Index Home About Blog
From: (Jobst Brandt)
Subject: Re: Splined Track Cranks
Date: 31 Mar 2000 20:29:27 GMT

John Dacey writes:

> We've been speculating about why, after waiting over a year and a
> half, Shimano's 7710 series track cranks (hollow and splined bottom
> bracket mount) still isn't available.

> Could it be that this attachment system distorts the aluminum arm
> splines (resulting in loosening) when subject to the push-pull loads
> of fixed gear use in a way that hasn't been a problem for the road
> models that are only stressed in one direction?

This has been my concern from the outset, but not for track.  On any
bicycle there is reverse torque on the spline when standing on both
pedals, right foot forward.  Because I naturally prefer to stand that
way when traversing rough surfaces, this could be a problem.  As many
people whom I have asked, none has understood what the concern is and
therefore, haven't helped me understand why these cranks don't fail in
that mode.  Without a press fit, there has got to be a lot of motion
under reversing loads and this can't be good.  Besides, the welded
hollow cross section is a huge expense and cannot be as fatigue
resistant as a solid forging.

I am still looking for a positive yes or no on these cranks.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Splined Track Cranks
Date: 4 Apr 2000 00:09:39 GMT

Phil Holman writes:

> Its claimed to be a snug fit with 200% more contact area than a
> conventional square taper spindle. The larger diameter spindle will
> also further reduce the bearing stress compared to the square where
> the bearing load due to pedaling is triangulated from the center
> over the width of the square higher at the corners). I got the
> impression when I installed them that they were more than snug and
> even had to be re-tightened after the first ride. I didn't think the
> square taper could be improved (no backlash) so I was surprised to
> see Shimano do this. So far I haven't experienced any play but time
> will tell if it holds up. No Coke Can fix with these.

All that said, the spline, not preloaded in a press fit, has clearance
under torsional load and therefore, backlash.  As you may be aware,
the square taper, like pushing and pulling spokes of a cross laced
rear wheel, transmit torque both by increased stress on the trailing
edge of the flats and by decreased stress on the leading edge.  If the
leading edge lifts off under load, the square hole will be damaged
because the trailing edge receives a canted load.

If the splines are not preloaded, they will be lifting off at each
torque loading, whether forward or rearward.  This must cause damage
and at least some perceptible backlash that exists and grows with use.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Splined Track Cranks
Date: 4 Apr 2000 00:25:44 GMT

Jesse Curl writes:

> Negative on the splined cranks.  IMHO they are much worse than
> square taper, press-fit style arms and BBs.  I put a pair of the new
> Ultegra cranks on my bike last year, and have had to tighten them 3
> times since install (approx.  4000 miles of riding ). Before you
> ask, yes...I have been using a torque wrench and tightening to
> Shimano's recommended specs.

I see no difference between Track and Road.  On the road, one spends
plenty of time standing on both pedals for bumps and the like,
something track bikes don't do except for track stands, while track
bikes backpedal now and then.  There is always reverse torque on
cranks and the backlash that I outlined is the cause of the retaining
bolt loosening.

Thanks.  I've been waiting to hear that this occurs because I
suspected it must.  This seems to be the existence proof I was looking
for.  This seems to be like Sturmey Archer syndrome where the user who
has problems is intimidated to believe he did it wrong and doesn't
want to talk about it.  If your AW Sturmey Archer hub jumps out of
gear, you will be chastised for misadjusting it.  Hence no one admits
to having the gear fail because they don't want to hear the diatribe.
The hubs are a hazard and I believe the splined cranks, without
preload (in the spine) must rotate and loosen the retaining bolt.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Splined Track Cranks
Date: 4 Apr 2000 02:25:07 GMT

Eric Salathe writes:

>> If the splines are not preloaded, they will be lifting off at each
>> torque loading, whether forward or rearward.  This must cause
>> damage and at least some perceptible backlash that exists and grows
>> with use.

> A splined fit has been used for attaching car wheels (eg Austin
> Healy).  Was backlash a problem with these? I think I recall reading
> of it in Austin Healy literature, but never noticed it in a sample
> of one.

If you will recall, the knock-off nuts had a double wedge so that
friction between nut and wheel was greater than even dry threads so
that the nut would screw back and forth with the motion of the wheel
on the spline.  In fact, if you didn't lube them with a graphite paste,
the rear wheels often chirped on torque reversal.  The amount of rouge
produces in these interfaces was enough to make any reasonable
engineer wince.

> By the way, I always have to rethink how cranks work when this comes
> up.  The crank spindle transmits torque only when the left crank is
> being pushed, so windup is opposite on each side. Which is why
> standing right-forward creates backlash in both cranks.

Yes and that happens.  It turns out that my preferred stance,
descending on bumpy roads is right foot forward although I do the
opposite sometimes just to see if I'm in a rut.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Loosening Ultegra cranks...
Date: 3 May 2000 22:57:23 GMT

Iain Mannix writes:

> Several asked if it was the drive/non drive side - when the problem
> manifests itself, both bolts are loose.

I would expect that because bolt looseness is caused by a different
mechanism than on a square taper where the taper is the limit of bolt
torque.  Here it is a solid end of the crank flush against the spindle.
The rotational backlash affects both cranks and their bolts.

>>> The splines are not damaged.

>> Well the splines ARE damages or there wouldn't be perceptible
>> backlash.

> Quite possibly so - but there is no visible damage to the spindle or
> cranks.  A friend at a FLBS(formerly local bike shop) suggested that
> there might be some differences in the spindle itself - he seemed to
> remember something about the early XTR splined spindles being
> slightly different than the current crop; not sure if that is the
> case yet.

I expect that the aluminum gives, not the steel.

>> They may still be carrying torque but a tight fit they do not have.
>> By the way, how about doing this without the chain in place and see
>> whether the cranks go beyond center.  That is, after taking up the
>> backlash, they should prefer to stop in that position when turned
>> freely whichever way you do it.

> Interesting.  They have never become loose enough to move by hand.

I don't mean by hand but by standing on the pedals and noticing
whether they droop below center from jumping on them horizontally.
Without a chain, they will prefer to stop in one orientation.  This
was a common test in the days of cottered cranks to see if they were
180 degrees apart.

> I guess if I stood on them when loose, then leaned over the bike and
> tried to pull the pedals "up," assuming they're "down," I might be
> able to move them - but they've not gotten floppy-loose yet(even on
> the ride w/o the Allen).

No, that's not what I mean.  I'm just trying to ascertain whether they
actually droop when you stand on them and then reverse droop when you
put the other foot forward and repeat.

> If the pedals are "down," and I rotated them w/o the chain 180d, they'd
> not move.  It would take considerable force to move them the other way(IE,
> me standing on them after 2 years of intermittent riding;).

Just stand on them and lunge on them when they are horizontal.  I
think that's what you said made them move on the spindle originally.

> A clunk I can feel on every revolution, even if considerable force
> is needed to replicate it, is "loose" in my book - when I say
> "loose," I don't mean "in the process of falling off the bike" like
> oh-so many bicyclists who wander into the shop wondering "if thems
> pedals is loose."

Well then we ought to see it in the chainless balance test.

>>> I have not replaced caps/bolts, however.  Maybe that's next.

>> That's cosmetic.

> That's my thinking, too.

>>> I've got a fair bit of experience with other types of crank
>>> configurations (one piece, cottered, 2d taper), but this is my
>>> first "new" bike after taking a couple of years off.

>> I hope you get to the bottom of this because it will answer some
>> questions I have has ever since Shimano introduced these cranks.

> I'll get to the bottom of it, even if it means putting a "real"
> BB/crank assembly on the bike.

Ah, but first I want to know what causes the effect you detect on these.

> I'll post whatever I wind up doing.

> As far as "incorrectly installed," that could be, if the factory
> installed them incorrectly; maybe Schwinn's assembly line leaves
> these cranks off for shipping, but IME, most mass produced bikes
> come with the cranks "installed," and only a handful of mechanics
> I've run into remove the cranks to look at the BB.

Incorrect assembly doesn't make much sense.  Either the spline is
engaged or it is not.  If not, the cranks don't transmit torque.  My
problem is that without pressed splines, there is always some backlash,
however small, and this should unscrew the bolt on the most pristine
cranks.  Yours seems to be an accelerated version of this.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Loosening Ultegra cranks...
Date: 4 May 2000 22:25:39 GMT

Jay Beattie writes:

>> What do you believe the torque value assures?  The spline bottoms
>> against the crank bore end.  More force at this place does not
>> improve >any parameter that is apparent from inspecting this joint.
>> Friction is not a valid means of preventing rotation.  Even with
>> taper-square cranks the interface creeps and these have far higher
>> contact pressure.

> Sorry to be thick, but wouldn't increased torque on the bolt better
> prevent loosening.

Maybe we could wait until Phil responds, since he was the one leading
the discussion at the time.

I don't know how mechanically inclined or how much experience you have
with machinery but a bolt holding down a part that has even a small
rotational motion WILL unscrew in the absence of some other holding
device.  The screw will generally not advance (tighten) in its thread
when the crank turns that way relative to the spindle, but it will
unscrew every time it moves the other way.  It's like shaking a
sloping table on which an object will gradually move toward the low

If the screw is prevented from turning, fretting motions alone will
wear away the preload that in this case is small because there is no
length of material for strain loading (such as a QR skewer).  A strain
screw (like a swaged spoke) can maintain preload as well as take up
torsional movement, but that still doesn't protect the spline from
backlash damage.

> Assuming that the bolt is strong enough and that the interface would
> not be destroyed by the high force, wouldn't "the tighter the
> better" apply with a splined crank?

I'm not sure what you mean by destroying the interface.  It won't
remain tight if the spline has backlash, and since the spline is not
tapered, there is no preload, which is the same as backlash when the
spline is torque loaded.  In this case, the trailing edge of the
spline lifts off its contact face causing rotation with respect to the
spindle.  Small as this motion may be it is enough to loosen the bolt.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: splined crank update
Date: 23 May 2000 16:03:07 GMT

Chris Fabri writes:

> If you have this on your crank, check the self-pulling crank arm
> bolt assembly.  I took it out, cleaned it up, put a little lube on
> it and re-installed.  This has cleared up my odd noise that wasn't
> fixed by cleaning and reinstalling the BB or cranks.

That may have solved the audible part of the problem for you but I am
still not clear on how these cranks sustain reverse pedaling torque,
that torque that occurs when standing on the cranks in the horizontal
position, right foot forward.  It is possible that Shimano assumed
that no one stands on both pedals, right foot forward.  I think if I
were stuck with these cranks, I would make an effort to not do this
although my normal horizontal pedal position IS right foot forward.  I
have tried the opposite and notice the difference as being an
unnatural position descending rough roads.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: splined crank update
Date: 23 May 2000 18:11:35 GMT

Jay Beattie writes:

> Would you be so kind as to refresh my memory: why isn't left foot
> forward also a problem?

Only left crank torque enters the spindle, the right crank being
attached directly to the chainwheel.  Reverse torque in the spindle
(and therefore the spline) occurs only when the crank that transmits
torque is used in the reverse direction.  That direction is when load
on the crank is applied when it is in the rear half of its rotation,
when the right crank is in the forward half of its rotation... "right
foot forward".

Reverse torque is the only effect that can make torsional backlash,
actual or elastic, move to loosen the retaining bolt and ultimately to
make sounds... and damage.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Bottom Bracket - splined or square?
Date: 25 May 2000 22:46:48 GMT

David T. Blake writes:

>> I haven't heard that anyone has put any of them to the test and
>> compared. All hear is that one can get by using them. I am
>> interested because I find the taper-square a spooky failure prone
>> assembly that can break any moment and cannot be inspected...  that
>> is, spindle failure can sneak up on you without warning. I have had
>> it happen.

> Well, the principle limitations are radius and angle of the faces.
> If it is made six sided, there is a loss because the angle between
> the faces and the force from torque decreases.

That is not so much a problem as hoop stress caused by torque.  With
the smaller angle of approach, radial pressure goes up as 1/tangent
of the angle between facets.  This added to the press fit of the crank
would make a destructive stress as the number gets larger than four.

> Also, square tapers are easy to make.  Therefore I propose a larger
> square taper. 50% larger should put failures out of the realm for
> everyone except large riders that like to turn big gears up lots of
> hills.

They are not easier than a spline but then a tapered spline with
rounded teeth is what's needed, something like a Torx profile but

> I don't think the spline is a cure for anything, except forced
> upgrades through incompatibility.  To make the spline work, you'd
> need the interface area orthogonal to the forces produced by torque
> to be LARGER than those of the square taper.

I don't think so.  As you see, the contact area is minuscule compared
to the square taper, yet it works (even if not perfectly).  The idea
is reasonable but not all the details are solved, as we have discussed
in these threads.  I'm not sure how to solve them either, given the

> And I don't think they are even close.  Since they are installed
> without preload, the interface area would need to be double that of
> the square taper.

Not so.  In fact if there were no reverse torque, this design would be
far better than the square taper.  That has been my concern since this
design was first presented.  What I like about it is that it does not
have these near damaging press fits that the square requires and that
torque and crank alignment are searate and that the crank bottoms when
installed so that it doesn't creep up the press fit to a fit tighter
than when installed.  It is this last feature that cracks some square
taper cranks.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Hollow Forged Crank Arms??
Date: 22 Aug 2000 23:43:39 GMT

Brewster who? writes:

>> Better yet, come up with a BB that does not have the flaw of the
>> current hollow splined end design, that assumes no reverse torque
>> on the spindle.  As we have seen here in this newsgroup, they come
>> undone if you stand, right foot forward, over rough stuff (the only
>> occurrence of reverse torque on a crank spindle).

> I have only seen one posting on this newsgroup that supports your
> theory that the current hollow splined design is flawed.  Since
> Shimano sells hundreds, if not thousands, of these cranks, is there
> any other evidence that this "flaw" is as problematic as you
> suggest?  Or, alternatively, could the occurrence of reverse torque
> on the crank spindle that you suggest be the result of the mechanic
> not properly tightening the crank bolt.

That the crank will move rotationally on the spindle is undisputable,
there being no preload on either face of the spline teeth that prevent
rotation.  This was apparent to me when I first saw the design at the
InterBike trade show a few years ago.  Subsequently, whenever I had
the opportunity, I interviewed bike shops to discover whether they had
a problem with it and found none.  However, most responses were so
vague that I am also sure the mechanics had no idea what I was talking
about or what it was they should have detected.  I see that one day I
may have to use this design and I want to know its problems and find
solutions before that happens.

By a small sampling here on the newsgroup and riders I meet, I take it
that few riders have the right foot forward when they stand on the
pedals.  Beyond that, most people never ride hard enough over rough
roads on a road bicycle to stand on both pedals.  The concept of
elastic backlash and fretting are so obscure to most people that these
problems go unrecognized for many years.

As I said, the junction between pedal and crank has been amiss for as
long as screwed in pedals have been used and the cranks have been
failing there ever since aluminum cranks came along.

Jobst Brandt      <>

Index Home About Blog