Filter.vi bug

[gslomo]

I'm trying to use Filter.vi provided with Labview 8.20 to apply a moving average on a signal. Applying the same input signal several times, sometimes the output generated by Filter.vi is not the same. Particularly, it sometimes generates a peak at the begining.

I would appreciate any help to solve this issue. Thanks.

Gabriel

[GerdW]

Hi Gabriel,

could you provide an example (vi!) showing this behaviour - this way we should be able to reproduce your problem...

[gslomo]

Hi GerdW,

I'm sending you the vi and a sample signal. Thanks for your help.

Regards,

Gabriel

[GerdW]

Hi Gabriel,

the problem is: you're using express vis...

I changed your vi to show the behaviour: the filter is now called in a loop. In the first iteration the result looks fine, but in the second you see the starting glitch! This is due to the buffering behaviour of the express vi! The express vi is buffering values to allow the running mean - when the next iteration starts you still have values from last iteration in this buffer resulting in this peak at start of signal...

To solve this:
- easiest: make your own running mean (see attachment) (and avoid express vis )
- prepend dummy data on your input array (twice the size of the running average...)

[gslomo]

Hi GerdW,

Great. Thanks very much for your detailed answer.

Regards,

Gabriel

[gerardo10]

GerdW,

I have the same problem but it is for labview 7.1.  Is there anyway you could repost your smoothing vi for labview 7.1? I would really appreciate it! Thank you in advance!

Gerardo

[GerdW]

Hi Gerardo,

here's a snippet of that VI:

The lower for loop was my suggestion for a running average. The upper while loop is for demonstration of the OPs problem...

[Gerardo09]

GerdW,

Thank you very much! This is a very helpful VI. I appreciate your help!