Netonix POE Smart Cable Check Error after unrelated change

[jeremylu]

We rebooted non-POE equipment on a different port and at that time it triggered some STP errors on that port (port 5) and then our device plugged into port 11 went offline. I tried re-enabling the 48V POE and am now getting this error:


Port 11 failed PoE Smart cable check (DEAD SHORT DETECTED), disabling PoE


A while ago port 12 stopped being able power POE as well, we stopped using it for power and do data only now.

[Julian]

Generally the cable diags are pretty reliable, you may have one or more blown ethernet transformers. Given that the switch still boots, it has about an 80% chance of being repairable.

RMA instructions can be found here: viewtopic.php?f=6&t=1259

Before you send it in, try:

Remove the cable in port 11, and run cable diagnostics again. If you get a short result, it's RMA time. If you don't, there's a problem with the powered device or cable.

Port 12 sounds like either a popped center tap or mosfet, both of those are generally repairable, but if it isn't broke in your application, don't fix it.

[sirhc]

OK your not providing needed information:

Model:
Firmware version:

"If" you are using Ethernet surge protectors that are designed for ACTIVE POE you could be damaging your switch as they work by clamping everything to ground which creates a DEAD SHORT and can damage your switch since PASSIVE POE has no way to deal with this and will fry components.

Run cable diagnostics on the ports in question with NO cable plugged in and POE OFF, you should get all OPEN on all 4 pairs.

If you get all OPEN then plug in your Laptop and see if you get a 1G connection.

Next you can test the POE function for that port by using short cable and airMAX radio to test 24V (never use 24VH with airMAX radios)

Next test 48V POE by powering an IP Phone or other 802.3af/at device with 48V

Next if the port supports 24VH or 48VH you can test that mode with an AFX radio as AFX radios support 24V, 48V, 24VH and 48VH.

When testing POE take note that proper watts are being reported on the Status Tab which will let you know if the current sensors are working and not damaged.

Passive POE is nothing magical:
24V applies 24V + to Pins 4 & 5 and negative to Pins 7 & 8 - 802.3af/at MODE B
48V applies 48V + to Pins 4 & 5 and negative to Pins 7 & 8 - 802.3af/at MODE B

24VH applies 24V + to Pins 1, 2, 4 & 5 and negative to Pins 3, 6, 7 & 8 - 802.3af/at MODE A & B at the same time
48VH applies 48V + to Pins 1, 2, 4 & 5 and negative to Pins 3, 6, 7 & 8 - 802.3af/at MODE A & B at the same time

So what can we learn from some simple tests explained above:
If current sensors are damaged then you will get lower or higher than expected wattage or even 0 watts
If Transformers are damaged on pairs 3 or 4 you will only get a 10/100 connection
If Transformers are damaged on pairs 1 or 2 you not get any connection
If MOSFETs are damaged you will not get POE power

If people are careful and investigate their issue they can determine what is the cause of their no happy.

Q) Can damaged ports be repaired?
A) Sometimes yes as long as there is no board damage.

Q) What does it cost to repair a port?
A) Usually first port is $50 then about $25 per port after that.

Q) What if the board is damaged in a way we can not repair the port?
A) We sale replacement boards at cost (includes programming and installation), see below:

WS-6-MINI - $75.00
WS-8-250-XX - $100.00
WS-8-150-AC - $100.00
WS-8-150-DC - $150.00
WS-10-250-AC - $125.00
WS-12-250-XX - $150.00
WS-24-400A - $275.00
WS-24-400B - $225.00
WS-26-X00-XXX -$225

So even if your switch is out of warranty we still repair it and it is always cheaper to repair it then buy a new one.

[jeremylu]

We are using the Ubiquiti ETH-SP Surge Protectors. Is that not recommended? What would you recommend we use, if anything?

I will likely do all your recommended tests tomorrow, as all my technicians did was move the equipment back to the PacketFlux injector, the cable and ETH-SP may still be connected.

First test since they left showed Short on 1 & 2, but subsequent tests show Open. I will have to do further testing on site. I'm 99% sure port 12 is toast though and has been for months, just waiting until I have a spare to RMA, we were just using the port for NON-POE function until I got the spare. It is definitely less than a year old.

WS-12-250 DC, 1.4.5rc4

[sirhc]

That is a very good question on what the UBNT surge arresters do or how they work as I do not use any in my WISP anywhere as I rely on good grounding which has protected me well.


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
viewtopic.php?f=17&t=1786&start=30#p13447
https://community.ubnt.com/t5/airFiber/ ... rue#M31070
I should order some in and tear them apart to see how they work and what they do.

[lligetfa]

The ETH-SP is designed to work with the all pairs 50V pinning of the AF24 so unless it took a hit and is now shorting to ground, it should be compatible with 48VH. Of course their raison d'être is to take a hit and short it to ground so if that happened, then after that, all bets are off.

[sirhc]

Hey Les,

I do know that most 802.3af/at Surge Arresters clamp all pairs to ground, I do not know what UBNT is doing in their arrestor, what are they doing with a presumed surge or excess voltage? How do they discharge the excess voltage?

If you have one Les you could perform some exploratory surgery on it and let me know what you find?

[lligetfa]

The principle they work on is to clamp only an overvoltage to ground and when shopping for a brand/model, one needs to be aware of what their voltage threshold is. I don't have any of the ETH-SP but as I said, they were designed with the 50V PoE that AF uses in mind so would not expect a properly working one to cause overcurrent issues. As for the diags, that is another story as they can cause false positives with the TDR function.

As I also mentioned, if they took a hit, they could be permanently shorted to ground or clamp at a lower voltage threshold so can no longer be trusted.

Surge protectors are a poor substitute for good grounding/bonding practices. They are mostly meant to meet NEC entrance protection requirements to protect life and buildings.

[sirhc]

Keep in mind that POE bricks provided from UBNT have short circuit protection so in the event all pairs are shorted to ground no damage is done, the POE brick simply shuts down and starts up when short is cleared.

Passive POE switches such as ToughSwitch and WISP Switch do not have short protection other than PolyFuses which are too slow to prevent damage, they are only there for safety measures to prevent personal injury and or fires so a short normally results in switch damage.

We have been slowly working on a version of passive POE that can react fast enough to prevent damage to components but this will not be implemented until we ever come out with the WS2 product line maybe late this year or next year.

The WS2 product line uses a custom chip to do current readings in real time with the ability to sense shorts fast enough to shut down MOSFET circuits to prevent damage to Transformers, Sensors, and Protection Diodes.

In reality it will be about the same cost to build as PolyFuses that we use today but only if we can get volumes up to about 1.5x to 2x what they are now.

[jakematic]

If you have one Les you could perform some exploratory surgery on it and let me know what you find?

Need anything else ?