9
\$\begingroup\$

I was reading about Serial communication in computers and embedded systems and found the following statement -

Computers connected by wires must have a common ground reference, typically provided by a ground wire in the cable. (Source)

What is the reason for this? I thought that the ground wire was only for protection against electrical leakage. What would be the need for a common ground wire in communication networks?

\$\endgroup\$

migrated from physics.stackexchange.com Aug 19 '12 at 18:28

This question came from our site for active researchers, academics and students of physics.

  • 6
    \$\begingroup\$ The short--short version is that you have to know what you should be measuring your signal against, of course. \$\endgroup\$ – dmckee Aug 19 '12 at 16:43
19
\$\begingroup\$

Voltage is relative. One voltage level only makes sense if you have another voltage level to compare it to. If you would look at the signal wire alone you wouldn't get any wiser, you wouldn't measure a voltage. You need a reference to measure against, so that you can say that the signal is so-and-so-many volts, referenced to ground.

That referenced to ground is implied when we talk about voltage levels, unless otherwise indicated. Balanced signals are often symmetrical about ground, so that when on 1 wire the voltage goes to +1 V on the other wire it will go to -1 V. Nicely symmetrical. But the receiver won't look at one wire and see the +1 V; it will look at the voltage difference between the two wires, and then ground is not relevant. Such signals will most often be transmitted over twisted pair cables, where one wire is the positive signal and the other one the negative signal. There's no ground wire.

edit
Kortuk wants to see a pretty picture. Again. (sigh)

enter image description here

Et voilà, une jolie image. Don't look at En-Ya, this is about my wall clock. How high does it hang? Most people will say "about 2 meters". That's because we usually will take the floor as reference level, our "ground". But my apartment is on the second floor, and if I take street level as reference the clock will be at 8 m above street level. A geographer may suggest 35 m above sea level. So "height" has no meaning without a reference, and with voltage it's the same thing.

\$\endgroup\$
  • \$\begingroup\$ birds on electrical wire =) \$\endgroup\$ – Onur Aug 19 '12 at 18:51
  • \$\begingroup\$ I notice your answer has no pretty picture. \$\endgroup\$ – Kortuk Aug 19 '12 at 20:28
  • \$\begingroup\$ @Kortuk - yep, I thought my nice picture was worth at least 5 votes... ;-) \$\endgroup\$ – Oli Glaser Aug 19 '12 at 21:18
  • \$\begingroup\$ @Oli - Yours may be animated, but mine has a very nice girl! :-) \$\endgroup\$ – stevenvh Aug 20 '12 at 7:16
  • \$\begingroup\$ @Kortuk - Has too! :-) \$\endgroup\$ – stevenvh Aug 20 '12 at 7:34
10
\$\begingroup\$

dmckee is correct (should add it as an answer really) it's because any electrical signal needs a return path, which the ground connection provides. Another name for Voltage is "Potential Difference", which means the difference in potential between two points. So you cannot just say a point is "Five volts", you must say it is "Five volts with respect to x" (in practice we do say the former, but it is assumed we know where the relative point is - e.g. a ground node)

Think of a light bulb - there are two terminals. A battery - two terminals. Could we light the bulb by just connecting one of it's terminals to one of the batteries terminals? No, we must complete the circuit.
Another way to look at it is a pump (battery) and pipes (wires) plus an (enclosed) water wheel (load) For the pump to turn the wheel there needs to be an outgoing pipe and a return pipe.

Here's a diagram to clarify:

water analogy

Note that the above explains the need for 2 wires - there does not necessarily need to be a "common ground" (i.e. in systems isolated with transformer/optically) as Wouter correctly explains. The transmitter and receiver still have a complete circuit though (e.g. the primary of the transformer which is magnetically coupled to the secondary - this is why the secondary can have a different "ground" reference) so things "look" the same to them.

\$\endgroup\$
  • 5
    \$\begingroup\$ Love that diagram! \$\endgroup\$ – dext0rb Aug 19 '12 at 18:44
  • \$\begingroup\$ @dext0rb - thanks, I thought it was pretty cool - I found it here where there are a couple more complex versions. Also Wiki and Hyperphysics both have good pages on the water analogy (and it's limitations) \$\endgroup\$ – Oli Glaser Aug 19 '12 at 18:49
  • 2
    \$\begingroup\$ I think it should rotate counterclockwise in the Northern hemisphere :-) \$\endgroup\$ – stevenvh Aug 19 '12 at 18:52
7
\$\begingroup\$

In general this is correct - the voltage must have a reference point. Think of voltage as distance: the concept of distance makes no sense unless you are agree from where.

But the claim is not strictly true. A wire does not need to carry electricity, it could be a glass fibre wire...

Another counter-example is balanced tranformer-coupled communication, which is used by most modern Ethernet connections, and some high-reliability systems like MIL-1553. For such communication the two communication systems need not have a common ground.

\$\endgroup\$
0
\$\begingroup\$

Communication network transmit data in terms of voltage difference. To measure any voltage there should be some reference point. Both computers or both equipments having serial connections bet them should understand the same meaning of the data ( read - same reference point for the voltage ) so providing same common ground is essential. This also helps to prevent noise corrupting the data exchange.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.