Sound Level test with Swept sine is error!! help

[chipie]

(I am using PXI-4461)

I use the example program "SVXMPL_Swept Sine FRF(DAQmx).vi" to test the frequency response of my DUT. Some frequencies whose level should lower than -40dBV,  tested with "SVXMPL_Swept Sine FRF(DAQmx).vi", their sound level is higher than -40dBV.

If use " SVXMPL_One Shot Gain and Distortion(DAQmx).vi" to test the frequency response one by one, the sound level is right.

I guess there are some error in the Swept sine toolkit or my setting is error. I have try my best to use the same setting when using the two VIs. Following is the setting comparision:

SVXMPL_Swept Sine FRF(DAQmx).vi                                  SVXMPL_One Shot Gain and Distortion(DAQmx).vi        
amplitude input:   51.2mV                                                       amplitude input: 51.2mV
propagation time [samples]: 108                                            propagation time [samples]: 108
sampling frequency [Hz]:50000HZ                                         sampling frequency [Hz]:50000HZ (set with "svx_override_DAQmx

                                                                                                                                                         Parameters.vi")
block duration [ms]: 500ms                                                    block size: 25000 (set with "svx_override_DAQmx Parameters.vi")
settling time:100ms                                                                settling time:100ms
settle cycles:10                                                                       settle cycles:10
integration time [s]:100ms                                                      integration time [s]:100ms
integration cycles:10                                                               integration cycles:10
sweep frequencies [Hz]:                                                         frequency [Hz]:100HZ
start freq=100HZ
stop freq=100HZ
number of steps=1

The output sound level of the two VI is:
SVXMPL_Swept Sine FRF(DAQmx).vi                                   SVXMPL_One Shot Gain and Distortion(DAQmx).vi
-38dBV                                                                                     -41.79dBV  

I found the time Record of "SVXMPL_Swept Sine FRF(DAQmx).vi" is abnormal, pls find the attached picture "swept.png", there are abnormal higher signal at the start and stop terminal. which caused the sound level lower than -40dBV can't be test rightly. However, the time record picture one-shot.png of "SVXMPL_One Shot Gain and Distortion(DAQmx).vi" is normal.  

Can't anyone help me?! Why one shot test and swept sine test at one frequency are different?

[dsb@NI]

Were you using IEPE to acquire your response signal from a microphone? There is a known issue with Sound and Vibration Toolkit 4.0 with DAQmx 7.5. The issue was seen when IEPE was enabled and the settling time for the IEPE was set to anything greater than zero. As a workaround to this issue, you can specify an IEPE settling time of zero. This issue is fixed in the most recent versions of the Sound and Vibration Toolkit and DAQmx. You should also pay special attention to units.

It is not clear how you converted the swept sine result or the gain result to dBV. If you are using the swept sine response rms level, the dB reference is in rms units (rms measurement divided by dB reference in rms units to convert to dB). The gain measurement uses the specified amplitude as the dB reference (detected peak amplitude divided by source amplitude in peak units to convert to dB).

If your measurement only requires one test frequency, it is recommended that you use the One Shot measurement as it is optimized for a single test frequency whereas the swept sine VIs are optimized to step through a set of frequencies. I would expect to see some differences (such as the total number of cycles of sinusoid generated) in the acquired signals but not the transient behavior shown in your screenshots.

Hopefully, one of these comments is helpful. If not, please indicate what versions of SVT, LabVIEW, and DAQmx you are using and attach a copy of your VI to this thread. Remember to save your configuration as the defaults.

[chipie]

Thank you for your attention!!!  This problem puzzles me for a long time. I had contacted NI of China by phone and  they supposed it was caused by my DUT. My DUT is a Amplifier with Audio processing by DSP.  I received that using one shot measurement one frequency by one frequency to replace swept sine measurement. It is so slow comparing to swept sine because each frequency point  I have to "Initialize", "setting", "start" and "stop", etc...   But I have to say that the result of measurement with one frequency by one frequency is right. 

By contraries, measurement with swept sine is error!! I had test the Amplifier with swept sine by "AUDIO PRECISION" , the result is ok.

Get to business. 

1. My DUT is a amplifier and the setting of IEPE is disable.

2. In the one shot measurement,I firstly get the data values of a waveform, then change the value to rms with "RMS.vi", lastly change the unit to dBV.  In swept sine, I directly read the rms value on the tap page "rms level".

3. However, I still want to use the swept sine to test frequency response for the reason of  test speed.

The versions of my software are as follows:(Readed in Measurement & Automation Explorer)

Labview 7.1 

NI-DAQmx   8.0.0f0

Sound and Vibration toolset   4.0.49153

Attached file is my using VI.

[dsb@NI]

I ran the examples that you posted connecting one output channel of the 4461 directly to both input channels. I did not find transient behavior at the edges of the acquisition period (as shown in your posted screenshot) but did find some unexpected DC offset and low-amplitude, high-frequency components in the acquired signals. Changing the terminal configurations to Differential for the AO channel and to Pseudodifferential for the AI channels improved the signal quality and the quality of the measurements. With the change to the terminal configuration implemented, the results returned by the swept sine and one shot gain measurements were equivalent (difference < 0.02 dB). Please let us know what you find.

[chipie]

The test results of swept sine and one shot are all right  when output channel of 4461 directly to input channel.   For the signal is connect directly, the output level is the same as the input level.

But when swept sine testing my amplifier, if the output sound level of the amplifer is low, the result is error. As per one shot test, the result is always right no matter the output sound level of my amplifer.

[dsb@NI]

Is this a different result than what you first posted? or is this more information on the same failure mode? If the acquired signal looks different than that first posted, please attach another screenshot or email me at doug.bendele@ni.com.

Did you try changing the terminal configurations as suggested? Did you see improved measurements? different failure mode or signals?

What is the input impedance of the amplifier - could it be overloading the output of the 4461 which specifies a minimum working load of 600 ohm?

[chipie]

It is the more information on the same failure mode. The screenshot of the first post is the test result of my Amplifier, and when the 4461 test itself(AO0 connect directly to AI0 and AI1), the result of swept sine and one shot is near.  So I  guess  maybe the failure was caused by test parameter setting, I am not sure.

Pls find the attached pictures of 4461 output directly to 4461 input. The input level is -60dBV. There is DC offset in the AI1, but not in AI0. and I didn't find high-frequency components that  you said.  I also changed the terminal configurations to Differential for AO channel and Pseudodifferential for the AI channels as you said, the result is similar.

Note: When testing my Amplifer, I can't change the AI0 setting to Pseudodifferential, because Amplifer is balanced output. Testing my Amplifer with pseudodifferential will damage the amplifer.

[chipie]

The input impedance of my Amplifier is 1M ohm.   Because the result of one shot test is right, so I think it may not cause by the input impedance of my amplifier.

[dsb@NI]

Thanks for clarifying the amplifier input impedence and the need to set the response channel terminal configuration to differential.

We need to verify that the one shot measurement is indeed returning correct results. In the One Shot* example, connect the 'all acquired data' output from the svx_Write Excitations_Read Responses (DAQmx).vi. Verify that you do not see the transient behavior in the response signal at the beginning or end of the excitation window.

On my end, the low-amplitude, high-frequency components and the DC offset were actually caused by a bad cable. Replacing the cable fixed both problems. The bad cable was not as obvious as I would have guessed (the measurements were not all bad and intermittently I could reproduce a bad measurement or a good measurement.

The DC offset shown in your screenshot is small (~0.5 mV). Because the signal level is so small, any DC offset will affect the measured results. Enable AC coupling on the stimulus and response channels by using SVT Set Swept Sine Coupling and IEPE Excitation (DAQmx).vi from the Swept Sine>>Swept Sine Configuration>>Configure Advanced Settings subpalette.