Dear John,
JC wrote:the mechanisms you have seen in the video can react completely differently with the tiniest adjustment
It is exactly here the most important point.
The mechanisms where the efficiency is dependent from a 1/10th of millimeter within one part, must be rejected because
it is the sign of an evident lack of torque
Only the designs with a consistent torque have a little opportunity to be successful, at the minimum for counteracting the friction losses.
In addition there is no reason (if the start has been laborious) why the mechanism could be more efficient few seconds later.
I see only one case of this slow start could be not disconvenient: the clock. On a certain point of view, your design is similar with some clock escapement.
I still continue to believe in the 'parametric pendulum' solution (one of several possibilities).
But I prefer for the moment to explore further some other more universal solutions based on the eccentricity.
If it can be useful, my suggestion for implementing a parametric pendulum is given in the drawing hereafter.
The fan of Bessler will recognize the famous 'A with legs' frame, wich is IMHO an important and more recognized clue than the pentagram, and can be eventually combined with the famous scissor.
In any case, many thanks again for sharing your ideas.
I'm not surprised by some answers from few systematical negationists: even with an legal officially stamped attestation they will continue to claim for a fraud.
These persons are the same unable to share anything neither to publish any fruitful design nor to build any test unit.
So far don't care so much of such as negative comments.
I cannot imagine why nobody though on this before, including myself? It is so simple!...