While there are tons of articles analyzing bikes or presenting companies, there is not so much in depth information on how bicycle frames are developed. Since we do not see any secrets in how we do ours, we wrote this extensive piece that tells the full story of how we developed the Extra.
Every product stats with an idea of what we want to do. It is important to put this into words, which we do in the requirement specifications. This document describes what we want to achieve with this bike, but not how. It is an important document to make sure everyone involved in the project has the same understanding and works towards the same goal. It is not like it is set in stone though: the requirement specification will be adapted along the development process several times.
The geometry is the foundation for any good bike. If you get it wrong, it will never ride well, even if the rest is perfect. There are many ways to find your geometry, but in our case, riding lots of bikes with different geometries, trying extremes with the help of prototypes or anglesets built a wealth of experience based on which we can design a geometry on paper and already have a quite precise idea of how the bike will feel.
For the Extra, it was clear that the number one goal in terms of geometry is building confidence in steep, technical terrain. Placing the rider far and low behind the front wheel, with a slack head angle, long reach and low BB gives the rider enough room to find his position.
Recently there is a lot of talk about weight distribution and I have a different view on it than many. In popular belief, the center of mass of the rider is always directly over the BB. But in my observation, that is not true. Especially with too short reach, riders struggle to get forward enough, hanging onto their bars (a huge factor for arm pump, too). There is simply a limit to how far forward you can get with your shoulder over your bars and still hold the impacts. Only recently the industry came to reach numbers where you can comfortably place your center of mass over the BB without the fear of getting ejected through the front door when hitting a hard compression. Now you can even move your body from that central position slightly forwards to push the front wheel into the ground while cornering. Assuming that the position of the center of mass is defined by the position of the handlebars rather than the BB (the truth might be somewhere in between), surprisingly a longer reach without any change in chainstay length actually puts more weight onto the front wheel. Riding the Extra confirms that: if a wheel starts to go wide in fast corners, it is the rear wheel first.
As below illustration shows, even though our reach is considerably longer and the headangle slacker than that of a leading enduro bike not too long ago, there is actually more weight on the front wheel with our geometry. Why not push that even further with long chainstays? Because as you can see below, even if we went for 450 mm chainstays we would get less than 1% more weight on the front tire, but the bike would loose dramatically in agility, ability to lift the fron wheel when you need it, and most important be less fun.
Last but not least the Extra should get up the mountain with ease, so you can enjoy as much trails downhill as possible. The long reach already helps for efficient pedalling, but even more so does our steep seat angle, putting you right over your pedals and your center of mass further to the front, so you have good front wheel traction and you can pedal up steep section in a relaxed position. Our seat angle is measured to a seat height of 750 mm, which is where an average rider sits and therefore really matters. Others measure to the height of the top of the head tube, which is meaningless. We often get asked about the resulting seat angle at other seat heights, so we provide you the chart below.
Once the geometry was set, the next was finding the best layout for our frame. This includes: Suspension system, shock placement, kinematics, tube dimensions and packaging.
When talking about suspension we hear a lot of talk about leverage ratio curves, anti sag, pedal kickback etc. But there is a lot more that needs to be considered when designing the layout of a suspension frame: Mechanical efficiency (can the frame be built strong and stiff enough without excessive weight), packaging (is there enough space for shock, suspension pivots, rear wheel, chain, cables through all the travel?), assembly, maintenance (is the shock protected from dirt?). So finding the right layout takes a considerable amount of time, going back and forth, trying different options, before deciding for one direction
We simplified that process by a script that integrates all kinematics calculations into Solidworks. We have a simple base sketch of our frame with which we can define kinematics while seeing all clearances for components, tubes, pivots, and with just one click, within seconds we get the kinematics data.
Talking about the kinemetics, we aimed for a leverage ratio curve that works well with air and coil shocks. While with an air shock, you can tune the progression by volume spacers, this is more complicated for coil shocks. There are longer end bumpers and progressive coils, but they are hard to get. Therefore, we aimed for a moderately progressive leverage ratio. Generally speaking, the leverage ratio follows part of a sine wave. Long levers create a flatter sine weave (which is diserable), but are heavier, less stiff and more difficult to package into the frame. Therefore it is important to find the right balance. We opted for a linkage that is long enough to give us a falling rate all the way from sag point to where the shock touches the internal bumper. There the shock gets hugely progressive anyways, so we could live with that slightly regressive hook at the end of the leverage curve, as confirmed by Pinkbike's Dan Roberts, who could only bottom out the shock with a massive huck to flat. We also accept the slightly (3%) raising rate at the start of the travel, as this counteracts bobbing from pedalling.
That nicely leads us to anti-sag. Again, it is a decision about taking the best compromise. A high anti-squat would also mean big pedal kickback. The last few years (remember Aaron Gwin's win in Leogang?), and some rides without chain, tought us that low pedal kickback results in a very smooth ride on the decents. So there are clear benefits to use leverage ratio, low speed compression damping and chain tension together to limit bobbing, instead of just letting chain tension balance the forces. We went for 100% anti-sag at sag point with a 30/28T gearing, which is a realistic climbing gear.
The result of this stage is a rough side view of our frame, with all hard points (suspension pivots, component interfaces) placed in 3D space.
Constructing technical detail solutions like pivot points, RD hanger or frame inserts are often a project in itself. Our philosophy is to develop a good solution, and then keep using it over all of our frames. Luckily we could benefit from our work we have done on the Essential, and keep using the same construction.
The only new part was a headset that was developed through a separate project with the headset supplier Prestine. It allows us to have the cables enter the frame through the headset. The cables run in between the steerer tube and the upper (1.5") bearing of the headset. This brings multiple benefits: Since the cable entrance turns together with the bars, cables can be kept super short and tidy. Because the steerer tube is much smaller in diameter than the head tube, cables influence steering much less than normal. Even X-Ups or barspins are possible if you are into that kind of thing. We do not need holes or cable ports in the frame. And since the cables exit through the headset bearing seat, there is a huge hole to get the cables out. Putting in new cables is actually an extremely simple task: just push them through from the back, and they almost automatically end up at the right place in the head tube.
What typically would happen next is loads of hand sketches, finding the best way to connect the dots in a way that makes sense from a technical point of view, but also in a shape that is not offending our eyes.
But since we bought a SubD add-in for Solidworks (Power Surfacing by nPower), we can almost leapfrog that step. SubD is a technique originally developed by Pixar for animation films, and is easiest described as digital clay modelling. Instead of building a frame surface by surface, connecting and cutting them, we can manipulate our clay model and have immediate see the result. Starting from a simple lump of digital clay we can form a frame within less than an hour and then change its shape almost any way we want.
A few sketches of the sideview of the frame, the master sketch with all hardpoints, and we go straight into 3D space. Starting from a simple lump of digital clay we can form a frame within less than an hour and then change its shape almost any way we want. This gives us the freedom to try a lot more solutions and shapes within a short time. Also, SubD allows us to create shapes that would be so difficult to achieve by traditional modelling we would probably not even try to create them. And last but not least, SubD creates perfectly curvature continuous (G2) surfaces, which are not just aesthetically pleasing, but also do create less high stress zones than surfaces connected by traditional round fillets would.
We are in the lucky situation to have Laminate Tools, a software specifically made for layup development. Widely used in Formula 1, Laminate Tools allows us to virtually build the layup ply by ply. For each ply, we see whether there is a risk for wrinkles or gaps in certain areas, how fiber angles change when draping the ply over curved surfaces, and in what shape the material needs to be cut. After stacking the plies, we see the resulting wall thickness and weight of the frame.
Once the basic layup is done, we can export it into Femap and run FEA simulations of various load cases, showing us strength and stiffness properties of the frame. Back in Laminate Tools, we can analyze the results in even greater detail. We can look at a single ply to determine whether there are areas where we can leave away material. Or we can see for a specific location which layers are taking the biggest load, and which one will fail first (not necessarily the same, we will get to that in a minute!). Now this whole process until here took a big chunk of time (it can easily be a week worth of work), but now modifying that layup and getting results is a matter of minutes. Compare that to building physical prototypes and testing them, and it becomes clear that we can try a lot more layups in a shorter period, at a lower cost, and even get a much better understanding of what is going on inside the frame.
Now to an important myth and misunderstanding surrounding carbon fibers: High modulus means better carbon fiber. High modulus simply means that a fiber is stiffer, but does not say anything about its strength. Typically, a high modulus fiber is weaker than one of lower modulus (though often also called high strength fiber). To make matters worse, high stiffness and lower strength means there is very little elongation before a high modulus fiber fails, or in other words, they are incredibly brittle. Not exacly what you want in a product like an MTB frame that is ridden hard, with big impact forces.
This gets even worse if people are using high modulus fibers to reinforce critical areas of a frame. Imagine a stack of a cotton cloth and a sheet of paper, and trying to tear it. What will happen? First the much stiffer paper will take all the load, and then fail at a force the cotton cloth can easly cope with. So by adding the paper, you actually lowered the force at which the you get initial failure. Same is true if you just add some high modulus fibers for reinforcement.
So you will not find us boasting about high modulus fibers in the Extra. Instead, we are using high strength and intermediate modulus type fibers. Intermediate modulus fibers are expensive, but provide roughly the same maximum elongation as high strength ones.
Once all the theoretical work is done, it is time to make physical prototypes. More than anything it is important to choose a capable production partner. Being based in Taiwan is invaluable here. Being able to visit all significant manufacturers in person, talking to them in their language and getting all the latest news from rumor mill is of immense help when choosing the right partners. And of course for a fast and efficient development too. Only like that we were able to go from drawings to rideable prototypes in less than 4 months.
A lot of the testing happens on testing machines. ISO testing is the baseline, although surely not enough to replicate the forces in enduro racing, so we employ a number of additional tests with a lower number of cycles at a much higher load. Unfortunately we cannot go into details of these, but be assured those tests are brutal.In the end, nothing can replace real world testing, where unforseen things happen. Crashes, vibrations, rocks flying, you will not see these in a lab. Our prototypes saw a lot of hard riding in the alps under a viariety of riders. Brake rotors were bent, rims dinged, bars bent, but the frames held strong.