Does anyone know off-hand how long you can run Cat.5e cable and still maintain Gigiabit speeds? I am trying to decide to wire my house with 5e or 6.

Here is a summary of the various maximum segment length defined for Ethernet. length are valid for networks with only one repeater or repeating hub. Further analysis is required for networks with more than one repeater.

Standard
Maximum Cable Segment Length in Meters

10Base5
500

10Base2
185

10Base-T
100

10Base-FL
2000

100Base-TX
100

100Base-FX
400

100Base-T4
100

100Base-T2
100

1000Base-LX
550 if 62.5 µ or 50 µ multimode fiber3000 if 10 µ singlemode fiber

1000Base-SX
250 if 62.5 µ multimode fiber550 if 50 µ multimode fiber

1000Base-CX
25

Maybe I'm missing something, but I don't think this answers my question. I'm talking about Gigabit over copper.

Yeah. None of those gigabit standards are for UTP. There's 1000baseT and 1000baseTX, I believe, with 1000baseT being much more common. (I'm not sure TX has even been ratified yet.) The max length for a run is 100m using Cat5 or greater, but it requires all four pairs, so make sure your installer doesn't cut corners by running only two pairs.

I'm getting the strong impression that it is limited to 100m irrelevant as to whether you use 5e or 6.

http://compnetworking.about.com/library/glossary/bldef-cat6.htm
http://www.foundrynet.com/services/faqs/GigCopper.html

damn, slow typing

There are more stringent limits on bend radius and how much you're allowed to untwist at the ends when doing gigabit over copper.

The max length for a run is 100m using Cat5 or greater...

Hmmm. I tried using plain-old Cat5 once and it simply wouldn't link. This was only about 3ft. cable length too.

Then again, all my gigabit stuff is bottom of the [price] barrel.

Are you sure all four pairs were connected and working properly? Only two pairs are used for 100baseTX and 10baseT.

It also matters if the pinout was correct. Remember, Pins 1-2, 3-6, 4-5, 7-8 should be twisted together. Pin 1 is on the left when looking at an end with the retention clip away from you.

Eh?

1-2 3-6 4-5 7-8, ne c'est pas?

OK... I give up. What's that in english?

It means, is it not so?, ne c'est pas?

Ah, well I hope so, but I've seen too many people wiring improperly. And 10 or 100 BT is more forgiving.

It means, is it not so?, ne c'est pas?

Actually, it's n'est-ce pas, not "ne c'est pas".

I can't be certain that the cable was still good, but when I made it it was.

My point is, how do you know? If it worked fine for regular or fast ethernet, that doesn't prove anything about half of the cable, and if it once worked with gigabit, you wouldn't be asking the question. Unless you tested with a tester that checked all four pairs, there's no way to know if it was a good cable or not.

Regardless, the spec (802.3ab) says Cat5. Period. Not even Cat5e. As Trevor pointed out, the ends of the cable may have been untwisted too much for gigabit. Or if it was a homemade cable, the data pairs may not have been the twisted pairs.

That said, it's possible that Cat6 cabling might give you more leeway than Cat5. That is, it may be that you can make zero mistakes with Cat5, but you can make a few with Cat6. I don't know. That's just conjecture.

Unless you tested with a tester that checked all four pairs, there's no way to know if it was a good cable or not.

That's what I meant when I said it was good when I made it.

I think I should probably just go with Cat.6 to be safe. Unfortunately, I haven't seen something like this with Cat.6 in it.

The cheap cable testers aren't to be trusted too much. They only check for continuity through each individual wire. I know somebody that spent days fiddling with a cable run because it would refuse to run at 100baseT. It turns out the person that installed it had gotten some of the wires mixed up and they weren't paired properly. The cheap cable checker said it was fine and was wired up correctly however.