How to evaluate stability of a non-causal System?

3 replies · 2K views · last reply
rudra
Messages
14
Reaction score
0
Multiple-choice question text: impulse response h(t) = exp(αt)u(t) + exp(βt)u(−t) of a linear time-invariant continuous-time system, asking for the stability condition, with options A to D about the signs of α and β.




For this problem I have taken laplace(one-sided) transform of h(t) which gives me
H(s)=1/(s-α). From this I can state that α must be -ve for G(s) to be stable.
But my problem is while taking one-sided Laplace Transform the exp(βt)u(-t) part gives 0.
So in H(s) according to my calculation, β doesn't appear. I don't know what I am doing wrong. Please help
 
rudra said:
Multiple-choice question text: impulse response h(t) = exp(αt)u(t) + exp(βt)u(-t) of a linear time-invariant continuous system; options ask which of alpha and beta signs makes the system stable, (A) both positive, (B) both negative, (C) alpha positive and beta negative, (D) alpha negative and beta positive.




For this problem I have taken laplace(one-sided) transform of h(t) which gives me
H(s)=1/(s-α). From this I can state that α must be -ve for G(s) to be stable.
But my problem is while taking one-sided Laplace Transform the exp(βt)u(-t) part gives 0.
So in H(s) according to my calculation, β doesn't appear. I don't know what I am doing wrong. Please help

As I said, this problem deals with a nonexistent situation (anyone diosagree?).

But: the u(t) part is easy: what does exp(αt) do as t → ∞?

Now for the noncausal part: u(-t) = 1 for t < 0 and = 0 for t => 0. So for any negative value of t, what does exp(βt) do as t gets more and more negative, approaching t → -∞, with β positive or negative?
 
@rude man,

I think your approach gives the proper soln. α should -ve and β by your logic should be positive.
 
rudra said:
@rude man,

I think your approach gives the proper soln. α should -ve and β by your logic should be positive.

That's where I would put my chips. t → -t in the u(-t) term.

I still think you should ask your prof if that problem has any physical meaning. Unless you're a math purist I see no reason to worry about it.