buddypoor wrote on Aug 13th, 2008, 10:12am:Conditional stability is a specific term from the control theory describing a system which can be stabilized by increasing the loop gain.
No, you are wrong. Maybe you are not familiar with RF circuit theory.
Terms, "Conditionally Stable" and "Unconditionally Stable" are also used in RF circuit.
But these can not be completely related to stability theorem in control system theory.
http://www.designers-guide.org/Forum/YaBB.pl?num=1190272820We can judge whether an amplifier is unconditionally stable or not by well-known K-factor or Mu-factor.
See stab_fact() and mu() in "Chapter 9: S-Parameter Analysis Functions" of the following.
http://cp.literature.agilent.com/litweb/pdf/ads2005a/pdf/expmeas.pdfbuddypoor wrote on Aug 13th, 2008, 10:12am:Presently, I cannot see how one can come to the conclusion that a negative real part of the input conductance of a system indicates this kind of "conditional stability". Perhaps something else is ment ?
Assume a simple amplifier. Here I define followings.
Gamma_S=Source Reflection Coefficient
Gamma_In=Input Reflection Coefficient
Gamma_L=Load Reflection Coefficient
Gamma_Out=Output Reflection Coefficient
StabFactor_S=real(Gamma_S*Gamma_In)
StabFactor_L=real(Gamma_L*Gamma_Out)
If both StabFactor_S<1.0 and StabFactor_L<1.0 are satisfied, this amplifier is stable. This is conditionally stability.
And this conditionally stability could be satisfied even if input impedance or ouput impedance have negative real part.
This stability is dependent on source and load impedances.
For example, when an input of amplifier is left open or short, abs(Gamma_S)=1.0. If this amplifier has real(Yin)<0 or real(Zin)<0, abs(Gamma_In)>1.0. So this amplifier could not be stable.
buddypoor wrote on Aug 13th, 2008, 10:12am:What does this mean for the system under discussion ? What has to be done to destabilize the system or to keep it stable ?
For example, an amplifier of real(Yin)<0 could be stable if shunt conductance is inserted at input of amplifier so that it cancels real(Yin)<0.
Of cource, this too simple stabilizer degrades gain of amplifier.