I do not have time right now to dig into your VI and data but here are a few observations.
- The baseline is about as high in some places as peaks in other places.
- Some of the peaks, such as around 2.4 on the Time axis, are not much different from the noise level.
- You mention that the peak amplitudes and locations move when threshold and width are changed. This is expected. Read the detailed help file for the Peak Detector VI. It explains how a polynomial is fitted to points around the peak and the calculated locations and amplitudes are derived from the coefficients of the polynomial. If you expand the graph around the peak near 1.65 you can see that the measured data points are not symmetrically distributed around the largest value. This suggests that the true peak is probably between the two highest data points and that is what the peak detector tries to find.
Because the baseline varies so much you may need to do an adaptive peak finding process where the threshold varies according to the baseline.
If you want the highest value in each peak rather than the estimated amplitude from the peak detector, use the peak detector to find locations and then take an Array Subset with length 2*width centered at the peak location. Use Array Max & Min to find the highest value and its index within that subset.
Anytime you are measuring data there will be noise and uncertainty associated with the measurement. In addition when samples are taken at discrete intervals almost always no sample occurs at the exact maximum point. So you need to decide based on the expected behavior of the system you are measuring, what form of approximation or estimate is best suited to the characterization of your data.
Lynn
