“The nice thing about standards is that there are so many to choose from.”
-Andrew S Tanenbaum Ph.D.
So Dr. Tanenbaum wasn’t thinking about bikes, but his comment certainly applies. There are a few measurable aspects of a frame that can tell you how it performs. But, if everyone is measuring these aspects differently, how do we know what we are comparing? It would be great if there was one agreed upon standard so that you could be certain that you are in fact comparing an apple to another apple.
There is no such thing as a wrong or bad standard. Different standards, or different measures, will tell you very different things about a bike. Essentially, these different measures will answer different questions. Knowing how to compare bikes accurately means knowing what questions you are asking, and what question a standard is answering.
Cervélo’s mission is to make our athletes, people like you, faster or more efficient in their riding. Therefore, we base our tests on answering ‘does it make you faster?’ The performance of one frame can be compared to another based on weight, stiffness, aerodynamic drag, and rider comfort. Over the next few articles, we will try to break down why different manufacturers will come up with different results and which results paint the truest picture. In Industry Standards Part 1, Damon Rinard tackled how we measure and report weight. In part 2, Will Chan explains how we measure stiffness.
Part 2: Stiffness
Comparing the stiffness of frames sounds like it should be very easy to do. However, the answer is more complex than one might expect. The closest we come to an ‘industry standard’ for stiffness testing is called the Zedler test protocol. It is used by Tour Magazine to compare the stiffness value of frames. But, manufacturers do not universally apply even this test protocol standard. In fact, at Cervélo, we have developed our own.
In order to have stiffness tests that represent real life riding, the test must have the same boundary conditions as real life riding. Boundary conditions are the constraints and forces acting on a system. If the system is the bicycle, then the constraints and forces exist at the rider’s contact points, at the tire contact patches with the road, and to a lesser extent, gravity and aerodynamic forces acting on the bike. The challenge is to mimic the boundary conditions of real life while keeping the test simple enough to conduct thoroughly and accurately while being capable of easily testing as many different frames sizes and geometries as possible.
First we need to define what aspects of stiffness are important to the rider and record what happens in reality. What forces is the bike under? Where in the frame? Through a long process of real life testing, we were able to ‘record reality’ as a part of Project California. To find out how we achieved this, read the following:
Lab vs Reality Part 1
Lab vs Reality Part 2
With the knowledge of what is happening when you ride in real life, we determined that there are 3 types of stiffness that matter: torsional stiffness, bottom bracket stiffness and vertical stiffness.
The torsional stiffness of a frame can be directly correlated to the handling characteristics of a bike. The rider steers the bike with input on the handle bar, which connects to the front wheel, and the front wheel to the ground. The steering action creates lateral forces on the front tire contact patch and those forces cause the rest of the bike and rider to lean over in a turn.
The torsional stiffness of the frame is needed to connect the front wheel to the rider and to the rear wheel when the bike is steered and leaned into a corner. Generally, higher torsional stiffness leads to more responsive handling by reducing the lag time between input from the hands and reaction in the bike and rider. Even while cornering smoothly, there is continuous feedback between rider and front wheel. We are always making minute adjustments to keep the bike going in the path we want it to follow.
In the case of cornering, this results in a set of forces (the ‘load case’) applied at the handle bar, the saddle, and the tires’ contact points on the road. Some of these forces are in opposite directions, essentially twisting your frame. The load path from the handle bar flows into your frame through the head set bearings, and the load path from the saddle flows into the frame through the seat post. For our stiffness testing purposes, we mimic these boundary conditions by either supporting or applying known forces at exactly these points. We even mimic the forces at the contact points of the tires, which as it turns out, is especially important to make our lab test match reality.
When a rider pushes down on the pedal, the frame deflects. Stiff frames deflect less, so more of your energy goes into turning the rear wheel, rather than into deforming the frame. The Zedler protocol is generally accepted as the industry standard to measure bottom bracket stiffness. This protocol supports the frame at the drop outs, leans the frame over at about 10 degrees and applies a downward force on a pedal attached to a crank at 45 degrees below horizontal, pulling on a simulated chain connected to the rear hub. Much like you stepping on your pedal.
To check this test protocol against reality, we followed the same procedure described in the Lab vs. Reality. We found that it matches pretty closely, except we measured slightly greater lean angle. Based on this test, we developed a bottom bracket stiffness test that is very similar to the Zedler protocol, but more accurately reflects riding conditions.
Figure 1: Pedal loads and direction vectors. (Measurements courtesy of our friends at Garmin.) The results are consistent with our testing in Project California.
Power transfer is one aspect of BB stiffness, but the loss is at most 0.3% to 4% depending on if you’re cruising along or a world class sprinter. We also know, from subjective rider feedback, that a stiff enough bottom bracket makes for a great feeling bike. Such a frame behaves as expected under even maximum efforts. Based on subjective feedback from our pros, we engineered BBright™ to produce frames with very high bottom bracket stiffness while simultaneously saving weight in the frame.
Vertical stiffness is the inverse of vertical compliance. In fact, TOUR Magazin uses this test to assign a frame a “comfort” score. More compliance, or less stiffness, on the riders contact points at the hands, saddle and pedals will reduce the peak force transmission from road inputs to the rider. However, vertical compliance does not paint the complete picture when evaluating comfort, especially for a rigid bike with inherently high stiffness but relatively low damping. Recent scientific research suggests that the correlation of a rider’s perception of comfort is much more involved than just vertical compliance. If not just vertical compliance, then what? We will discuss this further when we tackle comfort comparisons between frames.
How We Compare
At Cervélo, our goal is to improve the riding experience. Therefore, we include the rider – just as we do with our aerodynamic testing. To accurately assess differences in how bikes actually ride, every lab test must have the same boundary conditions as real life. After all, the purpose of testing is to validate the success of a design. And the design is only successful if the bike performs well for the rider in real conditions. If there is a standard test that achieves a realistic boundary condition, we will use it. Otherwise, we develop our own.
Testing for vertical stiffness is the simplest case. Apply a force straight down on the saddle and measure how far it deflects. The Zedler test protocol achieves this accurately so we stay very close to the same process. We’ll elaborate more on comfort measurements in the near future.
When measuring bottom bracket stiffness, common approaches measure deflection under a force applied at either a horizontal or vertical plane. In our case, we apply force at the same lean angle as you do when pushing on the pedals at peak torque, 15 degrees. The head tube is fixed to simulate out of saddle sprinting and measurements are taken in the same direction of pedal force vector to get an accurate measurement of pedaling efficiency.
Our torsional stiffness test is what truly sets us apart. Traditional test for torsional stiffness calls for the frame to be fixed to a jig at the rear dropouts and supported in the center of the head tube. A torsional load is then applied to the head tube and the frame is essentially twisted. While this does put the frame under torsion, it is not a realistic load case.
Figure 2: Cervélo Torsion Test simulates the cornering loads from the tires as well as the rider's inertia.
An example where realistic test boundary conditions benefit our bikes performance is when I first started designing layups for Cervelo. I studied older layup schedules and noticed there were always 0 degree plies running down the side of the head tube. Everyone just assumed that it was beneficial for head tube stiffness because we used to test using the Zedler protocol which supported the middle of the head tube (the 0-degree fibers stiffen the HT in bending). It was obvious then that having a point load in the middle of the head tube was not realistic at all and that those plies were there solely for that test protocol. So as we developed our own Inertia Torsion test [figure2], with realistic boundary conditions, I was able to remove those plies with no detrimental effects on torsional stiffness.
That opened our lay-ups to a more efficient use of material for increasing stiffness in real riding conditions while keeping a low weight.
What you need to know:
When you hear claims of a frame being stiffer or the stiffest, ask yourself how it was tested. How were the forces applied? Do those measurements reflect real world riding conditions? How were those conditions defined? How can you trust a company when they say it’s so? In Cervélo’s case, we’ve shared our engineering process and the data linking our test standard reality.
The stiffness of a frame is important to the rider to provide responsive handling and efficient transfer of power. Another importance is that a stiff frame under the pedals and during cornering will feel better and give the rider more confidence by behaving predictably and react quickly to your inputs. While the ‘feel’ of a frame is subjective, exhaustive data collection and realistic testing is drastically improving our understanding of how to make the best riding frame possible.