LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Best Filter to use to get specific frequencies

Solved!
Go to solution

Hi All,

 

Objective: I am trying to build a program which will give me just the octave band frequencies from a waveform.

 

I tried to use a butterworth filter on a wave form using a bandpass option to get between 22KHz and 44KHz freq. I am able to get that, but the problem is, if I look at the octave analysis of the whole signal, I am still getting data in 16KHz band which I don't want.  I tried to use two butterworth filter in series but I cannot seem to eliminate the 16KHz band.

The level in 31KHz band is 120dB and in 16KHz band is 108dB. Is there any way I can get the 16KHz band to go down to 90dB, that would be terrific. I have tried to play with couple of other filters,  but I can't seem to find a correct one or any other way to get what I need.

Also more thing, does anybody knows any tweeters which will be able to produce 31Khz freq band with atleast 110dB or more. I have no luck trying to find one which will produce such high frequency noise.

 

Best Regards,

Nitin

 

0 Kudos
Message 1 of 2
(2,494 Views)
Solution
Accepted by topic author npai

Nitin,

 

If you want to pass 22 kHz and attenuate 16 kHz by about 18 dB relative to the 22 kHz, you will need a high order filter.  I do not have my filter handbook nearby but a quick check of the book I do have available suggests that you may need an eighth to tenth order filter.  Your normalized bandwidth is about 1.4.  At that point a 7th order Butterworth filter attenuates the undesired signal about 12 dB more than the desired signal.

 

An elliptic filter or a notch filter may be more useful if you have a single interfering signal at a known frequency.  If you have broad band signals such as noise or vibration, you will need a high order filter.

 

You may also be abel to use transform techniques and work in the frequency domain, but there is no magic trick to isolate two signals close together in frequency and far apart in magnitude.

 

Lynn

Message 2 of 2
(2,489 Views)