Hi,

I built an ideal balun as in "Measuring S-Parameters of a Differential Mixer" and I am using it to measure the differential/common return loss of a 2 port device. I use spectre. The device is modeled as just 2 50Ohm resistors connected to ground in their center point (so 100 Ohm differential and 25 Ohm common mode). I have connected two ports to the balun:
port1 r=100 connected to terminal "d" of the balun
port2 r=25 connected to terminal "c" of the balun
resistors connected to terminal "p" and "n" respectively.
I chose sp analysis and I selected both ports.
The result is correct: S11=-300dB S22=-300dB so means perfect matching.
If I insted change the value of the two resistors to 80 and 20 respectively (so differential =100 and common mode = 16) the result are wrong: S11=S22=-20.1dB.
I was expecting S11=-300dB and S22=-13.1dB.
My conclusion is that I can use the balun to measure the S-Parameter only if the DUT is balanced in both the two inputs. Or, is not possible in one analysis calculate the differential/common return loss.

Any comments/suggestions are welcome,

Giacomo

hi Ken,

To simulate differential s-parameter between two nodes,
We only need to place a port cell between those two nodes, and then check s11.
Am I right?

From some meterial, I saw Cadence used a complex circuit which contains cccs and vcvs to simulate differential s-parameter. I believe there must be some reasons for Cadence to use that complex circuit. Could you give me some idea?

Hi Ken,

I am using this spectre netlist:

simulator lang=spectre
subckt balun (d c p n)
T1 (d 0 p c) transformer n1=2
T2 (d 0 c n) transformer n1=2
ends balun

I will send you the entire testbench netlist on monday.

Thanks,
Giacomo