spectral measurement

[engcad]

Hi, 

I use ni Mydaq to measure the voltage of 9V battery. I use the "Spectral Measurements" Express VI to get the magnitude "peak" of the signal in the ferquency domain 

Then I use a low pass filter and also show the filtered signal in the time and frequency domains

I have some questions

1- Is the X axis of the FFT "Peak" waveforms represents the true frequency in Hz or it need to be scaled?

2- The Y  axis of the FFT "Peak" waveforms is in dB.  what is the reference signal that used to compute the dB values 

3- Is the noise outlined by red ellipse in the picture represents the noise floor? I'm confused that this noise is reduced after low pass filter

Sorry it may be basic questions but I'm new to signal processing 

[GerdW]

Hi engcad,

1 - Usually you get an x axis in Hz when you used the correct function in a proper way. Did you?

2 - Which function did you use? What does the LabVIEW help say about that very function?

3 - Yes, it is signal noise as well as math noise (due to limited precision of algorithm and datatype). Why does it confuse you when a filter reduces noise?

Please attach your VI!

[engcad]

Thanks for your reply

my VI is attatched

[GerdW]

Hi engcad,

1 - Yes, x axis is in Hz. No need to scale…

2 - It looks like the main peak is used for reference…

3 - Are you still confused?

[engcad]

Thanks

[johnsold]

When looking at the spectrum of the noise on a large DC signal (~2 mV pp on 9.2 V) it is often better to remove the DC component. If you do this in software, it will eliminate the large (125 dB) DC component in the frequency domain. If you do it in hardware by inserting a suitable capacitor between the battery and the AI input terminal, you can use a much more sensitive input range such as +/-50 mV or 100 mV.

Also it is helpful when posting code to save some typical raw data as default in an indicator first. We generally do not have your devices or your equipment to duplicate the data.

When I ran a similar test with a USB-6009 and DAQmx Base, I get these results:

I supect that what you are seeing is the internal noise of your DAQ device and the limits of numerical calculations.

Lynn