January 01, 2015 4 Comments
[Note: this is #2 in a series of 4 about bearings]
Ball bearings are everywhere in your bike (you’d be surprised) but their greatest impact comes from four positions: hubs, headsets, bottom brackets, and pedals. Each is a different situation but low friction trumps other considerations.
IN HEADSET WE TRUST
Headset loads are huge and sideways to the bearing. But at the same time, a single-track vehicle requires low steering friction. Why? One of the most intriguing aspects of bike function is steering stability. To remain upright, it's necessary to steer gently in the direction of a lean, creating a gradual turn. The centrifugal force that appears with the turn is enough to support and even undo the lean. Ka-ching, stability.
To make staying upright nearly automatic, we have an autopilot system to initiate that turn that relies on:
(1) Gyroscopic force of the front wheel
When the bike leans, the wheel’s wants to turn in that direction. Instead of falling, we turn. But this urge would be useless if the headset interfered.
(2) Trail
The relationship between the steering axis of rotation and ground contact of the wheel is called trail. In a bicycle, as with casters on a shopping cart, ground contact is behind the rotational axis of steering. With forward motion, the wheel automatically prefers to stay straight.
(3) Moment of mass
It has recently been proved that the mass of the bike that is forward of the steering axis, induces a turn in the direction of lean. I touched on this topic here. We owe Jim Papadopolous and Andy Ruina of Cornell for the insight. More recently researchers at Holland’s University of Delft have provided physical testing.
These three effects are small compared to weight, energy, and outside influences when riding, so the headset must have very low friction to respond instantly. A free-running headset plus an ingenious system of stability makes riding upright easy. In fact, safer than walking for stablity. Falls while walking or running occur more often than when riding. An improvement to our evolved bipedal mobility?
A nasty test of the point is to over-tighten a headset. A wonderfully balanced and easy-to-ride bike becomes a jerky, uncooperative donkey. No, thank you. So, ball bearings are a good solution to the headset’s low-friction demands. But how do they deal with all that side force? The contact points for the balls are angled, rather than radial, as in hubs. This angle allows bearings to brace against side loads.
OPEN OR CLOSED
Bike mechanics all know the distinction between traditional "open" cup-and- cone bearings, and modern “closed" cartridge bearings. Today, the trend leans heavily toward the latter. But it wasn't always so.
A cup-and-cone system has two big advantages, One is retaining the balls: they're neatly captured as the cone moves toward the cup. The assembler can easily install them and then pick the amount of bearing play (looseness) by adjusting the inward movement of the cone. A modern cartridge bearing offers no such adjustability. To run with minimal looseness, the cartridge bearing must be made to very high precision, which is not an option for most bicycles.
The second big advantage of the cup-and-cone system is its high tolerance for misalignment. In the case of poorly made, bent, or worn elements, a cup-and-cone assembly can run smoothly where a cartridge system would seize. For 150 years, cup-and-cone bearings have proven their ruggedness and, for much of the world, they remain the best choice. In the hands of Shimano and Campagnolo, who continue to offer cup-and-cone hubs, this practical system approaches high art.
Why, then, are Cartridge bearings taking over? They have the advantage of mass production. By setting international standards, not only are ball sizes predictable, the entire mechanism is standardized. Mass production delivers huge economy as machine designers around the world share cartridge dimensions. For mid to high-level quality, cartridge bearings offer the most value and variety.
Maxicar (FR) and Phil Wood (US) were pioneers to install industrial cartridge bearings into hubs and bottom brackets. Chris King (US) was among the first with a sealed-bearing headset. Widespread use of cartridge bearings in the bike industry waited until the early 1980s with Specialized's sealed hubs that came on many of the earliest mountain bikes.
SEAL OF APPROVAL
However, these early cartridge-bearing components weren't exploiting their economy. Heck, these were products at the highest end. Ease of maintenance, “sealed for life,” weatherproof performance - this was the early dream of cartridge bearings. After all, they were known as "sealed" not "cartridge" bearings. Of course, such performance wasn't always possible, but the dream lived on. WTB developed Grease Guard to enable weather-challenged users to replenish the lubrication in "sealed" components.
To retain lubricant, which tends to leak, and to repel contamination, which brings rust and abrasive wear, bearings need protection. Shields, labyrinths and seals are all employed to extend bearing life. Shields fit so that only a very narrow gap remains through which contamination can enter. That works for large particles, but not for fine grit or moisture. Labyrinths interlock without contact and effectively provide a tortuous path to keep grease in and contaminants out. But they can be heavy and still not totally effective. Shields and labyrinths work best when used in addition to seals.
The most popular bearing seals are contact type. A very thin lip of urethane material makes light contact with the bearing's inner raceway. If the raceway finish is fine and the contact lightly lubricated, such a seal does a great job of repelling dirt and retaining lubrication. However, some friction is produced by the contact. But until recently, little attention has been paid to that.
FRICTION MYOPIA
For nearly a century, most engineers assumed that rolling resistance with ball bearings was negligible, about 1 percent. But bearings ran on oil when those early measurements were made. Modern bearings rely on thick grease, which adds rotational resistance. The efficiency estimate also fails to include bearing seals, since early bicycles used non-contact seals with the oil. In fact, this was a very effective system. Oil is a great, low-friction lubricant, but without contact seals, it leaks. That flow of oil did a fine job keeping contamination out, but it also made a mess of the bike's exterior.
Modern riders are only too glad to break the cycle of oily bikes (and stained trousers). Grease was a first step, followed by sealed cartridge bearings. Unfortunately, no one did the rolling-resistance calculation. A huge change was on the way.
Next Dirty Secrets will delve into a long period of stiff, low maintenance bearing options. Stay tuned for Dirty Secrets #3.
November 02, 2021
My experience has been that most cartridge bearing headsets are a total pain- either too loose, or too tight / too much friction. The exception being a Chris King which manages to be perfect, and never needs to be adjusted. I haven’t touched the one on my mountain bike for 10 years. Not sure how they do that. A loose bearing headset, on the other hand, can always be adjusted to perfection unless the races are worn out.
An interesting thing to ponder is that some claim that a needle bearing headset is necessary to avoid shimmy on bike with a low trail geometry. I have a Boulder Brevet that is wonderful in every way but I was strongly encouraged to install a Miche needle bearing headset. I did, and it doesn’t shimmy. But I wonder what would happen if I had opted for the King headset.
November 02, 2021
Ric – is it ok to adjust bearings so that there is a slight bit of play (not noticeable when riding)? Many have warned me about this practice because of safety concerns. I have always practiced this because it’s a safeguard against over-tightening and I have never broken anything and my bearings seem to last forever this way. Also many cartridge beairings are manufactured with specified “clearances”. Thanks again for Wheel Fanatyk.
Tom
November 02, 2021
Ric – I read the referenced Cornell report (all 63 pages), fully understood all concepts and priciples, checked and deduced all equations, and agree with everything. I thought I discovered a problem, but upon a fourth review I found I had an incorrect sign on one of the eigen values in a matrix. Just kidding! Happy New Year and thanks for Wheel Fanatyk! Tom
Comments will be approved before showing up.
Anonymous
November 02, 2021
Well done Ric. A nice piece. I remember a couple of my headsets that developed indexed steering! That was fun.