LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Should transducer observe" Nyquist rule" in DATA ACQUISITION?

because the expected high frequency signal, we have bought a new DAQ board with high enough frequency, but the displacement transducer we use is still the old one, To be exactly, the frequency range of teh signal is about 20-90k HZ,  But the transducer has a sensitivity frequency range of 100K hZ, just a little bit above the expected highest frequency of teh signal. as I know, Nyquist rule requires that the Data acquisition rate be at least twice as high as expected highest frequency of the signal, for practical use 10 times as high is suggested for good result.
 
I'm wondering, if this required data acquisition rate the sampling frequency of the board only, or it is the requirement for the whole data acquisition system( including the transducer)?
 
 
 
 
 
0 Kudos
Message 1 of 11
(3,856 Views)
As long as you are sure your received signals meet the Nyquist condition, then you are OK.  You want your components to have better specs than the signals you are acquiring.  Of course if they are much better, you would want to add some filtering to minimize noise.
Randall Pursley
0 Kudos
Message 2 of 11
(3,844 Views)
The answer would be a resounding YES.

 Not only the transducer, but the whole chain of components between the transducer and the final display  have to  honour the Nyquist criteria.

Raghunathan
Raghunathan
LabVIEW to Automate Hydraulic Test rigs.
0 Kudos
Message 3 of 11
(3,844 Views)

THNAKS!

I will propose a better transducer to my boss!

 

0 Kudos
Message 4 of 11
(3,837 Views)

If you've got a chain of instruments all sampling at different rates, shouldn't the sampling rate of each instrument double as you go up the hierarchical chain?  Ie- if the source signal is 10k, then shouldn't the transducer be at least 2X the source sample rate, and then whatever the device measureing the transducer be at least 2X the transducer... etc?  So really the DAC card should then be sampling at source signal X 2^(number of chain elements?) whoa, i think i just confused myself.

 

 

0 Kudos
Message 5 of 11
(3,833 Views)
No. If the transducer is producing an output in proportion to the signal it's measuring, then a 10kHz vibration will output a 10kHz signal that is measured by the DAQ board. The sampling rate of the DAQ board can be the same as the transducer.
0 Kudos
Message 6 of 11
(3,822 Views)
If you think about it, the DAQ card is a sampling device that is monitoring the output of a transducer. But the transducer is, in a sense, a sampling device too--except it's monitoring a physical process... Admittedly the anology isn't perfect, but you will see similar types of effects.
 
Mike...

Certified Professional Instructor
Certified LabVIEW Architect
LabVIEW Champion

"... after all, He's not a tame lion..."

For help with grief and grieving.
0 Kudos
Message 7 of 11
(3,821 Views)
The Nyquist theorem applies to SAMPLED data systems. The continuous or analog portions of the system need only have sufficient bandwidth to pass the signal components of interest. Most transducers (e.g., strain gauge-type pressure transducers) are continuous devices. If a system has multiple sampling points, which is unusual but not impossible, the Nyquist criterion should be applied at each samplng point. In such cases round-off and truncation errors should also be considered. The design of any measurement system should include an error analysis which determines the amount of error contributed by each component or at least determines that a given component will contribute much less error than others.

Remember that there is no universal definition of bandwidth. Many amplifiers are specified at the -3dB points, but if you want to follow the amplifier with a 16-bit A/D converter, the bandwidth for 16-bit accuracy is much smaller.

Lynn
0 Kudos
Message 8 of 11
(3,815 Views)
Just to reiterate what Lynn said to clear up confusion. Nyquist frequency ONLY applies when you acquire data in discrete fashion. In practice you usually want to sample at least 5 times the highest freq. of interest.
 
anolog electronics also have response times which effectively limit your frequency analysis. This however, depends on their architecture and material design and has nothing to do with sampling theory.
 
Sam
0 Kudos
Message 9 of 11
(3,807 Views)
Nyquist applies to sampling data. Transducer bandwidth is a separate issue. For your 100KHz xducer, 100KHz is where response to stimulus starts to drop off at the 3dB point. If signal of interest is 90KHz, then suggest taht you sample at least at 180KHz rate.
~~~~~~~~~~~~~~~~~~~~~~~~~~
"It’s the questions that drive us.”
~~~~~~~~~~~~~~~~~~~~~~~~~~
0 Kudos
Message 10 of 11
(3,803 Views)