Sign in with
Sign up | Sign in
Your question

Ethernet Line fault detection

Last response: in Networking
Share
November 14, 2004 1:59:35 PM

Archived from groups: comp.dcom.lans.ethernet (More info?)

Hi Guys and Gals,

I am faced with a requirement to be able to hardware detect line
faults (ground faults and shorts) on an ethernet link. Since this
line is behind a balanced isolation transformer (and I new to the
ethernet game) I am uncertain if this can even be done. Does anyone
have any experience with this situation? It has to be done real time
(could be coordinated to ensure no data traffic if necessary).

Thanks for any inputs in advance..

Frodo
Anonymous
November 15, 2004 12:39:31 AM

Archived from groups: comp.dcom.lans.ethernet (More info?)

Bonsoir Frodo,

> I am faced with a requirement to be able to hardware detect line
> faults (ground faults and shorts) on an ethernet link. Since this
> line is behind a balanced isolation transformer (and I new to the
> ethernet game) I am uncertain if this can even be done. Does anyone
> have any experience with this situation? It has to be done real time
> (could be coordinated to ensure no data traffic if necessary).

If an Ethernet port doesn't receive the link test pulses (those existing
between data frames) during 50 up to 150 ms, due to any failure in the link
segment, it results at least 2 events:

- the transmission is deactivated,
- the associated green LED (often called Link), if it exists, becomes off.

For me, there is no relationship with the balanced isolation transformer.
Michelot
Anonymous
November 15, 2004 10:46:08 AM

Archived from groups: comp.dcom.lans.ethernet (More info?)

Michelot wrote:

> Bonsoir Frodo,

>>I am faced with a requirement to be able to hardware detect line
>>faults (ground faults and shorts) on an ethernet link.

(snip)

> If an Ethernet port doesn't receive the link test pulses (those existing
> between data frames) during 50 up to 150 ms, due to any failure in the link
> segment, it results at least 2 events:

Shorting of one wire of a pair to ground likely wouldn't stop
link pulses, and maybe not ethernet. It might greatly increase
RFI emission, though.

I did once have a short in a coaxial ethernet system, which is
silent, unlike an open which is detected quickly.

-- glen
!