Some notes to THD and bandwidth limits.
THD is often referred as to be an appropriate number that tells something about the amount of signal saturation/clipping.
Here you are again with a plot that I already posted above
output power 300 Watts
At the top edge of the signal it can be obtained that the signal is already at the edge of clipping, the signal's rounded shape at the top also tells that there will be already some amount of harmonic distortion produced.
An FFT analysis and (and doing the THD math by hand) predicts the THD number to be equal to ~5.6% for the given signal and its 300 Watts of power.
Besides the slightly distorted top edge shape that already does generate harmonic distortion there are other artifacts as well that does contribute to the THD.
Once you look carefully from the signals top to its bottom end you may observe signal distortion that is even present in the middle area of the total voltage swing.
These obtainable artifacts are not uncommon for "high power" amplifiers. Although a strong negative global feedback may help to keep those artifacts small, jet they do happen (almost always, and more or less) with high power tube amplifiers.
Another option was to just reduce output power, or increase bias current, just to reduce those artifacts.
At the other hand side, although those artifacts very likely (and almost always) are probably unwanted with HiFi amplifiers, a musical instrument amplifier may even benefit from this artifact cause it may help to add some musical sounding texture, a more rich sounding timbre at higher levels respectively at overload range. May be similar like it was for a loudspeaker that was designed to add "musical harmonics" and an enriched timbre/texture at overload range.
IMO/IME the meaning as a summary is, for musical instrument amplifiers its not valid to draw any conclusions about the amount of signal saturation/(clipping) from provided/measured THD numbers.
Not every amplifier is equal, some may produce more THD in this regard while other ones may do less.
At least for the LT-spice 6x6550 amplifier it was not a big deal (nothing at all) to reduce THD down to about ~2..3% at ~300 Watts rather than ~5.6% THD at just the same output power.
40Hz output power ~150 Watts (lower bandwidth limit -3dB)
THD ~8%
In this plot there is additional artifacts noticeable, and these are caused by the primary inductance.
For the OT Spice model I have chosen 10H for the primary. It was possible to enlarge lower bandwidth, yet I leave it as is.
I have to mention that this artifact that does increase THD is caused by pure inductance only. No saturation of transformer iron materials involved in this model.
Yet the artifact in the modelling looks already very close to the artifacts that can be observed with real amplifiers, respectively with real OT in real practice.
At some content this artifact is already present even at 100Hz, and likewise observable with the SVT in real practice at 100Hz.
My guessing is about ~10% of THD at lower bandwidth once there was additionally saturation of iron materials in the spice modeling involved.
Well, for future use of this spice model I might implement an OT spice model that also gives respect to iron saturation in regards of "low bandwidth limit" THD.
Although this artifact may (almost always) be not wanted with HiFi amplifiers, at some content it may contribute some musical sounding benefit to MI amplifiers.
I think, whenever there is any attempt to evaluate an MI amplifier with principles that may better and preferably apply to HiFi amplifiers, then (IMO) there is a good bet that something turns wrong when drawing conclusions.
