A bicycle, it goes without saying, is the best Christmas present ever, but with the possible exception of a puppy it’s also the most difficult to wrap. So pity the poor frame designer who has to wrap and drape carbon around its complex curves such that, when baked and finished, the frame delivers the desired ride feel. The construction of a carbon fibre frame is a complex 3D puzzle that eclipses the Rubik’s Cube.
The beauty of carbon is that, unlike metal, multiple pieces can be layered at varying degrees of intersection and overlap to give very tight control over the performance attributes and strength required at any given point of a bike frame. The downside is that carbon is anisotropic – it’s stronger in one direction than another in a similar way to wood – which means strength is dependent on the direction of the fibres. For carbon to carry significant loads the forces must be directed along its fibres, which makes fibre direction absolutely crucial. A bicycle frame’s constituent sections experience forces in several directions, meaning the carbon fibres must run in several directions too. It’s why different layers have their fibres at different angles, commonly 0° (in line), +45°, -45°, +90° and -90°, and indeed any angle chosen by the designers if it will create the desired attributes.
In the depths
That’s how it is for all carbon frames. Beneath the lustrous exteriors are many layers of carbon fibre pieces whose stiffnesses, strengths, shapes, sizes, positions and orientations have been painstakingly planned, usually by a combination of computer software packages and engineers’ expertise. This is known as the lay-up schedule, or just the lay-up. When the carbon jigsaw is completed the bike must be light, responsive, cost-effective and able to endure the most extreme forces of cycling.
Professor Dan Adams, director of the composites mechanics laboratory at the University of Utah in Salt Lake City, himself a keen cyclist and who was involved with the development of Trek’s first carbon frames, says that constructing anything from carbon is all about the correct lay-up schedule. ‘It specifies the orientation of individual plies or layers of carbon/epoxy prepreg, stacked up to make the final part thickness,’ he says. ‘Some frame parts are easier to lay up than others. The tubes are relatively simple but the junctions between them are some of the most complex ply lay-ups you’ll see in production parts in any industry that uses carbon structurally, including aerospace and automotive.’
Carbon’s anisotropic nature also makes choosing the right carbon crucial. At its simplest, there are two ways that carbon is supplied. Unidirectional (UD) has all the carbon fibres running in one direction, parallel to one another. The alternative to UD is a woven fabric, or ‘cloth’. It has fibres that run in two directions, going under and over each other at right angles to give the classic appearance of carbon fibre. In the simplest fabric, known as plain weave, the fibres lace under and over at every crossing (called ‘1/1’) to produce a grid-like pattern. There are many other possible weave patterns. Twill (2/2) is a little looser so easier to drape and easily recognisable by its diagonal pattern, which looks like chevrons.
The modulus (a measure of elasticity) of the fibre is also fundamental to a given lay-up. Modulus defines how stiff a fibre is. A standard modulus fibre, rated at 265 gigapascals (GPa) is less stiff than an intermediate modulus fibre rated at 320GPa. Less of a higher modulus carbon is required to make components of the same stiffness, which results in a lighter product. Higher modulus fibres might therefore seem like the preferable choice, but there’s a catch. An analogy can be made with a rubber band versus a piece of spaghetti. The rubber band is very elastic (has a low modulus) and can be flexed with very little force applied but will not break, plus it will return to its original shape after bending. The spaghetti, on the other hand, is very stiff (high modulus) so will resist deformation to a point, and then simply break. Marketing departments often boast about the inclusion of a certain fibre modulus in the latest frame design, but in most cases a bike frame is a careful balance of several types of modulus within the lay-up to deliver a desirable combination of stiffness, durability and flex.
There’s one more variable to consider. A single strand of carbon fibre is extremely thin – far thinner than a human hair, so they are bundled together to form what’s called a ‘tow’. For bikes, a tow can contain anything between 1,000 and 12,000 strands, although 3,000 (written as 3K) is most common.
Fibre this, fibre that
Those are the basics, but creating a lay-up gets complicated. ‘From a pure strength and stiffness point of view the ideal composite would have the highest proportion of fibre to resin possible and the least bend in the fibre,’ says Dr Peter Giddings, a research engineer at the National Composites Centre, Bristol, who has worked with bikes and raced them for many years. ‘Unidirectional fibres, theoretically at least, are the best choice for this. UD materials have an increased stiffness-to-weight ratio in the fibre-direction. Unfortunately UD composites are more susceptible to damage and, once damaged, are more likely to fail than woven fabrics.’
To build a frame exclusively from UD carbon layers would create a bike that was dangerously brittle, not to mention prohibitively expensive owing to the material and man-hour costs. Hence woven carbon dominates and is the obvious choice for any areas where there are tight curves and complex joint shapes. What’s more, people like its appearance. ‘Aesthetically, woven materials are considered to look better than unidirectional materials and the public’s perception of a composite is a woven fabric,’ says Giddings. ‘In fact, many manufacturers paint [therefore concealing] areas where the frame construction prevents a smooth, woven appearance.’
Ease of fabrication also has to be factored in to a lay-up schedule to take into account labour costs. For complex joints and shapes it will take much longer to create the ideal lay-up with UD fibres. It’s another reason why woven fabrics are the preferred choice of most carbon bike manufacturers. ‘Woven cloth is easier to work with than UD and requires less skill to fit it to a required shape,’ says Giddings. ‘UD has a tendency to split or kink around complex shapes. Loosely woven fabrics conform more easily and the structure’s overall strength is less affected by minor manufacturing defects.’
Manufacturers are likely to opt for a lay-up with woven carbon at the most complex areas, such as the bottom bracket and head tube junctions, but it’s still not as simple as it sounds because there’s another factor to consider. ‘You want to keep continuity of the fibre orientation not just around junctions, but through and beyond them,’ says Paul Remy, a bike engineer at Scott Sports. ‘There can be complex curvatures at a junction such as the bottom bracket so you have to think of a way to continue the fibres’ orientation, to transfer the loading throughout them.’
It’s here that frame engineers such as Remy are thankful for the assistance of computer science. In the past the only way to know how the various lay-up schedule alterations might affect the end result was to build and test multiple prototypes, but now a lay-up schedule can be tested with a very high degree of accuracy by computers before a single strand of fibre has touched down in a frame mould.
‘Previously it was really difficult to know what effect changing just one part of the lay-up would have on the performance of the frame,’ says Remy.
Bob Parlee, founder of Massachusetts-based Parlee Cycles, remembers those old days before computers did all the number crunching rather fondly: ‘If you understand the loads on a truss structure such as a frame, lay-ups are straightforward, so initially I could work them out myself in my head.’ Parlee has since conceded computer finite element analysis (FEA) has its place. ‘Originally I wouldn’t put holes in frame tubes [for cable entry points or bottle cage mounts] because they were potential weak spots, but now FEA tells us what to do to reinforce that hole,’ he says.
Sometimes the computer spits out ideals that are far from ideal. ‘Most of the time I look at it and say, “There’s no way we can do that,”
Increasing computing power together with ever-more sophisticated software is allowing engineers to analyse many virtual models in a short time and push the boundaries of design and materials. According to Specialized design engineer Chris Meertens, ‘Iteration is the name of the game. FEA tools create a representative model of the frame and the goal is to get every fibre accounted for. The software allows me to design each ply, based on an optimisation model for the 17 load cases that we have for a model frame.’
What that means is the software instructs Meertens how much carbon should be in each area of the frame, and the optimal orientation for the fibres. The skill, though, is in knowing what is and isn’t possible with carbon lay-up. Sometimes the computer spits out ideals that are far from ideal. ‘Most of the time I look at it and say, “There’s no way we can do that,”’ Meertens says. ‘So then I get busy in laminate draping software to cut virtual plies and drape them on a virtual mandrel, basing it on manufacturing feasibility and laminate optimisations.’
Even using computer software this can take days to decipher, and there’s still a long way to go before the lay-up is finally defined. One aspect where the human element is essential is in making sure the right fibre grade is used in the right place. Meertens says, ‘0° fibre is very stiff but doesn’t have good impact strength so, to keep the composite damage tolerant, we have to avoid putting too much in places like the bottom of a down tube. I’ll know by this stage what ply shapes I need, but now I want to know how many of each ply. So I run another optimisation program that tells me how thick I should make them – essentially the number of layers. It will analyse anywhere from 30 to 50 combinations of plies. We’ll run through the cycle of virtual draping and optimisation four or five times, fine-tuning the plies a bit more each time. But at some point we need to hit “Go” and send it off.’
Definitive guide
The lay-up schedule is like a 3D map, detailing each piece of shaped carbon in each layer. ‘The frame is divided up into nine zones: two seatstays, two chainstays, bottom bracket, seat, top, head and down tubes,’ says Meertens. ‘We specify the datum, which is an axis, for each zone. The orientation of every piece of carbon in a zone is then related to that datum. A down tube may have plies at 45°, 30° and 0° relative to the local datum. In general, the higher strength material is used off-axis, at an angle. The higher modulus material we use axially, at 0°.’
The resulting file can be up to 100Mb in size and is eventually passed to the factory floor. Each worker in the factory receives only the portion relevant to the part of the frame they are responsible for creating. This is still not the final production run. The built frame is a prototype at this stage and it needs to be tested to ensure the digitally designed lay-up results in a frame that performs in practice. Ultrasound, X-ray inspection and physical dissection reveal laminate thicknesses. Elsewhere the resin matrix will be burned away to expose the quality of the lamination and whether material or fibres have migrated. Bending tests should show the same results as the FEA analysis. In the end, though, it’s a human who takes it out on the road.
‘Riding the bike is the only way we can truly quantify it,’ says Bob Parlee. ‘We can do the bending and load tests but we need to get out and ride it to see if it performs how we want.’ When the model passes muster, production is finally given the green light.
Most bike production happens in the Far East, and this places even greater importance on the lay-up schedule. The finely detailed plan, if followed to the letter, should ensure the products coming out of those large factories are identical twins of those tested and passed at the final prototype stage. Of course most brands continually test and re-test production frames to ensure consistency so that bikes reaching the shops meet customer expectations. In most cases manufacturers can also trace a frame’s entire journey, right back to the origins of the very first fibre strands. Which is something to think about next time you’re standing and admiring your pride and joy.