Thanks for the clarification on the rounding aspect.

I do understand the benefits of the low pass filtering but wish to clarify :

- I normally am required to acquire hydraulic flow and pressure signals. In the applications that I cover, you can take that flow plots are almost at DC levels and pressure plots alone can at times touch about 70 Bar / sec rise or fall.

- Even when handling the input signals as differential ones, I normally get noise glitches superimposed on the base plot. My requirement is to remove these glitches and only plot the trend as a neat curve.

In such an even how do I decide on the low pass filter value ? ( currently I oversample - say at about 2.5kHz - acquire 125 samples every 50mS and then average. The resulting mean value is then plotted. )

Thanks

Raghu