Octalink BB: Keep tightening left bolt, save a crank?

Hi all:

Without understanding the physics of the situation, I'll accept as
correct the assertion that a design defect in the Shimano Octalink BB
results in the left crank bolt loosening when riders stand on the
pedals with the right foot forward. Eventually this loosening of the
bolt can result in the trashing of the crank arm.

Apparently this process is different then the "loosening" of crank
bolts with square-tapered cranks: in this case the crank arms (both of
them) "elbow" their way up the taper within a few miles of riding,
relieving pressure on the bolt. The uninitiated say "these bolts have
loosened, I'd better tighten them up" which then allows the elbowing to
resume, followed by more bolt tightening in a vicious circle until the
cranks are destroyed.

Going back to the Octalink setup: it's not clear when sufficient relief
of bolt pressure starts to result in crank arm destruction. I'd assume
that this occurs when the torque falls below the minimum range
specified in Shimano's literature. It's also not clear how quickly
this loosening occurs, though I'm kind of assuming it doesn't typically
happen within a very short time. At least I didn't find a post that
said "I put new Octalink cranks on my bike and within an hour of my
left crank arm was loose."

Anyway, it strikes me that one strategy for getting good life out of an
Octalink setup would be to check the torque of the left crank arm after
every ride. As long as the bolt torque was above the Shimano minimum
you would be assured that the crank arm hasn't been damaged. If the
torque was reduced from the torque you initially set upon installation,
you'd re-torque to the installed value.

Could this work? I don't see why it wouldn't.

TIA

Tom Young

Tom Young writes:

Without understanding the physics of the situation, I'll accept as
correct the assertion that a design defect in the Shimano Octalink
BB results in the left crank bolt loosening when riders stand on the
pedals with the right foot forward. Eventually this loosening of
the bolt can result in the trashing of the crank arm.

Apparently this process is different then the "loosening" of crank
bolts with square-tapered cranks: in this case the crank arms (both
of them) "elbow" their way up the taper within a few miles of
riding, relieving pressure on the bolt. The uninitiated say "these
bolts have loosened, I'd better tighten them up" which then allows
the elbowing to resume, followed by more bolt tightening in a
vicious circle until the cranks are destroyed.

Going back to the Octalink setup: it's not clear when sufficient
relief of bolt pressure starts to result in crank arm destruction.
I'd assume that this occurs when the torque falls below the minimum
range specified in Shimano's literature. It's also not clear how
quickly this loosening occurs, though I'm kind of assuming it
doesn't typically happen within a very short time. At least I
didn't find a post that said "I put new Octalink cranks on my bike
and within an hour of my left crank arm was loose."

Crank attachment is not dependent on torque, only position of the
retaining bolt. Therefore, loosening causes no damage until the crank
and bolt back out far enough that there is insufficient spine
engagement with the spindle and the profile in the crank shears off.
That is why Shimano made a second version with a deeper spline,
erroneously thinking there was insufficient strength in the interface.

Anyway, it strikes me that one strategy for getting good life out of an
Octalink setup would be to check the torque of the left crank arm after
every ride. As long as the bolt torque was above the Shimano minimum
you would be assured that the crank arm hasn't been damaged. If the
torque was reduced from the torque you initially set upon installation,
you'd re-torque to the installed value.

Just take up any slack that occurs. Torque, as you see, has no
meaning here.

I find interesting that the right crank bolt doesn't also unscrew
because the spindle has the same torque reversal at both ends. That
it doesn't unscrew seems to indicate that its thread tends to tighten
the thread similarly to the left hand thread of the right BB cup or
left pedal. In Shimano's position, I would have changed the left
retaining bolt thread to be opposite from what it is rather than make
the spline deeper. Whether that had a positive effect would readily
be demonstrated quickly if they tested it with goofy footed riders.

Could this work? I don't see why it wouldn't.

You could change your leg position when standing to not cause
loosening. That may be asking too much but the alternative is a
different crank assembly. I'm not sure you can keep after that crank
bolt enough to prevent damage. I'm surprised that Shimano did not
recognize the problem sooner... or at all. Of course they don't read
wreck.bike so where would they discover such things.

Jobst Brandt

wrote:
Tom Young writes:

Without understanding the physics of the situation, I'll accept as
correct the assertion that a design defect in the Shimano Octalink
BB results in the left crank bolt loosening when riders stand on the
pedals with the right foot forward. Eventually this loosening of
the bolt can result in the trashing of the crank arm.

Apparently this process is different then the "loosening" of crank
bolts with square-tapered cranks: in this case the crank arms (both
of them) "elbow" their way up the taper within a few miles of
riding, relieving pressure on the bolt. The uninitiated say "these
bolts have loosened, I'd better tighten them up" which then allows
the elbowing to resume, followed by more bolt tightening in a
vicious circle until the cranks are destroyed.

Going back to the Octalink setup: it's not clear when sufficient
relief of bolt pressure starts to result in crank arm destruction.
I'd assume that this occurs when the torque falls below the minimum
range specified in Shimano's literature. It's also not clear how
quickly this loosening occurs, though I'm kind of assuming it
doesn't typically happen within a very short time. At least I
didn't find a post that said "I put new Octalink cranks on my bike
and within an hour of my left crank arm was loose."

Crank attachment is not dependent on torque, only position of the
retaining bolt. Therefore, loosening causes no damage until the crank
and bolt back out far enough that there is insufficient spine
engagement with the spindle and the profile in the crank shears off.
That is why Shimano made a second version with a deeper spline,
erroneously thinking there was insufficient strength in the interface.

Anyway, it strikes me that one strategy for getting good life out of an
Octalink setup would be to check the torque of the left crank arm after
every ride. As long as the bolt torque was above the Shimano minimum
you would be assured that the crank arm hasn't been damaged. If the
torque was reduced from the torque you initially set upon installation,
you'd re-torque to the installed value.

Just take up any slack that occurs. Torque, as you see, has no
meaning here.

I find interesting that the right crank bolt doesn't also unscrew
because the spindle has the same torque reversal at both ends. That
it doesn't unscrew seems to indicate that its thread tends to tighten
the thread similarly to the left hand thread of the right BB cup or
left pedal. In Shimano's position, I would have changed the left
retaining bolt thread to be opposite from what it is rather than make
the spline deeper. Whether that had a positive effect would readily
be demonstrated quickly if they tested it with goofy footed riders.

Could this work? I don't see why it wouldn't.

You could change your leg position when standing to not cause
loosening. That may be asking too much but the alternative is a
different crank assembly. I'm not sure you can keep after that crank
bolt enough to prevent damage. I'm surprised that Shimano did not
recognize the problem sooner... or at all. Of course they don't read
wreck.bike so where would they discover such things.

Jobst Brandt

Jobst,

I am confused regarding goofy-footedness and loosening bolt/loss of
preload. I've read the info you wrote at
http://www.sheldonbrown.com/brandt/i...ng-cranks.html, but I guess
I'm missing a few definitions.

Under which following conditions does spindle transmit reverse torque
(asking for clarity on goofy-footednes and loss of preload in bolt
relationship)?

1. "Coasting/standing" with right crank arm forward. Both feet engaged
in pedals
2. Starting the bike from a full stop with the right crank forward.
(i.e. left foot isn't fully "engaging" the pedal _yet_, but right foot
is.

You wrote "... Besides that, the right crank transmits torque to the
spindle only when standing on both pedals. Doing this with the right
foot forward (goofy footed) is the only time the spindle transmits
reverse torque." Does this _really_ mean only (1) applies?

wrote:

You could change your leg position when standing to not cause
loosening. That may be asking too much but the alternative is a
different crank assembly. I'm not sure you can keep after that crank
bolt enough to prevent damage. I'm surprised that Shimano did not
recognize the problem sooner... or at all. Of course they don't read
wreck.bike so where would they discover such things.

Changing leg position is fine on the road. When on a trail, I stand on
the pedals in such a manner that whichever side of the bike is nearer
hillside/obstacles/brush is the side where the pedal is in trail
position. So, if the hillside is on my left, I'm right-foot-forward.

I also mount the bike from the left, and start pedalling/clip in on the
right pedal.

All this time on an XT Octalink, and no problems with loosening. I'm
not of clydesdale weight anymore, though. Maybe that is a factor?

E.P.