0
\$\begingroup\$

When designing microwave amplifiers, stability is an important factor to consider. Stability is usually measured with the K factor. In a unilateral network, S_12 = 0, and this makes the K factor infinite.

However, we also know that being unilateral means Gamma_in = S_11, and Gamma_out = S_22. What if the network is unilateral, but either S_11 or S_22 are larger than 1? Will the amplifier be unstable despite being unilateral? Is this situation possible at all?

\$\endgroup\$
5
  • \$\begingroup\$ I've never used the term with bilateral which has a similar meaning as bipolar nor seen any stable s11>1 as this implies the feedback amplifies any input to the input. ( Return Gain not Loss) Hence unstable But I never thought of using s parms for an oscillator \$\endgroup\$ Dec 22 '19 at 18:38
  • \$\begingroup\$ @TonyStewartSunnyskyguyEE75, usually you use S-parameters to describe amplifiers, and when you design an amplifier you want to make sure you didn't accidentally design an oscillator instead. \$\endgroup\$
    – The Photon
    Dec 22 '19 at 18:58
  • \$\begingroup\$ @ThePhoton Γ_in = S_11 is true only when the port 2 is matched, not in the general case. \$\endgroup\$ Dec 22 '19 at 18:59
  • \$\begingroup\$ @Ali, OK, fair enough. But you could equally well talk about \$S_{11}\$ of the whole 1-port sub-system composed of the amplifier and its load. If you're going to parse the definitions of \$\Gamma_{in}\$ and \$S_{11}\$ this carefully, you should probably also give us the exact definition of "unstable" you're using. \$\endgroup\$
    – The Photon
    Dec 22 '19 at 19:04
  • \$\begingroup\$ it will certainly be distorted if unilateral and s11 or s22>1 and not useful \$\endgroup\$ Dec 22 '19 at 19:17
0
\$\begingroup\$

It depends whether your 2-port is an arbitrary black box where you can dial in any S-parameters, or something feasible like a transistor where the values are constrained to be physical.

If the former, consider a back box with |S11|>1, and |Sxy|=0 for x,y !=1. It will be unstable for some input impedances regardless of output load, yet it's unilateral by your definition.

Computing stability circles for a port means considering the network as a whole, so for the input it involves both the S11, and the S12/21/22/load. What the ideal of unilaterality does is reduce the second term to zero. Often the |S11| is <1, and then you've automatically got stability with passive inputs. But it isn't always. Look up a Pierce oscillator for instance. It's quite easy to turn an emitter follower into a Pierce with some strays.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.