|
|
|
Hi
Unusual case here, need to place outdoor 4G aerial 150m from router.
Criteria:
Router in basement of building
Aerial to go on roof (150m up)
We think no power on roof
Is this possible? � POE?
Any ideas appreciated.
Thanks
|
|
|
|
Definitely don't run an RF cable 150m from the router to the aeriel. It'll absolutely murder the signal.
I'd have a router on the roof with a short cable run to the aerial, then have ethernet down into the building. You're going to need some sort of switch and PoE injector halfway, as ethernet can only really can do 100m reliably (especially with PoE).
Any reason the router needs to be in the basement?
|
|
|
|
Thats not going to work because you only have 100m for an ethernet run. You are going to have to use a fibre link from somethink like Mikrotik SXT with a media converter powered locally down to a router with either a SFP or another media conveter. Either that or work out how to get the ethernet run down to 100m, then you could power the SXT using PoE from say an EdgeRouter X SFP.
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
|
150m? Do you live in the Shard?
Anyway, as others have said, 150m is a long way for ethernet.
There's two parts to the problem, power and signal.
Power is easier, the Microtik SXT LTE needs 11-57v, so some power loss on a long run will probably be OK (it only need 5w).
But you will need to break the distance with a switch - so a PoE pass through switch would probably be ideal so that you can power both the switch and the router with the single ethernet cable. You will need a suitable place to locate the router though.
|
|
|
Ethernet over coax maybe?
also maybe give this a read
https://www.cablinginstall.com/articles/2015/05/blac...
Edited by epyon (Mon 24-Dec-18 22:14:08)
|
|
|
|
A hi-gain 4G aerial would boost the signal by around 20dB, but sadly even a good 4G coaxial cable (e.g. LMR400) would lose around 2dB for every 10 metres that means you would lose around 30dB for a 150 metre run. The net loss would be 10dB which is not good.
|
|
|
150m? Do you live in the Shard?
Anyway, as others have said, 150m is a long way for ethernet.
For copper ethernet it is too far. You are strictly limited to 100m. Well strictly speaking you could use the old 10Base2 which is good for a 180m run or the original 10Base5 which is good for 500m, but then you would be limited to 10Mbps and well finding the kit would be hard.
There's two parts to the problem, power and signal.
Power is easier, the Microtik SXT LTE needs 11-57v, so some power loss on a long run will probably be OK (it only need 5w).
But you will need to break the distance with a switch - so a PoE pass through switch would probably be ideal so that you can power both the switch and the router with the single ethernet cable. You will need a suitable place to locate the router though.
You could use a hybrid fibre optic cable. That is one that has copper power and fibre all integrated into the single cable.
https://www.commscope.com/catalog/networking_systems...
You would need however to pay someone to fusion splice some connectors on each end of the cable.
The other option would be to use PoF, which uses a high power laser to litterly send power down the fibre optic cable, though that is going to be very expensive.
https://en.wikipedia.org/wiki/Power-over-fiber
At a pinch I guess you could run some pre-terminated fibre though I have never seen pre-terminated fibre at 150m lengths, and a separate cable for power.
|
|
|
|
A place I used to work had the fibre cut, and we had a 175M ethernet cable as a temporary measure. Apparently the cable was installed that way years prior... Surprisingly it connected at 1Gbps and bridged the two sites without any complaints for around 2 months... I was quite shocked it worked, but I guess 100m is for the cable to perform within spec, and we probably didn't have much interference or whatever else.
Just to be clear, I am not recommending the OP uses a 150M cable! BAD IDEA!
|
|
|
The reason that worked, is majority of the loss is at the junctions/connectors. The distance is expecting you to have:
switch -> patch panel = 1 patch cable
patch panel -> floor box = 1 structured cable installation
floor box -> computer = 1 patch panel
Each of those 3 RJ45 connectors loses signal and allows for noise to enter.
If you have switch -> switch with a very long cable, then you will almost always exceed the specification in length.
But as you wrote in bold - don't plan for it to work
plusnet 80/20 (2/jun/14) at 470m - sync 19/Sep/18: 61,689 / 8,831 - G.INP & 3.0 dB SNRm
19 years of broadband, from 1999's ntl:cable modem trial - Live BQM
|
|
|
The reason that worked, is majority of the loss is at the junctions/connectors. The distance is expecting you to have:
switch -> patch panel = 1 patch cable
patch panel -> floor box = 1 structured cable installation
floor box -> computer = 1 patch panel
Each of those 3 RJ45 connectors loses signal and allows for noise to enter.
If you have switch -> switch with a very long cable, then you will almost always exceed the specification in length.
But as you wrote in bold - don't plan for it to work 
True but as I recall also the issue is also to do with propagation delays and how the waveform gets distorted over distance. Which reducing the connectors involved won't really help.
I guess you could try with a top quality screened Cat5e with plugs directly on both ends. Ironically Cat6 and Cat6a would be worse as more twists mean a longer length of copper for a given run distance. However again do not plan on it working.
However the quoted 150m is going to be something like a 30 storey building. Which means a plemium rated cable too. I would go with hybrid fibre unless the distance can be shortened.
|