Picture of Carbon disc wheel for a track bike without custom tooling
This instructable describes how to make a one-off carbon track disc wheel without having to fabricate jigs, tools or moulds. A lathe and a drill press/milling machine is needed as well as some common composite lay-up supplies. Basic knowledge of machining and laminating is required.

The wheel is made up of a rim, two parallel side plates and a custom hub. They are joined using adhesive bonding. The flat side plates can be customised with, for example, patterns and text either by modifying the mould before laminating or by attaching decals etc. afterwards.

The project has a website at http://www.ideas2cycles.com and future versions will be added there.

Step 1: Design

Picture of Design
Selecting a rim

Almost any rim can be chosen as long as it has a two planar and parallel surfaces that the side plates can be bonded to. Features such as steel eyelets and double wall construction, deep cross section etc. only add weight. A cheap and light rim is recommended. Spoke hole count does not matter. Clincher and tubular rims both work. An aluminium clincher rim with machined side walls is used in this example.

Selecting material and manufacturing method for side plates

The side plates could be made from thin alloy (steel, titanium, aluminium, magnesium) sheet without a terrible weight penalty and with several benefits. However, using a low density composite laminate gives a better stiffness-to-weight ratio. Any reinforcement and resin system will work, but the best options are either aramid fiber ("Kevlar") or carbon fiber in an epoxy matrix. A laminate thickness of 0,6-0,9 mm is sufficient when using carbon fiber as reinforcement. Even thinner side plates are possible, but a core material may have to be added to ensure rigidity of the wheel and stability of the planar surfaces. A couple of plies of bi-directional carbon fiber weave works nicely. For the unidirectional pre-preg lay-up shown here, a [0/90/-45/+45]SE laminate was used. With eight plies the total thickness was 1,0 mm. 

The laminates can be made in several ways depending on available equipment. A regular wet lay-up requires minimal supplies, but a higher fiber volume fraction can be obtained by applying pressure on the laminate either mechanically or using a pressure difference (vacuum bagging, autoclave). No tooling or moulding is necessary since the side plates are flat. A non-porous planar surface, such as a sheet of metal, is sufficient. A table top could be used as well, but glass will result in the best surface finish. Mould preparation is the same regardless of material and manufacturing method and is described in the next step. Debossed features can be made into the laminate easily by placing thin two dimensional shapes on the mould before applying resin and reinforcement. A water/laser cutter or vinyl cutter is very useful for making debossed text, for example.

Designing the hub

A custom hub is required. The side plates are bonded to the hub and therefore special flanges are needed. The hub flanges must be as far apart as the selected rim is wide. If using, for example, a rim that is 20 mm wide, the outside distance between flanges must also be 20 mm. This way the flat side plates can be bonded using a uniform bond thickness. A three-piece hub was designed for a couple reasons. Firstly, the large diameter flanges can be cut from bigger stock, while the smaller diameter features can be cut using smaller stock resulting in less material wasted. Secondly,  the side plates can be joined to the hub not just by adhesive bonding, but also by clamping them between the hub bodies. Thirdly, the whole wheel is not ruined if the threads for the cog are stripped, because the threaded part can be detached and replaced. Finally, manufacturing is easier in the sense that the same program can be used for both sides of the hub (if using a NC-lathe) and the part does not have to be flipped. A drawing of the hub is included, but it does not have to look exactly like that. The only difference to a regular track hub design is the flanges. Spoke hole count should match the hole count in the selected rim, but not every hole on the rim needs to be populated. For example, a 32h rim can be laced with sufficient accuracy using 16 spokes, which means 8 holes on each hub flange. A hard aluminium alloy is recommended for hub material.
Hi guys, I have potential access to necessary tools/processes but would love to hear some of your ride feedback/durability feedback as I'd be making some to use while training for national level competitions. Have you used it for many miles since you made it? Has it been through racing conditions yet?
Hi! Sorry for the late reply. We haven't put that many laps on the wheel since it was built. For training purposes I'd get traditional steel spoked wheels and for comps a commercial disc wheel because it will have better performance than this design (mainly due to weight). There is no reason why this build could not be reliable, but a mass produced product is usually of higher quality compared to something built in a garage :) This instructable was intended as a way for people to ride something they customized and built themselves, not as a cheaper alternative to Mavics and Corimas.
AronC08162 years ago
What is the spoke count on this rim?
ideas2cycles (author)  AronC08162 years ago
We used a Mavic Open Sport 32h rim for this prototype but only used 16 spokes to true the wheel before bonding the sides onto it. You can recycle spokes from a disassembled wheel for this, since the length of the spokes is not critical when using a left side leading/right side trailing "radial" lacing pattern.
congrats on being a finalist. I liked your work despite it being well beyond me at this point
ideas2cycles (author)  veryrealperson2 years ago
The composite lay-up is something that everyone can do in their garage (or kitchen ;) ) albeit the materials are expensive. The hub is more difficult without prior experience and access to the right machines. I don't know how many DIYers are willing to outsource part of a project to a machine shop...
Excellent, I have wanted to know how this is done for a while.