Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

What's the system if the impulse response is h(t)=1/t?

Status
Not open for further replies.

No one

Full Member level 5
Joined
May 11, 2005
Messages
287
Helped
4
Reputation
8
Reaction score
2
Trophy points
1,298
Activity points
3,425
Hello everyone;
Suppose that impulse response of a system is:

h(t) = 1 / t
What is the system? What is the Fourier transform of it?
 

steve10

Full Member level 3
Joined
Mar 26, 2002
Messages
175
Helped
32
Reputation
64
Reaction score
0
Trophy points
1,296
Location
Los Angeles (Chinese)
Activity points
2,538
Re: 1/t System [hlp]

The Fourier Transform of 1/t can be easily obtained as follows (I'll use the notation in Mathematica):

F[1/t]=Integrate[Exp[-itw]/t,{t,-Inf,+Inf}]
=Integrate[Cos[tw]/t,{t,-Inf,+Inf}]-iIntegrate[Sin[tw]/t,{t,-Inf,+Inf}]

The first integral is zero when the Cauchy main value is applied. Actually, it's integral of an odd function. The second integral is a famous integral

iIntegrate[Sin[tw]/t,{t,-Inf,+Inf}]=Pi when w>0, and -Pi when w<0. Therefore,

F[1/t]=-i*Pi*[H(w)-H(-w)].
 

v_c

Advanced Member level 2
Joined
Oct 11, 2005
Messages
511
Helped
113
Reputation
226
Reaction score
37
Trophy points
1,308
Location
Chicago, IL, USA
Activity points
8,224
Re: 1/t System [hlp]

I think the Fourier transform for \[f(t)=\frac{1}{t}\] is \[F(\omega) = -j \pi sgn(\omega)\], where \[sgn\] is the sign function. See top of page 2 in **broken link removed** and note that you can multiply both sides of the transform integral by constants. I think this is consistent with what steve10 reported above.

Is this signal valid for all values of time (positive and negative t)?

If the signal does exist for positive values of time and h(t) is indeed an impulse response (with the impulse applied at t=0) then this would be a non-causal system. This just means that there is a response before the signal is even applied!! If this is a causal system then you would have to make the negative portion of the respose equal zero or just h(t) = (1/t) u(t), where u(t) is the unit step function. This would make the evaluation of the integral much easier as you would only have to integrate from zero to \[\infty\].

Best regards,
\[v_c\]
 
Last edited by a moderator:

neils_arm_strong

Full Member level 5
Joined
Jan 5, 2006
Messages
288
Helped
22
Reputation
44
Reaction score
3
Trophy points
1,298
Location
Bangalore, India
Activity points
3,247
Re: 1/t System [hlp]

The fourier transform of 1/Πt is -jsgn(f)
So FT of 1/t will bo -jΠsg(f)
sgn(f) has the property that
sgn(f)=1 f>0
= -1 f<0
The system is a hilbert transform system(i.e) it produces a phase shift of + or - 90 degrees based on wetherthe frequency is - or +.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top