Excuse any blatant mistakes, I'm very clearly learning this new. Anyway from a purely transmission line perspective, it seems you can have a single ended line and a differential pair. I see the single ended (like coax line), where the return path is the shield/ground; at first looking as if this is DC, it would mean I would read 0 potential difference between the shield and the ground of say the radio transmitter. But in the RF world I would assume for the field between the shield and center piece to propagate in coax or on a PCB, the line would sort of induce current on the return path/shield meaning at each point the shield wouldn't actually be 0 V (if you pronged into where the skin effect is) between it and say the radio transmitter but in sort of opposite to the center feed / trace's voltage.
Now when I think of differential pairs, I don't actually see any difference in my mind. When I see images online of waveforms its pretty clear it is trying to get to the point that these opposing signals swing around a reference point but I don't see how that is any different.
I noticed on the an RF transceiver (cant remember the name sorry) had its transmit outputs as N and P requiring a balun for coax, so I'm clearly missing something. My only thought is say the N would instantaneously output +4V (wrt ground), whilst the P outputs -4V (wrt ground), but again I don't see how this changes thing along the transmission line.
Essentially I'd like to know the exact difference between the single end and differential pair, and if this is dependent of thinking in hf terms or simple dc cases.
Best Answer
A load is balanced if both terminals of the load have the same impedance to ground. A signal generator / transmitter / line driver is balanced likewise, that is, both terminals of the signal generator / transmitter / line driver have the same impedance to ground.
A transmission line is balanced if it is the case that if the two conductors were to be driven in parallel with a single voltage signal, the current in each conductor would be the same. Examples would include twisted pair and twin-lead.
A transmission line is unbalanced if it is the case that if the two conductors were to be driven in parallel with a single voltage signal, the currents in the two conductors would be different. An example would be a coaxial cable.
simulate this circuit – Schematic created using CircuitLab
I am using the definitions for balanced and unbalanced transmission lines used in imbalance difference modeling of mode conversion.
I believe that what you are referring to as "single end" and "differential pair", are unbalanced and balanced transmission lines.
The physical distinction between them is related to how current would be divided between the conductors if the conductors are driven in parallel.
An center-fed dipole antenna is a balanced load. A coaxial cable is an unbalanced transmission line. When connecting them together, a balun is used.
In normal operation a high frequencies, a coaxial cable has current on the outer surface of the inner conductor, and on the inner surface of the outer conductor. If a balun is not used to connect the coaxial cable to the balanced load, current will flow on the outer surface of the outer conductor. This will make the coaxial cable an unintended antenna. This will radiate power, but not necessarily in the desired direction or polarization of the dipole antenna. Similarly, the currents in arms of the dipole antenna will not be equal, and so the antenna will radiate less power.
The effects of transmission line imbalance become much more pronounced at higher frequencies.
However, twisted pair and coaxial cables have different noise properties at audio frequencies. Twisted pair is good for rejecting magnetically induced noise. Twisted pair will accept electric field induced noise but this will primarily be common mode noise. If differential signaling is used, this common mode noise can be attenuated.
Coaxial cable is susceptible to noise from electric fields, but if the outer conductor is grounded, the noise will be shunted away. Thus coaxial cable works better with single-ended signaling. However, if 2 coaxial cables are used, differential signaling is obviously possible.