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

[sirhc]

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.

Yea what happens is Earth Ground Current will try to equalize between the 2 grounding systems and the path of this current looks like this:

EARTH GROUND===[Electric Service]===100' cord===[cabinet + switch]=====Ethernet Cable====[Tower]===EARTH GROUND

Now no matter which side has the better ground the current will try to equalize and flow THROUGH your switch. - NOT GOOD

Very possible you have damaged your switch, I would bench test it as I outlined above.

So many people think Ground is Ground and they could not be further from the fact/truth.

Keep in mind that any AC equipment located on the electric service wants to clamp excessive voltage/current to Earth Ground so if the Earth Ground is better at the tower then all the excessive voltage/current will flow through that 100' cords to and through your switch to get to the tower Earth Ground.

[scott@airgrids.com]

Ok. I climbed the tower yesterday and put up one RF shielded AP and one Titanium AP. I grounded the APs and the AC power with double AWG 6 to 1/4" copper plate. Then connected to the tower's electrical ground ring directly. The tower has ground on each post to the ground ring, so all is bonded, and the electrical is straight copper to the ground. I picked up the ground to the Netonix box (at the base of the tower) right with the electrical ground to the ground ring. So it is the shortest/thickest path possible.

I did all the grounding before adding any power. I used a brand new (still rev E) WS-12-250-AC. I bench tested the switch in the office and made sure all the values read correctly (per your earlier post in this thread). Then, after the grounding, powered up the equipment.

HERE ARE THE RESULTS:
The two APs and the incoming backhaul took a few minutes of bouncing down/up and speed changes. Then finally settled on the two APs at 10M instead of 100M.

The power readings always show errors on all POE devices.

Not sure what to try next, except maybe put up a toughswitch to see if it changes any of this behavior.



---LOG---
Dec 31 17:00:05 netonix: 1.4.2 on WS-12-250-AC
Dec 31 17:00:09 system: Setting MAC address from flash configuration: EC:13:B2:61:41:A0
Dec 31 17:00:12 udhcpc[511]: udhcpc (v1.19.4) started
Dec 31 17:00:12 udhcpc[511]: Sending discover...
Dec 31 17:00:13 udhcpc[511]: Sending select for 192.168.103.5...
Dec 31 17:00:13 udhcpc[511]: Lease of 192.168.103.5 obtained, lease time 259200
Dec 31 17:00:15 admin: adding lan (eth0) to firewall zone lan
Dec 31 17:00:16 system: starting ntpclient
Aug 13 09:41:15 Port: link state changed to 'down' on port 4
Aug 13 09:41:16 Port: link state changed to 'down' on port 2
Aug 13 09:41:16 Port: link state changed to 'up' (10M-F) on port 4
Aug 13 09:41:16 Port: link state changed to 'down' on port 4
Aug 13 09:41:16 Port: link state changed to 'down' on port 6
Aug 13 09:41:18 Port: link state changed to 'up' (100M-F) on port 2
Aug 13 09:41:18 Port: link state changed to 'up' (100M-F) on port 4
Aug 13 09:41:18 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:41:20 Port: link state changed to 'up' (1G) on port 1
Aug 13 09:41:20 Port: link state changed to 'down' on port 4
Aug 13 09:41:20 Port: link state changed to 'down' on port 1
Aug 13 09:41:21 Port: link state changed to 'up' (100M-F) on port 4
Aug 13 09:41:21 Port: link state changed to 'down' on port 4
Aug 13 09:41:21 Port: link state changed to 'up' (100M-F) on port 1
Aug 13 09:41:23 STP: set port 1 to discarding
Aug 13 09:41:23 STP: set port 2 to discarding
Aug 13 09:41:23 Port: link state changed to 'up' (100M-F) on port 4
Aug 13 09:41:23 STP: set port 4 to discarding
Aug 13 09:41:23 STP: set port 6 to discarding
Aug 13 09:41:23 STP: set port 4 to discarding
Aug 13 09:41:24 Port: link state changed to 'down' on port 4
Aug 13 09:41:25 dropbear[810]: Running in background
Aug 13 09:41:25 Port: link state changed to 'up' (100M-F) on port 4
Aug 13 09:41:25 STP: set port 4 to discarding
Aug 13 09:41:26 STP: set port 6 to learning
Aug 13 09:41:26 STP: set port 6 to forwarding
Aug 13 09:41:26 STP: set port 2 to learning
Aug 13 09:41:26 STP: set port 2 to forwarding
Aug 13 09:41:26 STP: set port 1 to learning
Aug 13 09:41:26 STP: set port 1 to forwarding
Aug 13 09:41:26 Port: link state changed to 'down' on port 4
Aug 13 09:41:28 Port: link state changed to 'up' (100M-F) on port 4
Aug 13 09:41:28 STP: set port 4 to discarding
Aug 13 09:41:29 Port: link state changed to 'down' on port 4
Aug 13 09:41:29 switch[840]: Detected warm boot
Aug 13 09:41:29 Port: link state changed to 'up' (100M-F) on port 3
Aug 13 09:41:29 STP: set port 3 to discarding
Aug 13 09:41:29 STP: set port 3 to learning
Aug 13 09:41:29 STP: set port 3 to forwarding
Aug 13 09:41:30 Port: link state changed to 'up' (100M-F) on port 4
Aug 13 09:41:30 STP: set port 4 to discarding
Aug 13 09:41:31 Port: link state changed to 'down' on port 4
Aug 13 09:41:33 Port: link state changed to 'up' (10M-F) on port 4
Aug 13 09:41:33 STP: set port 4 to discarding
Aug 13 09:41:35 STP: set port 4 to learning
Aug 13 09:41:35 STP: set port 4 to forwarding
Aug 13 09:43:34 Port: link state changed to 'down' on port 6
Aug 13 09:43:34 STP: set port 6 to discarding
Aug 13 09:43:35 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:43:35 STP: set port 6 to discarding
Aug 13 09:43:36 Port: link state changed to 'down' on port 6
Aug 13 09:43:37 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:43:37 STP: set port 6 to discarding
Aug 13 09:43:40 STP: set port 6 to learning
Aug 13 09:43:40 STP: set port 6 to forwarding
Aug 13 09:43:46 Port: link state changed to 'down' on port 6
Aug 13 09:43:46 STP: set port 6 to discarding
Aug 13 09:43:47 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:43:47 STP: set port 6 to discarding
Aug 13 09:43:48 Port: link state changed to 'down' on port 6
Aug 13 09:43:50 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:43:50 STP: set port 6 to discarding
Aug 13 09:43:51 Port: link state changed to 'down' on port 6
Aug 13 09:43:52 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:43:52 STP: set port 6 to discarding
Aug 13 09:43:55 STP: set port 6 to learning
Aug 13 09:43:55 STP: set port 6 to forwarding
Aug 13 09:43:55 Port: link state changed to 'down' on port 6
Aug 13 09:43:55 STP: set port 6 to discarding
Aug 13 09:43:56 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:43:56 STP: set port 6 to discarding
Aug 13 09:43:57 Port: link state changed to 'down' on port 6
Aug 13 09:43:58 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:43:58 STP: set port 6 to discarding
Aug 13 09:43:59 Port: link state changed to 'down' on port 6
Aug 13 09:44:00 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:00 STP: set port 6 to discarding
Aug 13 09:44:02 Port: link state changed to 'down' on port 6
Aug 13 09:44:04 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:04 STP: set port 6 to discarding
Aug 13 09:44:05 Port: link state changed to 'down' on port 6
Aug 13 09:44:07 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:07 STP: set port 6 to discarding
Aug 13 09:44:09 STP: set port 6 to learning
Aug 13 09:44:09 STP: set port 6 to forwarding
Aug 13 09:44:13 Port: link state changed to 'down' on port 6
Aug 13 09:44:13 STP: set port 6 to discarding
Aug 13 09:44:15 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:15 STP: set port 6 to discarding
Aug 13 09:44:18 STP: set port 6 to learning
Aug 13 09:44:18 STP: set port 6 to forwarding
Aug 13 09:44:20 Port: link state changed to 'down' on port 6
Aug 13 09:44:20 STP: set port 6 to discarding
Aug 13 09:44:22 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:22 STP: set port 6 to discarding
Aug 13 09:44:22 Port: link state changed to 'down' on port 6
Aug 13 09:44:24 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:24 STP: set port 6 to discarding
Aug 13 09:44:27 STP: set port 6 to learning
Aug 13 09:44:27 STP: set port 6 to forwarding
Aug 13 09:44:33 Port: link state changed to 'down' on port 6
Aug 13 09:44:33 STP: set port 6 to discarding
Aug 13 09:44:35 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:35 STP: set port 6 to discarding
Aug 13 09:44:38 STP: set port 6 to learning
Aug 13 09:44:38 STP: set port 6 to forwarding
Aug 13 09:44:41 Port: link state changed to 'down' on port 6
Aug 13 09:44:41 STP: set port 6 to discarding
Aug 13 09:44:42 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:42 STP: set port 6 to discarding
Aug 13 09:44:43 Port: link state changed to 'down' on port 6
Aug 13 09:44:45 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:45 STP: set port 6 to discarding
Aug 13 09:44:45 Port: link state changed to 'down' on port 6
Aug 13 09:44:47 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:47 STP: set port 6 to discarding
Aug 13 09:44:47 Port: link state changed to 'down' on port 6
Aug 13 09:44:49 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:49 STP: set port 6 to discarding
Aug 13 09:44:49 Port: link state changed to 'down' on port 6
Aug 13 09:44:51 Port: link state changed to 'up' (100M-F) on port 6
Aug 13 09:44:51 STP: set port 6 to discarding
Aug 13 09:44:52 Port: link state changed to 'down' on port 6
Aug 13 09:44:54 Port: link state changed to 'up' (10M-F) on port 6
Aug 13 09:44:54 STP: set port 6 to discarding
Aug 13 09:44:57 STP: set port 6 to learning
Aug 13 09:44:57 STP: set port 6 to forwarding
Copyright 2014-2016 Netonix

[sirhc]

I do not know what to tell you other than send it in for RMA however I will tell you that we test every single unit to determine that no current sensors are damaged and are working prior to serializing them.

If a current sensor is "blown" it is visible damage on the board and we always send pictures of visible damage to the customer with an explanation of what probably happened as the chip will be literally blown and a black mark will be present.

Current sensors can "only" be damaged by 2 thing:
1) Exposure to voltage > 80V which our switch is incapable of generating
2) A ground potential difference / shift

A ground potential shift is cause from 2 separate grounding systems that are not properly bonded such as one ground system for a tower and one for the electric service that are not bonded with a heavy wire such as #2. Then when you plug in the switch to AC socket then run a wire to a radio on a tower the Ethernet cable becomes that bond and if there is excess current on either grounding system it will flow through the Ethernet cable, radio, and switch. This current is often low voltage but high current and will over heat the components and POP however the excess current can sometimes also be high voltage especially from Static buildup from wind blowing over the antenna at the proper humidity or an ESD discharge that can sometimes be a very high voltage low amperage.

So if we examine a switch board and find no visible physical damage or damaged Ethernet transformers which usually is not visible but easy to diagnose it is fixed under WARRANTY but if there is physical damage as described there is a repair fee.

Most times when a current sensor is "damaged" not failed it normally only affects that single port however "sometimes" if it gets blown is just the right manor that prevents the I2C bus from working or the damage is severe enough that it actually takes out the I2C bus it can cause all sensors to report ERROR as they all share the same bus.

I was trying to follow your grounding system as best I could and I think you have it right but if your interested in my opinion could provide a diagram and I will review it?

What is the MAC address of your unit and could you please post up a screen grab of the Status and Device Status Tab

[scott@airgrids.com]

Sirc! - I'm sure your product is tested and is quality. Just trying to figure out what is happening and what I should do about it.

I was able to get back up there and do some additional configuration and testing. Also, I'll attach a really bad graphic on my grounding system. I'm not sure it will help clear up my info, but you did ask. I know I'll get slammed on this one. Remember this is all on an FM tower and about 25' up from the ground and about another 20' to FM antennas.

A) WHAT I DID:
Added metal boxes (along with the RF armor), used high-quality STP cat5e. Plugged in an AC (APC battery surge suppressor - all I had on hand) at the base of the tower powering the WS-12-AC-250.
A) RESULT: Still had errors on all POE ports, regardless of ports used. However, port down/up and speed negotiation cleaned up significantly.

B) WHAT I DID:
Swapped out the WS-12-AC-250 Rev. E for a WS-8-DC-150 Rev. F. Powered from 12v battery at the base of the tower.
B) RESULT: WithOut a metal box around switch/router, All POEs show wattage. Full 100M connections, some stp port down/ups.
B) RESULT: WITH a metal box around switch/router, All POEs show wattage. Full 100M connections, NO stp port down/ups. Working as expected.

C) WHAT I DID:
Took the WS-12-AC-250 Rev. E into the outbuilding (~60ft away from the tower) and then plugged in one POE device.
C) RESULT: Full 100M connection to the internet, ZERO ports down/up. Shows watts used (no more error).

D) CONCLUSION:
Maybe FM signals were entering into the switch via the AC power cord, and causing the POE errors? FM seems to be the cause of the problem.

Use more significant Faraday type cage and extra shielding around ALL electronics (including the Netonix). I may just buy copper tubing as conduit/ground for any cat5e runs from AP to Switch.

Piss Poor Diagram! Hope this helps
https://docs.google.com/presentation/d/ ... sp=sharing

[Dave]

Scott

Thanks for you nice, detailed feedback.

You did all of the correct steps, narrowing down issue & finding a solution.

Our AC & latest Rev DC switches have good FM immunity, but no switch is perfect. The AC & DC power supply's have filtering caps that shunt the FM range frequencies to the chassis ground. The caps can only shunt so much FM energy, and it is only shunted to chassis. Now if the chassis ground wire also has lots of FM noise on it, it gets hard for the shunted FM noise to go anywhere....kind of a cat & mouse game going on.

Sometimes the only real solution is to indeed go the Faraday cage solution like you are doing.

Well done.

I will let Chris comment on your diagram.

Thanks.

Dave