Pete Lockhart/Anixter, Inc.
What was Cat 5 designed for?
Cat 5 was developed to run 100Mbps CDDI on pairs 1-2, 7-8 with the other 2 pairs terminated to GROUND as a common mode termination to reduce intersymbol interference and NEXT caused by unbalances in the cable plant and the environment.
This ANSI Standard is called TPPMD for Twisted Pair Physical Media Dependence. It had sufficient design 'headroom' to run the MLT3 encoded copper version of FDDI.
What happens when you run 100 Base T on cat 5?
When you run 100 Base-T you violate all of these original design parameters. You run the data on pairs 2 and 3, the absolute worst pairs from a NEXT and Return Loss standpoint because they are the SPLIT and Crossed over pairs. You also do not terminate the other 2 pairs into ground. All the design headroom is shot!
There is one other big "GOTCHA". The original Cat 5 only intended for one pair to be transmitting at a time, no full duplex was ever considered. The Ethernet guys did add full duplex when considering the spec for 100 Base-T.
Cat 5E was supposed to correct all of that, but it doesn't. They still only require testing to 100MHz. There is no component test procedure to qualify Cat 5E Patch Cords to. Therefore, they CANNOT really exist as a product! There is a test procedure for a Cat 5 cord which includes NEXT and RL using a 'Standard' Cat 5 Jack, but no similar test jack is defined or even contemplated for Cat 5E. You are supposed to plug the Cat 5E patch cord into the installed Cat 5E link and test it in place. If it works then that is the only cord test approved for that node, if you change the cords out you have to completely re-test. The Cat 5E spec is a patch on a Cat 5 spec that is being used for the wrong application.
Why test beyond 100 MHz?
The original CAT 5 design only went to 100MHz because that was the best Baluns we had to test with in 1991.
Now we can test reliably with 350MHZ Baluns and therefore, should be testing any cable to at least 125% of its rated speed (per IEEE request to both IEC and TIA in 1998!) That would mean that the MLT3 signal, which is 4b/5b encoded FDDI signal, really runs at 125 MHz and it should be tested to 155MHz!
In Level 6 that much better than Cat5e?
Anixter's Level 6 is the Specification for the REAL Cat 5 'Enhanced' product that the Gig-Alliance talks about. Take Belden as an example, they upgrade their existing 1585 Cat 5 cable to be Cat 5E by tweaking the pairs to pass the RL limits. Their TRUE enhanced product has always been DataTwist 350, which blows away the 1585 performance by 10dB in NEXT and 4 dB in RL - no contest!
Anixter's Level 7 is what CAT 6 should be. The NEMA group published WC66 2 years ago and is the ONLY current Standard for a CAT 6 Cable. The hang up now at both TIA and ISO is this issue of interoperability, which says you should get CAT 6 connector performance no matter what piece of junk patch cord you buy. We say, buy the patch cord from the guy that makes the jacks and there is no issue! If you do that you can blow away the proposed CAT 6 connector spec. The latest gems for the TIA is that in order to get the CAT 6 interoperability they want, they will probably have to REDUCE the design performance by as much as 3 dB.
Will gigabit Ethernet run on Cat5e?
Gig Ethernet will run on Cat 5E if you know who's cable and connectors to buy and you get very good patch cords. Anixter does not support or even claim to sell a Cat 5E only Channel. We will only sell higher end patch cords with connectors that exceed the spec by a bunch. We really see only Level 6 as being the answer.
I heard cases where the cabling runs Gigabit, but not 100 Base T?
The issue of the electronics has to do with the 2 circuits that are in the Gig-E Chips. (1) Noise canceling and (2) Echo canceling. This is where the 4 watts of energy requirement comes from. The 100 Base-T chips have an Adaptive Equalizer that tries to chase ATTN bumps of less then 3 dB over frequencies of 16 - 40 MHz. We can show you in the lab stuff that runs Gig but won't run 100 Base-T all day long!
|