bandstop fir filter

[Pauliexcluded]

I am attempting to filter out 120htz from some data.  I thought I would just use a simple FIR filter and chop out 110 to 130 Htz band.  I don't know much about FIR filters however I'm not trying to do anything particularly complicated.  The issue is I don't seem to know what PB and SB are.  I have read thought the forums and they seem to only repeat what the help implies...that Upper SB is the greater of the two stopband frequencies and lower SB is the lesser of the two stopband frequencies.  But that doesn't seem to be the case...or I don't understand what is going on.

If I am trying to use a bandstop type filter, I should ONLY need Upper and Lower SB to be set.  For my application 110 should be the lower SB and 130 the upper SB....but this doesn't work.  So obviously this VI is not doing what I think it is doing.  Can anyone help or point me twards and example?  Or is there an easier way to fliter out a narorw frequency band from a data set?

And yes I know this is a LOT of 120htz noise.  I can reduce it quite a bit.  This has been intentionally increased so I could play with the bandstop filter.  Also I am using the current student version of labview...and don't have the ability to purchase the filter development toolkit.

Kevin

[johnsold · Accepted solution]

Kevin,

1. The very large offset in the data shown on your graph (1.8E8) is part of the problem.  Remove that and you can get some filtering.

2. The Topology must be something other than the default (Off). The default does no filtering.

3. Because the decay function (is this what you want?) has frequency components in the same range as your interference, no filter will remove all of the 120 Hz component.

4. The filter transient may prevent you from getting much information about the first part of your signal.

5. While your sampling rate satisfies the Nyquist criterion, you only have about 8 sampes per cycle of the 120 Hz signal.  A higher sampling rate might give you more data to work with.

Can you tell us more about the real problem?  What is the nature of the desired signal? Is the 120 Hz interference stable or from a source which can be monitored independently of the desired signal?

Lynn

[Pauliexcluded]

First of all, thank you so much.  The offset was absoutely the issue.  This was just me not knowing enough about the math behind the filter. I am aware that I am filtering very near the area of interest.  I'm just looking for a little bit of cleanup.

Yes, I may be able to monitor the source.  Let me work with what I have and if I need a little more cleanup I may ask for help with an adaptive method.

Kevin

**update:  Yes, even with this grossly exagerated noise, I can fit that data just fine.  Thank you.  I think this will work for me just great!