Recently, we’ve been measuring vibration in wheels, and building on the complexity of the measurements we are gathering to further our understanding of the role that vibration potentially plays in determining impedance break point. Our original work used a unidirectional accelerometer, which measures accelerations in one plane and simplifies the output. This unidirectional sensor was mounted to a skewer and measured the vibration in the yaxis, or up and down—like the up and down of a bike going over bumps.
Vibration & The 3Axis Of Acceleration
3Axis Accelerometers
Last week, we discussed adding a 3axis accelerometer to the fork and to the wheel. This week, we’re breaking down the the three axes and how they align. Below is an image of the accelerometer and the alignment of the axis.
Fork Mounted Sensor
For the fork mount, the concept is simple since the accelerometer is stationary. Below, we highlight the direction of each acceleration and what causes each one:
 XAxis: acceleration up and down caused by bumps and imperfections in the road, or popping wheelies, or hitting sweet jumps.
 YAxis: acceleration forwards and backwards caused by going faster or slower
 ZAxis: acceleration side to side caused by rocking the bike left to right.
While the image below shows a sensor mounted to the fork at an angle, the concept stays the same.
Wheel Mounted Sensor
A wheel spins, so here things get a little more complicated. Let’s talk through the acceleration of each axis and its cause.
 XAxis: acceleration in rotation in the direction of rotation; caused by spinning the wheel.
 YAxis: called centripetal acceleration in the direction of the hub; caused by spinning the wheel.
 ZAxis: acceleration side to side caused by rocking the bike.
The image below shows the sensor mounted to the rim of a wheel.
The Complexities Of A Wheel Mounted Accelerometer
The only axis that is predictable is the zaxis. While the wheel is rotating, the side to side motion is relatively unaffected. Now, if you are sprinting and really rocking the bike, of course the motion would be affected more, but in a regular riding situation, things stay fairly consistent. Let’s talk about the xaxis and yaxis.

XAxis: The xaxis is acceleration that follows the direction of the rotating wheel. Below highlights the different key positions.
 12 o’clock Position: The sensor is at the top of the wheel, and the acceleration is forward with the direction of travel.
 3 o’clock Position (front of wheel): The sensor is in front of the wheel, and the acceleration is down.
 6 o’clock Position: The sensor is at the bottom of the wheel, and the acceleration is actually zero, since when in contact with the road, the velocity at that point is 0.
 9 o’clock Position (back of wheel): The sensor is in the back of the wheel, and the acceleration is up.
At first glance this may seem simple, however, you still have to factor gravity into the equation. At 12 and 6 there is no gravitational acceleration since the orientation of the xaxis is perpendicular to gravity, but at 3 and 9 you are either working with (moving down) or against (moving up) gravity. This complicates your output, and the position is tough to determine.
 YAxis: A good way to visualize the yaxis is to imagine the feeling of being in a car and going around a sharp corner at high speed. That acceleration you feel is the same acceleration felt by the yaxis in the wheel. The faster the wheel goes, the higher the acceleration. Gravity is opposite of the xaxis, and comes into play at 12 and 6.
Output From Each Axis
Below is the output from for the xaxis, yaxis, and zaxis on both the fork and wheel at 120 psi. You can see the differences in each. Note the graph below is a Fast Fourier Transform (FFT) plot of the acceleration felt which looks at frequency excluding time.
Resultant Vectors
Finally, we can look at the resultant vectors. This means we are looking at a combination of the x, y, and z axis results.
Final Thoughts
Wheels definitely make things more complex. However, we are interested in looking at the data from the wheel since it is the closest thing to the tire we can measure. Next week we will discuss sensor data from rollers to understand what happens in a more controlled environment.
3 comments
Thanks Jon, this is very interesting! I come back here every week waiting for the next installment. Keep up the great work.
Michael,
Good question. We have plans to study what happens at the rider level. Our bigger question relates to understanding if we can help control this at the wheels. To know if this is true, we are starting by trying to get a solid picture of what happens at the wheels.
Ride Safe,
Jon
Why not put a sensor on the handlebars, saddle and pedals? We know that hysteresis loses come mainly from the rider so shouldn’t we be analyzing the vibrations being transferred to the rider and minimizing the amount of power absorbed by the rider?
I see the idea of looking for resonance, but not sure what the end goal would be. Resonance damping designed into the the wheel, therefore decreasing the total vibration energy transferred to the rider?