From: Ingraham, Andrew (Andrew.Ingraham@compaq.com)
Date: Wed Jan 03 2001 - 15:00:03 PST
> First, is this statement valid?
> - In a differential pair, one 'leg' of a signal's return current path is
> through the complementary 'leg' of a differential pair and not through the
> ground or power planes (Assuming equal trace lengths, Zo=50 single-ended,
> Zo=100 diff. impedance - using ECL logic as an example).
The return current flows through any nearby conductors; at high frequencies,
most of the return current is through the closest conductor(s).
If this happens to be a reference plane, that's where it goes. If it is the
complementary signal line, that's where it goes.
You can construct a differential pair out of two separate traces that are
far apart from one another, in which case each one has a return current in
its neighboring plane(s) (which happens to roughly equal the signal current
in the opposite trace). The two currents in the plane are opposite and
would cancel at low frequencies, but at high frequencies the return current
is confined to the space "under" each trace and does not cancel, so in
effect you have a circulating loop of return current in the planes, of
magnitude equal to the signal current.
In the other extreme, you can also construct a differential pair out of two
traces that are tightly coupled to one another (such as broadside coupled)
and further away from reference planes, in which case each line's "return"
current is the other line's signal current and there is no significant
current in the planes.
Differential pairs have both a differential (odd-mode) impedance and a
common-mode (even-mode) impedance. The two examples could have the same
differential impedance but very different even-mode impedances.
If you route two 50 ohm traces over ground planes, you'd get slightly less
than 100 ohms differential impedance. To get closer to 100 ohms, you need
to keep them further apart, in which case most of the return current remains
in the planes. To get more cancellation of the plane's return current, you
need them closer together, which reduces the differential impedance (you
would need to use >50 ohm traces to get 100 ohms differential).
**** To unsubscribe from si-list or si-list-digest: send e-mail to
firstname.lastname@example.org. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
This archive was generated by hypermail 2b29 : Tue May 08 2001 - 14:30:33 PDT