What you are describing is what is known as a TDR - Time Domain
Reflectometer.
Basically works pretty much as you describe and are easy enough to play
around with providing you can generate a fast enough risetime.

They are used for testing both coaxial and paired cables and, yes, to
determine the distance to a cable fault. You do need to know the velocity
factor of the cable however (0.67 for soilid PE down to 0.84 or so for foam.
PVC can be anywhere and will change with frequency anyway so the result can
be pretty messy).

As you get a reflection from any impedance mismatch it can be used for
roughish cable impedance measurements as well as poor joints or connectors.
- all that is required is that the length of the disruption is of the same
order as the risetime.

A closely related device is the OTDR - Optical... This is a major test item
for optic fibre cables as it can provide a good indication of the fibre
attenuation as well as the length - or distance to a fault. We also used to
use them (in a previous employment) as a source of light for direct
(cut-back) attenuation tests as the power output was constant & the launch
conditions well defined. The main operational difference between a TDR and
an OTDR is that the former applies a step input while the latter provides a
narrow pulse. Again, cable losses, poor splices and damage show up clearly -
as can different fibre grades and stressed cable sections.

As an example - we used to use thenm to measure the attenuation of 20km
fibres with losses down to about 0.16dB/km at 1550nm, and 0.33db/km at
1300nm. That was about 10years ago so I guess things have improved somewhat
since.

Richard



> -----Original Message-----
> From: Des Bromilow [mailto:Des.Bromilow@CITEC.COM.AU]
> Sent: Monday, September 20, 1999 1:12 PM
> To: PICLIST@mitvma.mit.edu
> Subject: reflectometry.. measuring the length of a cable
>
>
> Anyone got any bright ideas on a device which can be plugged
> into the end of a single wire (say one core in a power cable)
> and the device will return a length from the end of where the
> cable ends.
>
> the use would be for finding the break in a cable, so if you
> have a 20m cable, and the device says the end of the active
> (hot) is 20m, and the length of the neutral is 15m, and the
> length of the earth is 20m, then you would know that the
> neutral has a break in it 5m from the other end of the cable.
>
> You could also use it to measure (approximate) the length of
> a spooled cable, or the kink in an underground cable.
>
> anyone know how to build such a beastie? Anyone ever built one?
> I envisage something along the lines of a pulse with a delay
> which is then translated to length, but the detection of the
> echo, and processing speed may make it harder. Would a PIC be
> able to keep up?
>
> Any ideas?
>
> Thanks ,
> Des Bromilow
>