magnitude response from swept sine

[titpoul]

Hello,

I want to measure, with a PCI 6225, the magnitude response of the vibrations applied to my device and measured with accelerometers (which have an offset of 2.5V). I have tried two methods. One uses SignalExpress with in particular the boxes "Subset and Resample (Time Domain)" and "Tone Measurements" inside a sweep on the frequency, which is adapted from the example SVA_Amplitude Swept THD (DAQmx). The other method uses the LabVIEW example "Getting started with with Swept Sine (DAQmx)".

I tried to configure them in the same way (subset start=settle time, subset length=integration time) but I obtain different magnitude response, as you can see in the pictures attached. However, both methods give the same magnitude response for a simulated 2nd-order bandpass filter.

I think that the result from SignalExpress is closer to my expectations, but I am not sure and I would like to understand why the results are different.

Best regards,

Guillaume

[Frank-L]

I believe that what you are currently doing is different measurement types, which is why we are getting the different results. If you are indeed trying to do a Swept Sine, the best method would be by using Labview as currently there is no direct way to do a Swept Sine measurement in SignalExpress. Also, did you mean that they give the same frequency response? Also, where did you find the SVA_Amplitude Swept THD (DAQmx) example that you were referencing?  

[titpoul]

Hello,

I found SVA_Amplitude Swept THD (DAQmx AO-AI).seproj in C:\Users\Public\Documents\National Instruments\Sound and Vibration Assistant\Examples. I changed in this example the sweepable parameter "Amplitude" of  "Create Analog Signal" to "Frequency (Hz)" of the same step. In this way, I think to obtain a discrete swept sine measurement.

To try a simulated filter in the LabVIEW example aforementioned, I modified sv_Extract Swept Data to Process.vi to replace the response waveform from DAQmx with the output of an Express VI filter taking the stimulus waveform as input. The magnitude response was similar to that obtained with SignalExpress (where I replaced, in a similar way, the DAQmx Acquire step with a Filter step).

Also in the LabVIEW example, I substracted the DC offset of the response measurement, but it did not affect the result.

I agree that signal processing in the two methods seems to be different:

- in the LabVIEW example, in sv_Swept Sine Processor.vi, the stimulus and the response measurements are windowed with a sine, then RMS levelled and multiplied by a complex sine, lastly, in sv_Compute Swept FRF.vi, the magnitude computed is the real part of the division of the response by the stimulus;

- in SignalExpress, the Express VI "tone measurements" computes the gain from the FFT spectrum of the response in svc_Extract Single Tone Information (complex).vi which is password protected...

The magnitude response from the LabVIEW example seems to be the mirror of that from SignalExpress, plus a decreasing slope added...

[Frank-L]

The "SVA_Amplitude Swept THD (DAQmx AO-AI)" is actually doing a swept sine and extracting the Total Harmonic Distortion from the signal whereas the Labview example is just executing a swept sine measurement of the signal. The choice of which you should use is dependent upon whether you would like to find the Total Harmonic Distortion.  Getting Started with Swept Sine (DAQmx) has some additional information on what a Swept Sine Measurement is actually doing, whereas Total Harmonic Distortion (THD) (Sound and Vibration Measurement Suite) and Harmonic Distortion Measurement provides some additional information on what the THD is doing, as well how it is done in labview using the sound and vibration measurement suite.

[titpoul]

---

Frank L a écrit :

The "SVA_Amplitude Swept THD (DAQmx AO-AI)" is actually doing a swept sine and extracting the Total Harmonic Distortion

---

Thanks for the links to the descriptions.

However I am not interested in distortion measurement, but in gain evaluation. In the Tone Measurements step, I unselected distortion extraction and only selected gain evaluation. My initial question was why the gain evaluated by Tone Measurements in a frequency sweep is different from the magnitude response given by the Swept Sine example of LabVIEW.

[Frank-L]

Sorry about that, I was just making sure that we were not working with different measurement methods. Since we are no longer doing the Total Harmonic Distortion measurement, are we sure that all the setting for FFT type, ranges, and other configuration settings are identical to those set in Labview? If they are all taking measurements from the same source signal, the differences in output have to be either configuration or the different steps that are being used to get your signal. The different software (LabVIEW or SignalExpress) should not make a difference in the analysis if it is indeed doing an identical analysis calculation. 

[M.GH]

Hello,

I am quite new in LV and going to start to do some measurements. First, I need to stimlate my vibrator with Swept-sine signal (modulated) and later measure the sound pressure to get the Frequency Response Function. I am using NI USB-6218 (BNC).

I have somehow generated the modulated signal but I cannot send it to Daq. Maybe, it is so easy but I don't really know! In the forum mostly they use Daq assistant!

Another question; Should I think of buying Sound and vibration toolkits ot this can be done without! Ofcourse, I have some feeling about coming problems like sampling limitation and synchronization of inputs.

I appreciate your time and consideration in advance!

Best regards,

Ghoncheh

[Cavarval]

The DAQ assistant express VI is the easiest way, but I would recommend you to use the lower level functions if you have additional time to spend. If you need a quick deployment, use the assistant.

If you use LabVIEW 2013, you can experiment with the Voltage Continuous Output.vi / Voltage Continuous Input.vi (depending on if your signal will be generated or received by your 6218 correspondingly) located on Help > Find Examples > Browse (by task) > Hardware Input & Output > DAQmx > Analog Output / Analog Input. If you are using an older LabVIEW version, there are equivalent examples.

You mostly don't "need" the toolkit. You can do almost everything the toolkits do by creating your own applications, but this will probably require a lot of time. The toolkits idea is to provide the APIs you will need for certain tasks, thus saving you that time, as well as providing advanced analysis options which would be quite hard to develp from scratch (then again, time).

You can check the following link for more information about the functions you can get with the toolkit:

http://www.ni.com/white-paper/7415/en/

However, as this is a very old thread with almost three years of inactivity, I strongly recommend you to post your questions on a new thread. There is a Sound and Vibrations specialized forum on www.ni.com/forums, under Special Interest Boards > Reference Design Discussions > Sound & Vibration Applications, or you can post on the LabVIEW forum here:

http://forums.ni.com/t5/LabVIEW/bd-p/170

[M.GH]

Thank you! I will try to find more info there!

Best regards,

Ghoncheh