Re: Ethernet Crossover Cables

Hi Scott,

Done some more digging around and testing so thought I would share what I
have found

1000Base-T
==========
IEEE 802.3ab
Approved it as a standard on June 28, 1999 by Institute of Electrical and
Electronics Engineers Standards Association
Requires four pairs of CAT-5, CAT-5e or CAT-6
Transmit and receive 250Mbps over each of the 4 pairs = 1Gbps full duplex

1000Base-TX
===========
TIA/EIA-854
Requires two pairs of CAT-6
Two pairs transmit at 500 Mbps, and two receive at 500 Mbps = 1Gbps full
duplex

Good reference here
http://www.cablinginstall.com/index/display/article-display/150060/articles/cabling-installation-maintenance/volume-10/issue-8/contents/standards/the-case-for-category-6-as-a-gigabit-ethernet-infrastructure.html

Definitely plugging a cable missing connections on pins 4,5,7,8 will not
get a 1Gbps link but a 100Mbps link.

Using a crossover only crossing 1/3 and 2/6 and disabling auto mdix a
1Gbps link between 2 switches is successful.

Using a crossover crossing 1/3, 2/6, 4/7 and 5/8 and disabling auto mdix a
1Gbps link between 2 switches is successful.

So back to the original question regarding the requirement to cross or not
to cross 4/7 and 5/8 on a 1000Base-T crossover cable....seems both work.
Some Cisco documents say cross only 2 pairs
(http://www.ciscosystems.or.at/en/US/docs/switches/lan/catalyst2950/hardware/installation/guide/hgcable.html#wp1020386)
and other that say 4 pairs
(http://www.cisco.com/en/US/docs/app_ntwk_services/data_center_app_services/css11500series/v8.20_v8.10/installation/guide/Pinouts.html#wp1009423)

There is a patent from Texas Instruments here
http://www.freshpatents.com/-dt20091203ptan20090296570.php that is to
"detect and resolve connections for cables that are either fully aligned,
fully crossed or semi-crossed" that could explain things....

Ivan

> The 1000Base-T4 became 1000Base-T. 1000Base-TX exists as well.
>
> I wouldn't call it a "commercial failure" at all, though I suppose it
> depends on who wrote the wiki page and who they worked for!
>
> There was a long pull at the beginning towards using all the pairs so
> that people didn't have to purchase Level 6 cable (prior to Cat6
> standard) or Level 7 cable... They could get by with Cat5 or Cat5e.
>
> In short runs, it likely doesn't matter all that much what you use...
> But at full length, it will make a difference! I haven't really paid
> attention all that much, but Cisco switches are listed as TX.... My
> macbook is listed as T. And yet it works fine at 1000M with a two-pair
> cable. *shrug*
>
> Not sure what to tell you as I really haven't read the specs to see who
> is or isn't compatible with the other! Or if there's a downshift in
> speed (still above 100?).
>
> Remember that wikis are written by anyone. So there's room for social
> commentary and not 100% accurate information 100% of the time!
>
> Scott Morris, CCIEx4 (R&S/ISP-Dial/Security/Service Provider) #4713,
>
> JNCIE-M #153, JNCIS-ER, CISSP, et al.
>
> JNCI-M, JNCI-ER
>
> evil_at_ine.com
>
> Internetwork Expert, Inc.
>
> http://www.InternetworkExpert.com
>
> Toll Free: 877-224-8987
>
> Outside US: 775-826-4344
>
> Knowledge is power.
>
> Power corrupts.
>
> Study hard and be Eeeeviiiil......
>
> Ivan Walker wrote:
>
> Hi Scott,
>
> Thanks for the quick response. Again looking a wikipedia
> ( http://en.wikipedia.org/wiki/Gigabit_Ethernet ) I can find no
> reference to
> a 1000Base-T4 standard.
>
> I am specifically interested in why a *gigabit* link between two
> switches
> using a crossover with only 2 pairs crossed actually works.
>
> The wikipedia link above states "1000BASE-T requires all four pairs to
> be
> present." and also mentions a standard 1000BasseTX which only uses 2
> pairs
> but "has been a commercial failure".
>
> Perhaps the Cisco switches are using 1000Base-TX???
>
> switch#show interfaces capabilities
> GigabitEthernet0/1
> Model: WS-C3560G-24PS
> Type: 10/100/1000BaseTX
> Speed: 10,100,1000,auto
> Duplex: half,full,auto
> Trunk encap. type: 802.1Q,ISL
> Trunk mode: on,off,desirable,nonegotiate
> Channel: yes
> Broadcast suppression: percentage(0-100)
> Flowcontrol: rx-(off,on,desired),tx-(none)
> Fast Start: yes
> QoS scheduling: rx-(not configurable on per port basis),
> tx-(4q3t) (3t: Two configurable values and one
> fixed.)
> CoS rewrite: yes
> ToS rewrite: yes
> UDLD: yes
> Inline power: yes
> SPAN: source/destination
> PortSecure: yes
> Dot1x: yes
>
> Thanks
>
> Ivan
>
> Because there are two different standards. T4 uses all four pair,
> it's
> designed for use with lower-quality cables. TX only uses two pair
> like
> all other ethernet variants, but is supposed to be higher level cable.
>
> HTH,
>
> Scott Morris, CCIEx4 (R&S/ISP-Dial/Security/Service Provider) #4713,
>
> JNCIE-M #153, JNCIS-ER, CISSP, et al.
>
> JNCI-M, JNCI-ER
> evil_at_ine.com
> Internetwork Expert, Inc.
> http://www.InternetworkExpert.com
> Toll Free: 877-224-8987
>
> Outside US: 775-826-4344
>
> Knowledge is power.
>
> Power corrupts.
>
> Study hard and be Eeeeviiiil......
>
> Ivan Walker wrote:
>
> Wikipedia has a detailed explanation of crossover cables here
> http://en.wikipedia.org/wiki/Ethernet_crossover_cable#Crossover_cable_pinouts
> covering the pins outs for 1000Base-T and 100Base-TX.
> 1000Base-T
> crossover cables cross all four pairs where as 100Base-TX crossover
> cables
> cross only 2 pairs (can cross four but only 2 pairs are used
> anyway).
>
> Looking at some crossover cables I found I did indeed find some with
> all
> pairs crossed and some with only 2 pairs crossed. When testing
> these in
> some Cisco switches both worked fine at 1Gps. This was kind of
> unexpected
> as I anticipated that the crossover with only 2 pairs crossed would
> not
> work.
>
> I tried disabling mdix and speed/duplex negotiation etc but could
> not
> break
> it. Can anyone explain why a crossover cable with only 2 pairs
> crossed
> still works for 1000Base-T.
>
> Cheers
>
> Ivan

Blogs and organic groups at http://www.ccie.net
Received on Fri Feb 05 2010 - 17:00:55 ART