h a l f b a k e r y
I think this would be a great thing to not do.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
or get an account
Flywheels are essentially a rotating mass most commonly
used to smooth out the pulses of a piston engine. Car
engines have them, usually a few kg worth on the end of
the crankshaft. In practice, they smooth out the power
pulses from the engine to reduce shock loads on the
transmission in addition
to adding to the total rotational
momentum of the engine at low RPM. This gives a little bit
of leeway when engaging the clutch, a slight mismatch
between engine output and clutch friction can be drawn
from energy stored in the flywheel.
At high RPM this is much less important, the power pulses
are much closer together, the rotational momentum of the
engine without the flywheel starts to become significant in
its own right and the power output meets or exceeds the
energy demands of accelerating up the drive-train. At this
point, a heavy flywheel becomes a liability. At high RPM
the engine has far more energy stored in the flywheel than
it needs and RPM changes become difficult. For this
reason, racing engines have much lighter flywheels. This
destroys much of the low RPM benefits needed in regular
The solution. A two part flywheel. Closest to the engine,
mounted to the crank is the driven flywheel #1. This
contains weights on radial tracks held close to the center
by springs. As the rotation speed increases the weights are
driven centrifugally outward moving levers. Through
significant mechanical advantage the levers retract pads,
similar to brake pads, into the flywheel, so that they no
longer contact flywheel #2. The second flywheel, mounted
in front of the first is not driven by the crank. It is bearing-
mounted and now may spin freely. As the engine speed
falls again, the weights move inward and the pads emerge
a small distance from the face of flywheel #1 where they
contact flywheel #2 and accelerate it up to engine speed.
This engine will perform like one with a heavy flywheel at
low RPM and an engine with a light flywheel at high RPM.
Smooth low speed driving and good throttle response at
high speeds. I'm aware that dual mass flywheels already
exist, however these are radially coupled by springs. The
dual masses cannot decouple. The second sprung mass
functions to dampen torque oscillations. This idea is
essentially a second flywheel actuated by a reverse-
Obviously inspired by
[bs0u0155, May 11 2016]
||// As the engine speed falls again, the weights move inward and the pads emerge a small distance from the face of flywheel #1 where they contact flywheel #2 and accelerate it up to engine speed. //
||That's going to release as lot of heat, wasting energy.
||Also, as the engine slows, the second flywheel starts to engage, increasing drag and slowing the engine further, thus inducing the second flywheel to engage more. Hard to avoid a "snatch" effect on slowing down. The rate of re-engagement would need to be damped (slower reconnection, hence more slippage, more wear, more energy wasted as frictional heating).
||Some sort of viscous coupling ?
||Or a magnetic coupling perhaps. Potentially yields the
option for selectable coupling/decoupling and limiting
||//that's going to release as lot of heat, wasting
||Hmm, this isn't like the brakes, we're talking a few kg
which is free to spin up.
||//as the engine slows, the second flywheel starts to
engage, increasing drag and slowing the engine further,
thus inducing the second flywheel to engage more. Hard
to avoid a "snatch" effect on slowing down.//
||This would be true if the second flywheel was a fixed
object. It isn't though. As the engine speed slows, and the
first flywheel engages the second, the second begins to
move, the differential speed between them decreases
and therefore the first flywheel experiences less torque.
||Put a manual transmission in neutral, then dump the
clutch. There's a tiny change in engine note, but it
equalizes quickly. The second half of the clutch and the
input shafts of the gearbox are in the kind of mass
territory I'm talking about... Similarly, the synchromesh
in such gearboxes happily uses friction to match a few kg
of differentially spinning metal all the time, with minimal
heat or wear.... So, beefier than a synchro, much less
beefy than the clutch.
||// The second flywheel can also be engaged at high RPM
to drop engine speed faster between gear changes//
||And the beauty of this is that as you lose rpm in
anticipation of a gear change you GAIN rotational
momentum which translates to free torque when you
engage the next gear. The beauty of that is that it works
when you are changing up or down (provided there's
||In fact, there's no reason you cant have more nested
flywheels. You could have 2-3 nested flywheels which
engage/disengage at strategic points in the rev range
where gear changes normally happen.
||// the differential speed between them decreases and therefore the first flywheel experiences less torque. //
||But in your design, the amount of brake pad pressure is only proportional to the rotation of the primary flywheel. Once the "clutch" starts to bite, any further slowing makes it bite harder, irrespective of the relative speed of the driven plate.
||It seems to me that if the first flywheel physically moves
part of its mass away from the axle, then that increases the
average moment-arm of each radial section of the
flywheel, somewhat counteracting the attempt to
disengage the second flywheel.
||Forget the mechanical centrifugal clutch. What you need
here is a ferrofluidic clutch.
||I like the nested flywheels.
||There's an effect (can't remember what it's called) where it takes a bit of time for a magnetic field to establish a "full lock", so if the relative rotational speed of a flywheel is vastly different than the one next to it, it won't drag as much. (Or vague blather to that effect.) So the system might not require calculated shift points at all.
||// vague blather to that effect //
||If there's no conductive path to remove the energy, you're going to generate eddy currents, and a lot of heat. If the permanent magnets get heated above their Curie point, they'll lose their magnetism.