All things being equal, how can a unidirectional winding be more efficient than bidirectional?

[drunken monkey]

Hmm, it could be something to do with the difference in torque required to work a gear system in one direction compared to the torque required to wind a gear system that has attached to it, a reversing wheel. My brief reading on bi-directional winding suggests to me that the gears are constantly meshed (i.e the reversing wheel isn't suddenly engaged only when the rotor goes the other way) so there is simply more mass (and therefore more inertia to overcome) to turn in the first place.

Of course, this doesn't give indication on just how much better one is compared to the other.
0.001% is better.
10% is better.
10000% is better.

It would also explain why Seiko's Magic Lever is often cited as a very efficient bi-directional winding mechanism as it doesn't use a reversing wheel.

[Torrid]

Yeah, I more just meant that, in a thread about whether unidirectional or bidirectional winding is more energy efficient, offering as an answer the fact that one is cheaper than the other is not helpful. Especially given that the OP took great pains to specify that everything else is held constant for the purposes of this discussion.

Anyway, synthesizing what factoroftwo and dnslater said, I think this is a pretty plausible argument (in my non-sciency mind): On a unidirectional movement, the rotor gets spinning really fast in one direction. Then, when your wrist twists such that forces push it in the opposite direction, all that energy comes to bear on the mainspring, giving it a really good crank. On a bidirectional rotor, since it's harder for the rotor to get going fast in either direction, it never gets those really hefty cranks.

Scientifically sound or shockingly stupid?

I think it is negligible either way. I've had both uni-directional and bi-directional and never had a watch that didn't stay fully wound or had a movement with the accuracy affected by the type of winding mechanism. I wouldn't be able to state which type would be better nor do I see any possible real benefit of one over the other.

[CitizenM]

I can't imagine that expense is currently a factor, due to the existence of Seiko 5s for instance, and the fact that JLC has apparently decided that the majority of their new calibres are and will be unidirectional.

The reversing mechanism argument seems interesting. So perhaps the mechanism involved allows for superior efficiency in one direction and the other allows for inferior efficiency but in two directions. Maybe I'll call up JLC tomorrow.

[lvt]

Very tricky question, it's like asking how could the Hayabusa (248mph) be faster than a Porsche GT3 (193mph) while it only has 2 wheels (only 1 drive wheel) and the Porsche has four wheels (and all of them are drive wheels).

[drunken monkey]

Very tricky question, it's like asking how could the Hayabusa (248mph) be faster than a Porsche GT3 (193mph) while it only has 2 wheels (only 1 drive wheel) and the Porsche has four wheels (and all of them are drive wheels).

actually, car vs bike is easy to explain because we know the factors that affect them.
power, torque, weight/mass, cd, etc.

by the way, GT3 isn't 4WD.

[v76]

The analysis won't be as straightforward as some have suggested already. You need to take into account the orientation of the wrist, analyse impulse and quasi-steady forces separately (taking into account inertia, drag and frictional forces), in-plane and out-of-plane forces and associated moments as well (and the additional influence of acceleration due to gravity). Since this would change day-to-day, the results may also vary ... on some days one kind of winding might prove to be better than the other. The best way to do this would be to set up control experiments for watches with both styles of winding - a statistically valid sample which is monitored over a number of days with the necessary instrumentation to measure forces/work done (accelerometers and the like).

And again, simulated lab conditions might differ from individual wear patterns/orientations.

Edit: Change in orientation (number of times, extent) and the associated impulse or quasi-steady forces are also important to consider.

[Sean779]

I read the article "Automatic winding: one direction or two" the other day and glanced at it again recently. It says that many companies, including major movement authorities like JLC, say that their unidirectional winding systems are more efficient in the real world.

I think perhaps that when off the wrist and you wave the watch back and forth to get it going that bidirectional is more efficient but when worn on the wrist unidirectional is more efficient due to the biased movement of the wrist.

[CitizenM]

Even assuming a movement bias in one direction, as long as you move at least one time ever in the other (sufficient to wind the spring, of course), it will be less efficient overall. There's got to be more to it. Somehow, there's an efficiency cost in the bidirectional winding mechanism that arguably isn't being overcome due to a movement bias.

[v76]

I have also heard about a couple of earlier studies that concluded that bi-directional winding is more efficient (Certina and Lemania), and one that concluded that uni-directional winding was (AS). The JLC study might be one of the latest, but it definitely won't be the last!

Edit: Also, in the thread linked to ... it looks like for limited arm motion, bi-directional winding seems more efficient while for more vigorous arm movements, uni-directional might be better.

Even assuming a movement bias in one direction, as long as you move at least one time ever in the other (sufficient to wind the spring, of course), it will be less efficient overall. There's got to be more to it. Somehow, there's an efficiency cost in the bidirectional winding mechanism that arguably isn't being overcome due to a movement bias.

[v76]

Also, this thread is interesting in that respect ...
IWC Schaffhausen | Fine Timepieces From Switzerland | Forum | Bidirectional vs. Unidirectional winding