanked: 1317 times

Re: PoE is enabled and port has link, power should not be 0

Mon Dec 28, 2015 3:18 pm

It "could" also just be a "bad" SFP module?
Support is handled on the Forums not in Emails and PMs.
Before you ask a question use the Search function to see it has been answered before.
To do an Advanced Search click the magnifying glass in the Search Box.
To upload pictures click the Upload attachment link below the BLUE SUBMIT BUTTON.

User avatar
prot0fein
Member
 
Posts: 4
Joined: Tue Jul 28, 2015 9:52 am
Location: Sudbury, ON
Has thanked: 0 time
Been thanked: 0 time

Re: PoE is enabled and port has link, power should not be 0

Wed Feb 03, 2016 12:02 am

I have a problem very similar to this, but the ports that report '0' power, do actually show active wattage i would say half the time.
I too have an SFP Module in port 26 on a WS-24-400A model switch. I need to confirm what the brand of SFP is in the port, but I can tell you it isn't a 3com. Also, the SFP in port 26 is in fact currently working and in production. So until I read this thread, I would have never considered the optics as a possible problem to a false power reading on physically separate ports. Given this information, I will look at trying different optics.

The devices plugged into the ports that report '0' power are also up, active & in production at this time, providing power to the device plugged into it.

I guess the most annoying thing about this is that my device log is littered with these false reports of '0' power.

User avatar
sirhc
Employee
Employee
 
Posts: 7347
Joined: Tue Apr 08, 2014 3:48 pm
Location: Lancaster, PA
Has thanked: 1597 times
Been thanked: 1317 times

Damaged Current Sensor

Wed Feb 03, 2016 1:08 am

This could be an SFP module interfering with the I2C bus or it could be damaged current sensor(s).

If you remove the SFP module and all is good then you know it is either a bad or incompatible SFP module.

If the problem persists then the current sensor(s) are damaged. Current sensors are damaged from exposure to more than 80V either positive or negative.

Read this post on my recent damage because a co-tenant (Sprint Cellular) of a tower/water tank left and their crew removed the tower grounding system for the copper scrap.
viewtopic.php?f=30&t=1429

You can also read this post on good grounding practices
viewtopic.php?f=30&t=188

A bad current sensor will either read no or lower then expected watts or read watts when nothing is there. If a current sensor is damaged in the right way it can affect other sensors as they all share the same I2C bus.

It is important that ESD and Static charges at the Radio/Antenna have a shorter less resistant path to the grounding system through the intended grounding wire than through the Ethernet cable.

I always shake my head when people ground the base of a steel tower and think that is good. First off towers are usually multiple pieces of steel bolted together. Forget the fact that steel has a higher resistance value then copper but all those tower sections bolt together and every section joint is not a great electrical connection meaning they insert resistance.

From your radio/antenna you need a shorter heavier copper ground path with less resistance to Earth Ground than the Ethernet cables. This is another reason why you use a service loop in your Ethernet cables to insure the Ethernet cable path is longer and of a higher resistance than the intended ground wire path to Earth Ground.

If there is no dedicated copper wire for grounding from your radio/antenna to Earth Ground that is a shorter path and has less resistance than through the Ethernet cable then ESD and Static charge will follow your Ethernet cable.

If your tower ground potential is different than your Electrical service ground because they are not properly bonded then your Ethernet cable is the bond and the ground potential difference will constantly try to equalize through your Ethernet cable.

If your ground rods are not sufficient and in virgin clay or your ground potential is greater than 5 Ohms under normal conditions then rain events or ESD or Static discharges can overwhelm your grounding system.

Look the cellular guys put in those expensive grounding systems for a reason because with THOUSANDS of tower sites all over the place they would have towers dropping out all the time all over the place but they do not because they do GOOD grounding systems. Normally cellular sites take all but direct hits and keep on ticking.

You need good grounding. Look I have been preaching this for many years way back on the UBNT forums long before I ever started building switches. Do a search for my grounding posts over there and they preach the same things I am preaching today.

The only difference now is people expect me to replace or repair switches that get damaged from poor grounding for free.

We warranty our switches for defects or failures for 1 year but not from surges, ESD, Static discharges or user errors such as shorted cables or applying the wrong POE to an incompatible device.


Today I did 10 RMA's
(3) Were repaired and returned for FREE because of component failure
(2) Were blown Ethernet Transformers - Charged for repair
(2) Were simple firmware upgrades - DUH
(4) were visible Damage to boards from grounding faults or simply bad luck because sometimes you can do everything right and still take damage. - Charged for repair.

The pictures below are from 3 of the damaged boards I charged to repair or replace the boards today:

This is a blown current sensor "probably" from grounding issues, either inadequate grounding and or tower grounding system is not bonded to service ground properly.
current sensor damage.jpg


The next 2 pictures are from either poor grounding and or just bad luck but damage is the ESD protection circuit that is on the inside of the Ethernet Transformer which means it "probably" was a sudden high voltage surge as it had to jump the Magnetic coupling of the Ethernet Transformer and be high enough voltage to pop the chip like that. Now it is possible to be a low steady excessive current and slowly heat that chip up until it splats but more likely it was sudden but my crystal ball is no better then yours but I do know this is NOT a failure due to defect or simple parts failure but rather the result of environmental circumstances.
ESD 2.jpg

ESD 1.jpg
Support is handled on the Forums not in Emails and PMs.
Before you ask a question use the Search function to see it has been answered before.
To do an Advanced Search click the magnifying glass in the Search Box.
To upload pictures click the Upload attachment link below the BLUE SUBMIT BUTTON.

User avatar
sakita
Experienced Member
 
Posts: 200
Joined: Mon Aug 17, 2015 2:44 pm
Location: Arizona, USA
Has thanked: 87 times
Been thanked: 80 times

Re: PoE is enabled and port has link, power should not be 0

Mon Feb 08, 2016 3:32 pm

Had the same issue on a WS-12-250-AC (see viewtopic.php?f=17&t=1433). Upgrading to 1.3.9 final didn't fix it but...

...power cycling seems to have (had tech take different brand SFPs along to try but didn't need them).

Will be monitoring and share update if this changes. :cheers:

Repeat the tech support mantra: "have you tried rebooting" :???: and the 2nd mantra which is "have you tried power cycling it?" :P
Today is an average day: Worse than yesterday, but better than tomorrow.

scott@airgrids.com
Member
 
Posts: 11
Joined: Sat May 21, 2016 5:05 pm
Has thanked: 5 times
Been thanked: 1 time

Re: PoE is enabled and port has link, power should not be 0

Sat Jul 02, 2016 5:16 pm

Although this looks like an old thread, I'm having the same issue with the POE error.

The switch also powers down the link and cycles the radios. Below is the log. I'm using a WS-12-250-AC (Rev D). It is at the base of an FM tower (although I thought the AC units did not have issues, only the older DC units).

The problem started as soon as I installed the unit (with the shipping version of Firmware as of 12/2015). I moved radios (UBNT RM5, NStream 5M) from ports 1/2 to 7/8. Upgraded the firmware to 1.4.2. Retested, same issue. Powered off again moved to ports 9/10. Powered up -- Same. Rebooted -- Same. I turned off port 10 and just left the BackHaul on (port 9). The port power cycling seems to have stopped, however, still getting the errors.

I also have two DC units that are Rev E. I've not tried to power them up there, as it seems they would need to be upgraded to Rev F or better?

Thoughts?

------------- copy of log ---------------
Dec 31 17:00:05 netonix: 1.4.2 on WS-12-250-AC
Dec 31 17:00:10 system: Setting MAC address from flash configuration: EC:13:B2:61:36:08
Dec 31 17:00:11 system: starting ntpclient
Dec 31 17:00:12 admin: adding lan (eth0) to firewall zone lan
Dec 31 17:00:21 STP: set port 9 to discarding
Dec 31 17:00:23 dropbear[732]: Running in background
Dec 31 17:00:24 STP: set port 9 to learning
Dec 31 17:00:24 STP: set port 9 to forwarding
Dec 31 17:00:27 switch[769]: Detected warm boot
Dec 31 17:00:28 system: starting ntpclient
Jul 2 14:54:06 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 14:54:32 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 14:59:29 Port: link state changed to 'down' on port 9
Jul 2 14:59:29 STP: set port 9 to discarding
Jul 2 14:59:31 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 14:59:31 STP: set port 9 to discarding
Jul 2 14:59:33 STP: set port 9 to learning
Jul 2 14:59:33 STP: set port 9 to forwarding
Jul 2 15:03:22 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 15:03:37 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 15:06:00 Port: link state changed to 'down' on port 9
Jul 2 15:06:00 STP: set port 9 to discarding
Jul 2 15:06:02 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 15:06:02 STP: set port 9 to discarding
Jul 2 15:06:05 STP: set port 9 to learning
Jul 2 15:06:05 STP: set port 9 to forwarding

User avatar
sirhc
Employee
Employee
 
Posts: 7347
Joined: Tue Apr 08, 2014 3:48 pm
Location: Lancaster, PA
Has thanked: 1597 times
Been thanked: 1317 times

Re: PoE is enabled and port has link, power should not be 0

Sat Jul 02, 2016 5:50 pm

That is correct, AC models do not have a problem with FM towers sites so that os not your issue.

Jul 2 14:54:06 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0


My "guess" is you more then likely have damaged current sensors on Port 9 which is why is says power should not be 0.

The switch is saying POE is enabled but it is reading 0 watts used so it thinks the device does not need POE so it is disabling POE so as not to damage a non POE device.

Current sensors get damaged from surges or ground potential differences. To blow a current sensor you need to have voltage > 80V DC as that is the upper limit rating of the current sensor or a ground potential difference which forces excessive "current" through the port. Surges can come from Static or ESD discharges.

If you manually turn POE back on look and see if any watts are being reported as being used, my guess is it shoes 0 watts.

Move the cable to another port and see if the problem goes away.

We could think about taking this feature out, I will discuss with Eric and Dave.

Here are some good posts on grounding:
viewtopic.php?f=30&t=1816
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429
Support is handled on the Forums not in Emails and PMs.
Before you ask a question use the Search function to see it has been answered before.
To do an Advanced Search click the magnifying glass in the Search Box.
To upload pictures click the Upload attachment link below the BLUE SUBMIT BUTTON.

User avatar
sirhc
Employee
Employee
 
Posts: 7347
Joined: Tue Apr 08, 2014 3:48 pm
Location: Lancaster, PA
Has thanked: 1597 times
Been thanked: 1317 times

Re: PoE is enabled and port has link, power should not be 0

Sat Jul 02, 2016 5:55 pm

I asked Eric to not turn the POE OFF but to just put a log entry in and alert you to the issues with red explanation mark.
Support is handled on the Forums not in Emails and PMs.
Before you ask a question use the Search function to see it has been answered before.
To do an Advanced Search click the magnifying glass in the Search Box.
To upload pictures click the Upload attachment link below the BLUE SUBMIT BUTTON.

scott@airgrids.com
Member
 
Posts: 11
Joined: Sat May 21, 2016 5:05 pm
Has thanked: 5 times
Been thanked: 1 time

Re: PoE is enabled and port has link, power should not be 0

Sun Jul 03, 2016 11:51 pm

The only damage I can think of from your examples would be ESD. Although, I just ran it in my office for a few different weeks. burned it in. Then did the one install. This all occurred during its first install and as soon as I hooked it up. It occurred on all 6 ports that I tried. I captured a little longer log....
Do you still think its just using so "little energy" that it thinks it should shut the POE off? If so, should I just wait for a firmware update?

Do you have a way for me to quickly test to see if I have a blown board?


ul 2 15:42:30 STP: set port 9 to discarding
Jul 2 15:42:32 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 15:42:32 STP: set port 9 to discarding
Jul 2 15:42:33 Port: link state changed to 'down' on port 9
Jul 2 15:42:34 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 15:42:34 STP: set port 9 to discarding
Jul 2 15:42:35 Port: link state changed to 'down' on port 9
Jul 2 15:42:37 Port: link state changed to 'up' (10M-F) on port 9
Jul 2 15:42:37 STP: set port 9 to discarding
Jul 2 15:42:39 STP: set port 9 to learning
Jul 2 15:42:39 STP: set port 9 to forwarding
Jul 2 15:42:46 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 15:43:34 Port: link state changed to 'down' on port 9
Jul 2 15:43:34 STP: set port 9 to discarding
Jul 2 15:43:36 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 15:43:36 STP: set port 9 to discarding
Jul 2 15:43:38 STP: set port 9 to learning
Jul 2 15:43:38 STP: set port 9 to forwarding
Jul 2 15:43:56 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 15:44:01 Port: link state changed to 'down' on port 9
Jul 2 15:44:01 STP: set port 9 to discarding
Jul 2 15:44:03 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 15:44:03 STP: set port 9 to discarding
Jul 2 15:44:05 STP: set port 9 to learning
Jul 2 15:44:05 STP: set port 9 to forwarding
Jul 2 15:44:14 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 15:44:39 switch[751]: Port 9: PoE is enabled and port has link, power should not be 0
Jul 2 15:44:48 Port: link state changed to 'down' on port 9
Jul 2 15:44:48 STP: set port 9 to discarding
Jul 2 15:44:49 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 15:44:49 STP: set port 9 to discarding
Jul 2 15:44:50 Port: link state changed to 'down' on port 9
Jul 2 15:44:52 Port: link state changed to 'up' (100M-F) on port 9
Jul 2 15:44:52 STP: set port 9 to discarding
Jul 2 15:44:52 Port: link state changed to 'down' on port 9
Jul 2 15:44:54 Port: link state changed to 'up' (10M-F) on port 9
Jul 2 15:44:54 STP: set port 9 to discarding
Jul 2 15:44:56 STP: set port 9 to learning

User avatar
sirhc
Employee
Employee
 
Posts: 7347
Joined: Tue Apr 08, 2014 3:48 pm
Location: Lancaster, PA
Has thanked: 1597 times
Been thanked: 1317 times

Re: PoE is enabled and port has link, power should not be 0

Mon Jul 04, 2016 6:58 am

scott@airgrids.com wrote:The only damage I can think of from your examples would be ESD. Although, I just ran it in my office for a few different weeks. burned it in. Then did the one install. This all occurred during its first install and as soon as I hooked it up. It occurred on all 6 ports that I tried. I captured a little longer log....
Do you still think its just using so "little energy" that it thinks it should shut the POE off? If so, should I just wait for a firmware update?

Do you have a way for me to quickly test to see if I have a blown board?


ESD or Static discharge is a common damage we see but most common is Ground Potential Shift from no or improper bonding between the tower and service ground and no dedicated bond wire from cabinet to radios. Also we see a lot of damage from people getting power from a distance and using the Earth Ground from the Electric service from a Distance.

If using AC power from a distance do what is described in this post:
viewtopic.php?f=17&t=1786&start=30#p13447

Sure, here is how you test a switch, take it to your bench:

1) Factory default it and log into it with your laptop.

2) Verify each port gets a 1G connection to your laptop or another 1G device that is not POE such as a switch or another laptop. Any port that does not get 1G do a cable diagnostic with nothing plugged in to the port. You should get OPEN, OPEN, OPEN, OPEN for all 4 pairs. Anything else and you have a damaged Ethernet Transformer. Or if the test hangs you also have a damaged transformer and possibly damaged PHY.

3) Use any airMAX radio known good such as a NanoStation and plug into each port and turn 24V POE ON. Now observe the Status Tab and you should get 1-6 watts, if you get 0 or HIGH watts you have damaged current sensors.

4) Verify all sensors on Device/Status Tab, if they are wrong chances are I2C bus is damaged somewhere probably a current sensor.

Here are some good posts on grounding:
viewtopic.php?f=30&t=1816
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429
Support is handled on the Forums not in Emails and PMs.
Before you ask a question use the Search function to see it has been answered before.
To do an Advanced Search click the magnifying glass in the Search Box.
To upload pictures click the Upload attachment link below the BLUE SUBMIT BUTTON.

scott@airgrids.com
Member
 
Posts: 11
Joined: Sat May 21, 2016 5:05 pm
Has thanked: 5 times
Been thanked: 1 time

Re: PoE is enabled and port has link, power should not be 0

Mon Jul 04, 2016 10:22 am

Sirhc. I did use a 100' 12/3 extension cord for power to the switch/tower. I didn't change the ground to the local tower earth ground as you described - so that likely didn't help. I will tie the earth ground as you suggest and run the tests above in a few days when I get back to the tower. I'll update then. THX.

PreviousNext
Return to Hardware and software issues

Who is online

Users browsing this forum: Bing [Bot], Google [Bot] and 13 guests