Multifunction DAQ

cancel
Showing results for 
Search instead for 
Did you mean: 

Beat occurs while analog voltage measurement

I configured the NI 9205 module to analog voltage measurement in RSE mode. The sample rate is set to maximum, 125.000 samples per channel. Now I put a sine-waveform with 60KHz to the analog input channel. The problem is that the analog channel measures a wave which is not like a sine wave but which is like a superposition of two sine waves which have similar frequency.

 

The Nyquist criteria is complied, the used wire for measurement is an BNC-wire which is plugged in a banana-plug adaper. The generated sine wave is generated on a function generator and the sine wave is checked on a oscilloscope. I do not see other points which I might have overlooked. What causes the phenomena that a single sine wave is sampled like two sine waves which are forming a beat? Smiley Happy

 

NI9205_Sample_60KHz-Wave

 

Sine wave sampled with osci

0 Kudos
Message 1 of 6
(4,704 Views)

That apparent beat frequency looks to have a period of around 200 microsec corresponding to around 5 kHz.

 

You *could* be picking up something at 5kHz I suppose, but much more likely you're simply seeing the result of sampling too near the Nyquist minimum.  With a 60 kHz sine wave and 125 kHz sampling, each sample moves you almost a half-cycle along the real 60 kHz sine wave.   This has the effect of slowly working through different +/- pairs of similar amplitude that produce the apparent "beat" at 5 kHz.

 

You can verify with an experiment where you feed in a 1 kHz sine wave instead while continuing to sample at 125 kHz.  I predict the "beat" will disappear because it's caused by the sampling phenomenon I described rather than any real signal at 5 kHz (or any high freq real signal that aliases back to look like 5 kHz).  

 

-Kevin P

CAUTION! New LabVIEW adopters -- it's too late for me, but you *can* save yourself. The new subscription policy for LabVIEW puts NI's hand in your wallet for the rest of your working life. Are you sure you're *that* dedicated to LabVIEW? (Summary of my reasons in this post, part of a voluminous thread of mostly complaints starting here).
0 Kudos
Message 2 of 6
(4,694 Views)

My rule of thumb is to sample at least 5x of your signal that you actually want to measure.  This avoids a lot of the aliasing issues and gives you many points on your sine wave that.  So if you are trying to measure a 60kHz signal, I would recommend at least 300kHz sampling rate.

 

Nyquist is just a theory saying that you can figure out the frequency content at 2x the bandwidth.  It doesn't say anything about the sampled signal looking pretty.


GCentral
There are only two ways to tell somebody thanks: Kudos and Marked Solutions
Unofficial Forum Rules and Guidelines
"Not that we are sufficient in ourselves to claim anything as coming from us, but our sufficiency is from God" - 2 Corinthians 3:5
Message 3 of 6
(4,686 Views)

Now I generated sine waves with lower frequency. The "beat" phenomenon slightly occurs at 10kHz and becomes more distinct when the frequency rises. It seems that my NI9205 module is suitable for just lower frequency analog voltage measurement.

0 Kudos
Message 4 of 6
(4,660 Views)

Any ADC will look like that.  Just take piece of paper, draw out the sine wave in and where the samples are taken.  You will see the same effect just sketching it out.  The problem is that you are looking at the output without any hard filtering of any output frequencies above the Nyquist/2 frequency. The mixing of the frequency below the Nyquist/2 and above the Nyquist/2 creates the beating that you see.  That is why the frequency changes with delta in sampling rate. 

 

This has always hurt my head thinking about it, but somehow it all comes out correct in the end.

0 Kudos
Message 5 of 6
(4,654 Views)

@J! wrote:

I configured the NI 9205 module to analog voltage measurement in RSE mode. The sample rate is set to maximum, 125.000 samples per channel. Now I put a sine-waveform with 60KHz to the analog input channel. The problem is that the analog channel measures a wave which is not like a sine wave but which is like a superposition of two sine waves which have similar frequency.

 

The Nyquist criteria is complied, the used wire for measurement is an BNC-wire which is plugged in a banana-plug adaper. The generated sine wave is generated on a function generator and the sine wave is checked on a oscilloscope. I do not see other points which I might have overlooked. What causes the phenomena that a single sine wave is sampled like two sine waves which are forming a beat? Smiley Happy


Let's do the math. You are samplig at the max rate, which is 250kS/s aggregated sample rate. How many channels are you measuring? The frequency you are trying to measure is 60kHz. Since the signal rate is not an integer multilple of the sampling rate, the FFT will fill two adjacent bins. Two close frequencies always create a beat pattern, so, as other have already said, this is completely expected. Try to move the sampling rate to an integer multiple of the frequency you are trying to measure (e.g. 240kS/s) and things should look cleaner. Still, you are pushing the limits if you are tring to get more details on e.g. the actual shape of the signal. There might be harmonics but you won't be able to detect them.

 

You can easily reproduce this. It  is a theoretical, not an instrumental limitation. The following code creates a single sine wave similar what you are doing. As you can see, there is a beat. 😄

 

 

 

(Note that I did not calibare dt of the x axis according to the sample rate. The units are arbitrary)

 

Message 6 of 6
(4,640 Views)