Bearing binding due to thread direction and bottom bracket shell width
Why bottom bracket cups and pedals normally do not unscrew while pedaling
Regarding Nov. 11 post on Campy Ultra-Torque cranks on an Italian-threaded frame:
Dear Lennard,
With an Ultra-torque crank in an Italian-threaded bottom bracket, you have to make sure the shell is the correct width, the bottom bracket is faced, and the cups are torqued enough or you’ll damage the bearings. This actually happened to me on my Italian-threaded Calfee BB. With an Ultra-torque crank, the bearings are fixed against the inside of the crank arms (see attached photos, but note that the alloy crank has an English RH BB cup). The bearing cups are threaded into the BB and face outwards, pressing against the bearings. Anything that increases the spacing between the cups increases the force against the bearings, tightening them.
On my bike, there were two forces in play. First, the BB shell was too long left-to-right (perpendicular to direction of travel), forcing the cups to press too hard against the bearings. It would shred bearings every 3000 to 4000 miles.
I finally had my LBS reface the BB, and that solved the problem. Facing the BB shell removed some material, decreasing the perpendicular length of the BB shell, reducing the spacing between the cups, thereby reducing the pressure against the bearings, letting them spin more freely.
Second, the drive side bearing cup, which is right-hand threaded in the Italian BB, started to back out. It had turned a bit counter-clockwise when viewed from the starboard side of the bike. This also increased the perpendicular distance between the cups, effectively making the BB shell longer. This also increased the pressure against the bearings, making them rough after a few thousand miles.
Michael
Dear Michael,
Thanks for that. Very important points you make.
In the old school of loose-bearing bottom brackets, if a cup unscrews, the bottom bracket gets looser, not tighter, but that is not the case with threaded bottom brackets on integrated-spindle cranks, and not just with Campagnolo ones.
Once the lateral play has been adjusted out of an integrated-spindle crankset, if a bottom bracket cup unscrews, it will press outward on both cartridge bearings. And an Italian-threaded drive-side cup is prone to unscrew while pedaling if it has not been tightened into the shell with enough torque and/or threadlock to prevent it from unscrewing.
Similarly, the bearings can also be damaged with some integrated-spindle systems if the cups are too far apart when the left crank is tightened on. This is a common mistake people made with SRAM GXP bottom brackets, which were often packaged with extra 2.5mm-thick spacers. The spacers were intended for installing GXP MTB cranks on 68mm-wide bottom bracket shells, since the spindle length was designed for use with 73mm-wide bottom bracket shells. With a road GXP crank (which has a 5mm shorter spindle designed for a 68mm-wide shell), if either or both of those spacers were installed between the cup(s) and the shell, it would bind the bearings when the left crankarm was tightened on. Since tightening the crank bolt on a GXP left crank forces the inner race of the left bearing against a shoulder on the left end of the spindle, if the bearings were too far apart, the center races of the bearings would be getting forced inward further than the outer races, frying the bearings rapidly with use. Hopefully this exploded diagram from Zinn and the Art of Road Bike Maintenance clarifies this issue.
On a Campagnolo Ultra Torque (UT) crank, since the cartridge bearings are pressed onto the spindle against the crankarm, there is little room for error in cup spacing, especially with an Italian-threaded bottom bracket shell. The drive-side Ultra Torque bottom bracket cup has a retaining clip hooked behind the bearing, establishing the chainline (see book illustration photo).
The non-drive UT bearing can float a bit in the left cup, but only if the shell is not too long. And an Italian-threaded shell, already being 2mm longer at 70mm than an English-threaded one (68mm wide), that lateral float is already taken away. So if the shell is longer than 70mm or not faced so that cup is slightly cocked, tightening the short central crank bolt (see illustration photo) that pulls the two halves of the spindle together in the center of the bottom bracket will jam the cups against the bearings.
― Lennard
Dear Lennard,
Why do my pedals on our tandem keep unscrewing?
I couldn’t find a 180mm captain’s crank, so I put a standard road crank on backwards. The lineup of the chainring with the left chainring on the stoker crank is perfect, and since my wife and I use Crank Bros. pedals, it seemed like an elegant solution. I just put the left pedal on the right and vice versa.
But the first ride after putting the crank on, my right pedal fell off before we’d even ridden 5 miles. We were lucky to be on a dirt bike path and not on a road with cars on it, because my foot slammed into the ground with the pedal still connected to my shoe! It happened again a few more miles later after I tightened it back on. After that, we stopped every few miles to retighten before it came off again. And each time, the left pedal was about ready to fall out, too. I had put my wife’s pedals on at the same time when I put the bike together, and they have never loosened up.
Now I tighten them on with a super long 8mm hex key to put my pedals on. I used the stubby little one on my multi tool the first time and on that first ride. I even put a tube over the hex key and push on it as hard as I can. Knock on wood—my pedals have stayed on ever since.
It got me thinking about the tightening direction. It seems like the normal way with right hand threads on the right and left hand threads on the left would be the way that would unscrew while riding and they would get tighter while riding with my crank turned around backwards. Can you explain?
Christoph
Dear Christoph,
The pedals would unscrew from a normal crank when pedaling only if their bearings were seized. That is essentially the same thing as putting a wrench on them—it is analogous to having the pedal body fixed to the spindle and grabbing it with a giant wrench. As you apparently understand, pedals unscrew in the pedaling direction with a wrench.
However, when the bearings are working properly, each ball bearing rotates the opposite direction of the pedal body. That is why pedaling on standard cranks tightens the pedals, or at least doesn’t tend to loosen them up if they are not tightened on very tightly.
An early intention of the right-hand threads on the right crank and left-hand threads on the left crank was to avoid horrific crashes on penny farthings if the pedal bearings froze. On a penny farthing (high-wheeler bicycle), the crank is connected directly to the huge front wheel in the same way as on a unicycle or kid’s tricycle. But if the pedal were to seize up with the rider’s foot strapped tightly to it, the helmetless rider would be flung headfirst onto the road from a great height unless the pedal unscrewed instead.
The same logic applies to a drive-side Italian-threaded bottom bracket cup. Even though the drive crank turns clockwise and the right cup tightens in a clockwise direction, it will loosen up while pedaling if it is not super tight. That’s because the ball bearings turn the opposite direction of the crank.
We use your crank-reversal solution on some e-bikes when the rider wants a crank length we can’t get in e-bike cranks. In that case, we use a Direct Drive crankset (which has a lobed interface with three or eight bolt holes to attach the spider or single chainring) and put the right one on the left and the left one on the right (see photos).
We can’t put the right crank on the right, because the spider interface would interfere with the motor sprocket lockring; unlike on an “analog bike,” the chainring is not connected to the crankarm. There is room for the right crank’s spider interface on the left, though, so interchanging the crankarms works like a charm. As you did, we put symmetrical pedals (flat pedals or Crank Bros. clip-in pedals) on them so we don’t have to interchange the spindles in the pedals (which would also work). But, as you learned, we tighten the pedals very tightly so they don’t loosen up while riding.
― Lennard
As the comment Robert Morris says, mechanical precession is unrelated to bearing rotation. That wiki link illustrates it nicely.
I credit Jobst Brandt posts on rec.bicycles.tech decades ago for my first understanding of precession. Brandt claimed that precession forces were so great that no amount of tightening, threadlock, etc would keep (backwards) pedals from loosening. That was surely speculative, as I can't imagine he measured precession force. It would seem to depend on the tolerance in a given pedal/crank threaded joint (which varies), applied load, # of cycles, etc. Your use of backwards cranks super tightened suggests that enough applied torque will work.
Pedal threading and precession are widely misunderstood in bicycle engineering. I was an engineering expert witness in a lawsuit in which an ill-designed pedal/adapter system from Bell Sports caused some horrific crashes when pedals unscrewed. Like the designer of those pedals the PhD ME experts for the defense also failed to understand precession. It's not good when injuries result from such misunderstanding.
I've looked for info (patents, period literature, etc) on the origins of pedal threading, including the mention you have of high-wheel bicycles -- but never found anything. If you have I'd love to see it.
For a nice illustration of this effect see:
https://en.wikipedia.org/wiki/Precession_(mechanical)
The effect is due to rotating radial forces and has nothing to do with bearing friction.
Bob Morris