2-port network conversion from S-parameters to Voltages and Currents

For the general 2-P network in Fig. 1, considering the input/output lines (that model finite connection lengths) effect and assuming loss-less propagation with real ZG, ZL:

Fig.1

The incident/reflected waves and voltage/current at the ith (i=1,2) port are by definition and as a function of distance z from the source:

(1)

The average power associated with the incident wave on the input/output line (z=0, z=l3) connected to the ith (i=1,2) port is:

(2)

Similarly, the reflected power at the same line is:

(3)

because the lines are loss-less, the incident and reflected power at port ith are the same as in (1) and (2).

If both input and output line are matched i.e. and ZG= Z01 and ZL= Z02 then:

(4)

and:

(5)

then PAVS=P1+(0) when ZG= Z01 and because the line are loss-less this is also the the power available from the source at port 1, which is independent of ZIN. If ZGZ01 then the expression for V1(0) in (5) must be modified accordingly and the resulting incident power, which is provided by (2), will no longer be equal to the power available from the source. In the first case, i.e. ZG= Z01 and (5) is valid, because:

(6)

then subtracting the two gives the power delivered to port 1 (or to the line, because the line has no loss):

(7)

Then:

(8)

which shows that the generator sends to the input line or port (if the line is loss-free) the power PAVS that is independent of the input matching condition (i.e. ZIN). However if the input impedance is not matched to the input line ( ZINZ01) part of the incident power PAVS=½|a1(0)|2=½|a1(l1)|2 is reflected and this reflected power is just ½|b1(l1)|2 while the net power delivered to port 1 is P1(0)=P1(l1)=PAVS-½|b1(l1)|2 .

In the particular case with ZG=Z01=Z02=ZL= Z0 the input/ouput lines have no-influence because the input impedance of a transmission line closed on a load impedance equal to its characteristic impedance is just equal to the load (i.e. the of characteristic impedance):

(9)

and in the input size Thevenin analysis shows that the generator through the input line is equivalent to a generator with internal resistance Z0 and voltage VG that differs from the original one only by a phase factor. Then the circuit in Fig. 1 simplifies to the one in Fig. 2:

Fig. 2

From the incident/reflected waves, and voltage/current definition:

(10)

and considering the matched termination at port 2:

(11)

therefore the input impedance and the current at each port as a function of the same port voltage are:

(12)

the port 2 parameters can be expressed as a function of port 1 voltage by noticing that:

(13)

hence all port voltages, currents and the total current flowing into the 2-port network Itot=I1+I2 can be expressed as a function of port 1 voltage which is useful when one wants to extract Itot and this is the current flowing in the shunt branch of a T-network such that the current might not be physically accessible while V1 and the scattering parameters are generally measurable:

(14)

Notice that in this case (2-port network is simply a shunt Y) from (7) in the section on even/odd mode decomposition of symmetrical networks S21-S11=1 or S21=1+S11 thus (14) reduce to:

(15)

where particularly the first and the last one are intuitive from the shunt configuration forcing the same voltage on both ports. As pointed out in the section on Even/odd mode decomposition, if |Y| is small, which is the case when R is in the tens of kΩ range and C is in the pF range, then S11≈ -YZ0/2 and S21≈1. Hence Y can be evaluated by reflection-type measurement if S11 or the reflected power are available.