What is the proper way to analyze a sample sinewave to get the amplitude in 'Linear Vpk' for frequencies present?

[JackMetcalfe]

The frequencies present returned from AutoPowerSpectrum are fine but their amplitude is way too high.

The logic is based on that provided with the LabWindows/CVI spectrum analyzer sample application. Unfortunately it displays the exact same problem.

The Spectrum Analyzer LabView vi provides the results that I was expecting. To me this confirms that the signal is ok.

Sample Code that produces significantly too high amplitude values as output from AutoPowerSpectrum.

I have screen shots of the results being plotted as well as an Excel spreadsheet of values I thought were relevant. If interested please provide me with an e-mail add
ress to send them to.

Real64* pVoltBuffer = ( Real64* )
::GlobalLock( pScanInfo->Sample->m_hSample );

CNiReal64Vector oVoltageReadings(
pScanInfo->Sample->m_lCount, pVoltBuffer );

CNiReal64Vector oSpectrum;
CNiMath::WindowConstants oWindowConstants;
double dNIDAQFreq;

// Apply window
CNiMath::ScaledWindow(
oVoltageReadings,
oWindowConstants,
CNiMath::FlatTop
);

// calculate the power spectrum for a
// time domain signal
CNiMath::AutoPowerSpectrum(
oVoltageReadings,
oSpectrum,
dNIDAQFreq,
(Real64)
1 / pScanInfo->Sample->m_dScanRate
);

[JackMetcalfe]

This gave me what I desired but now I need help understanding in detail why it solved it.

CNiReal64Vector oConvertedSpectrum( Spectrum.GetSize() );
CString oUnitsString = "V";
CNiMath::SpectrumUnitConversion(
oSpectrum, oWindowConstants,
oConvertedSpectrum, oUnitsString,
CNiMath::Power, CNiMath::LinearScaling,
CNiMath::VoltsPeak, dNIDAQFreq
);