I agree that something like 9451 might be more appropriate and that higher twist rates are better. However, you're talking 4.8 twists per
foot versus 6.86 twists per
foot, so it is not that it is not that the 5301FE is not twisted, just that it is not as tightly twisted.
I'm not
clear on how the third
conductor would unbalance anything since it would be unterminated. Twisting works by making the individual conductors more likely to be equally affected by any interference, interfering sources can be directional or physically closer to one side of the cable so by varying the physical relationship of the conductors they are more likely to receive the same interference. A third
conductor that is unterminated may limit the twist rate due to simply having more conductors to twist and I could see it increasing SCIN by increasing the variations in distance from the
shield, but don't see that it would otherwise significantly affect the resulting CMRR. Am I overlooking something?
Just to clarify, the noise rejection is achieved by the use of a differential input
circuit that looks at the difference between the signals on two conductors and thus any signal common to both is rejected. To optimize the noise rejection the concept is to try to make both conductors equally affected by any interference. One major component of this is
impedance balancing where both conductors are equal
impedance to
ground, this is the actual "balanced" aspect and without that a common interference source could result in differences in the signal induced in each
conductor. A second major component is twisting, varying the physical relationships of the conductors to try to keep both equally exposed to any interference. Note that the shielding is not really relevant here, as long as any interference affects both conductors equally it will be canceled via the differential input. With such a situation the shielding has minimal impact on the resulting common mode noise rejection, this is one reason that
CAT5 and other
UTP cabling schemes work so well for analog signals, the tighter twist rates are quite effective when used in conjunction with balanced I/O and differential inputs even though the cable is unshielded.