Fourier Transfrom of scaled periodic impulse train

AI Thread Summary
The discussion centers on the Fourier transform of a scaled periodic impulse train, where the amplitudes of impulses alternate between 2 and 1. The initial solution proposed was deemed incorrect due to the oversight of alternating values in the pulse train, which should be represented as separate summations for each amplitude. A corrected expression was suggested, indicating the need to account for the distinct values of the impulses at different intervals. Participants emphasized the importance of accurately describing the input pulse train to clarify the Fourier transform. The final consensus acknowledged that the modified equation effectively captures the periodic nature of the impulse train.
neg_ion13
Messages
29
Reaction score
0

Homework Statement


Fourier transform of scaled pulse train from -inf to +inf. Starting at 0 the first impulse is scaled at 2, the second impulse at 1 is scaled as 1, third at three scaled at 2, etc...

Homework Equations


Does my solution look correct?

The Attempt at a Solution



X(jw) = 2\Pi\Sigma\delta(w - \Pik) + \Pi\Sigma\delta(w - \Pi(k + 1))
the sigma notation is from k = -inf to inf. My thinking behind this is to represent a pulse for each scaled delta funciton in freq domain
 
Last edited:
Physics news on Phys.org
neg_ion13 said:

Homework Statement


Fourier transform of scaled pulse train from -inf to +inf. Starting at 0 the first impulse is scaled at 2, the second impulse at 1 is scaled as 1, third at three scaled at 2, etc...
I'm having trouble seeing any pattern in the pulse train you describe. Is there no pulse at t=2? What about the pulses at negative times?

Homework Equations

The transform of a simple pulse train might come in useful here.

Does my solution look correct?

The Attempt at a Solution



X(jw) = 2\Pi\Sigma\delta(w - \Pik) + \Pi\Sigma\delta(w - \Pi(k + 1))
the sigma notation is from k = -inf to inf. My thinking behind this is to represent a pulse for each scaled delta funciton in freq domain
If you can describe the input pulse train better, it will be easier for us to see what is going on.
 
The pulse train is periodic so it runs from plus -inf to +inf.
At t = -2 the impulse = 2
At t = -1 the impulse = 1
At t = 0 the impulse = 2
At t = 1 the impulse = 1
At t = 2 the impulse = 2
At t = 3 the impulse = 1
At t = 4 the impulse = 2
Etc... from -inf to +inf
 
neg_ion13 said:
The pulse train is periodic so it runs from plus -inf to +inf.
At t = -2 the impulse = 2
At t = -1 the impulse = 1
At t = 0 the impulse = 2
At t = 1 the impulse = 1
At t = 2 the impulse = 2
At t = 3 the impulse = 1
At t = 4 the impulse = 2
Etc... from -inf to +inf
Thank you for clarifying that.
neg_ion13 said:
Does my solution look correct?

The Attempt at a Solution



X(jw) = 2\Pi\Sigma\delta(w - \Pik) + \Pi\Sigma\delta(w - \Pi(k + 1))
the sigma notation is from k = -inf to inf. My thinking behind this is to represent a pulse for each scaled delta funciton in freq domain
No, it does not look correct. Can you show your work?
 
The Fourier transform for a pulse train is 2*pi*sigma*delta(w - 2*pi/T). Here T = 2. the problem is that a regular pulse train has a constant value of 1 for the impulse while this one does not. This means the above equation needs to be modified to alternate between value of 2 and a value of 1. That is how I arrived at the above equation. The first part of the equation is for value of 2. When k = 0 the first part 2*pi. The second part = pi when w = pi, etc... This was my thinking behind this X(jw). Because it's linear I could split the problem into two summations to represent each of the 2 values of the delta function.
 
Oops! I am overwriting my impulses. It should be:

X(jw) = 2*pi*Sigma[delta(w - pi*(2k))] + pi*Sigma[delta(w - pi*(2k + 1))]
 
Okay, I agree with your solution now. I had in mind expressing it a different way, but it is equivalent to what you have.

What I had in mind:
X(ω) = π(∑δ(ω-πk) + ∑δ(ω-2πk))

where k runs from -∞ to +∞.
 
Cool. Thanks a lot.
 
Back
Top