FFT analysis: SNR and Noise level vary with FFT Window size

[VirtinsTech]

Thanks for the information. This crash issue has actually been fixed internally during our MI 3.9 development. It only happens when the spectrum analyzer is open but oscilloscope is closed, and FFT size has just been changed while ADC is not running. A temporary workaround is to change FFT size with ADC running always. If you change FFT size with oscilloscope window opened, then Record Length will not change with FFT size. If Record Length < FFT Size, zero-padding will be used. If Record Length > FFT size, intra-frame average will be used. The number of averages will be displayed as "FFT segments" at the bottom left corner of the spectrum analyzer window.

[VirtinsTech]

As one can see, the Noise level varies quite a lot with the FFT Window size when level is at -2.4dBFS (pretty similar to my previous measurements)
But it's stable when you lower the level.

Another interesting thing is the variation with the number of averaged frames:

Thank you so much for these observations which are very helpful. We have thus done a detailed investigation and found that these strange behaviours occur only when the signal level (S) is roughly 165 dB or more above the "apparent" noise level. This condition will be satisfied only when FFT size is very large, the noise level of the test is extremely low, and the signal level is close to the maxium. The cause of the issue has something to do with how to maintain the required precision during numerical computation. We have improved the code in our internal version and it will be released soon in MI 3.9.

[RJA4000]

Hi

Thanks for taking the time to investigate that !
And thank you even more for finding the problem

My preferred measurement software will become even better

[RJA4000]

Fixed with version 3.9

Thanks