FIR Low Pass Filter Response

Status
Not open for further replies.
E

Eminent.Engineer

Member level 3
Joined
Oct 8, 2012
Messages
54
Helped
10
Reputation
20
Reaction score
10
Trophy points
1,288
Visit site
Activity points
1,630
Hello Everyone

I designed a FIR Low Pass Filter using MATLAB's fdatool, filter specifications are:
Fpass = 10K
Fstop = 12.5K
stopband attenuation = 200dB
but the designed filter doesn't show the 200dB attenuation ? The filter's specs and its response in fdatool is attached as images.

Thanks in Advance.

 

This is far beyond precision that matlab could provide I think.
By the way, 200 dB stopband attenuation is not a real world value. Please come back to Earth.
 

Mityan said:
By the way, 200 dB stopband attenuation is not a real world value. Please come back to Earth.
Click to expand...

Thanks for the reply. Yes I know but we are doing this just for analysis/experimentation in simulations. Any suggestion how I can design the filter with specs as mentioned above ?
 

Reducing the extreme Fs/Fc ratio would be a step towards a reasonable design implementation. I'm not sure about the 200 dB number, but a higher stop band attenuation than about 45 dB can be surely acheived.
 

I tried your model but were tired of waiting the end of estimation.

Why do you want such a great sample rate? It is not needed for that kind of filter.
Downsample 100 times (Fs = 65 khz) and it works great, you may get your 200 dB.
 
Reactions: FvM

    FvM

    Points: 2
    Helpful Answer Positive Rating
Mityan said:
I tried your model but were tired of waiting the end of estimation.
Click to expand...
Thanks for your time and try.

Mityan said:
Why do you want such a great sample rate?
Click to expand...
Since this filter (I am designing) is acting as Decimation Low Pass Filter for Delta Sigma Modulator so Fs of the Delta Sigma Modulator is "6500000".
 

Since this filter (I am designing) is acting as Decimation Low Pass Filter for Delta Sigma Modulator so Fs of the Delta Sigma Modulator is "6500000".
Click to expand...
Read about multi-rate filters.
 


Hello Everyone

Can someone guide me, Its my fault or MATLAB's FDA tool ? Are there any other tools for FIR filter design ?
Thanks in Advance.
 

Hi,

This FIR synthesis uses the Parks-McClellan procedure, that is based in an iterative method (the Remez exchange algorithm). This algorithm not always converges to the specification.
Situations when convergence can fail are, for instance, too high order, specifications with extremely high attenuations in stopbands or too low ripples in passbands, very narrow bands or sharp transitions (relative to sampling frequency), etc. This happens in the present case.
The "Density Factor" is a paramenter that can help, but I'm not sure how much.
Another issue is that the filter order estimated with the option "minimum order" could be insufficient.

In order to met your specification, you can design the filter as the cascade ot two or more filters (e.g. two identical filters with Apass=0.5dB, Astop=100dB, or 4 cascaded with Apass=0.25dB, Astop=50dB). Although the total order would be higher than if designed as a single filter, it would be very reasonable.

Even with a nice design, in a practical implementation finite-precision arithmetic and numerical errors can be the dominant limitation.
Regards

Z
 
Last edited:
Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…