Troubleshooters.Com and Steve Litt's Guide to Transportational Bicycling Present

   Overhauling a Coaster Brake Bike

Bicycle icon

Copyright (C) 2006 by Steve Litt

See the Troubleshooters.Com Bookstore.



The information in this document is information is presented "as is",  without warranty of any kind, either expressed or implied, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose. The entire risk as to the quality and performance of the information is with you. Should this information prove defective, you assume the cost of all necessary servicing, repair, correction or medical care.

This page discusses the maintenance of coaster brake bicycles. A loss of braking power could lead to death or severe injury, especially on hills or in traffic. Your coaster brake may not match the one described in this document. This document may not clearly express the assembly of the coaster brake it describes, or may even contain errors. We are not responsible. If you use the information in this document, you take full responsibility for the outcome. If that is not acceptable to you, please do not read this document.

In all cases, I strongly recommend you install a front brake so if the back brake goes out, you can still stop. If such a front brake is installed, it must be installed in such a way that it will not fall off, which in itself could cause death or severe injury (through flipping the bike).

In no event unless required by applicable law or agreed to in writing will the copyright holder, authors, or any other party who may modify and/or redistribute the information, be liable to you for damages, including any general, special, incidental or consequential damages or personal injury arising out of the use or inability to use the information, even if such holder or other party has been advised of the possibility of such damages.

If this is not acceptable to you, you may not read this information.


I'm Steve Litt. I created the Universal Troubleshooting Process (UTP). I create and license UTP courseware, as well as teaching the UTP onsite. I've written five books on troubleshooting: Past professions include software development, electronic repair, corrosion engineering, and bicycle repair.

I currently ride my bike about 30 miles per week. That probably doesn't sound like much to someone reading this website, but it adds up to 1500 miles per year -- enough to wear out a lot of components and even whole bicycles in a few years.

In 2006 and most of 2007, my daily driver bicycle was a one speed with coaster brakes. I chose such a bike for one reason -- it's always ready. My one speed needs very little repair, parts are easy to find, and it functions quite well with wobbly wheels.


You can get most of these benefits with a seven speed with coaster brakes using a Shimano Nexus multispeed hub. I now recommend using a multispeed bike as a daily driver, because doing so provides more exercise, keeping the rider in the kind of shape necessary to run twenty mile (each way) errands.

This document describes my experience overhauling the coaster brake hub of my bicycle. Note that most pictures on this page can be clicked to see a full sized version of the picture.


Here are all the things I needed. I made sure they were available before I started: A few comments. Be SURE you have a full can of engine cleaner. You might not use the whole thing, but you might. You'll for sure use a lot.

Theoretically you can make due with the engine cleaner alone and not use the "orange" cleaner. I found that my "orange" cleaner got some dirt left behind by the engine cleaner. Also, it's hard to get engine cleaner out of the can in sufficient quantities to really soak the parts. I used "orange" cleaner -- one could use any water based cleaner that really cuts grease, as long as it doesn't have grit. Hand cleaner is NOT ACCEPTABLE -- the grit in it will destroy bearings and bearing surfaces if even one piece of grit remains.

I got by without the needlenose -- using a screwdriver to disassemble the chain's master link, but it would have been easier with a needlenose.

Preliminary Evaluation

Before beginning work, evaluate the wheel as a whole. Is it missing a lot of spokes? Are a lot of spokes bent but not yet broken? Are parts of the rim flared or dented? Is the rim extremely out of true, to the point where straightening it would cause undue tension on some spokes and looseness on others? Are the teeth of the sprocket so chainworn that it detracts from your speed or causes the chain to jump?

Remember that a brand new 26" rear wheel with coaster brake costs somewhere between $30.00 and $50.00, depending on where you get it. Ask your local bike shop to see one. My alloy rimmed coaster brake rear wheel has a part number of 65312. I think that's a J & B Importers part number.

With the markup bike shops get on spokes and other parts, if your wheel has serious defects, trying to repair the wheel might waste time and save little money.


The main challenge in dissassembly is remembering how it came apart so it can be put together in the same way. Also, at some point the brake pads will fall out. The hot tip is to remove the wheel slowly so you can see how they fit in.

In my opinion the easiest way to start is to bolt the left side of the wheel to the outside of the right rear fork.

Mounting the wheel on the outside of the rear fork         To the left you see the wheel mounted on the outside of the fork, with the sprocket facing outward. This gives a stable working environment, and to some degree immobilizes the axle. I disassembled from the sprocket in. The next step is to remove the locknut and cone holding the sprocket on the wheel.

Removing the locknut
        I removed the locknut by placing the adjustable wrench on the locknut and the cone wrench on the cone. The cone provides a riding surface for the bearings. Because both the cone and locknut unscrew counterclockwise, the locknut is twisted counterclockwise while the cone is held steady. The locknut and cone are first separated, and then removed.

Locknut removed, locknut unscrewed
        Here you see the locknut is gone and the cone has been unscrewed to the end of the axle. The bearing has been removed to make it visible, and to reveal the bearing cone surface in the sprocket.

Removing the sprocket
        The next step is to remove the sprocket. Twist it counter clockwise, and it comes out. Once it's removed, you can see why twisting counter clockwise removed it. The back side of the sprocket has thick, course threads. In the picture on the left you can also see the bearing that rides between a cone on the back side of the sprocket and the cup in the hub. The sprocket and bearing can now be completely removed, but the wheel SHOULDN'T be removed at this time.

WARNING!!! When the wheel is removed, the brake pads will fall out. Once they fall out, it's difficult to figure out where they were placed, and their orientation. Therefore, the wheel should be removed VERY slowly while looking at the side of the hub closest to the fork. If removed slowly enough, the brake pads should be visible in their original configuration, before they fall. By placing one's hand below the brake pads, the brake pads' original orientation can be preserved, to some extent, when they fall out.

After wheel is removed
        This is what it looks like after the wheel has been removed and the brake pads have fallen out. You now see the axle, with (toward fork) the clutch assembly, the brakepad locking assembly (barely visible though the grease), the bearing, the torque arm, the left side locknut, and the fork. For the purposes of this document, we'll consider the brakepad locking assembly, the torque arm and the left side locknut to be a unit, because I did not separate them. They have no moving parts, so separating them would have made cleaning slightly easier, but would have lost the information concerning where to lock the brake locking assembly.

All the parts of the hub
        To the left is a picture of all the parts comprising the hub, except the wheel, which goes in the large gap in the center. The names of the parts are listed below:

The preceding picture shows all the parts. Several parts will later be shown individually for a closer look. For the time being, in the preceding picture, the parts are, from left to right:
  1. The left wheel nut.
  2. The brakepad locking assembly, which remains on the axel. To the left of the brakepad locking assembly is the tork arm and the locknut.
  3. The left bearing, which rides between the cone on the brakepad locking assembly and the cup on the wheel's hub.
  4. The brakepads and the clutch assembly. The clutch assemblyh is the part between the two brake pads.
  5. The wheel hub (not shown, but there's a large space where it should be.
  6. The right bearing, which rides between the cup on the hub and the cone on the sprocket assembly.
  7. The sproket assembly.
  8. The outside bearing, which rides between the cup in the sprocket assembly and the cone.
  9. The cone.
  10. The locknut for the cone.
  11. The right wheel nut.

How it works (quickstart)

The first thing to understand is that the course threads on the back side of the sprocket assembly thread into the course threads inside the clutch assembly. When you pedal forward, the threads pull the clutch assembly toward the sprocket assembly until the clutch assembly butts up against the lip inside the hub shell and transmits power to the hub shell.

When you pedal backwards, the clutch assembly is pushed away from the sprocket assembly. The brake pads are pushed up the slanted surface of the clutch assembly, causing them to rub on the interior of the hub. Because the brake pads are locked by the brakepad locking assembly, which is in turn anchored to the torque arm which is in turn screwed onto a fork stay, the brake pads cannot rotate, hence they stop the rotating hub as they press against it.

The brakepad locking assembly serves to prevent the brakepads from rotating along with the wheel when you put on the brakes. The outer right cone prevents the sproket from coming out when you put on the brakes, so instead the clutch assembly is forced in.

Detailed pictures

Here are some detailed pictures of some of the parts.
Outside surface of brakepads   The outside surface of the brakepads. The grooves help it grab the inside of the hub, and also shed impurities like grease.

Inside surface of brakepads   The inside surface of the brakepads. On one end, the ridge interlocks with the brakepad locking assembly, preventing them from turning with the hub and thus enabling braking. On the other end is a slanted surface over which the clutch assembly glides, so that the brakepads can be pushed out to rub the inside of the hub..

Brakepad locking assy with axle and torque arm
  Here's the brakepad locking assembly with axle and torque arm. You can see one of the two raised parts that interlock with the brake pads to prevent the brake pads from turning. Farther back you see the cone surface on which ride the bearings that also ride the cup of the hub shell. The following picture shows both raised parts:

Brakepad locking assy, straight on view
  Here's a straight on view showing the two parts that lock with the brake pads.

Side view of clutch assembly
  Here's a side view of the clutch assembly. On the left you see two little pawls that lock into the brake locking assembly. They're on a little "platform" that's connected to the main body via a spring loaded clutch mechanism. When the main body turns clockwise with respect to the "platform", there is a constant friction. It's that friction that enables the rider to pedal forward and "suck" the clutch assembly toward the sprocket assembly. Otherwise, the main body would just rotate with the sprocket assembly and go nowhere. Once the main body of the clutch assembly is engaged with the lip inside the hub shell, the main body rotates while the "platform" doesn't.

The clutch mechanism does not allow the main body to rotate counterclockwise with respect to the platform. That way back peddling quickly pushes the clutch assembly away from the sprocket assembly (one wouldn't want the clutch to rotate with the sprocket, even a little, or braking would be delayed).

The spring loaded clutch assembly also allows a moderate force to push the main body toward the platform. Such compression is necessary to achieve braking. More to the point, the spring's tendency to thrust the platform and body away from each other enables the rider to switch from braking to coasting, without peddling forward.

The little pawls on the "platform" always remain inserted in the brakpad locking assembly. The left and right movement of the clutch body is achieved by compression and uncompression of the spring.

Also visible in the picture is the slanted ribbed surface in the middle of the clutch body. In braking, as the clutch body is thrust toward the platform, the brake pads ride up this slanted surface and are thrust into the interior diameter of the hub shell, initiating braking.

Interior of the clutch assembly, showing course threads
  Here's an interior view of the clutch assembly. I highly processed this image in Gimp in order that you could see the threads. These threads mesh with the external threads on the projection on the back of the sprocket assembly so that when you pedal backwards, the clutch assembly is forced toward the left and pushes the brake pads out and into the interior walls of the hub.

Hub shell interior
  This is the hub shell, photographed through the left (wider) side. You can see the bearing cup around the edge -- this is where the bearing rides. There's another one on the right side.

The important part of this photograph is the difference in diameters between the wider left side and the narrower right side. The diameter change occurs abruptly at a lip inside the hub shell. This lip can be seen upon careful examination of this photo. You might want to click the photo for a full sized version. It is this lip that the clutch main body butts up against during forward peddling. The sprocket assembly transmits power to the clutch assembly, which then transmits power to the hub shell via the contact between the clutch main body and this lip.

Note that to make the lip even somewhat visible, I had to use Gimp to lighten the interior, which is why the interior appears much lighter than the ball cup, and appears almost as bright as the flange. If you look carefully you'll also notice it has less than 36 spokes. I took this picture after my 26" wheel was already rebuilt (and had 40 miles on the rebuild), and so I took a picture of a hub off a thrift-store girl's 20 inch bike my wife bought for $7.00, for our daughter who since outgrew 20 inch bikes,  leaving it available for cannibalization.

Back view of sprocket assembly
  Here's a view of the back side of the sprocket. Note the external threads, which mesh with the internal threads of the clutch assembly. Notice also that below (in this picture) the threads is a cone on which ball bearings race. The ball bearings run between this cone and a cup in the hub wall.

How it Works (the gory details)

This article describes a modern Shimano 1 speed with coaster brake hub. By "modern" I mean in the last 10 years or so. The Bendix brakes of my youth were built differently.

The hub can operate in one of three functional states:
  1. Propelling
  2. Coasting
  3. Braking
Propelling occurs when the rider pedals forward. Coasting occurs when the rider keeps the pedals stationary. Braking occurs when the rider pedals backward until the brakes engage, and then pushes harder or softer to achieve more or less braking, respectively.

A moment of thought makes it clear that the only way to get to the propelling state is through the coasting state, and the only way to achieve the braking state is through the coasting state. The rider must stop peddling one way before he pedals the other. The coasting state can be entered either from the propelling state or the braking state.

In my opinion, the best way to explain the functioning of a coaster brake is to examine how these states and how state changes occur. View this coaster brake state table, which might need to be viewed landscape on a small device.

Now that the steady states have been described, a description of state transitions is in order.

Coasting to Propelling

The rider turns the sprocket in the forward direction. This causes the projection on the back of the sprocket assembly to "screw into" the clutch main body, which in turn causes the clutch main body to move toward the sprocket. Because the clutch main body is coupled to the (always stationary) clutch platform by a clutch with some friction, the "screwing into" occurs quickly and surely. If the clutch were to lose its friction, the clutch body would simply rotate with the sprocket, and no screwing in would occur, and therefore either propelling would never occur, or would occur slowly and only with a great deal of pedal spinning. I had that happen once, and the problem was fixed by replacing the clutch assembly.

One other point should be made. While propelling, a part of the rider's energy is consumed by the clutch's designed-in friction. Even though the wheel might be able to spin an entire minute with the bike upside down and hand cranked, the propelling state consumes energy overcoming the clutch's intended friction. It's not much, and it's tiny with respect to wind resistance (and maybe even tire resistance), but it's there. If there's one design flaw in a coaster brake, this is it.

As the screwing in occurs and the clutch main body moves toward the sprocket, the clutch main body is stopped by the lip on the inside of the hub shell, and no further "screwing in" can occur. Further peddling simply turns the clutch body (via friction in the course threads), and the clutch body then turns the hub via friction with the lip.

This puts a rather large force on the bearing between the sprocket assembly and the right hub cup. However, because the sprocket, hub and clutch body all spin at the exact same speed, the balls in the bearing assembly don't rotate with respect to the cup or cone surfaces, thereby saving the wear and tear you would otherwise expect given the increased forces.

The balls that DO rotate are those in the small bearing assembly between the outside of the sprocket assembly and the outer cone on the right. However, during propelling, this bearing does not have increased force on it.

Propelling to Coasting

When the rider stops peddling, leaving his feet on the pedals, pedal rotation ceases. Therefore sprocket rotation ceases. However, the clutch body is still pressed tightly against the hub shell's lip, so it continues to rotate. This is the equivalent of rotating the sprocket backwards when the clutch body is stationary -- it "unscrews" and forces the clutch body farther from the sprocket. This in turn causes the clutch body to disengage from the hub shell's lip, so the clutch body no longer rotates. Once it no longer rotates, there is no further "unscrewing", so the clutch body never goes so far to the left as to engage braking. The bike coasts. The wheel continues turning, but both the clutch body and the sprocket don't turn.

One might question what would happen if friction in the system caused the clutch body to keep rotating with the wheel. It would take a lot of system friction to overcome the purposeful drag of the clutch mechanism. It would then take even more friction to overcome the sum of the clutch drag AND compress the spring between the clutch main body and the clutch platform. If there were that much friction, the brakes would become engaged to the extent of the system friction. In practice, I've never seen that happen.

Coasting to Braking

The rider pedals backward, "unscrewing" between sprocket assembly and clutch main body. Thus the sprocket assembly and clutch main body separate. Because the outermost right side bearing and cone prevent the sprocket assembly from moving to the right, the clutch body is pushed forcefully to the left. As the wider diameter part of the clutch body slides under the brake pads, the brake pads are forced outward against the rotating hub shell, achieving braking. The harder the sprocket is pushed backward, the harder the clutch main body is pushed to the left, and the harder the brake pads are forced outward into the hub shell.

The clutch mechanism does not allow the clutch main body to rotate counterclockwise (backwards) with respect to the clutch platform, which of course is held stationary by its prongs into the brakepad locking assembly. What this means is that the "screwing out" between sprocket assembly and clutch main body is sure and swift. If the clutch were to break so that the clutch main body could turn counterclockwise with respect to clutch platform, braking would be delayed (possibly requiring several backward turns of the pedals) or possibly braking would be completely disabled.

It should be noted that in order to move far enough left to push apart the brakepads, the clutch body must compress the spring between it and the clutch platform. This spring is important in the braking to coasting transition, described later.

To achieve braking, the brake pads must be held stationary so they don't simply rotate with the hub shell. This is accomplished by the brakepad locking assembly, which has two heavy duty raised surfaces to prevent the brakepads from moving. This in turn puts a huge torque on the brakepad locking assembly, which would cause it to unscrew, or possibly the entire axle might rotate. To prevent that, a torque arm is direct coupled to the brakepad locking mechanism, and the torque arm is fastened to a fork stay via a thin strap of metal (coaster strap) and a screw.

Braking to Coasting

The rider stops pushing backward. The clutch spring now pushes the clutch body away from the brakepads (toward the sprocket). This causes the sprocket slightly in the forward direction, thereby rotating the pedals slightly in the forward direction. Once the clutch body is pushed away from the brakepads, the brakepads no longer rub on the hub shell, and coasting is achieved.

The question could be asked "what if the rider doesn't let the pedals move in the forward direction?" The only way the rider could do that would be to apply a significant backward pressure, in which case, by definition, he would still be braking. In fact, riders intuitively turn the pedals forward very slightly to transition from braking to coasting.

If the spring between clutch platform and clutch main body became severely weakened, the spring would not push the clutch main body away from the brakepads upon cessation of backward pressure. In that case, braking might continue, at least to some degree, until the rider consciously pedaled forward. This would result in reduced drivability and could, under certain circumstances, be dangerous.

The Brains Are in the Clutch Assembly

As this article makes clear, the brains and personality of the coaster brake are in the clutch assembly. If the designed in forward drag is too strong, it will be hard to pedal (but still easy to coast). If the designed in forward drag is too little, the rider might need to spin the pedals forward several times to engage. If the clutch breaks so as to allow reverse rotation of the clutch main body, braking might require several backward turns or might not work at all. If the compression spring between clutch platform and clutch main body became severely weakened, braking might continue after the rider stops applying backward pressure, in which case the rider would need to consciously pedal forward to cease braking.

If your clutch fails, you can either get one out of a dumpsterized kids' bike or buy one new. Obviously, either way, the part must be identical and must be functioning correctly. Fortunately, most kids' bikes are thrown away because they were outgrown, not because they were ridden thousands of miles.

I haven't investigated thoroughly, but a casual inspection of appears to imply that they will sell you the entire guts of a Shimano E type coaster brake hub for $4.00 plus shipping. So if you finally wear out bearings or whatever, this might be an easy answer.


After disassembly, the parts are caked with gobs of dirt and grease. The big gobs can be removed with a screwdriver or a rag. Then comes the cleaning.

My favorite cleaning method is to use engine cleaner. At about $5.00 per can, it's cheap enough to use occasionally, even though this job will probably require almost half a can. Engine cleaner foams and sticks, so it can be sprayed on, and then a toothbrush can be used to loosen hard to remove dirt. Afterward, the parts can be rinsed under a hose (careful not to lose a part) or in a large bucket of plain water. This procedure can be repeated as necessary.

I like my final cleaning step to involve a pure water-soluble cleanser such as "orange cleaner". However, as long as the parts have all reasonable dirt removed and all the engine cleaner is rinsed away, this isn't necessary. NEVER use a gritty cleanser such as hand cleaner, because it will scar bearing surfaces after a very short usage.

Post Disassembly Evaluation

As discussed in the Preliminary Evaluation section the point was made that if the wheel has serious defects requiring the purchase of parts, you're often better off buying a brand new wheel. Now that parts have been cleaned, inspect all parts for signs of wear. If the bearing surfaces of the hub shell (cups) are pitted, a new wheel is in order. If cones are pitted, it might be possible to cannibalize these parts from other hubs. Lots of old cheap childrens' bikes have coaster brakes. It's absolutely essential, however, that the parts be identical, or else the brakes could fail at exactly the wrong time, putting the rider's life in danger.

If all parts are reasonably in good repair, or if identical good parts can be replaced through (inexpensive) purchase or cannibalization, the overhaul can continue. Otherwise, it's best to purchase a new wheel.

I haven't investigated thoroughly, but a casual inspection of appears to imply that they will sell you the entire guts of a Shimano E type coaster brake hub for $4.00 plus shipping. So if you finally wear out bearings or whatever, this might be an easy answer, always assuming your rim, spokes and hub shell are in excellent condition.


If the wheel is missing spokes, or if the some spokes are severely bent or gouged, this is the time to replace the broken or bad spokes. Without the sprocket in the way, installing new spokes is relatively easy. After removal of the tire, tube and rim strip, the wheel can be brought to a competent bike shop that can measure the existing spokes and sell correctly sized replacements (including new nipples).

When I replace spokes, I do them one at a time, following the path of the spoke I just pulled out. Wheels are typically "cross spoked", meaning each spoke goes under two and then above one, or vice versa. On four cross wheels it's under three and over one, or vice versa. Cross spoking makes the wheel more ridgid, which is a good thing for both speed and manuverability.

I always put a little grease on the spoke threads, the part of the nipple that sits on the rim, and the spoke head. The grease on the threads and the nipple make it easier to turn the nipple and true the wheel. The grease on the head of the spoke, in my opinion, lessens wear, stress and corrosion, giving the spoke a longer life.

Once again, remember that if spokes cost 75 cents apiece, and you need to replace 15 of them, a new wheel might be more economical. Interestingly, a couple days ago a bike shop charged me $1.49 per spoke. I was surprised, but gasoline being what it is (remember, my bike was out of commission), I didn't want to drive to a different bike shop to save a couple bucks.

Dry Fitting

Before final reassembly, I always make sure to put it together dry, without any grease. This gives me several advantages:

It's easy to do and do over

When reassembling anything, I often screw it up the first, second or even third time. By reassembling several times, I understand how it works, plus I can try it out. When I reassemble the wheel without grease, I can turn the sprocket forward and verify that the wheel turns. I can turn the sprocket backwards and verify that the wheel quickly stops.

I found that I can dry assemble and unassemble the hub in about 7 minutes. That means I can do it three times in 21 minutes. That 21 minutes is time well spent, because when I put it together with grease, I do it correctly the first time.

It's much less messy

Performing a trial assembly with grease would be a messy affair to say the least. Dry assembling it is neat and clean.

You see baseline performance without grease

My bicycle repair philosophy is usually "there's no such thing as too much grease". After all, I can always wipe off the excess after assembly.

This philosophy is NOT appropriate for coaster brakes. Because the brake pads ride so close to the hub wall, excessive grease between the clutch and brake pads could actually push the brakepads lightly into the hub, creating drag. Excessive grease between the brake pads and hub wall could also create drag, and could also decrease braking effectiveness.

Putting the wheel together dry gives an opportunitye to see how it operates without grease. If there's excessive friction after final assembly with grease, and the cones are not overly tight, then the grease is slowing the wheel. Dry assembly yields a baseline to diagnose problems caused by excessive grease.

Final Reassembly

I use voluminous grease on bearing surfaces. I don't want these running dry. I use a VERY light coat of grease on the outer side of the brakepads -- just enough so that the brakepads don't quickly cut up the interior of the hub. I use a light to moderate amount of grease between the clutch and brakepads, so that it won't run dry and cause notches to be worn in the slanted sliding surfaces.

Here's the reassembly method I use:
  1. Lightly grease all the exposed threads on the axle. This helps in all reassembly.
  2. Moderately grease up the cone on the brakepad locking assembly, avoiding getting grease on other parts of the assembly.
  3. Completely inundate one of the large bearing assemblies with grease, placing it on the brakepad locking assembly's cone with the ball side facing away from the torque arm and the holder side facing toward the torque arm.
  4. Grease the threads inside the clutch assembly with a moderate amount of grease.
  5. Slide the clutch onto the axle with the two little prongs toward the brakepad locking assembly.
  6. Push lightly and rotate clutch until the two little prongs go into the two mating holes on the brakepad locking assembly.
  7. Put a light to moderate amount of grease on the brakepad locking assembly and the exterior of the clutch.
  8. Cover the outside of the brakepads with a VERY light coat of grease -- just enough so that metal on metal contact won't cut up the metal surfaces.
  9. Fit the brakepads over the brakepad locking assembly and clutch so that the heavy metal on the interior of the brakepads fit between the raised metal on the brakepad locking assembly. This configuration keeps the brakepads from spinning while braking.
  10. Very slowly, gently and carefully fit everything into the wider end of the hub. After it goes in, gently push and rotate until the bearing completely seats and bearing cannot be seen. If it doesn't go in right, or parts appear to become dislodged, remove, reseat the brakepads, and try again.
  11. Put the wheel down on the axle so that the wheel is kept in place by gravity. Put the axle on a soft surface so the threads aren't damaged.
  12. Put moderate grease on the cone at the back of the sprocket and the cup on the front side of the sprocket. Put moderate grease in the course threads on the projection at the back of the sprocket.
  13. Heavily grease the second large bearing assembly and put it on the cup part of the sprocket assembly so that the balls face away from the sprocket and the holder side faces toward the sprocket. The idea is for the full balls to run on the cup in the hub.
  14. Holding the tork arm so the guts don't fall out of the hub, push the sproket assembly, projection first, over the axle and into the hub. When it meets resistance, turn it clockwise and it will go in further. When it goes as far as it can, once again put the wheel down so that the torque arm side of the axel rests on a soft surface, keeping the wheel snugly over the clutch, brakpads and brakepad locking assembly.
  15. Heavily grease the small bearing assembly, and place the balls side into the cup in the center of the sprocket assembly.
  16. Moderately grease the cone, and screw it down into the ball bearing assembly to the point where there's no axle play.
  17. Screw on the cone's locknut, hand tightening it onto the cone.
  18. Place a cone wrench on the cone, and an adjustable wrench on the locknut, and tighten the two against each other.
  19. Test for play. There should be just the tiniest amount of play, like maybe a half a millimeter at axle's end. As with all cup and cone bearing arrangements, one walks a tightrope between tightening it to binding and loosening it so there's too much play to true the wheel.
  20. If there's too much play, or serious wheel binding, loosen the locknut away from the cone, readjust, and retighten. Do this until there's the right amount of play (almost none).
  21. You're done with the hub rebuild.

True the Wheel

Now that the hub has been adjusted for minimal play, the wheel can be accurately trued. This involves first loosening spokes on the side the rim should move away from, and tightening them on the side the rim should move toward. Spokes are loosened and tightened by twisting the spoke nipples. This can be done with a spoke wrench, or in a pinch, a very carefully adjusted adjustable wrench.

Spoke nipples should never be twisted more than 1/2 turn at once. The wheel should be spun, and the part exhibiting the most sideways displacement identified, and then the spokes around that part should be adjusted. Then the wheel is spun again, and the procedure repeated.

One should never try to adjust more than one part of a wheel at a time -- that will lead to an ever more egg-shaped wheel in the hands of all but the most competent wheel truers.

Wheel truing is one of the hardest tasks in bicycle maintenance. Few can do it well, and even amongst those who do it well there's a spectrum. I can true a wheel to maybe 1/8 inch, which is great for transportational bikes but is not acceptable for racing bikes. Wheelmakers can true a wheel to the point where wobble cannot be perceived with the human eye.

Without knowledge of wheel truing, the best move is to pay a bike shop to true it. Those wanting to learn to true wheels, can learn the skill by practice on old discarded wheels pulled out of dumpsters. A detailed treatise of wheel building and truing is beyond the scope of this document.

Reinstall the Wheel

Obviously the wheel must be reinstalled before the bike can be driven.

If the tire, tube and rimstrip were removed, they should be reinstalled. Most tires are intended to roll in a certain direction, so care must be taken that they're installed in the proper "polarity". Care must be taken in inflating the tire that there is not a blowout. As the tire is slowly inflated (I do it with a gauge equipped hand pump), the tire should be checked to make sure no part is starting to "come out" of the rim. I often underinflate the tire by about 5 to 10 PSI, and then after my first ride I pump it up to the correct pressure. The correct pressure might be less than the max pressure printed on the tire, but should never be more if you want to minimize the chance of blowouts and the crashes they create.

The torque arm must be fastened to the chain-stay via a coaster strap and a safety screw that won't come loose. The nuts on both sides must be very tight so a quick and heavy push of the type used to out-accelerate a bus won't pull it loose. The chain tension must be adjusted to account for the eccentricity of the front chainwheel. No part of the revolution should tighten the chain so much that there's absolutely no slack, but  there should be no extra slack, to prevent the chain coming off. On a coaster brake bike with no front hand brake (I recommend installing a front hand brake as a "plan B"), a thrown chain means no brakes and disaster. The chain tension must be gotten right, and it must be checked often.


After any bike repair I test thoroughly and vigorously. If the repair was inadequate, I want it to fail in controlled conditions, not when I'm racing through the intersection of 436 and 434. In the case of a coaster brake bike, I repeatedly perform controlled stops, and then perform one skidding stop. I ride for several minutes. Then I go on a hill in a lightly trafficked area and brake there. I also test for easy rolling.


I've named many of my bicycles. My current one speed is named "Always Ready". With no derailleur worries, and minimal handbrake worries (the front can be adjusted fairly loose, after all, it's just a "plan B"), life is easy. My wheels needn't be all that true. Even a serious wheel pretzelling experience on a ride doesn't prevent my return home.

Always Ready certainly isn't the fastest bike I've ever had -- I'd need to spin 100 rpm to go 20 miles per hour, and over 120 to go 25 miles an hour. It has 46 teeth in front, 18 teeth in back, and 26" tires. I can't push it up a 4% grade. But for average riding around the city, if I'm not in a hurry, it's pretty good. And it works every time.

Top of Page