<PRE>

</PRE>

What about something like Shimano's Silent Clutch design? No pawls, no ratchet and supposedly instant engagement. Or is the design flawed?

EDIT, found this:
"Shimano silent clutch freehubs will function at 60 below with no special treatment. Instead of pawls, they drive through about two dozen eighth inch cylinders that roll into pockets and jam to drive, backing out a little to coast.

They’re called RO-80 and cost and look about the same as LX. They engage with less pedal motion at warm temperatures than pawl freehubs and take a little longer to engage at lower temperatures. While they always work at low temperatures, winterization brings them back to near-instant engagement no matter how cold it gets. They’re pretty much cold proof since they depend on the viscosity of the grease to move the rollers into engagement. They’re about 120 grams heavier than LX."

So it is currently heavier, but a redesign can probably be lighter.

True Precision uses a similar system in their Stealth hubs

Ah-hah! Don't have an answer for that one, do you?

Ok, just kidding. On the silent clutch thing, there have been a number of attempts at this in the past from Winner, Machine-Tech, Shimano and currently Gravity Zero. They all seemed to have a weight penalty because of the nature of the transmission of force from the casette body to the hub shell. In a roller clutch design, the force is transmitted radially outward from the cassette body against the hub shell to provide the holding force. The harder the cassette body is driven forward, the harder the rollers push against the hub shell to maintain a firm grip. This neccessitates a much stronger hub shell to accomodate the large radial forces driving it. In the case of more traditional "toothed" engagement systems, the majority of the engagement force is tangential to the hub shell and is carried by a tooth and pawl that are designed to withstand intense localized pressure.
So, would some kind of filament wound carbon spool be light and strong enough to contain the roller clutch system? Or, are teeth the way to go and could engagement lag be reduced by increasing the number of teeth and perhaps putting a positive angle on the faces of the two pieces that actually engage one another? It seems to me that this positive engagement angle would encourage complete engagement every time the two parts got close enough to touch. MAVIC and DT/Hugi have or have in the past had the engagement occur laterally, while all other versions seem to actuate radially. Which way makes more sense for material stresses and design? As you've got the axle and the Q/R already fighting against lateral dimension changes, it seems to me that you could make a simpler, lighter hub shell by actuating laterally. On the other hand, the radial actuation layout seems to do quite well with a small, light ring of material capable of withstanding the actual engagement forces to transmit them to the hub shell.
I can't wait to see the next improvement on this old design quandry. I hope somebody figures out the silent, instant engagement thing most of all.

...