Signal reflection on long cable

Thread Starter

bug13

Joined Feb 13, 2012
2,002
So another question, does the characteristic impedance of transmission cable effect the speed of the transmission signal?

PS: I am able to recreate the experiment at home, even the rise time of my function gen is ~45ns, so I must have a long long extension cable?
 

MrChips

Joined Oct 2, 2009
35,296
We have covered the first half, the problem, what causes reflection in transmission lines.

We have not discussed the second half, how to minimize reflections. To do this we need proper termination on both ends of the cable. We do this by installing resistances that match the characteristic impedance of the cable.

If we assume the output impedance of the signal driver is low compared to the impedance of the cable, we need to install the resistor in series between the output of the driver and the cable.

At the receiving end, we install the terminating resistor across the wires of the cable.

For example, if we are using lamp cord with a characteristic impedance of 150Ω and if the driver output impedance is relatively small, we put a 150Ω resistor in series with the output.

At the receiver we put 150Ω across the cable.

If a 5V pulse is sent down the cable, the signal received will be attenuated by a factor of 2 to 2.5V.

You may want to read:

Transmission Lines, Matching and Crosstalk - Kenneth L. Kaiser
 

Thread Starter

bug13

Joined Feb 13, 2012
2,002
We have covered the first half, the problem, what causes reflection in transmission lines.

We have not discussed the second half, how to minimize reflections. To do this we need proper termination on both ends of the cable. We do this by installing resistances that match the characteristic impedance of the cable.

If we assume the output impedance of the signal driver is low compared to the impedance of the cable, we need to install the resistor in series between the output of the driver and the cable.

At the receiving end, we install the terminating resistor across the wires of the cable.

For example, if we are using lamp cord with a characteristic impedance of 150Ω and if the driver output impedance is relatively small, we put a 150Ω resistor in series with the output.

At the receiver we put 150Ω across the cable.

If a 5V pulse is sent down the cable, the signal received will be attenuated by a factor of 2 to 2.5V.

You may want to read:

Transmission Lines, Matching and Crosstalk - Kenneth L. Kaiser
Thanks a lot MrChip, that's more than I asked for, I learned something once again. I don't know until now I need a serial resistor on the input too. (I only know about the I need one on the output).

edit:

do I use this formula to calculate the impedance characteristic of a cable: Z(0)=sqrt(L/C), L and C are inductance and capacitance of a cable.
 

crutschow

Joined Mar 14, 2008
38,936
...................................
We have not discussed the second half, how to minimize reflections. To do this we need proper termination on both ends of the cable. We do this by installing resistances that match the characteristic impedance of the cable.

If we assume the output impedance of the signal driver is low compared to the impedance of the cable, we need to install the resistor in series between the output of the driver and the cable.

........................
This not absolutely required by theory. Theory states that if the load end impedance is perfectly matched to the cable impedance, then there will be no reflections from the load back to the source. Thus there will be no reflections in the cable, even if the source impedance does not match the cable impedance.

But typically the source impedance is also matched to the cable impedance to compensate for any small mismatch between the cable and the load due to impedance tolerances which, in practice, will create a small reflection from the load. This is then mostly absorbed by the source matching impedance so it doesn't get re-reflected back down to the load.
 

t_n_k

Joined Mar 6, 2009
5,455
This not absolutely required by theory. Theory states that if the load end perfectly matched to the cable impedance, then there will be no reflections from the load back to the source. Thus there will be no reflections in the cable, even if the source impedance does not match the cable impedance.

But typically the source impedance is also matched to the cable impedance to compensate for any small mismatch between the cable and the load due to impedance tolerances which, in practice, will create a small reflection from the load. This is then mostly absorbed by the source matching impedance so it doesn't get re-reflected back down to the load.
In RF transmission it is unusual for the antenna load to correctly match to the feed line impedance. That is anticipated by transmitter designers who include a transmatch network before the line feed point. In that circumstance the returning wave from the antenna is also reflected at the transmatch load side interface.
 

nsaspook

Joined Aug 27, 2009
16,518
This not absolutely required by theory. Theory states that if the load end impedance is perfectly matched to the cable impedance, then there will be no reflections from the load back to the source. Thus there will be no reflections in the cable, even if the source impedance does not match the cable impedance.
Correct for perfectly matched cables and loads and but as you say, how many PCB trace 'transmission lines' are perfectly matched? The series resistor is important to provide a stable resistive load for the line driver to the cable when driving long cables or traces with signification capacitance. Many high speed digital line driver chips have built in resistors.
 
Top