obtaining general trend from noisy data

[aa2105]

Hi

I am measuring the bending strain of a wing which will be 'plucked' to cause it to vibrate. Firstly, I need to record the strain measurements versus time and ouput them to a file. In LabView I can't seem to adjust the graph so that it gives amplitude versus time in seconds - its in 'absolute time'. (I'm very new to LabView...)

Next thing I need to do is to filter the data so that it gives a nice smooth curve (from which the damping ratio can be found). i.e. for example, if I have a really noisy sine wave, I want a filter that will remove the noise leaving the trend behind which represents the amplitude and phase accurately. Is there an in-built filter for this? There are various filters for certain applications which affect the amplitude and phase and I am really not sure which one would be best for my application. Thanks in advance.

[aa2105]

Sorry, let me clarify: its not actually noise that I'm getting (I think ...). Having thought about it a little, the data I'm getting is 'erratic' and seems noisy because of the sensitivity of the gauges. So the voltage output changes by 'x' amount all the time causing some deviations - however, I still need to 'filter' this data to get the trend. I guess, the best way is to get an average at every timestep or something similar. Any ideas?  

[johnsold]

The time axis of the graph can be changed by using the properties.  Context-click on the graph and the dialog should allow you to change from absolute time to relative time or to a numeric format for the axis.

The trend issue depends on the nature of the signal and the nature of the noise.  Averaging is easy and often the first thing to try.  Smoothing filters or sliding window filters can be used to smooth the data and reduce the effects of noise.  If you know what the signal waveform should look like, fitting a function to the data and deriving the parameters from the fitted function is good.  From your description I suspect that you may have combinations of several types of signals and fitting may not work well.

Lynn 

[aa2105]

I am only measuring the voltage output from the strain gauges using the NI 6229 data acquisition unit. In LabView, the the DAQAssistant is configured to continously read 100 samples at a rate of 1kHz (not sure if this information helps...).

I'll try the smoothing filters. Thanks.