I hope you are well.
I just came across your measurements of the Arcam SA30, and in your description of the chart labeled “THD ratio (unweighted) vs. output power at 1kHz into 4 and 8 ohms,” you commented: “Of note is that this dual ‘knee’ behavior in the plots above is unusual; however, THD values are consistently very low for any amplifier up until the second ‘knee.’”
I believe this “dual knee” behavior is just an indication of the SA30’s class-G operation. The SA30 uses two power-supply rails for the power-amp section; one lower-voltage rail for low-power outputs, where the output transistors are biased in class A, and a second, higher-voltage rail for higher power, where the transistors are biased in class AB.
If you consider this and look at your graphs, you can clearly see the class-G behavior, with the amp operating in class A up to about 27W into 8 ohms and 40W into 4 ohms. Above these levels, the amp switches to the second power-supply rail, effectively switching to class-AB biasing, hence the increased THD ratios.
And as a side note, I believe your measurements are excellent and comprehensive—an extremely good job! People are often quoting measurements on the Audio Science Review (ASR) forum, but they do not compare to yours. You even measured frequency response with a real-life load (i.e., a speaker), which is very uncommon. Sometimes, this might reveal differences in sound quality and signature [between components], which are often ignored on ASR because “they measure the same.” Yep, they do—with a purely resistive dummy load. Thank you very much for doing that!
I also thought the SA30 measured really well, considering it’s a budget all-in-one device. I’m impressed [by the results], although, as an owner of this device already, perhaps I shouldn’t be!
Thank you for your comments, insights, and kind words of praise. You’re absolutely right, and I should have taken more time to investigate the SA30’s amplifier topology when I was interpreting my measurements. This would have provided some context around the SA30’s measured THD results compared to more typical class-AB amps. Thanks to you, I have now added a few lines on class-G operation in my explanation under the “THD ratio (unweighted) vs. output power at 1kHz into 4 and 8 ohms” chart.
And I agree, the measured performance of the SA30 is exemplary, especially considering its price and exhaustive list of features, which, along with what Roger Kanno wrote in his review, led to the SA30 receiving a 2021 SoundStage! Network Product of the Year award.