15
\$\begingroup\$

What are the mathematics behind twisted pair cabling? I know that the twists are there to help eliminate cross talk and interference, but I want to know about the math of how they find the specific twist rates.

Is it just experimentation ("Oh looks, it works better now"), or is there actual math behind it that says based on the size of the cable you should twist it this many times per inch/foot to get the desired effect?

\$\endgroup\$
  • \$\begingroup\$ +1 Good question. If it were me I would do the "Oh look, it works better now" method. \$\endgroup\$ – Kellenjb Feb 2 '11 at 21:21
12
\$\begingroup\$

Its actually a fairly complicated topic, I don't know of any "rule of thumb" calculation but there is an article from EDN here on transmission line parameters of twisted pair cables including the effects of pitch (twist rate). The twist rate will change the effective dielectric constant between the 2 conductors which does impact impedance and dielectric losses vs frequency.

There are also other things going on. For instance not many people know that the 4 twisted pairs in an ethernet cable all have slightly different twist rates. This prevents the same conductors from constantly and periodically being next to each other in the cable which would cause increased non-common-mode interference.

\$\endgroup\$
  • \$\begingroup\$ I thought it boiled down to, as you increase twist rate you are effectively reducing distance between the wires and decreases cross-sectional area. \$\endgroup\$ – Kortuk Feb 3 '11 at 2:32
  • 4
    \$\begingroup\$ and the more twists the longer the wire resulting in more attenuation \$\endgroup\$ – Matt Williamson Feb 3 '11 at 5:43
  • \$\begingroup\$ It's not just ethernet; the 50-pair telephone trunk cables did this exact same thing (slightly different twists on the pairs) well before ethernet was even a concept. \$\endgroup\$ – akohlsmith Feb 3 '11 at 15:37
  • \$\begingroup\$ @kortuk I think that is certainly a consideration also, reduced cross-sectional area = smaller loop area = less inductive coupling. But of course pulling the conductors closer together changes the characteristic impedance also which has to be adjusted for. Ultimately i think there is a lot that goes into the final specification but its not something we'd often hear about as for the most part, UTP/STP is bought to a specification (Cat1 through Cat7) and the twist rate is part of the specification so all we ever see is the final performance of the twist rate chosen. \$\endgroup\$ – Mark Feb 3 '11 at 19:09
-1
\$\begingroup\$

the twist step depends on the wave length of the signal transmitted

that is why cables for different frequencies have different twist steps

\$\endgroup\$
  • 1
    \$\begingroup\$ Your correlation vs causation is off. For instance gigabit Ethernet operates at the same frequency that 100mbit Ethernet does but the required specifications for gigabit operation are much tighter in several regards including cross-talk. One of the ways used to meet the specification is to use a higher twist rate. As a byproduct this does improve performance at higher frequencies but thats not why it was done as higher signaling frequencies were not needed to operate gigabit Ethernet. \$\endgroup\$ – Mark Feb 3 '11 at 19:45
  • 1
    \$\begingroup\$ -1. Different twist rates within the same cable are used to reduce crosstalk. \$\endgroup\$ – Johan.A Sep 5 '13 at 11:49

Your Answer

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