1
\$\begingroup\$

Why do differential signal traces on a PCB require a reference plane to have a controlled impedance? Whereas, in an unshielded twisted pair there is no reference, there are only the two conductors? Thank you for any answers that you can provide.

\$\endgroup\$
2
\$\begingroup\$

Why do differential signal traces on a PCB require a reference plane to have a controlled impedance?

They don't is the short and long answer. However, if you are wanting to fully utilize the space on your PCB for other components and other non-connected circuits you need to use ground planes to avoid upsetting the controlled impedance and avoiding unnecessary cross-talk or interference.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ So how would you calculate the controlled impedance without a reference? Every calculation I see for controlling the impedance requires there to be a distance from the reference (microstrip, stripline). \$\endgroup\$ – Swifty Dec 11 '19 at 13:00
  • 1
    \$\begingroup\$ Without a reference plane you resort to the general formula for a transmission line. That formula is (for high frequencies) Z0 = \$\sqrt{L/C}\$. So if you can evaluate the loop inductance and capacitance of the conductors, you can plug them into the Z0 formula. \$\endgroup\$ – Andy aka Dec 11 '19 at 13:05
  • \$\begingroup\$ Thank you. So how does using a ground plane avoid upsetting the controlled impedance? \$\endgroup\$ – Swifty Dec 11 '19 at 13:32
  • 2
    \$\begingroup\$ The ground plane disengages electric fields extending into unwanted areas where other components might be. Regard the ground plane as a wall in a house or building; a little noise or sound may leak through but, in the main, any activity in one room is segregated from activity in the next room. \$\endgroup\$ – Andy aka Dec 11 '19 at 13:35
2
\$\begingroup\$

The coupling between differential pairs is different when differential traces are run over a ground (or power) plane on a PCB than when they are run using unshielded twisted pair (UTP) wires.

In the first case, assuming edge coupled traces on (or in) a PWB, each trace couples much more strongly to the ground plane than it does to the other trace. This is why in most cases the differential impedance of a pair of traces is a little less than twice the single ended impedance of each trace. For example, if you design a pwb with two traces, each one of which has a single ended impedance of 50 ohms, then the differential impedance of those two traces ends by being approximately 90 ohms. The ground plane in this case is used to establish the impedance of the traces, not upset them.

In the case of UTP wires, the wires couple to each other much more strongly than they do to anything else.

| improve this answer | |
\$\endgroup\$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.