Ethernet woes over 50m cable

I’ve spent the last two days in ethernet land and scratching my head at this, maybe you guys can help. we are running a cable perhaps 40-50 meters long from our satellite modem to our office.
Now we use 2 different brand cables along this length for troubleshooting and attempting to eliminate the cable as the cause. here are some observations-

Our wrt54g router wont connect to the modem over these 50m ethernet cables, but will if it is close and plugged in with a 3m cable. cable testers say all cables work.

3 mac’s with osx tested, wont connect over those 50m cables to the satelite modem, but will connect to each other using the same cable without a problem. they work fine if plugged in closely.

A dlink dir-825 router will connect to the modem with one of the long cables (haven’t tried the other), however speed tests only give 1Mbit, while down there at the modem we get 4Mbit with speed tests.

Testing for packet loss we get no issues.

Also, if we plug the WRT54g down at the modem and run the long cable up from the router, we can’t connect, but if we plug in close to the router then it works fine. So this should rule out the satelite modem as not being in spec and providing enough power over the ethernet line, because the dlink router should be doing this in this situation.

I dont see why a cable that long should give issues when the limit is supposed to be 100m. are all these devices not following the spec on their ethernet ports?

Tommorrow I will try plugging in the dlink router closer and see if we observe speed loss and also how it behaves with a computer plugged in to the other end of the long cable.

Are you testing this cable installed into the building or on your desk?
What type of cable is it?
Are the ends crimped into a plug or socket?
Who did the crimps?

Has the cable been tested with a Network tester?

I’m thinking you might have some noise on the line when installed.
The receiver might not handle this noise and thus it’s not working.

My Recommendation:
Get shielded Cat5e or better cable and ensure it’s not ran parallel with 240 cable if possible, or separate it by placing on the other side of the beam if it’s not.

Is there anything that could be causing signal interference on the line? Are you running it over a fluorescent light, next to power cables, or kinked tightly?

If I read the post correctly… you have joined two different shorter cables to make one longer cable?

IIRC joins in shielded cables can cause a lot of noise issues, so are to be avoided if possible.

The spec is for cables to be run with 300mm segregation from mains/LV. You’d have to have some pretty serious interference to stop the devices from linking.

As has been previously mentioned, if you’re joining cables, that’s not a great idea. You can get couplers to properly do this. Otherwise terminate each of the cables and put a switch in between.

We are testing the cable on the desk, its not gone inside a wall. we are not joinging cables. I just have two for elimination of a broken cable as a potential cause.

It’s cat 5e- not shielded, however we are on top of a mountain and there is nothing but grass and trees between the two points. There’s no 240 cables nearby, I don’t know what opportunities there’d be for interference.

The ends have been crimped by me and one other person, we used plugs on both ends.

The cables have been tested with a cable tester and they were fine.

Can you post some decent (in focus) pics of the connections? Something wrong with the termination seems the most likely cause given what you’re saying. Joins are fine of course (there are 4 joins in every typical cat5 structured cable run - each end and wall plate/patch panel - but they need to be properly terminated with the twists maintained as close as possible etc.
That said, I’ve seen (and done) some really dodgy things without any problems - ethernet over cat5 is pretty damn solid. The worst things are smart switches which down ports when they get confused as you swap things around during testing.

Also, when you say “cable tester” do you mean something that cost less than fifty dollars with 8 LEDs on it, or something that can tell you how long the cable is by timing electrons doing laps?

How exactly is the cable installed? by the sounds of it its run outside somehow… if it isnt gell filled water ingress into conduit or whatever and then into the cable may be the issue…
its worth noting that the sheath on some communication cables can be hygroscopic…

It’s too small for me to photograph it properly with any camera I have.
Our cable tester is just the cheap kind with 8 leds.

These cables have been freshly cut and tested on the same day so water is not the issue.

I went overboard and found a cat7 cable on amazon for $60 thats the same length so we shall see what happens with that. id still love to know how these other two cables aren’t working though.


Understand that it is too late for an advice, but most likely NIC’s chip in the modem is a crap and could not sustain long distance, what you can do is connect modem to an other switch (4 port cheap one) and then that switch connect to a long cable… Of course you have to test cheap switch as well if it works on a long distance…

Hope it helps.

1 Like