When I was in college, I was introduced to composite design and manufacturing through our formula student project team- basically a miniaturized hybrid F1 style racecar. Since then, I’ve been continuously experimenting with carbon fiber design and manufacturing at home, and I’d like to chronicle my own experiments and provide a general overview of frame design and manufacturing constraints. Buckle up, this will be a long one.
The most logical place to start is with the material itself. There’ve been hundreds of articles just within the cycling industry about carbon fiber, so I won’t beat a dead horse- I’ll just provide an overview.
Carbon is one type of reinforcing fiber among many options- including Glass, Polymer (Aramid, Nylon, etc.) Basalt, even organics- I built a bamboo composite road fork for my senior project in College. It’s ubiquity is due in large part to it’s optimal mechanical characteristics for bicycle fabrication. Carbon comes in many different grades of fiber too, in a range of strength and modulus (stiffness). Generally speaking, super high modulus fibers are too compromised in other ways to be useful in the bulk of cycling applications, and due to this most bicycles are made of Standard or Intermediate modulus fibers. Here is a useful link from Toray, one of the world’s largest fiber manufacturers.
The other half of the composite equation is of course the matrix, or resin that holds the structure together. Currently, epoxy based thermoset resins are the high performance matrix material of choice across aerospace, automotive, and cycling applications due to it’s combination of mechanical performance and ease of processing. However, there are significant advancements being made in thermoplastic resins, such as PEEK prepregs, that have advantages in toughness, recyclability, and ease of automated manufacturing (my current area of expertise, actually…). There is actually some low volume cycling production here and here, but it is far from being ready and scalable.
There are many ways to process carbon cloth for structural applications, but these are a few common processes you will see in hobby building that are helpful to know.
- Wet layup- saturate dry carbon cloth with resin, place into a mold, then cure at room temperature under pressure. Can post-cure in an oven for more strength and high temperature stability.
- Infusion- similar to the wet layup process, except you use pressure to pull resin through dry cloth laid onto a pattern. Generally speaking, atmospheric infusion (VARTM) only works on single sided pieces, and is unsuitable for bicycle frames. In addition, dry cloth is more difficult to process by hand than wet fabric as it unravels easily and has no tack (stickiness).
- Prepreg- fabric that has resin already applied consistently in an optimal amount. This material is then placed in a mold and needs a specific temperature/pressure cycle to cure correctly (oven, autoclave, heat press, or heated tooling).
The other half of the processing equation is how you compress the laminate. Not all composites need compression, and in fact many types of composite are unsuitable- an example being Fiberglass Chopped Strand Mat (CSM) used in pattern making. Compressing CSM to remove resin and consolidate the fibers will actually reduce the laminate strength. However, almost any carbon fiber/epoxy composite performs best at approximately 70% fiber/30% resin, which is only achievable with a consolidated cure cycle (compression). The compression also helps to drive out voids, or air pockets, which create stress concentrations within a laminate and, over time under cyclical loading, will propogate into interlaminar cracking/planar voids, and eventually failure.
There are a multitude of ways to achieve pressure (many not listed here), and many composite parts use a combination of methods which are tailored to suit the individual locations in question.
- Bulk compression- using a tool/mold to create pressure directly. Examples include matched tooling pressed together (imagine two plates squeezing together) or using a mold with a different Coefficient of Thermal Expansion (CTE) than carbon fiber to create pressure when heated.
- Trapped rubber- a type of bulk compression that uses a hard outer tool and a silicone rubber insert with a very high CTE- can provide hundreds of PSI if needed.
- Compression tape- plastic tape that wraps around the part and is either elasticated for room temperature cure or shrinks in heat. This method only works with convex curvatures.
- Vacuum- uses the pressure of the atmosphere to push down on a laminate held under vacuum under a bag. It’s important to note that this can only achieve 14.7PSI maximum at sea level. This is one of the most common processing methods as it removes air from the laminate to cut down on void content. It is also what most hobbyists use (including me!)
- Autoclave- this method uses a pressure vessel in combination with a standard vacuum bag to add additional “atmospheric” pressure and consolidate the part with more force. Common in aerospace and F1.
- Bladder molding- an inflatable “balloon” is inserted inside a closed mold, then inflated with compressed air to press the composite out into the mold. This is the method most commonly used in the bicycle industry.
Now, we can move into the peculiarities of carbon fiber processing in the bike industry. The first, and perhaps most important thing about carbon is that it doesn’t like bends. It also doesn’t like holes. Planes, trains, cars, pressure vessels, spaceships, etc. all take this into account, and if there’s a part taking near isotropic loading that’s highly shaped, it’s usually made from aluminum, titanium, or steel. Bikes have all of these problem features, which makes using carbon for a bike significantly more complicated than it might seem given it’s prevalence in the marketplace.
The first constraint to consider with designing a frame is… machining? I thought we were doing carbon layups! One of the most important constraints on bike shaping is your tooling. Most tools are two piece clamshells, meaning that your parts need to avoid undercuts. Another constraint is radii of corners- which is dependent on whether you intend to use a female master or a male master mold.
There are two ways to make a female “closed” mold like the ones used for bikes- either machine directly into a block of material, or machine a master “male” part and then create a female mold from that piece. Most bike companies use the first method, which results in a strong, stable metal tool- with the compromise that they can usually only achieve 5mm corner radii. I’ve used both methods, and for a little bit more work, making a composite mold is much cheaper and can be very high performance.
The other thing to consider when designing tooling is the eventual mechanical interface points. For example, chainstays have to attach somewhere. With a closed mold and an internal bladder, only the external surface will be toleranced well and have a smooth finish. Therefore, to use a bonded lap joint to connect two parts, a manufacturer has to post machine and prep the inside of the bonding surface… which then begs the question- what hard points (e.g. a clamping surface) can be fixtured off of to make sure the cut is accurate? These decisions drive the number of sections a “monocoque” frame is divided into, the types of curvatures seen, and are sometimes in direct opposition to best practices related to composites processing or design goals, either stiffness to weight or aerodynamic performance.
On that note, Part 2 of this article will cover aspects of design, optimization, and a few sordid manufacturing failures. Stay tuned!
OTB Engineering 