Hi all,
I've just found the strangest problem I've ever encountered, and I am completely out of ideas. I'm hoping someone here can help me out
I've got a wireless antenna that I'm installing in a factory to improve wireless coverage. For the record, it's a UniFI AP. I'm using a UniFI 24VDC 0.5A power over ethernet adaptor to power it. I have proved this unit on the bench - all the settings are OK, it connects as it should, no issues whatsoever when I put it together on the bench.
In the field, what I've got is an ethernet switch at ground level, and the antenna up near the roof, with a 25m (80ft) run of CAT5E cable. I've plugged a patch cable from the switch into the POE adaptor in the ethernet switchboard, and my 25m cable then runs from the POE in the board up to the antenna. The antenna powers up and does it's self check, and the "I'm a happy antenna" light comes on, and you can connect to it with a wireless device, e.g. my iPad. However, the antenna can't connect to the network. So although my iPad is connected to the wireless signal, I can't join the network, and on the switch, there's no lights come on (not even the green "I've got something plugged into me" light). I thought maybe the port on the switch is dodgy, but I tried other ports, including ones that I know to work, because they were in service at the time. I even tried a different switch entirely. Same result.
I thought, obviously I've got a dodgy cable, and the power cores are OK, but the data cores are not. But they all meter out fine, at exactly the same resistance - there's no crossovers, shorted cores, etc. I thought that maybe 25m was too long a run for voltage drop considerations etc, although it seemed unlikely as the device powered up fine, just like it did on the bench, the signal is strong, I can connect to it, etc etc. But to test both these theories out, I ran a power lead all the way up to the switch so I could use the POE adaptor close by. So now what I've got is my 25m run plugged straight into the switch, carrying just the network up to the POE adaptor. I've then got the patch lead plugged from the POE adaptor to the antenna. Same two cables, just the other one now has the power. This way, it worked. So I know now that the data cores in both cables are intact and fine. Must be just the length over which I'm trying to run the power, I thought. I rang the supplier, and he said he'd never heard of this happening. He says he's run the POE three times the length I'm running it with no issues whatsoever. So I figured maybe the POE adaptor or the wireless antenna were at fault. I tried different ones of both - same result. When the power is on the long cable, power but no network; when the power is on the short cable, no issues. I decided it must therefore be the length or installation conditions of the cable run. Both seemed unlikely, as the run is the best run I've seen in the whole factory - it goes nowhere near any power cables, and the supplier said he'd run way further than me. But I was out of likely options, so I started on the unlikely ones. I remembered that when I'd run the cable, I'd cable tied it to another ethernet cable that was plugged into the same switch. i traced it out and found that it followed exactly the same route - mine was cable tied to it all the way, except for where mine terminated in the antenna, the other one ran another 10m to another device. So I figured, if there are any issue with length, or interference, or installation conditions, that cable will have it exactly the same - or even worse, because it goes that extra bit further, and it's an older, cheaper style of cable. So I ran patch lead from switch to POE, and the long lead from POE to antenna. It worked without a hitch.
So by now, my only conclusion is that despite my cable metering out ok, and being able to transmit both power and data, it for some reason can't do it at the same time even though one exactly the same can, therefore my cable must be the culprit. I cut the plugs of both ends and very carefully re-fitted - same result. So I ran in a new cable. High quality CAT6 this time, just to be sure. The result was the same - still power but no network, although once again it worked fine if I used the short cable for the power. I dragged the cable all the way back to the ethernet switchboard, so now all I've got is a 25m patch lead coiled up at my feet. No chance of interference. (Also, as the engineer suggested, no issues with the smoke not being able to run uphill). Still no network. And still, not even a green light on the switch to say "I've got something plugged into me". So (and I have no idea what posessed me to try this), I switched ends on the long cable. I.e., the end that was in the switchboard plugged into the POE adaptor, I now plugged into the antenna, and the end that had been up near the roof in the antenna, I now plugged into the POE adaptor.
It worked.
Repeatably.
As much as I tried to wiggle plugs, bend cables, and generally make life hard for the signal to get through, it was flawless. And no matter how many times I switch the ends back and forth, it always works instantly one way, and never works the other.
has anybody ever seen this before? Can anyone suggest anything?
Other information, in case it's relevant:
My connection sequence in the plugs is:
1-white/green
2-green
3-white/orange
4-blue
5-white/blue
6-orange
7-white/brown
8-brown
This matches up with all the patch leads I've ever seen, and conforms to the T568A standard.
The POE adaptor uses pins 4&5 for positive, 7&8 for negative.
I'm out of likely ideas, I'm out of unlikely ideas, at the stage I'm even out of ridiculous ideas. So any ideas, however ridiculous, are welcome
Join CF
You are here: 
