This has got so far from the original that we may as well carry on. The others will be well advised to ignore this lot.
You are right that sometimes it does pay to have a look at things in a new light. Many issues here are not exactly straightforward and little of them can stand examination in detail without issues arising that are different from common practice.
Now I have got to grips with what you are saying I see where you are coming from. You have obviously been involved with power supplies with capacitive input filters and this is also true with a battery. I follow your ripple figures now. I spent my life on motors and rectifier loads involving inductance and resistance. A motor can behave as a capacitor in some respects in that it takes current when the supply exceeds the emf but the inductive bit is never absent.
This accounts for our differences in the value of ripple. That is probably why you don't recognise my figures for mean dc and ripple as they are based on rectifiers with included inductance.
Now we come to the issue about cut in into a battery load and how this ties in with things we normally measure and it certainly gets confusing. When we measure volts from a coil and predict the dc voltage we are predicting what we would measure with an average reading dc voltmeter ( which normal moving coil and digital multimeters are)
This may be where the confusion started as I believe you use a scope in preference to meters. You would instinctively measure peak and the mean dc would be tricky to measure.
The average is 1.35 x vrms. We start conduction at 1.414 v rms and I agree with you that full conduction does not occur into a battery until we reach the crossover point on the waveform. Throw in a few unknown diode drops, non sinusoidal waveforms and a range of battery volts from well discharged to float and we have something that takes some sorting out.
As few people have a scope and the thing can't have any degree of accuracy because of the many variables then I think we can stick to accepted method but I am inclined to revise the factor to 1.35 from 1.4.
I hope we can agree on this at least for air gap alternators with no significant inductance, I am not entirely sure that this will apply to all motor conversions, some of which exhibit a fair bit of leakage inductance.
Flux