# Coaxial-cable and computer networks

#### Niles

Joined Nov 23, 2008
56
Hi all.

This is not a textbook problem, but something I thought about today when riding home from school on my bike.

A few years ago, our computers at home were connected by coaxial-cables just like in this image: http://www.muellerelectric.com/images/BU5050B.jpg

On the last computer in the network, one side of the T-BNC (as seen here: http://media.digikey.com/photos/Amphenol%20Photos/31-8,%2031-208.jpg) had an end attached to it (a picture of the end piece is here: http://www.morecomputers.com/openimages/Belkin/OR1230000054155.jpg).

Now my question is: As far as I have understood, the end piece terminates the coax.-cable, so no signal is reflected at the end. But in order for this to happen, the resistance of the end piece must be exactly the same as the total impedance of the coax.-cable (please correct me when I am wrong).

But doesn't the impedance of the cable depend on its length? So how can the end piece, which has a fixed resistance, be used with a cable, which has an impedance dependant on its length?

#### plutarch

Joined Dec 19, 2008
1
This can only be understood with the "transmission line" theory. See the wiki page of transmission line for some pointers.

#### Niles

Joined Nov 23, 2008
56
Hmm, is there no short answers? E.g. "the resistance of the end piece can vary" or something else ?

Thanks for replying so fast.

Last edited:

#### KL7AJ

Joined Nov 4, 2008
2,208
Hi all.

This is not a textbook problem, but something I thought about today when riding home from school on my bike.

A few years ago, our computers at home were connected by coaxial-cables just like in this image: http://www.muellerelectric.com/images/BU5050B.jpg

On the last computer in the network, one side of the T-BNC (as seen here: http://media.digikey.com/photos/Amphenol Photos/31-8, 31-208.jpg) had an end attached to it (a picture of the end piece is here: http://www.morecomputers.com/openimages/Belkin/OR1230000054155.jpg).

Now my question is: As far as I have understood, the end piece terminates the coax.-cable, so no signal is reflected at the end. But in order for this to happen, the resistance of the end piece must be exactly the same as the total impedance of the coax.-cable (please correct me when I am wrong).

But doesn't the impedance of the cable depend on its length? So how can the end piece, which has a fixed resistance, be used with a cable, which has an impedance dependant on its length?

The termination has to match the CHARACTERISTIC IMPEDANCE of the transmission line, not the DC resistance (which for all practical purposes is almost nothing). Generally 75 ohms cables were used for ethernet....though I've seen some installed with 50 ohm cables.

It doesn't have to be all that good a match, either....within a couple of percent never caused any grief.

eric

#### Niles

Joined Nov 23, 2008
56
Ahh great, characteristic impedance is of course just constant for each cable. So I was kinda right.

It's funny that my generation has reached the age, where I/we are able to say : "When I was a kid, our computers were connected by coaxial cables"... perhaps some still are ?

Thanks.

#### KL7AJ

Joined Nov 4, 2008
2,208
Ahh great, characteristic impedance is of course just constant for each cable. So I was kinda right.

It's funny that my generation has reached the age, where I/we are able to say : "When I was a kid, our computers were connected by coaxial cables"... perhaps some still are ?

Thanks.

You're a mere whippersnapper. I've worked on ANALOG computers!

Eric

#### subtech

Joined Nov 21, 2006
123
Yes, some still are.
In some applications where ethernet cable (5,5e, or 6) may not be available, or the length of run very long, co-axial cable may be applied.
I am sure however that fiber optic cable and related equipment are fast replacing such installations.