01-02-2019, 06:04 PM
Bear in mind that your THD measurement is almost certainly THD+N. So it's not unusual for the apparent distortion value to get worse as you go down in level.
The way to be sure about this is to observe the post-filter signal on a 'scope (or listen to it if you can). If I'm right, then that post-filter signal will be mostly noise, rather than 2kHz and 3kHz.
With practice, it's easy to hear what sort of distortion dominates from listening to the "residual" (the post-filter signal). Also, mains hum and other problems can often be heard. We have to really encourage our students to listen to the residual - they simply don't realise how much useful information is contained in that signal.
Of course, I could be wrong about this specific case, but in general, watch out for that - it's a trap for the unwary. Always look at and listen to the residual.
I could also be wrong about your measurement - what bit of kit are you using? It might genuinely do THD (rather than THD+N) if it's based on spectrum analysis in some form. THD is still subject to noise problems, however.
Sorry I haven't commented earlier - I've been really busy lately - but it's really looking the business now! Really nice work
The way to be sure about this is to observe the post-filter signal on a 'scope (or listen to it if you can). If I'm right, then that post-filter signal will be mostly noise, rather than 2kHz and 3kHz.
With practice, it's easy to hear what sort of distortion dominates from listening to the "residual" (the post-filter signal). Also, mains hum and other problems can often be heard. We have to really encourage our students to listen to the residual - they simply don't realise how much useful information is contained in that signal.
Of course, I could be wrong about this specific case, but in general, watch out for that - it's a trap for the unwary. Always look at and listen to the residual.
I could also be wrong about your measurement - what bit of kit are you using? It might genuinely do THD (rather than THD+N) if it's based on spectrum analysis in some form. THD is still subject to noise problems, however.
Sorry I haven't commented earlier - I've been really busy lately - but it's really looking the business now! Really nice work
