The Amber 5500 residuals, as seen on the scope. Hor: 10uS/div Lower trace Vert: same arbitrary units as here. Note the missing 3rd harmonic. Amber 5500 THD digital display was fluctuating 6-7-8 ppm.

The Amber 5500 10KHz spectra. 0dB=1Vrms Note the 3rd armonic well below (about 20dB) the 2nd and the noise floor (non differential) of our measurement setup.

The Amber 5500 20KHz spectra. 0dB=1Vrms Note the 3rd armonic well below (about 30dB) the 2nd, at par with the 4th and the noise floor (non differential) of our measurement setup. The reason why the 3rd harmonic is even lower than at 10KHz is is the Amber 5500 construction. 10KHz is a cusp point in the hardware; over 10KHz the state variable oscillator automatically changes the configuration to adapt to a new frequency range (10KHz-50KHz). The 3rd harmonic component (and the THD) increase monotonically over each frequency range.