node A and node B don't have DC path to gnd - basically floating, which is odd.
unless this ckt may be used for it's impedance between A and B?
Looks like differentially Zab is very small (two parallel Rds, with vds=0, overdrive vgs decided by M4a/M4b with current source M3a/M3b) due to the negative feedback
however, common mode Rab is very extremely high, essentially the input impedance of the source follower.
Since the Rab is one-port device, you can build following test bench.
Provide a combination of common mode and differential mode voltage between node A and node B.
From AC sim, find Rab.
Calculate noise: output noise is Iac into the differential voltage source, input source is the differential voltage source.
Result: Iac is equivalent current noise, input noise is equivalent voltage noise.
Let me know if this works. |