It’s natural for an engineer to change a bike’s design to try to improve its performance. A common engineering technique is to measure a performance attribute in the lab, for example frame stiffness, then change the design and measure the improvement.
Of course, for this “improvement” to be felt on the road, the lab tests have to match reality. Do they?
To make sure the lab test matches reality, you have to know how the bike behaves in real life. Years ago (can you remember when the Soloist was introduced?) Cervélo created a “strain gauge bike” (above) to measure and record the load and flex (bending & twisting) in each area of the bike when ridden. The laboratory load case should match this data.
Let’s look at the real data recorded for different load cases.
The chart above shows how each frame tube responds to different load cases while riding on the road. This is old data (naturally we aren’t showing our latest data), taken from a ride Jason McCartney did during a visit to Project California’s test track when he rode Cervélos for Team CSC. In each different load case, the strain gauges report that some frame tubes flex more (or less) than others.
This chart continues the ride, including a few more load cases (plus the real-world situation of slowing slightly for oncoming traffic).
We could create a separate lab test for every load case, but in reality, two cases are enough to cover all riding situations: - “Out of saddle climbing” covers seated and standing sprinting and climbing; - “Turning” covers sweeping turns, 180 degree corners, descending curves
So we need two lab tests to simulate these two groups of real riding situations. These tests are the bottom bracket stiffness test and the torsion test.
These tests already exist in the bike industry. TOUR Magazin’s test protocols have set the worldwide industry standards for years. How do TOUR’s protocols compare to reality?
Cervélo can test any protocol. We have lab test fixtures that can be configured multiple ways to follow different test protocols. So we used TOUR’s protocols to measure bottom bracket stiffness and torsional stiffness of our strain gauge bike. We then compared the strain distribution from the lab tests to the strain distribution from the road: the acid test of lab vs. reality.
TOUR’s bottom bracket stiffness test does correlate well with reality, but TOUR’s torsional stiffness test does not correlate well with reality. The bottom bracket test is okay, so how can we improve the torsion test?
We redesigned the torsion test to better match real world conditions. The Cervélo Torsion Test closely matches the real-world data from the strain gauge bike.
We made three changes from the TOUR protocol, all leading to a more realistic testing process:
- Rear pivots at the road contact point (not fixed at dropouts)
- Front pivots at lower headset bearing (not middle of head tube)
- Seat post is fixed (not free). In real riding, rider inertia anchors the seat. (Out of saddle riding is covered by the bottom bracket stiffness test.)
As a result, the Cervélo Torsion Test closely reproduces the real-world data from the strain gauge bike.The outcome of the analysis: we were testing the wrong way! There is a significant difference in stresses throughout the top & down tubes, and higher bending stresses in the seat tube we were missing. “Improving” the lab test results the old way didn’t necessarily translate into improved performance in the real world.
What does this mean for frame design?
We use the Cervélo Torsion Test to optimize for a situation closer to reality. When we change the layup to improve the test results, we know it’s a change you’ll feel. We can add, remove and re-orient plies in specific areas of the frame, confident that these changes affect the real stresses in the frame. Bottom Line: You can feel the improved performance in the real world. Have a look at our Project California bike series page to see what happens when we put all this knowledge into one frame.
Update: Read "Lab vs. Reality - Part 2" in our Ask the Engineers section.