science of shift
DESCRIPTION
tech feature from Bike magzine on the intricasies of bicycle transmission designTRANSCRIPT
bike 085bike 085
gearhead [by chris lesser]
THE SCIENCE OF SHIFTINGWe shine light on the crucial,
untold secrets of shifting gears
MODERN MOUNTAIN BIKE TRANSMISSIONS LET USERS RIFLE THROUGH SHIFTS
without missing a beat. That tactile sensation between clicks and taps of the shifters
and the corresponding sound of the chain dancing across and slipping into exactly
the right gear happens hundreds of times during a long ride, and for
most mountain bikers shifting is an ingrained, if seldom examined action.
But here’s the rub: There’s a lot more to shifting than just shifters
and derailleurs. Break out the magnifying glass and hit the slow-
motion button and you’ll find that what happens in that narrow win-
dow of time between shifting and shifted takes place
within a moving mechanical microcosm of ramps, pins,
chamfers, bevels, recesses and mind-numbing
attention to detail—all working in concert to
produce seamless shifts. >
PHOTOS: MORGAN MEREDITH
BIKP-070700-GEAR lyot 4/27/07 7:49 PM Page 085
086 bike
gearhead
The Stakes are HighImagine if you will the wide world of gears, with
the precision engineering of a Swiss-made Rolex
at one end and a donkey-driven grinding mill at
the other.
The Rolex can synchronize seconds, minutes,
hours and days through a mechanical symphony of
variable gearing, but it can’t handle more torque
than what it takes to push a second hand around a
dial. The mill, however, can handle incredibly high
loads, but its single-speed gear ratio is locked in.
A mountain bike drivetrain, by contrast,
demands it all: Rolex-smooth shifting that can
simultaneously withstand the brute, donkey-
strength torque generated by the pedal mashing
of an out-of-the-saddle climb.
One of the first attempts to balance the twin
demands of precision shifting and torque was
Campagnolo’s Cambio Corsa system, circa 1940.
To shift between four gear choices, users first
had to loosen the rear axle quick release—while
riding—via a seatstay-mounted lever, then use
an adjacent lever to guide the chain to the next
gear before clamping the axle into its slotted
dropout. It was a revolutionary first step,
albeit a sketchy one.
In the ensuing 70 years of shifter
development, changing gears has
become less of a death-defying
feat and more of the instinctual
action we take for granted. The
evolution of parallelogram
derailleurs, which articulate
along a highly specified arc
to guide the chain onto the
next gear, has played a
major role in the evolution
of today’s high shifting stan-
dards. But there’s another
critical, if relatively unher-
alded, aspect of the shifting
story: the gears themselves. The
untold engineering hours imbed-
ded in the gears used by modern
mountain bikes—right down to the
shape, angle, radius and offset of each and
every tooth and trough—separate today’s chain-
rings and cassettes from practically every other
gear-driven contraption ever conceived.
“Most gears in the world today
are still round,” points out Ric
Hjertberg, FSA’s new technology man-
ager. “They’re made by taking big stacks
of steel rings and a cutter, which goes up
and down and cleaves out one notch at a time,
and then the big stack rotates.”
But bicycle gears’ complex shapes demand
extensive programming hours on a CNC mill,
Hjertberg says. “Cross angles for roller chains?
Gears that aren’t uniform? Lassoing spinning gears
with a roller chain? Nobody else does this stuff.”
To scratch the surface of what goes into
engineering these gears, just look at a cassette
cog or chainring and consider that the posi-
tion, shape and size of every scallop, rivet, pin
and recess owes its existence to decades of
engineers wrestling with the challenge of get-
ting steel roller chains to simultaneously
engage solidly with, and float fluidly over, a
range of gear combinations.
Seeing these static shift features is one thing.
Understanding how they work on a trail is
another matter entirely.
Anatomy of a ShiftOne relatively constant factor in the grand equation
of shifting is the actuation of the shift itself. Re-
gardless of whether the “click” of an indexed shift is
made by the push, pull, dab or twist of a
shifter, it has the same relative effect on the
derailleur, and that derailleur has the
same relative effect on the chain—
which is to say, it suggests the
chain jump to the next gear.
And this is where those
thankless hours engineering
chainrings come into play—
turning that “suggestion” into
an offer the chain can’t refuse.
In shifting from a 32-
tooth to a 44-tooth chain-
ring, for example, strategically
placed ramps and pins built
into the side of the big ring
mate precisely with the outside
profile of the chain, helping coax it
onto the larger gear. Look close and
you’ll see that certain teeth on the
outer ring are scalloped ever so slightly to
allow the chain to cross onto the bigger gear.
Now, pause the frame in mid-shift. The chain,
taut with torque, is simultaneously engaged on
two rings at once, allowing for constant power
transfer through the shift. But if either of the two
rings in question were rotated a single degree in
Different strokes for different folks. Clockwise frombottom left: Truvativ Noir, Shimano XTR, Race FaceAtlas AM and FSA K-Force.
This 30-year-old SunTour six-speed freewheel (left) hasinterchangeable cogs, tall, uniform teeth and no pre-deter-mined shift gates. The Shimano XTR cassette, by compari-son, is engineered as a system to shift up and down onlywhere the chain will transition smoothly to the next gear.
BIKP-070700-GEAR lyot 4/27/07 7:49 PM Page 086
either direction, it would spell disaster. The chain
would ride on top of the gear teeth and when the
rider mashes on the pedal, the chain would slip
off the rings and break away like a trapdoor—the
sudden release of torque sending the rider’s body
crashing forward.
Next, consider the reverse shift: Approaching
a hill in a high gear, our crash test dummy flicks
the front shifter to drop the chain from the big
ring to the middle ring. The spring of the front
derailleur is released and the derailleur cage
swings inward and thwacks against the outside
of the chain. But freeze the frame. The crank
arms are lined up horizontally and the leading
pedal is preloaded with as much leverage as the
rider can generate. Under such load, the chain is
not going to change gears easily, and if it did,
such a sudden change in gear ratio would result
in a violently altered cadence, sending the rider
crashing forward, again. But the teeth on the big
chainring that line up with the front derailleur
at this precise moment have been engineered
with a slightly taller profile to dissuade the
chain from dropping down just yet.
Now, slowly advance the frame and see how,
as the forward crank arm rotates past the 5
o’clock position, torque on the chain lessens.
Inch forward another couple frames, to the
point when the crank arms are aligned straight-
up-and-down, and notice how the teeth on the
big ring that line up with the front derailleur
are chamfered slightly. There also is a recess in
the chainring itself, opening up just enough
space to let the spring of the clattering front
derailleur push through its stroke and slide the
chain over—its rollers falling cleanly into the
troughs of the middle ring.
To Clock A ShiftThe “strategy” by which shift features are inte-
grated into chainrings relies largely on an under-
standing of how riders deliver power to the ped-
als. Aside from a very few mutant cyclists who
have an even spin, the rest of us pretty much
just mash down on our pedals.
Engineers who pay attention to this stuff
refer to this as sinusoidal power input, which
simply means that the human body’s muscu-
loskeletal system generates varying amounts of
leverage at different points in the pedal stroke,
with maximum power coming when the leading
crank arm is at 2 o’clock, and with the least
amount of leverage exerted when the crank
arms are oriented in the 12/6 o’clock positions.
Thus armed with an idea of how the motor is
behaving, shift engineers design the specific
gearhead
There are a limited number of places on the circumference of any two given rings where a chain can transitionsmoothly off one gear and neatly into the next. Finding those places and then getting the chain to move there iscritical to smooth shifting.
BIKP-070700-GEAR lyot 4/27/07 7:49 PM Page 088
gearhead
points where they want shifts to happen—and as
importantly, where they don’t want shifts to
happen. Each chainring configuration presents
unique challenges, but mountain biking’s fairly
standard 22-32-44 combination provides a solid
model to work from.
Most chainring manufacturers build two
sets of shift ramps into the inner surface of
the 32- and 44-tooth ring, and position them
so they line up with the front derailleur at the
moment either crank is at the 2 o’clock posi-
tion, the point in a pedal stroke where an
upshift feels best from a biomechanical stand-
point. Because chains consist of alternating
inner and outer link plates, shift ramps come
in pairs to account for both “chain phase” con-
tingencies—engineer-speak that describes
whether an inner or outer link will line up
at any given point on a gear.
Ramps and pins from different chainring
manufacturers come in all shapes and sizes—
some are hardened steel plates, others are
machined directly into the chainring—but in
each pair of ramps, one ramp is designed to
pick up an outer link and the other is designed
to pick up an inner link. And by positioning
pairs of ramps across from one another on the
chainring, it shouldn’t take more than a half a
crank revolution for the chain to catch a ramp
and complete the shift.
The ideal time to initiate a downshift is
when the least amount of chain torque is
being exerted on the system, which in terms
of our power-clocking schematic happens
when the crank arms are aligned vertically—
by extension, this is also where the derailleur
spring has the most influence on bending a
chain laterally. Chainring engineers can help
induce a downshift by chamfering the edges of
specific chainring teeth, creating angled tooth
profiles and shaving away material where the
shift needs to happen.
Locating those shift ramps and recesses is
always a compromise because the best place to
engineer a shift one way might get in the way of
the shift features you need to let the chain shift
back the other way.
Why Chains SuckOnce we have the up-shift and downshift fea-
tures dialed, we’re only two-thirds of the way
home. The last third of the equation, according
to Garrett Smith, Truvativ’s resident chainring
guru, is the relative clocking of each chainring
with adjacent chainrings. The goal, he says, is to
get the chain to slide nicely into the next ring’s
teeth as it rolls from gear to gear.
Timed perfectly, in the instant between
shifts, a chain will be engaged on both gears
simultaneously. If the chain doesn’t line up
perfectly in the ring it is shifting to, it will ride
on top of the next ring’s teeth while staying
locked onto the primary chainring. Tension will
keep the chain bridged across the two rings
until it binds up underneath the bottom brack-
et, which will stop the rider faster than a
squirrel in the spokes.
“In the end, shifting on a bike is always a
compromise,” Smith says. “Because the ideal
clocking for making the perfect downshift will
not necessarily be the perfect clocking for a
perfect upshift, or for preventing chainsuck.”
Changing the relative clocking of two adjacent
chainrings by as little as a quarter of a degree
can make the difference between chainsuck or
a clean shift.
If these subtly varying shift ramps still seem
trivial, consider this: the new Shimano XTR
crank comes in two gearing variations, a 22-32-44
and a 24-32-44. The only difference between the
two appears to be the 22- or 24-tooth chainring,
right? Wrong. Because of clocking considerations
and because the whole system is interconnected,
changing the granny gear by just two teeth
means Shimano needed to engineer a whole
new middle ring, too. >
The secret to shift-ing doesn’t comeeasily, and compa-nies protect thedimensions of everylast gear tooth withextensive patents.
BIKP-070700-GEAR lyot 4/27/07 7:49 PM Page 090
gearhead
Bringing Up the RearDespite having three times as many gears in the rear as in the front,
rear shifting is actually less complicated because engineers don’t have to
factor in the varying torque influence of the crank arms.
Just as with front chainrings, there are a limited number of places on
the circumference of a cassette cog where the teeth line up to allow the
chain to mesh cleanly from one gear to the next. By identifying these
critical geometries, and by lining them up on an indexed freehub body,
engineers can create “shift gates,” or sectors of corresponding cogs
where two or three tooth profiles are manipulated just enough to
encourage the chain to pass up or down the cassette.
Look straight down on a cassette and you’ll see a series of wave-like
scallops on opposing sides. These are the “off-ramps” and “on-ramps”
where the chain will, by design, want to shift. If all goes according to
plan, the shift will happen in the specified shift gate and the chain will
line up evenly in the troughs of the gear it’s shifting to. But lining up
nine cogs to work as a harmonious package is easier said than done.
Shimano was the first company to figure this out back in 1989, with
its Hyperglide system, and now a close view of a chain shifting through
a cassette looks a lot more like a snake slithering through the gears than
a chain riding up, over, and then down into the gears.
Where Shifting is GoingShifting may be one of the most overlooked aspects of cycling, but
the technology that lessens the time in between shifts is one of the
most scrutinized and hotly contested fiefdoms in the whole of
cycling’s intellectual property landscape. A good portion of the thou-
sands of claims Shimano has filed with the U.S. patent office relate
to the nuances of how a chain is shifted between gears. Because
Shimano’s engineers study shifting so closely, and because the com-
pany also makes chains, derailleurs and shifters to accompany its
chainrings and gears, “Systems Engineering” is unavoidable—much
to the chagrin of competitors trying to field compatible products
into the marketplace.
For the most part, other chainring and cassette manufacturers
have had to reverse engineer what Shimano has done so that their
components play nice with Shimano parts. But now that SRAM has
enough brands under its umbrella to engineer its own systems, and
now that FSA is reportedly working on its own shifters, competition
should heat up.
As mountain biking has developed as a sport, “standards” have
come and gone as axle spacing dimensions and freehub body designs
have evolved to make room for more gears. Where once there were
five gears, now there are nine. Freewheels have given way to freehub
bodies and, gear for gear, there is less real estate to work with. Chains
are narrower, spacing between gears is closer and tolerances are
tighter. Through this evolution, drivetrain engineers have had to not
just keep up with shifting speed and accuracy, but also improve per-
formance at the same time.
While there always will be stubborn retro-grouches who swear by their
30-year-old SunTour friction shifters, and while others flat-out prefer the
simplicity of single speeds, a cadre of shift engineers, employed by vari-
ous manufacturers and spread out across the world, are constantly test-
ing every combination of chainrings, cassettes and chains to see where
they can improve next. And ten-speed, they say, isn’t far away.
BIKP-070700-GEAR lyot 4/27/07 7:49 PM Page 092