Extracting a sinewave signal when characteristic phase and frequency are known

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
Sign up Log in
5 replies · 4K views
tchetch
Messages
1
Reaction score
0
Hi there,

i have this problem:
I have a sinewave signal contaminated by a large amount of noise. I already know the phase of the signal and its characteristic frequency, and I am searching for its magnitude.

If I do a basic Fourier transform, I would obtain the magnitude at its characteristic frequency, but the phase would be generally shifted. Does anyone know a way to extract the magnitude at its specific phase and frequency ?

Thanks in advance
 
Engineering news on Phys.org

Discussion

Filter out the noise and then do peak detect? Or just take the RMS of the signal, and calculate the peak value (adding or subtracting an offset as necessary)?
 
I don't understand how a sine wave can have absolute phase?
 
Is it truly noise or have you considered it could be an aliasing issue? Have you treid playing with the sampling rate?
 
I was also looking through some old notes I have. Tom Irvine over at Vibrationdata.com published a quick paper regarding a method he wrote a small program for, called sinefind. The paper is entitled "Sine Function Identification and Removal." I can't post it here because it is Tom's work. It pulls out amplitude, frequency and phase by a curve fit method.

If you have any interest in vibration analysis, take a look over at Tom's website.

www.vibrationdata.com
 
There is a type of numerical filter than can do this, but the name escapes me right now. Is the noise gaussian or normal?