Lost a PoE port

[TheHox]

Doing some testing in my home lab, I had a radio on Port 1 with PoE enabled (24V).

I was setting up some VLANS for some sectors, so I tested one on port 3. I realized I had to change some subnets so I unplugged that radio and plugged it back to my TS.
I did not disable PoE to port 3 when I unplugged it. Seconds later I notice my port 1 PoE has died.

Port 3 still works, other ports still work, port 1 LAN appears to work, just no PoE.

It smoked for good then eh?

I just read the post on the forum advising not to do what I just did. Too late I suppose.

[WisTech]

Disable 24V on that port, apply, then enable 24V on that port and apply.

[TheHox]

Did that.

Power cycled the switch. I even unplugged it then went to bed last night hoping I could sleep it off. Same today.

[sirhc]

Unplugging a device on port 3 would not effect port 1
So you are saying port 1 Ethernet communications works fine then?
See if 48V POE still works fine on port 1 please.
Also check port 2 for POE and communications.

Advising people to turn POE off on the port they are unplugging and plugging cables/radios on and to run cable diagnostics on new cables and radios before applying POE power is to prevent possible dead shorts from occurring that can damage the Ethernet transformers.

If you can not get 24V or 48V or 48VH that would indicate the center tap on the transformer got fried and the only way that could happen is if you dead shorted 48VH.

It is possible 24V dead short "could" damage the transformer but it normally took repeated attempts, not to say that 48V standard could not do it the first time but usually not (this is not an excuse to try) but 48VH always fried the transformer.

But we considered ours more durable to the ToughSwitch because most dead short tests we preformed on the ToughSwitches resulted in an entire switch loss.

But there is no way that unplugging something from or even dead shorting port 3 could fry port 1 as they are separate transformers.

And if it was 24V you would have to be very unlucky for the first dead short at 24V to take out the transformer but not impossible, 48V the odds are higher, and at 48VH almost certainly.

You can RMA it back and we can look at it and see what happened. If it just failed it is under warranty but if the transformer is fried from a dead short we will repair it for a fee, it is pretty easy to determine f the damage is from a dead short or not or just a part failure. Not sure what that fee is yet as we have not determined our costs but repairing boards is something we want to offer people as well as sell parts like power supplies and fans and even replacement boards for lightening damaged switches because why have to buy the whole switch when all you need is the board.

Depending on what the damage is obviously somethings are not worth fixing, cheaper to buy new board but transformers are some what easy to replace and diagnose and determine how they failed.

What firmware are you running?

[sirhc]

PS.

Another way to kill a transformer is trying to power an airFIBER or other radio that requires too much current for standard 48V.

Meaning if you plug an airFIBER in port 1 and turn on 48V you could fry the transformer as you would be trying to send too much power on a single pair over loading the Transformer.

NEVER try to power airFIBER or such radios with 48V standard on any ports, even the ports capable of 48VH as you are demanding too much current on a single pair which is over the rating of the transformer. We are able to supply that much power (1.2 Amps) by spreading it across TWO pairs. The Ethernet transformers are only rated at .72A per pair, the 48VH ports are 1.2 Amps buy combining 2 pairs and sharing the load across 2 pairs and across 2 channels in the transformer.

For high power radios like airFIBER you MUST use a port capable of 48VH and make sure to select 48VH not just 48V.

[TheHox]

I had a NanoBeamM5 plugged in Port 1 with PoE enabled. (24V)
I had its uplink to my router plugged in port 2. (both vland)
Those were working fine and live.

I had a vlan'd uplink for my sectors in port 3.
I then tested a sector using 24V in port 4 (not port 3 as stated above). I had to remove it to change some IP settings, so I unplugged it from port 4, and plugged it back into the TS. I sat down at my desk and noticed I was then disconnected from the internet. I then realized my PTP link died, look at the WispSwitch and see I have no PoE to port 1. So I quick throw a PoE brick and everything is working again.

I turned the 24V on/off, I've power cycled the switch, I even let it sit unplugged overnight.

I have since tested:
Port 1 lan communication works just fine
port 2 poe/lan works
port 3 poe/lan works
port 4 poe/lan works

I have not had any AF or any devices using 48V or 48VH in here at all, nor have I ever enabled any thing 48V.

I will dig out an AF from the garage and test 48VH now on ports 1 and 2 shortly.

[sirhc]

Yea please let me know if 48VH and or 48V works because if they do then maybe part of the 24V MOSFET circuit died and would definitely be covered under warranty.

As I said anything but a dead short, trying to power 48VH devices more than 25+ watts with 48V option, or powering more than 25+ watts at 24V is warrantied for 1 year.

[TheHox]

So I dug out a new AF24 out of the garage, tested on port 2 first, works fine.

port248.png (55.59 KiB) Viewed 7808 times

I then tested port 1 , again 48VH, it worked.

port148.png (54.79 KiB) Viewed 7808 times

I then tested 24V on port 1 confirm it didn't work, and, now it works again. ?

port124.png (55.4 KiB) Viewed 7808 times

I tested port 1 over and over it never worked. It only works now after I ran 48VH through it. What ever that did, seems to have fixed it.

[sirhc]

WEIRD?

What firmware are you running?

That airFIBER only draws 26 watts?

Maybe it only draws 26 watts since it is not paired and is only transmitting a beacon?

My airFIBERS pull 48+ watts

Glad it is working though!

Keep an eye on it and let us know!

[TheHox]

1.0.10rc13

Yea the AF was sitting in the box, not paired to anything so hopefully that is why it is low. The AF24 will get hung on this tower this month and the goal is for the WS to power it.

So everything is good for now.

Thanks