I am trying to calculate the impedance of a single pair UTP cable. In particular I want to calculate the even mode impedance so I can implement the correct common mode termination. I am having difficulty visualizing the reference plane. Intuitively it seems like it should be freespace but I am unsure.

 

UTP= Unshielded Twisted Pair, at least I am guessing that is what it means in this context. At that point the rest of the description is missing, which includes wire diameter, insulation thickness and dielectric constant, and the number of twists per inch or other unit of distance. And all of the standard formulas assume it is in free space.
Quite probably the same formula for a pair without the twisting will give a fair approximation.
What frequency range would the UTP be used with? That affects how important the impedance match will be. The actual power level will matter if the losses may be large.

 

I am trying to calculate the impedance of a single pair UTP cable. In particular I want to calculate the even mode impedance so I can implement the correct common mode termination. I am having difficulty visualizing the reference plane. Intuitively it seems like it should be freespace but I am unsure.

It should be in the neighborhood of 100-120Ω and it is easy to measure with a TDR or an Antenna Analyzer. Just terminate it and measure the return loss. Look for a minimum value.

 

UTP= Unshielded Twisted Pair, at least I am guessing that is what it means in this context. At that point the rest of the description is missing, which includes wire diameter, insulation thickness and dielectric constant, and the number of twists per inch or other unit of distance. And all of the standard formulas assume it is in free space.

I agree with you regarding differential impedance. Each wire of the pair references the other wire. Even mode impedance is less clear to me. With both wires at the same potential there is no capacitance between them so the reference plane (return path) is less clear.

Quite probably the same formula for a pair without the twisting will give a fair approximation.

I agree. If the return path of a single wire is freespace then the impedance is quite high. It would be higher than the differential impedance but that doesn't sound right to me.

What frequency range would the UTP be used with? That affects how important the impedance match will be. The actual power level will matter if the losses may be large.

The frequency is ~100MHz but power loss is not my concern. Differential signaling and termination are fairly easy to figure out. Common mode current is a radiated emission source so I believe that properly terminating it should reduce emissions.

 

It should be in the neighborhood of 100-120Ω and it is easy to measure with a TDR or an Antenna Analyzer. Just terminate it and measure the return loss. Look for a minimum value.

That is true for differential impedance. Common mode impedance is my true interest. Doing a VNA measurement of common mode impedance both wires (shorted) connect to the signal of the port. Where do I connect the ground?

 

That is true for differential impedance. Common mode impedance is my true interest. Doing a VNA measurement of common mode impedance both wires (shorted) connect to the signal of the port. Where do I connect the ground?

Do you think the results will be affected by a ground plane at some specified distance from the UTP? 100 MHz is a wavelength of approximately 3 meters. How about a ground plane 1.5 meters away. Then try closer and farther to see if there is any effect.

 

One possibly useful reference would be information on the CAT5 and Cat6 cables, which are bundles of four twisted pairs.

I agree with you regarding differential impedance. Each wire of the pair references the other wire. Even mode impedance is less clear to me. With both wires at the same potential there is no capacitance between them so the reference plane (return path) is less clear.


I agree. If the return path of a single wire is freespace then the impedance is quite high. It would be higher than the differential impedance but that doesn't sound right to me.


The frequency is ~100MHz but power loss is not my concern. Differential signaling and termination are fairly easy to figure out. Common mode current is a radiated emission source so I believe that properly terminating it should reduce emissions.

Avoiding common mode current at 100MHz is often done with ferrite beads and chokes around a cable. And as the impedance of wht would effectively equal a single conductor is fairly high, the chokes seem to do an adequate job. OR will this be a higher powered system, where 40 dB down from 1000 watts is till a fair bit of power?

 

Do you think the results will be affected by a ground plane at some specified distance from the UTP? 100 MHz is a wavelength of approximately 3 meters. How about a ground plane 1.5 meters away. Then try closer and farther to see if there is any effect.

I am fairly certain it is. For a transmission line Z0 = (L / C)^0.5. The further away the ground plane C gets smaller which will increase Z0. I am less sure if or at what point this equation becomes invalid.

 

One possibly useful reference would be information on the CAT5 and Cat6 cables, which are bundles of four twisted pairs.

The trouble is that 3 pairs could be a reference plane for the pair under test. An additional problem is that manufactures do not specify common mode impedance. I have measured a STP cable but I am less confident in correctly measuring UTP.

Avoiding common mode current at 100MHz is often done with ferrite beads and chokes around a cable. And as the impedance of wht would effectively equal a single conductor is fairly high, the chokes seem to do an adequate job. OR will this be a higher powered system, where 40 dB down from 1000 watts is till a fair bit of power?
[/QUOTE]
This is low power. I can eliminate the noise with a cm choke, preferred over ferrite core due to their size and cost. I could be completely incorrect but I think proper cm termination would reduce noise on my cable. I can test this but the problem is the value to target for cm termination.

 

I am guessing that the signal across the pair is also 100MHz, and that somehow external reactance at one end is what converts some of the signal to common mode. So the actual application is still a mystery.