CAT 5e will only transfer at 10 mb/s

I just moved into a brand new house within the last week. I paid the builder to install three CAT 5e lines throughout the house. I moved my four computers into the new home (I ran my own CAT 5e cable in the basement for the forth one).

The computer with my own CAT 5e runs fine at 100Base FD.

When I connected the other computers to the builder's outlets, the computers refuse to make a connection. The lights on the NICs flash and the router lights come on, but data won't transfer. Each computer shows data packets going out, but 0 for received.

I fiddled with the other three computers for a while and discovered this: if I go to each computer and change the NIC settings to force 10Base HD then they work fine, but any other settings refuse to work.

I have run my own CAT 5e cable through the house along the floor just to test each computer and the router. They all work fine at 100Base FD with MY cable.

Isn't CAT 5e supposed to support up to 1000 Mb/s, yet I can't even get his to run at 100 Mb/s. BTW, I did check the cables the builder used and they all say "CAT 5e" on them.

What would be causing this problem? Three bad cables? Bad ends? Incorrect wiring at the outlet? Too long of cables? (mine works fine with 75ft).

I'm going to have a talk with the builder soon about this (everything is warrantied for a year), but I don't think he knows much about computers and I need to explain to him what he needs to do.

If it was only one cable, it can be one bad crimping/punching, but across the board.

Some serious installation mistake was done.

10Mbit/s only uses 2 pairs, so it's possible the guy just screwed up the wiring on the other 2 pairs. I'd doublecheck all the wiring ends.

10Mb/sec, and !00Mb/sec are wired the same (i.e. both using only 2 pairs).

When cables that meant to work at 100Mb/sec. work at 10Mb/sec. it usually mean bad crimping or loose punching of Keystones.

Seems odd that it would be bad punching or crimping since it's on 3 different lines. Usually you only see that on a small percentage of line, unless someone really didn't know what they were doing in which case I'm surprised it works at all.

OP: What type of packet loss are you getting?

Using the Windows XP ping command from the command line:

100% Packet loss (everything times out) with full autonegotiate, autonegotiate 100/10/FD/HD, forced 100 Mb FD, forced 100 Mb HD.

0% packet loss with forced 10 MB HD and autonegotiate 10/HD/FD.

I used the -t option and let it run for several minutes.

Each computer is totally different and uses a different NIC, all of which have different settings for speed. I only ran this test on one computer since I didn't have enough time.

Well, the electrician is coming back today to fix something else at the house. I will let him know what's going on and hopefully he can figure it out, since he is the one who actually installed it.

Weird. I'm curious what the problem is if you're getting 0% loss. Keep us updated.

So, the electrician came back and I told him what was going on.

He tells me he will make me some cross-over cables tomorrow and bring a line tester. I asked him why I would need cross-over cables since they are direct lines going from the computers to the router. He said it depends on the hardware. I again told him that MY straight-through cables work fine with this setup, but he insisted they might need crossover cables.

I then asked him what kind of tester he was bringing and to describe it to me. He said it was a tester that checks for continuity and has lights on it to verify the lines are wired correctly, but it doesn't send data. I'm no expert, but it sounds like one of those cheap $30 testers that just checks if electricity is passing through the lines.

I'm starting to think this guy doesn't know what he is doing. I guess I will have to wait and see tomorrow.

You absolutely do not need crossover cables if everything is wired correctly. I've never even heard of such a thing. Crossover cables are for peer-to-peer setups or hub-to-hub (in the days before auto-mdi-x was the norm). You never need them from hub to client (unless you're plugging into an "uplink" port, but that's beyond this scope).

Assuming his crossover cable idea works, he has the thing wired completely wrong.

And those testers are just fine for most applications. It wouldn't take a genius with a $10 multimeter to do the same thing though.

Its hard to imagine given how foolproof the jacks and panels are now with the color coding that someone could possibly improperly wire them.

Test of continuation is Not good enough.

A good Network cable tester is comprise of a Module that inject RF on one side of the cable and a tester that detect and analyze it on the other side.

Quote:

---

Originally Posted by cat5e

Test of continuation is Not good enough.

A good Network cable tester is comprise of a Module that inject RF on one side of the cable and a tester that detect and analyze it on the other side.

---

I want one of those portable FLUKE network testers with certification.... But dont really have $10,000 to drop on one now :( But in my current role it would be more of a novelty item with bragging rights.

It may be possible he is using the wrong pairs for punching down the block. This can cause a lot of crosstalk and cause problems with longer runs...

Another possibility is that the actual connectors used are not cat5.
I've seen come cat3 rj45's for pbx voice connections still being sold in discount stores.
Still I suspect a wiring error at or near the patch panel.
If the wires all terminate in the same room, are they bundled with high voltage wires?

Yeah Fluke is used as a standard for certification, but there is RF injector Testers for about $200 or less.

Are you absolutely sure the cables are Cat5? Another possibility is the cables are Cat3. Obviously you're getting flawless 10Mbit (which is what Cat3 does) and no 100Mbit.

Cat3 and Cat5 cables look the same, they are just rated for different amounts of signal.

You'd have to go really out of your way to use Cat3 instead of Cat5 but I guess it could happen.