Detecting line faults in Ethernet

[Frodo]

Hi all,

This is my first post so if miss the format or something "kindly" let
me know.

I have a requirement to be able to sense line faults on biphase lines
(ethernet). The faults that I have to sense are shorts (the two lines
shorted together) and ground faults to either line. The problem, of
course, is that it is on the other side of an isolation transformer!
Does anybody have experience with this and have some suggestion? All
ideas regarding this issue would be most appreciated.

Frodo

 

[j.b. miller]

Try 'Googling' for CAT-5 cable tester ? One of those device might be
sutiable.
? Can you test while not in use or must it be '24/7' ?

That's one of the problems with Enet,etc., it gets kinda tricky checking for
'inuse' faults. If you could go with another form of communications, I've
done remote energy control systems with direct wire for the past 20 years.
Simple,cheap, effective and REAL EASY to check for grounds and shorts. Can
even run off one wire but NOT to be confused with the Dallas stuff.
Jay

 

[Frodo]

Try 'Googling' for CAT-5 cable tester ? One of those device might be
sutiable.
? Can you test while not in use or must it be '24/7' ?

That's one of the problems with Enet,etc., it gets kinda tricky checking for
'inuse' faults. If you could go with another form of communications, I've
done remote energy control systems with direct wire for the past 20 years.
Simple,cheap, effective and REAL EASY to check for grounds and shorts. Can
even run off one wire but NOT to be confused with the Dallas stuff.
Jay

Jay,

Thanks for your reply and search suggestion!

Unfortunately, it has to be 24/7 but the actual time it is tested can
be chosen and somewhat controlled. Also, Enet is not a choice. I
have to live with it. We have several other means of communication
but this one is mission critical.

Any ethernet guru's out there with a suggestion?

 

[Tim Shoppa]

I have a requirement to be able to sense line faults on biphase lines
(ethernet). The faults that I have to sense are shorts (the two lines
shorted together) and ground faults to either line.

I'm still a little unclear on what the medium for the "biphase line" is.
Coax, twisted-pair, ?

In any event, for Ethernet over Coax (thinnet) transceivers can detect
line open and line shorted. This works because for CDMA they sample
their own output while talking, and if they see a drastically wrong impedance
it is obvious. I don't know if twisted-pair transceivers
have this feature (I think some early ones did but the market may have
vetoed this feature because of cost and software drivers unable to handle
the error flags.)

The problem, of
course, is that it is on the other side of an isolation transformer!

Detecting a ground fault on a circuit that is supposed to be isolated from
ground may take some tricks. Something like a GFCI for AC power. You
won't be able to detect a fault until there are two ground faults with
AC current flowing between them. Maybe you can detect DC current between
two faults by the change in permeability as seen from another winding?

Tim.

 

[Rich Grise]

I'm still a little unclear on what the medium for the "biphase line" is.
Coax, twisted-pair, ?

In any event, for Ethernet over Coax (thinnet) transceivers can detect
line open and line shorted. This works because for CDMA they sample
their own output while talking, and if they see a drastically wrong impedance
it is obvious. I don't know if twisted-pair transceivers
have this feature (I think some early ones did but the market may have
vetoed this feature because of cost and software drivers unable to handle
the error flags.)

FWIW, I do know that they detect, and announce, in no uncertain terms,
when the cable is unplugged! But a short, or an unplugged condition
on the other side of a transformer, I do not know.

Thanks,
Rich
[remainder snipped]

 

[Ian Stirling]

Hi all,

This is my first post so if miss the format or something "kindly" let
me know.

I have a requirement to be able to sense line faults on biphase lines
(ethernet). The faults that I have to sense are shorts (the two lines
shorted together) and ground faults to either line. The problem, of
course, is that it is on the other side of an isolation transformer!
Does anybody have experience with this and have some suggestion? All
ideas regarding this issue would be most appreciated.

The right way is if possible to move it up a level, to regularly for
example ping the other end.

 

[Frodo]

The right way is if possible to move it up a level, to regularly for
example ping the other end.

Thanks for all of the help guys! It has helped direct my thinking
towards a possible solution. By the way, the system are connected by
CAT5 cable.

Thanks to all.