# Realisation conditions for impedances

Given a frequency function, which conditions should it fulfill to be realisable and implemented as a one-port impedance? I know that for it to be realisable with passive RLC components, the frequency function has to be positive real. But I'm also looking for needed conditions for an active realisation. In my particular case, the frequency function looks somewhat like (s^(n+2)+...)/(s^n); the degree of the numerator exceeds the degree of the denominator by two. I wonder if this makes its realisation impossible. I want to use it as a shunt filter impedance connected in parallel to a load.

• Did you ask this same question the other day? – Andy aka Jun 10 at 9:40
• Yes, I did. But yesterday I failed to formulate the question correctly. – Chris Jun 10 at 10:25
• So, do you remember what I advised? – Andy aka Jun 10 at 11:19
• Yes, to be more specific. But you didn't mention what kind of specific information you needed. – Chris Jun 10 at 13:13
• There is no converter that takes an arbitrary numerator and denominator of any order and churns out an impedance value that can be simulated. Different changes in numerator and denominator can produce vastly different impedance circuits. – Andy aka Jun 10 at 13:31