The bicycle as we know it came together from the innovations of many inventors, such as the inventor of the roller chain, and first person to put a crank on a Dandy Horse, the first person to invent pneumatic tires, and the first person to adapt ball bearings to bicycles wheels.
But enough of these features came together in one machine in 1885 to form a device we would recognize as a bicycle. It was built by John Kemp Starley of England. John moved to the big city, Coventry England, to work for his uncle, James Starley. James was an inventor, and was in the sewing machine business, had perfected the penny farthing high wheel bike, and had invented the first tricycle, which was sold as the Rover.
John Starley built an improved Rover, which was a two wheeler with a chain drive on the rear wheel, equal sized wheels, diamond shaped tubular frame, tangential spokes, ball bearings in wheels and cranks, and pneumatic tires. It was a truly modern bicycle. The photo below is of John Starley’s Rover of 1885. Other early bikes were Isaac Johnson’s folding frame bike, and Harmon Moise’s bike with a freewheel, both of which came after Starley’s Rover.
The Rover company also experimented with motorcycles, and also started a car company. The Rover Motor Car Company went on to build Rover cars, which run from luxury sedans to the famous Land Rover and Range Rover.
August Schrader immigrated from Hanover, Germany, to New York in 1843. Within a few years he started a small company making brass fittings for the rubber industry, which had been started only a few years before.
In l890, pnuematic tires were in use on the bicycle racing circuit, and soon bikes with pnuematic tires began winning the races. A tire manufacturer asked Schrader to design a better air valve than the one they were using, and Schrader did so. Schrader and his son George applied for a patent on their design in 1893, and made many improvements over the years. Every car today uses Schrader valves to keep the air in the tire, whether tube or tubeless tires. Most bicycles today use Schrader valves, with certain tubes using an alternative valve, the Presta valve. The Schrader valves used today are very similar to the 1893 version.
Spoked wheels have been around for millenia, and have reached a high degree of refinement in bicycle wheels. Bicycle wheels must operate for years with no maintanance, support hundreds of pounds of weight, and be as light as possible. There are a number of possible spoke configurations, but the one that is used on most bikes is the “triple cross” tangential pattern. If you look at bike wheels, the spokes are attached to the hub tangentially, and each spoke crosses another spoke three times, hence the “triple cross”. This tangential spoke pattern was invented by renowned bike designer James Starley, and was patented in England in 1874.
When properly made, each spoke has a certain amount of tension on it when unweighted, pulling the hub towards the rim with a force of about 50 pounds per spoke. When the wheel is weighted, by weight pressing down on the axle, the hub tries to move toward the ground. When it is does this, the three spokes directly below the hub become unweighted, and the rim deflects a small amount, about .001 inches. That is about the thickness of a piece of paper. The tire around the rim deflects a good deal more than that. Although the three spokes under the hub are unstessed, the other remaining spokes pull equally on the hub, and prevent it from moving in relation to the rim. As the wheel is weighted and rolled, the rim is thus constantly flexing at the spot under the hub, and the spokes are constantly being unstressed and stressed as they come under the hub. Below is an exaggerated view of the wheel under weight.
These spokes appear to cross more than three other spokes because the spokes on both sides of the wheel are shown. Each spoke only touches and crosses three spokes on its side of the wheel. When a rider hits a hard object, such as railroad tracks or a curb, the spokes most likely to break are the spokes that are above the hub and closest to vertical, and they are most likely to break where the spoke head enters the rim hole. The hub basically shears them off the head of the spoke.
This information has almost no practical value, but isn’t it fascinating! Of course some sites go into the physics of things like spoke strain and hub deflection, such as Held by Downward Force by John Forester, and Henry P. Gavin’s paper on Bicycle Wheel Spoke Patterns and Spoke Fatigue.
This how-to is aimed primarily at the Shimano-made cartridge-bearing rear hubs used in the majority of Catrikes for the last several years. Speeds, Expeditions and 700’s used the Shimano “Deore” branded hubs mostly, and those have cup/cone hubs, making their bearings easily adjustable for play or preload. Roads, Trails and Pockets used primarily the “Catrike” branded hubs that had cartridge bearings in them. There are many exceptions to which hubs were used in all of the models, but this is what will be found in the majority of cases. Other brands of cartridge-bearing hubs may be the same.
Replacing the bearings in a rear wheel hub that uses cartridge bearings is pretty easy, but not without pitfalls. If done correctly using the right tools, it should only take about five minutes.
Almost all of Catrike’s hubs, either cup/cone or cartridge, came with rubber conical boots over the ends of the axles. These are critical on cup/cone bearings because they keep the dirt out of the bearings. However, they are not necessary on cartridge-bearing-hubs as the bearings have their own seals. For simplicity I have left the rubber covers out of any pics. I also do not run those covers on my own cartridge-bearing-hubs.
There are two problems that can complicate removal of the cartridge bearings. First, some of the 130mm wide hubs have the inner lock nut on the drive side (the side where the cassette is) recessed so far inside the freehub that you cannot get a wrench on it. The second problem is being sure to remove the washer that spaces the bearing from the axle’s shoulder on the non-drive side. I will address each one of these problems separately.
First, the picture below shows the inner lock nut so far inside the freehub that you cannot get even a thin cone wrench on it.
The above is on a 130mm wide hub. On a 135mm wide hub, the nut would not be quite so recessed. A good, thin cone wrench can grab those flats on the 135mm hub, but on the 130mm hub, you must first back the inner lock nut off just a bit on the non-drive side (picture below), ….
…then, using a rubber or plastic-faced hammer, or a press (for ceramic bearings), drive the axle toward the drive side enough to expose the inner lock nut (picture below) so it can be loosened.
This is how to remove the drive side lock nuts first. It is recommended to use a good quality cone wrench (By Park Tools or Pedro’s, for examples) on the inner lock nuts because they are properly hardened to take the abuse of such a thin wrench being used to loosen and tighten these nuts.
Some of you may say, “Well, just remove the non-drive side nuts completely first and drive the axle out toward the drive side.”. This brings us to the second problem. There is a small washer between the non-drive side bearing and a shoulder machined on the axle. The picture below shows the whole assembly laid out as it assembles, with the non-drive side to the right in the picture.
There is no washer on the drive side. The Picture below shows this washer on the non-drive side, and the next picture below hows no washer on the drive side. This washer MUST be kept oriented exactly as it is from the factory!
Non drive side with washer
Drive side with no washer.
The two pictures below show that the washer is not symetrical. One side is convex and one side is flat. The convex side also has a small radius on it’s inner edge that matches the radius next to the shoulder on the axle (picture 10). If it gets flipped over during reassembly it will cause serious problems. Mostly, it will space the bearings farther apart, and this means the bearings’ outer race edges will no longer seat properly against the hub body’s bearing pockets. Under load, the bearings will “walk” side-to-side, slowly wearing the pockets larger until the bearings will be sloppy inside the pockets.
The next pictures show the groove for the washer, the washer assembled correctly on the axle, and a cross section of the axle and the washer, with exaggerated curvature of the washer.
The picture above show the distance between the bearings being 3.224″, which is correct for this hub, axle, and bearing assembly.
The pictures below show the washer and axle assembled with the washer reversed, showing a distance between bearings of 3.236″. For precision bearings, this a big difference.
It should be apparent to the reader that this washer MUST be kept oriented correctly during reassembly. The problem is that if you remove the non-drive side lock nuts first and drive the axle out toward the drive side, that washer will then be floating inside the hub body and can be flipped around freely. It is important to keep the axle, washer, bearing and inner lock nut together until the axle is removed and the washer is examined for correct orientation.
Again, the reader may ask, “Why not just place it with the convex side away from the bearing all the time, then?”. Because sometimes Shimano flips these washers themselves to obtain a correct spacing between the bearings to suit the hub body they will be going into. Both the axle and the hub body have tolerances that are kept within limits during manufacture, but the axles and bodies still have to be matched somehow, and these spacers (washers) are how they make everything fit correctly. Hence, we run into another small problem. Hub bodies, axles and washers must be kept as a “set” to be sure of a correct fit for the bearings, and that washer must be installed the same every time! If the washer is replaced with one that is thinner, you could actually put so much preload on the bearing to burn it out shortly. Ceramic bearings have been crushed from too much preloading due to replacing that washer with a thinner one!
Once the non-drive side bearing, washer and lock nuts are removed from the axle, the axle can be used to drive the drive-side bearing out of it’s seating.
Once the bearings are removed, if you want you can pop the coverings, clean out the old grease, put in new grease and reinstall the old bearings. We have experience with relubing ceramic bearings in this way with no ill effects, through at least 3 relubes. See how to do the relubing step in this article.
On reassembly, it is easiest to start by assembling the drive side. Slide the bearing onto the axle, then put the inner lock nut on. Using both lock nuts on the non-drive side, tighten them together for use to hold the axle still while you tighten the inner lock nut on the drive side against the bearing. Then, tighten the outer lock nut against the inner lock nut. The drive side is done. Take both non-drive side lock nuts off the axle. Insert the axle from the drive side. It should now look like the picture below on the drive side, and the second picture below is how the non-drive side should look.
Drive side of hub
Non drive side of hub.
No need to drive the bearing into it’s pocket at this time. It will slowly be pressed it in as we tighten the inner lock nut on the non-drive side. Now, install the washer (picture below) the bearing (second picture below), and the inner lock nut (third picture below).
Using this lock nut, and a wrench on the outer lock nut of the drive side, tighten this inner nut (picture 18) to draw both bearings into their pockets until this nut is tight against the bearing. Now install the spacer (picture below) and outer lock nut (second picture below) and tighten everything on the non-drive side.
Be sure that the amount of axle threads showing is even from one side to the other, and that neither side extends outside of the dropouts of the trike/bike. On cartridge bearing axles, the shoulders that the bearings ride against will almost always keep this axle protrusion correct if the correct bearings, original nuts and spacers are used.
To remove a “current generation” of Shimano’s Hyperglide cassette from the freehub of a rear wheel you will need two rather specialized tools besides a 12″ adjustable wrench. First, the lock-ring that holds the cassette on the freehub has 12 internal splines, as can be seen in the first picture.
Park Tool’s #FR-1 or #FR-5G will fit this lock-ring. FR-5G has an alignment pin that makes the job of loosening the lock-ring a lot easier. Without the pin, the tool is hard to hold in place while also holding the cassette with a chain-whip. Using a quick release skewer lightly snugged up will also hold the tool in place. In the second picture you can see FR-1 on the left and FR-5G on the right. The third picture shows FR-5G in place for loosening the lock-ring.
The next tool you will need is some kind of device to hold the cassette from rotating backwards (counter-clockwise) as you loosen the lock-ring. Large ChannelLock-type pliers will not do it as they tend to bend or nick the teeth of the cogs of the cassette. An old piece of chain used with a pair of Vice-Grips will hold the cassette well enough for loosening. The best way to hold the cassette, however, is to use Park Tool’s SR-1 chain-whip, or equivalent, positioned as seen in picture four, to the left.
A 12″ adjustable wrench, a 1″ open-end or box wrench, or even a 26mm open-end or box wrench will turn the FR-5G tool to loosen the lock-ring, as seen in the fourth picture.
Holding the chain-whip, turn the adjustable wrench counter-clockwise to loosen the lock-ring. The lock-ring and the smallest cog of the cassette have serrations on their mating faces to help prevent loosening while riding, so you will hear some clicking as you loosen the ring. Once it is loose, remove the chain-whip and wrench and also the ring. It should look like picture five.
Now, simply lift the cassette off of the freehub, but be careful to hold onto the two smallest cogs of the cassette because they will not be firmly attached as the rest may be. Take special care to notice how they fit onto the cassette for replacement later. Notice how their flanges face, and how they also have a larger spline tooth and smaller groove like the rest of the cogs on the cassette. These must all line up.
Most cassettes are held together by a single screw that keeps all but the two smallest cogs together for easier handling and assembly. Some cassettes have all but those two smallest cogs rivetted to an aluminum spider, too! But, some cassettes have all of their cogs and spacers simply assembled onto the freehub’s splines one-at-a-time. Be very careful about keeping everything in order and be sure you know exactly how all of the spacers and cogs fit onto the freehub. Each spacer has two small pins that MUST fit into the correct holes of the cogs. If they don’t, they will hold two cogs just slightly too far apart and that will ruin the alignment that allows for index shifting. If you are not absolutely sure you know how all of the spacers and cogs fit together, then if they all feel loose as you try to lift the cassette off the freehub, stop and take it all to your local bike shop for servicing.
The freehub with the cassette removed should now look like picture six.
The majority of freehubs are made from hardened steel. Some road freehubs were made with aluminum spline shells to help save weight. Once the cassette is finally removed, these softer splines will usually show where the cogs have dug into the edge of the spline over time and under heavy loading. It will also make it hard to remove the cassette on these freehubs.
To replace the cassette on the freehub, just line up the large spline tooth of the cogs with the large spline groove in the hub’s shell and slide the cassette onto the hub. Be careful to get the flange of each of the last two cogs facing the wheel. Screw the lockring on and tighten it with the special tool and wrench. It doesn’t need to be “stand-on-it” tight. Park Tools lists it as needing about 260-434 in/lb. (about 21.7-36.2 ft/lb). Since the freehub’s ratchet will stop the cassette from turning clockwise as you tighten the lock-ring, you do not need the chain-whip for installation of the cassette.
The Wright brothers’ Van Cleve mark lives on in a modern namesake, the Van Cleve bike built by Cycles Gaansari of Springboro Ohio. Here is what Gary Boulanger of Cycles Gaansari adds:
Much is known about the Wright Brothers’ aviation results, but little has been told about how the men designed and tested their theories, and how big a role bicycle technology played in their research and development. Like most self-sufficient and frugal bicyclists, the brothers scrounged discarded bike components to make something useful out of something lying around the shop. In this case, it wasn’t a fixed gear or townie bike, but the airplane that was created, born from Wilbur’s vision for flight in the 1890s.
Cycles Gaansari was born from the need to provide reliable service, durable goods, and exciting products to the Greater Dayton cycling community. We’re housed in a former livery stable/barn built in Springboro in the 1850s, just three miles south of the Wright Brothers Airport, and across the street from the Jonathan Wright House, now a popular bed & breakfast, built by the founder of Springboro in 1815.
To many, the bicycle is a tool for transportation, adventure, freedom, and recreation. Little did the inventors of the bicycle know what impact they’d have on millions of people. Then again, little did two bicycle manufacturers from Dayton, Ohio realize where their dream of manned flight would catapult both them and the fruit of their labor.
James Starley’s Rover of 1885 was the first successful bike in which pedals and a crank drove the rear wheel with a chain, but he was not the first with that design. In 1879 Englishman Harry Lawson designed and patented a version of a large front wheeled bike with a smaller rear wheel driven by cranks and a chain. Lawson’s bike was not very well received, and he went on to design bikes using levers for power transmission. The Bicyclette was a commercial failure, but he had hit upon a superior design feature.
This artwork of the Bicyclette is a version featured on cigarette cards. This and other bicycle art is found at bicyclegifts.com. Framed versions of these beautiful cigarette cards, posters , cards, and other bicycle art recognize that brilliant design is art.
Bruce went crazy with his drill press, and removed, if I read his notes correctly, about 13.7 pounds from the normally 30 pound Catrike Speed! This is Catrike #CS754, named Holey Spokes.
Now we need to see that thing assembled, a final weigh in, and a test ride to see if it whistles. Its just remotely possible that Bruce has too much time on his hands. One last picture:
Those old bike designers tried a lot of ways to cushion the ride of the safety bike on the rough roads found at the end of the 19th century. Here is a different way to employ springs on the front forks to cushion the ride.
Mikey says: Toe is a measurement of the horizontal diameters of the two front wheels, and how close to parallel they are to each other when the wheels are pointed straight ahead. If they’re not parallel, the wheels either “toe in” (the fronts of the wheels are closer together than the backs of the wheels) or “toe out” (backs are closer). I’ve seen people use a tape measure, piece of string, Catrike flagpole, trammel points, framing squares, or (last resort) a special tool. Once you’ve got a way to measure the inter-wheel distance front and rear, just fiddle with a tie-rod end to tweak the wheels to make the distances equal, or at least within about 1/16″ of each other. Loosen the stop nut, disconnect the tie-rod end from the bracket on the wheel, turn the rod-end in (adjusts the toe “out”) or out (adjusts the toe “in”). The finest adjustment you can make is a half-turn of the rod end, so you may have to settle for a tiny bit of toe. Some people prefer toe-in, others toe-out if they can’t make it exactly neutral, and some people don’t want it exactly neutral anyway. Once you’re happy, reconnect the rod-end to the bracket, tighten the stop nut, and go riding.
trikebldr adds this about toe:
Basically, toe-in gives better stability at higher speeds, but higher tire wear. Toe-out will give more nimble, power-steering feel, with more tire wear. Neutral toe will give a balance of the two, with very little tire wear.
To explain the stability issue, think of it this way; with toe-out, each wheel is trying to pull the trike in it’s direction, and when you try to go straight and hit even a tiny bump with one wheel, that wheel gets a traction advantage over the other and begins to turn the trike it’s direction. As it does so, the weight advantage transfers to the other wheel. And, this cyclic action happens over and over, making the trike feel very unstable. Example: left wheel hits a bump, gets more traction than the right wheel and the trike starts to turn to the left, shifting the weight over to the right wheel. Now, the right wheel has more traction than the left, so the trike starts to turn to the right, shifting the weight over to the left wheel. This cyclic action happens over and over, creating an unstable, unpredictable feel for the rider. It manifests itself as a tendency to wander from side-to-side.
With toe-in, when one wheel gets a weight advantage and tries to turn the trike, that action only adds more weight to that wheel and nothing changes, giving a feeling of stability. No wandering!
Absolute neutral toe tends to feel more like toe-out at very high speeds, so just a touch of toe-in is preferable for most riding conditions, but not so much that it causes tire wear. That’s where it gets tricky! I run all of my trikes at zero toe when unloaded (in the work stand), and that gives it just a hair of toe-in when I sit on the trike. My original set of Stelvios on my ’07 have over 9000 miles (YES! NINE THOUSAND!), and are still useable. I replaced them only because Tickle Pink was going to be riding that trike during the rally week this year, and I wanted it to be absolutely trouble free. I think this setting has well proven to be optimal for at least my ’07 Speed. Tire wear on my ’08 looks good so far, too, with this setting, and it has over 2000 miles so far.
Here’s a little bit about setting toe. The trueness of the wheels can kill an adjustment completely!!! If each wheel wobbles even 1/32″, and they are in just the right position to each other during the toe setting, that could affect the setting by 1/16″, and that is all I would recommend as a maximum toe-in value. So, even if you actually have neutral toe, it COULD look like 1/16″. The only way to eliminate that is to bind the wheels slightly by adjusting the brake pad just enough to hold the wheel from spinning freely, but allow you to rotate them during this operation. Now. working from the right side, take your toe reading, front and back. Write it down. Now rotate only the right wheel 1/4 turn and take the toe reading again. After doing this at the four “corners” of the right wheel, rotate the left wheel 1/4 turn and start over on the right wheel. This means you will be taking sixteen readings total to see how much your wheel trueness affects the real toe setting. An average of all sixteen values will be a very accurate reading!
This is all very tedious, but if you like a very accurately tuned machine, it is worth it. And, once done to this accuracy, it shouldn’t change unless you take things apart or bend something. My ’07 never changed, and it was never apart in over two years until I recently took it completely apart for “surgery”. My ’08 is now almost 18 months old and has also never changed.