I think, to be more exact, that only 10 Mbit hdx uses just the two pair
of the 4 that are present and I believe it (10 Mb) works if pairs 1 and 4
are
physically not present, but your post shows that it is necessary
to have validly connected 4 pairs. I think the testing spec is TSB-67
and they of course check all pairs in the cable.
It goes without saying that the cabling should be as up to date and
high speed capable as possible.
There is also a requirement of a minimum length for patch cables,
which I understand is one half meter and is related to the NEXT
(near end cross talk) margins being violated due to presence of the
nearby modular connectors which provide impedance bumps
and excessive coupling between adjacent pairs. I made a few 10-inch
patches when I started with all this, but they've been long gone by now.
Thanx for adding the extra info on higher speed ethernet which I wasn't
even thinking of when I posted. Bob
> Robert Foxworth wrote:
> >
> > As you see, only pairs 2 and 3 conduct signal.
> > Pairs 1 and 4 are electrically idle, but the spec
> > says they are all connected. When making cables you
> > need to connect them anyway to maintain proper
> > wire position.
>
> 100baseTX uses one pair of wires for transmission (125Mbaud 4B/5B
> encoding), and a second pair for collision detection and receive. Due to
> the 20% overhead of 4B/5B encoding, 125Mbaud becomes 100Mbit.
>
> 100baseT4 uses 3 pairs for transmission (33Mhz per pair) and the fourth
> pair for collision detection. The lower signalling rate makes it
> possible to use Cat3 cabling.
>
> 1000baseT uses all 4 pairs for transmitting and receiving (125Mbaud
> PAM-5 encoding). 4x250Mbit = 1Gbit.
>
> With 10baseT and 100baseTX, both pairs are electrically active - the
> balanced twisted pairs allow for detection and correction of errors
> induced in both lines.
>
> - Ian
>
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 19:02:08 EDT