Multifunction DAQ
amplitude of a periodic signal
Some points to get started.
1. The sequence structure is unnecessary. It makes no difference to the way the program runs. Remove it.
2. The length input to Extract Portion of Signals.vi is 500 but the length of the signal is 1. Setting that input to 0.5 gets half of the data.
3. The width input to the peak detector is in units of samples. With a setting of 500 it looks for a peak 500 samples wide. With that setting the output of the peak detector is almost always an empty array. Put an indicator on the output wire and watch. Also put an indicator on the # found output and watch it stay zero.
3.a. Widths of 14 t0 48 samples seemed to give me 5 peaks with 5 cycles of the waveform (500 samples) going into the peak detector. This was with amplitude = 1 and noise = 0.1.
4. Since the peak detector has array inputs it might be simpler (and easier to understand) if you just put Get Waveform Components followed by Array Subset to get the data you want to send to the peak detectors. Express VIs (and the Dynamic Data Type they insist on generating as outputs) tend to obscure what is going on. I try to avoid them.
Lynn
how do i link width of peak detector and a length of array, please tell me in three words if you may
Only three words? Use a wire!
There really is no relationship between the width input to the peak detector and the length of the array, except that obviously a width larger than the length of the array will never find a peak. Start by reading the detailed help file for the Peak Detector VI.
 |
width specifies the number of consecutive data points to use in the quadratic least squares fit. width is coerced to a value greater than or equal to 3. The value should be no more than about 1/2 of the half-width of the peaks/valleys and can be much smaller (but > 2) for noise-free data. Large widths can reduce the apparent amplitude of peaks and shift the apparent location. For noisy data, this modification is unimportant since the noise obscures the actual peak. Ideally, width should be as small as possible but must be balanced against the possibility of false peak detection due to noise. |
With the noise amplitude about 10% of the sine amplitude I found that width needed to be larger than 10. This was with a frequency of 10 Hz which produced 100 samples per cycle. Thus, the width is > 10% of the period of the waveform. For other frequencies and other signal to noise ratios, other widths might be better.
The width parameter must often be determined empirically. If both the noise and the frequency of the signal change while the program is running, it may be very difficult to find a setting which works well for all cases. And having the program adapt may also be difficult unless you can clearly define how many peaks should be present at any given time.
I put the (-x) in because it seemed that the threshold should be negative for the valleys. If you have a reason to use the same threshold for both directions, it can be removed.
why does this program refuse to work with certain parameters like no noise at all or 0 freq and amplitude of tones?
When you add two arrays of different lengths, the resulting array is the length of the shorter input array. Some of the settings you describe probably result in an output of the noise generator having zero elements. I do not see this effect for any settings of zero frequency amplitude. The Generator VIs generate 1000 element arrays with all zeros. I notice that the valley detector seems to be more particular about the width setting with the sawtooth and square waveforms. This is probably due to the way it fits a quadratic curve to the data. The sharp edges on those waveforms do not fit well to a second order polynomial.
Lynn
Here is something to play 🙂
ARRRG this new NI forum fail to load my png 😞
OK no more block cutting, the block size is generated 🙂
Assuming you know your base frequency I calculated some values for the peak valley stuff
and added my try with a fitting filter
