Hi,
I am hassan from siemens - Kuwait. We are currently implementing an upgrading project (SCADA SYSTEM) for our customer (Ministry of electricity and water).
We are currently having an issue on getting gigabit connection on our user interface.
SCENARIO :
1) Cable length is 22 meters CAT7a.
2) both are crimped to GG45 on both sides. For testing purposes I am using CAT6 patch cables.
I connect one end to the gigabit switch (Cisco 2960) using cat 6 patch cord.
I connect the other end to my laptop. All I get is 100mbps using cat 6 patch cord.
3) Please note that the wiring scheme I used in the GG45 connector is TIA/EIA-568-B series.
4) There are no issues with the cisco switch. Because when I hook my laptop directly to the switch, I do get gigabit conneciton.
Please advice as to where and what have I missed in this scenario.
Regards,
Hassan syed
SIEMENS -CCSO
KUWAIT
EMAIL : >removed<

When you made the connections... did you untwist the pairs beyond exactly what is needed to get it on the jacks?

You can get away with a lot at 100BaseT but for 1000BaseT its critical that you don't untwist any more than is absolutely needed to make the connection. Any extra untwisted length will degrade a connection.

You are using the correct cabling scheme. Have you verified the cabling visa the Cisco to hard loopback at the end of the cable.

Have you visually confirmed the cable has not been kinked at any point.

Well I have only twisted what was needed according to the T568B cabling scheme.
Let me tell you that I did make a 10 meter CAT7a cable with the same cabling scheme and it works. I do get gigabit connection when I connect it to the switch via cat6 cable on one end and via another cat 6 cable to my laptop on the other.
Please note that I am using RJ45 connectors. Is this an issue ?

Should not be... have you tried cutting off the old connectores and doing it again. Perhaps a single wire of the eight did not seat deeply enough. Gigbit Ethernet is difficult to get terminated correctly. Any excessive untwisted cable at the connector can cause problems unlike 100BaseT that's very forgiving. RJ45 Connectors are no issue at all, They are commonly used in 1000BaseT as well. Try a loopback plug and see if the router sees that. You should be able to get a Cisco to come up on a loop. I'm not a Cisco expert, but have done that to prove a problem is in house. If you have or can borrow a ethernet cable tester it can help a lot.

If the cable was pulled too hard and snapped a strand or it kinked when unspooling it it can be the source of the failure as well. Just to be certain... you do know what I mean by kinked? I can't assume english is your mother language that's why I am asking or that it's a term that would be used outside of American English. As many overseas English speakers Learn other english dialects and I don't know if its used in those dialects.

In case you were reading my last post I didn't edit that one.

I assume the 10 meter cable that does work is also a straight through cable same as the 22 meter cable... not a crossover cable.

Well I tried with the connectors. I redo them. But still the same problem. I have a 44meters cable laid. And the link on it won't even come up. Forget about getting 100 or 1000mbps. However I have used the LAN tester and I am getting the straight signals on both sides.
I do know what does kink means. You mean if the cable was pulled harder tightly curled somewhere. But I think CAT7a could bear that. Its very thick as you know.
Yes the 10meter cable is also wired acoording to the T568B scheme.

I have a 1000BaseT network wired with Cat6E cable with no issues. In fact it worked before I upgraded it from cat5 cable which is below spec for that application. But the old network was 100BaseT which is why it used that.

In fact part of one network at the office is still Cat5 but the equipment on each end is actually 1000BaseT, and it works, registers as 1000BaseT and in reality can't move enough data fast enough in its application to have problems.

I've never used Cat7a, we don't have any at my employer so can't comment on working with it.

Make sure you didn't misread the colors and have some wrong. Some brands of cable I have used over the years are very difficult to see clearly particularly when its in a connector, others are very easy. That would lead to split pairs and problems as ethernet is balanced pairs, and needs the correct wires to the correct pins.

I think the problem may be something simple that is being missed. You didn't kink the cable before it was run... you didn't pull the cable too hard when you ran it because of a snag so it should not have a broken strand or strands. You said your tester verified the cable is good on all 4 pairs (8 strands). I assume the cable was not run next to a strong electromagnetic source such as high current power cables or large transformers, balanced pair types are not sensitive to induced noise but not completely immune either.

I have a personal dislike for the Teflon insulated Cable (any brand). It seems to have a high degree of crosstalk even though its spec'd for the job.

If you have a Digital Volt Meter, I would make a lookback and attach it to one end that loops each color pair together... and test continuity with that meter of each pair. And go back and make sure yet another time that every connection is properly wired based on the color. Make no assumptions. When I have a situation I can't find a reason for not working, I sometimes have a coworker check my work in case I missed something looking myself.

I also assume any jack fields or cross connection blocks used are rated for 1000BaseT use, and not simply intended for phone use.

It sounds like you do know what you are doing, and it appears that you are doing it correctly, that leaves either a mistake or defect to be the root cause.

The fact the 10 meter cable works, but not the identically wired longer one tells me your application isn't one that would call for a crossover cable...