When actual distortion and noise are extremely low, the bandwidth used to calculate the fundamental's power becomes critical in THD+N, SINAD and SNR measurements. Multi-Instrument uses the width of the main lobe of a window function to calculate the power of the fundamental and harmonics.
In your case, the external signal generator does not share the ADC's sampling clock, which results in a wider skirt around the measured fundamental peak. This skirt can be viewed as leakage of the fundamental frequency or as phase noise caused by jitter. If you prefer the former interpretation, you can use a window function with a wider main lobe, and vice versa. In Multi-Instrument, Kaiser 20 has the widest main lobe in Kaiser window series. Other window functions that have a very wide main lobe include Cosine Sum 261 and Dolph Chebyshev 250. It is also possible to use the bandwidth used by AES17-2015 notch filter, which is probably too wide, please refer to
viewtopic.php?f=6&t=2325&hilit=AES for details.
Posted: Sat Oct 19, 2024 4:05 am