I am not sure why the efficiency degrades at the halfway mark. Do you have any ideas?
Be aware that this is a particular way to measure the efficiency, it's a relative and not an absolute efficiency.
In absolute terms, there is the best efficiency when the top spins at the lowest speed, and you can see this looking at the RPM drop, for example. You had the best efficiency, 94.99%, at minute 17; but in that minute the top lost 42 RPM, while at the end of the spin the top lost less than 20 RPM per minute, so, in absolute terms, the top is more efficient at the end of the spin.
Even considering the energy consumption, the lowest consumption is at the end of the spin.
This is normal, because both the air drag and the tip friction increase with speed, so tops become less and less efficient at the increasing of speed.
So, this is a relative efficiency, which does not tell the "real" efficiency of the top.
It is not very meaningful to compare differences of this "relative efficiency" along the same curve.
Anyway this relative efficiency data can be useful for comparing different tops;
one way is to plot in a graph, relative efficiency vs. RPM, (to compare differences of efficiency at parity of RPM).
To compare differences of this "relative efficiency" in different tops, makes sense, because a top which loses speed more rapidly than another one, at parity of RPM, is certainly less efficient in real terms.
The graph makes available various informations at a glance, which may be interesting.
For example:
This is about the "efficiency" of my tops Nr. 20, 23 and 33.
The starts and the ends of the curves tell the starting speed and toppling down speed;
The Nr. 20 can be started at the highest speed thanks to a quite long stem, while the Nr. 33 has a short stem and it is difficult to spin it hard. Nr. 20 and 23 topple down at very low speed because of their deeply recessed tips.
The efficiency at high speed, is especially related to air drag, (because air drag is generally very high at high speed), and moment of inertia. The best efficiency at high speed is obtained with the most compact, littlest shapes, having the highest moment of inertia.
Little shapes with high moment of inertia means to use dense materials, like metals.
So, the parts of the curves at the left, tell above all the ratio between the moment of inertia and the aerodynamic efficiency. Nr. 20 and 23 have flywheels with same shape and dimensions, but the Nr. 23 has three holes for the balancing screws in the flywheel, and for this reason its aerodynamics is not so good like that of the Nr. 20, which has not these holes; the curve of the Nr. 23 starts lower in the graph than that of the Nr. 20.
The Nr. 33 is a light top with large diameter, so it has poor aerodynamic efficiency, and poor moment of inertia, and in fact the curve of this top at the left is very low in the graph.
At very low speed, air drag nearly disappears. So, the efficiency at very low speed is especially related to the tip friction, and to the moment of inertia.
Tip friction is related to the shape and quality of the materials of both the contact points, the lubricant, and the balance of the top.
Best efficiency here is obtained with the largest radius of gyration, so to have the highest moment of inertia with the littlest weight.
At parity of the other factors, light tops are more efficient, because tip friction is not proportional to weight, but increases a bit more rapidly than weight.
The parts of the curves at the far right tell above all the ratio between the moment of inertia and the tip friction. The Nr. 23 has a slightly better tip friction than the Nr. 20, because of the better quality of the tip material, (carbide for the Nr. 23, HSS for the Nr. 20).
The Nr. 33 has very good efficiency at very low speed, because it is light, and it has a large radius of gyration.