If you want to make numerical comparison, you cant avoid defining in what space do you modell the network. Time? Frequency? Complex frequency? In time-domain and with generic waveform any relation is possible in specific moments. No way you can simply compare back EMF and applied voltage, ratio varies from -infinity to +infinity with time even with a simple sinus excitation.
In frequency domain it's advisable to specify what the assumed excitation signal is like (typically sinus) and how signals are represented (axsin+bxcos, or amplitude and phase, etc...).
Your first statement is true if you are talking about single frequency and effective values. But in this case there is no exception, back EMF will always be lower than excitation voltage, unless you make the speaker move from outside. But back EMF can be close in amplitude to the applied voltage and can have a significant phase-shift, so if you look at it in time domain, it can be reversed for a certain duration twice in every period. This is why a ClassAB amp must have a short circuit limit current much higher than 0, tipically 25...50% of max. current.
Summary: assumption of nonlinearity is not necessary to explain the exception in time domain, and not enough in case of frequency domain, unless you are talking about generation of harmonics, but this is simply the well known distortion, i dont think really related to this topic.
My comments are informed by real world experiences and observations, and confirmed by theory. This stuff actually happens in the real world and it can be duplicated on the bench whenever one decides to do the work.
For example this statement appears to be very general:
"But in this case there is no exception, back EMF will always be lower than excitation voltage, unless you make the speaker move from outside."
In the case of real world speaker drivers mounted inside real world enclosures, there is the slight matter of the air inside the enclosure.
Over a longer period of time the law of conservation of energy has to apply, but from instant to instant, not so much. Energy can be stored and released inside the system. The air can be a force that quite clearly makes the speaker cone move from a source that is outside of the speaker driver.
Therefore, the statement is true, but in the real world it does not explain away the role of nonlinearity in speaker driver motor assemblies.
And that my friend is why theory needs to be informed by practice. Theory may be 100% true as far as it goes, but there may be other stuff that is going on in the UUT that is before you.
Last edited: