Multifunction DAQ

cancel
Showing results for 
Search instead for 
Did you mean: 

unable to read accurate input values with NI-DAQ 6008


stefan57 wrote:

 what does "The peak voltages will be several times higher than the rms voltages with the exact range depending on the statistics of the noise sources."can you explain a bit more abotu this.


It's simple math.  You need to select your input voltage ranges based on the peak voltage you will see.  The RMS voltage will always be less than the Peak voltage.  For a sine wave the Vrms = 0.707*Vpeak.


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
0 Kudos
Message 11 of 15
(1,560 Views)

@GoviRe wrote:

Is there any way to make sampling better using DAQ-6008 (as it is very handy to use and I will use it only) with Terminal configuration setting as 'RSE".


As Dennis said, there are different wiring configurations.  The RSE means Reference Single Ended.  This type of measurement is good for measuring signals that are directly referenced to ground.  If you use this, make sure your AI GND is tied to the ground of your signal.

 

You might also want to give this paper a good read: Field Wiring and Noise Considerations for Analog Signals


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
0 Kudos
Message 12 of 15
(1,560 Views)

stefan57,

 

RMS is a measure of the energy in a voltage signal. As crossrulz pointed out the peak value of a sine wave is 1.414 of the RMS value. For noise things are more complicated. Because noise is random, the ratio between peak values and RMS values depends on the distribution of the random values. Gaussian distributions will be different from uniform distributions. When the noise sources are transitions in logic circuits, as is sometimes the situation in data acquisition devices, the peak to RMS ratio can be much higher. 

 

NI does not specify a peak value for that noise (partly because it is much more difficult to define the probability that a particular measurement is close to the maximum).  For noise I usually estimate that the peak values may be about 5 times the RMS values but rare instances could be higher.

 

GoviRe,

 

Using very low source impedances is the best thing you can do to eliminate the magnitude errors. About the only thing you can do about noise is take more samples and average. 

 

The images you showed with the 6016 and the USB-6008 are consistent with the resolution of the devices. The USB-6008 has 9.77 mV resolution in single ended mode and the step sizes in your image are approximately 10 mV. The 16 bit converter of the other device produces much finer resolution. 

 

Lynn

0 Kudos
Message 13 of 15
(1,548 Views)

Hi Lynn,

 

Can you explain a bit more detailed of how to take more samples and average fro removing noise.

 

Thanks.

0 Kudos
Message 14 of 15
(1,540 Views)

The basic concept is that the noise is uncorrelated with the signal. So averaging over the period of the signal reinforces the signal and reduces the noise.

 

For this to be effective two things are required:

1. You must be able to trigger the acquisition in phase with the signal or locate a reference point on the signal in post-processing.

2. The signal must have stable timing and an average of mutiple cycles of the signal must carry useful information.

 

Another option (which uses the same statistical principles, but different timing assumptions) is to oversample and then reduce the number of samples by averaging blocks. I think you are sampling at 500 Hz. Suppose that you sample at 5 kHz and take the average of every 10 samples. Those averaged values represent your 500 Hz sampled signal but now the noise on ten consecutive samples has been averaged. This should have lower noise than a single sample taken at the same rate.  This makes the assumption that the signal does not change significantly over the 2 ms interval between the 500 Hz samples so that the average of the 10 samples acquired at 5 kHz accurately represent the 500 Hz samples. If the 500 Hz sampling rate meets the  Nyquist criterion for your signal, this will be a valid assumption.

 

Lynn

0 Kudos
Message 15 of 15
(1,526 Views)