The characteristic equation is 1+L(s)=0, i.e. L(s)=−1. So a closed-loop pole sits wherever the open-loop L equals −1. Studying how L behaves relative to −1 is the same as studying 1+L relative to 0.
Now set F(s)=1+L(s) and let C be the Nyquist contour — the entire imaginary axis s=jω closed off by an infinite semicircle in the RHP (clockwise). This contour encloses the whole RHP.
Zeros of F enclosed =Z (unstable closed-loop poles — what we want to count).
Poles of F enclosed = poles of L in RHP =P.
Clockwise encirclements of the origin by 1+L=N.
But 1+L encircling 0 is identical to L encircling −1 (just shift left by 1). Hence:
N=Z−P⇒Z=N+P.
That's the whole criterion — falling straight out of counting angle changes.
What equation defines closed-loop poles? → 1+L(s)=0.
Why the point −1? → because L=−1 solves that equation.
State the count formula. → Z=N+P.
Which principle underlies it? → Cauchy's argument principle.
Stable condition in terms of N,P? → N=−P (so Z=0).
Recall Feynman: explain to a 12-year-old
Imagine shouting into a canyon. If the echo comes back weaker, the sound dies — safe. But if the echo comes back just as loud and pushing the same way you pushed, it stacks up and gets louder and louder — that's instability. The Nyquist plot is a treasure-map of how the echo behaves at every pitch of your voice. There's a "danger X" on the map called −1. We walk the whole map's edge and count how many times the drawn line loops around that X, and which way it spins. That count tells us whether the echo will explode — and we never had to actually listen to it blow up.
Dekho, feedback control me sabse important sawaal hai: system stable rahega ya blow up ho jayega? Nyquist criterion isko bina roots solve kiye batata hai. Idea simple hai — open-loop transfer function L(s) ki value har frequency par nikaalo, aur uska graph (Nyquist plot) banao complex plane me. Ab ek special point hai: −1. Ye point isliye khaas hai kyunki characteristic equation 1+L=0 hoti hai, matlab jahan L=−1, wahi closed-loop pole banta hai. Toh −1 ke aas-paas kya ho raha hai, wahi asli khel hai.
Formula yaad rakho: Z=N+P. Yahan P = open-loop ke RHP (right-half plane) poles ki ginti, N = plot kitni baar −1 ko clockwise ghera (encircle kiya), aur Z = closed-loop ke unstable poles. Stable tabhi jab Z=0. Agar plant already stable hai (P=0), toh bas dhyan rakho ki plot −1 ko na ghere — N=0 chahiye. Lekin agar plant hi unstable hai (P=1), toh ulta chahiye: ek counter-clockwise encirclement, taki Z zero ho jaye. Ye thoda ulta lagta hai par yahi correct bookkeeping hai.
Iske peeche maths hai Cauchy ka argument principle — jab tum ek closed loop ke around chalte ho, toh function ka angle kitna ghoomta hai wahi Z−P deta hai. Bas 1+L ka origin ke around ghoomna, wahi L ka −1 ke around ghoomna hai (bas left shift). GNC me — jaise satellite attitude control ya autopilot loops — yahi criterion se gain kitna badha sakte ho, gain margin aur phase margin sab decide hote hain. −1 se plot kitni door hai, utni safety margin.