good soundcards are ultimately limited by adc distortion and interference artifacts, the noise floor can be pushed very low with long records/averaging
to some degree interference from line power, sw supplies, ect can be "dodged" by choosing test frequency
but even the better monolithic adc chips have harmonic distortion in the 105 +/- dB range, so a large amplitude fundamental limits your harmonic measurement capability, also the soundcard input signal conditioning circuitry could contribute distortion at these very low levels with large inputs
presumably the adc/soundcard input distortion components are dependent on input amplitude to some power, so even a 20 dB null should be adequate to reduce the adc/soundcard harmonic distortion to negligible levels (or null + gain such that input < -20 dB at the adc )
of course the notch/highpass filter itself has to be very good, polystyrene caps are a really good idea, over size wattage and low tcr resistors may be helpful for the attenuator at power amp output voltages
to some degree interference from line power, sw supplies, ect can be "dodged" by choosing test frequency
but even the better monolithic adc chips have harmonic distortion in the 105 +/- dB range, so a large amplitude fundamental limits your harmonic measurement capability, also the soundcard input signal conditioning circuitry could contribute distortion at these very low levels with large inputs
presumably the adc/soundcard input distortion components are dependent on input amplitude to some power, so even a 20 dB null should be adequate to reduce the adc/soundcard harmonic distortion to negligible levels (or null + gain such that input < -20 dB at the adc )
of course the notch/highpass filter itself has to be very good, polystyrene caps are a really good idea, over size wattage and low tcr resistors may be helpful for the attenuator at power amp output voltages