# Re: differential impedance

J. Eric Bracken ([email protected])
Wed, 09 Apr 1997 13:48:12 -0400

>>>>> "WB" == Weston Beal <[email protected]> writes:

WB> We've seen specifications on differential impedance, but no
WB> one has been able to define what that number represents. We
WB> use the rule of thumb that the differential impedance is 2
WB> times the odd mode impedance. Does that make sense? Can
WB> someone give me a real definition of differential impedance
WB> and it's relation to odd mode impedance?

Exactly correct. Here's why:

Differential impedance is defined conventionally (see e.g. Gray &

Vdifferential
Zdifferential = ---------------
I1

where I1 is the current in ONE terminal of the diff pair.

Differential voltage is defined as

Vdifferential = V1 - V2.

Odd-mode impedance is defined as

Vodd
Zodd = -------
Iodd

Here you have to be careful. Let's let V1, V2 be the voltages on the
two lines and I1, I2 the currents.

V1 = Vodd, V2 = -Vodd
I1 = Iodd, I2 = -Iodd

So, Vdifferential = V1 - V2 = 2*Vodd.

Thus we have

Zdifferential = (V1 - V2)/I1 = 2*Vodd/Iodd = 2*Zodd.

You can also take a look at

```--