A New EC-Composite LM3886 Amp

This is why as a rule-of-thumb the measurement equipment must be an order of magnitude better than the DUT, otherwise you can't separate the measurement of the DUT from that of the measurement equipment.
Ideally, yes. In the real world performance of DUT and measuement rig can be on quite similar and very low levels, and then all you can do is to double- and triple-check with slightly changed parameters to see if a suspisciously low looking harmonic is real and not just a lucky cancelling.
 
Not correct. When you have odd harmonics from, say, a symmetic expanding transfer function then a symmetric but inversly compressing transfer function can cancel it. This no different than with even order harmonics.
Think twice before you say that and don't forget that the even and odd harmonics represent different types of non-linearity of the transfer function, and not any linearity can be brought to linearity just applying simple phase inversion and summing.

A simple example from the real world. If you're right then a very well balanced bridge amp would have infinitely low THD. But unfortunately it is not so. In this case the even harmonics can be significantly reduced, but not odd harmonics. If you have experience in THD measurements most probably you've seen that.
 
6xLM3886 PB Composite Amplifier.
 

Attachments

  • Fig189 Min 6xLM3886 PB.jpg
    Fig189 Min 6xLM3886 PB.jpg
    132.1 KB · Views: 150
  • Like
Reactions: userguide