Andreas said:
Hi,
does anyone know, how to interpret the results of the Virtual Lan Cable Test
in the Bios ???
Must the end of the cable be connected to the switch or not ???
Andreas
The VCT function is apparently based on TDR or Time Domain
Reflectometry. A pulse is sent down the line, and depending
on how the line is terminated or not, determines whether a
right-side-up, an upside-down, or no pulse gets returned.
The amplitude and shape of the returned pulse tells you
things about the condition of the line. No returned pulse
means the line is perfectly terminated, as it would be if it
was connected to a switch/router.
An Ethernet cable has eight wires in it, arranged as four pairs.
For 100baseT, only two pairs are used (wires 1,2,3,6). The other
four wires can be in any state, and still be valid (some boxes
might ground them, others use various forms of termination).
If the Marvell is plugged into a GbE switch, all four pairs would
be used. If the BIOS function says any of them are "open", what
that means is a pulse with a significant amplitude is bouncing
off an open in the line. If you leave the cable dangling off the
back of your P5GDC, all four pairs will read open, and the software
will tell you the length of the Ethernet cable, to within some
resolution. If you plug the cable into a GbE switch, the software
will tell you they are terminated and no length measurement will
be possible, or necessary for that matter. This does not guarantee
the hardware works - only that the VCT has sensed an appropriate
load on the end of the line, one similar to another Ethernet
interface.
So, the purpose of the tool, is to show you whether the cable is
connected to a load at the end of the line. If the cable itself
is defective (broken wire, or short between members of a pair),
the BIOS will tell you the distance to the defect in the cable.
This is only important for in-building wiring, and the info can
be used for maintenance purposes.
Note that, if you have a lot of connectors, lengths of wire made
by different vendors etc., the function could easily fail and be
useless. The best situation would be a SOHO application, with
two pieces of Ethernet equipment and one cable, as the number of
impedance discontinuities is limited and the software is more
likely to classify the results correctly.
TDR is also available as an (expensive) lab instrument, and the
best classifier of pulse shape, is a trained human eye. The software
in the BIOS will only know a fraction of all there is to know about
TDR and how to interpret the results.
HTH,
Paul
