Eliminating vibration effects from the output of a loadcell?

[seyyah]

As you know loadcells generally work in an environment with lots of mechanical vibration. So the output signal of the loadcell also vibrates very much. Averaging the output slows the output very much. Can we eliminate/ filter out this vibration from the output wihout much loss of the response rate? Thanks in advance.

 

[FvM]

Re: Eliminating vibration effects from the output of a loadc

Obviously, the problem is about adjusting the filter specification according to the interfering signal ("vibration"). The filter
characteristic must not be too steep, because it would create an oscillating response on it's own. Moderate filter characteristics
are bessel or butterworth. Also a 2nd or 3rd order CIC decimator can be suitable. If the vibration has a dominant frequency,
decimating/averaging to the period of the interfering signal can help. The analog signal filtering should meet the Nyquist criterion
not to undersample the interfering signal.

 

[seyyah]

Re: Eliminating vibration effects from the output of a loadc

So you propose a low pass filter. I saw some devices filter outs oscillations over 0.25Hz (their datasheet says). But obviously the actual output should change or response faster than this. How can we separate an actual load variation and an oscillation?

 

[karesz]

Re: Eliminating vibration effects from the output of a loadc

... How can we separate an actual load variation and an oscillation?

Hi,
I think your only possibility is to know the spectren & do filter for the falsh lines!
Some other technique, per algorithmic, can be an adaptive median filtering...
K.

 

[seyyah]

Re: Eliminating vibration effects from the output of a loadc

... How can we separate an actual load variation and an oscillation?

Hi,
I think your only possibility is to know the spectren & do filter for the falsh lines!
Some other technique, per algorithmic, can be an adaptive median filtering...
K.

What is spectren & do filter ? can you give some more information please?

 

[FvM]

Re: Eliminating vibration effects from the output of a loadc

So you propose a low pass filter. I saw some devices filter outs oscillations over 0.25Hz (their datasheet says). But obviously
the actual output should change or response faster than this.

Sounds like you are rather guessing about possible problems than observing real oscillations in your measurement. Otherwise you
should know about oscillation frequencies.

Generally, I actually don't expect that other filter types than a low-pass are of much use in load-cell applications. There
may be exceptions. But the procedure is basically simple. Digitize the load cell signal with sufficient resolution and try to
filter it with a signal processing tool. Determine, if you get a significant improvement by this filtering. If so, implement
it in your application.

 

[karesz]

Re: Eliminating vibration effects from the output of a loadc

..What is spectren & do filter ? can you give some more information please?

Also; I mean: you must know what are your frequencies (spectral lines in frequency domain) & if their are from your system mechanically given_its frequency(or spectrum)is stabile = you can filter these unwished vibration frequencies (or spectral lines)... (do filter, for me, = eliminate it, filter it out electronically/per software...)
K.