How We Test
Our original products were designed via trial and error and tested on the road. Our new products are engineered and rigorously tested. We employ a variety of ways of testing. Each method has pros and cons. Together, they represent a level of rigor that no competitor comes close to.
Testing on the Road
Testing while cycling allows us to evaluate the performance of our products in the actual aerodynamic and acoustical environment they are used in. We do this in two ways. One is simply riding with and without the products and recording impressions. The other is to instrument cyclists with in-ear microphones, record wind noise at known speeds by incorporating detailed Gamin data, and calculate sound pressure level reductions. Our experience shows that while our acoustic measurements cannot predict exactly how much wind noise level reduction a particular individual cyclist will experience, they give us a good idea about what the average cyclist will experience, and a very good indication of which product designs perform the best. We have asked many cyclists for their feedback... so we know that our acoustic measurements and the psychoacoustic perceptions are similar.
Because wind noise varies significantly by speed, we capture wind speed data during our test runs and extract exact cycling speed information from our Garmin 520. The digital anemometer and Garmin data is matched to the recorded sound files to help ensure that our road testing conclusions are as accurate as possible.
And, of course, periodic equipment testing and re-calibration is essential for accuracy.
We use probe microphones for accurate real in-ear measurements (REM) - similar to those used by audiologists. The tip of the probe tube is placed approximately 5 millimeters from the eardrum. This allows us to measure wind noise more accurately. We have read about people measuring wind noise by placing a microphone at the center of the rider’s outer ear. Anyone who has held a microphone in the wind or even just breathed at one will know that they can be susceptible to small variations in air pressure, causing false wind noise signals known as 'pseudo-sound'.
However, instrumented road testing is not without its challenges. Wind direction and wind turbulence levels can change rapidly. Exact orientation of the cyclist's head (yaw / pitch) is difficult to control. Because of these issues, we augment our road testing with aeroacoustic testing performed under carefully controlled laboratory conditions.
Testing in the Lab
Most wind tunnels are designed for testing of aerodynamic effects like drag, and are very noisy. Aeroacoustic testing requires extensive noise suppression, and these quiet wind tunnels are rare, specialized, and very expensive. Cat-Ears has built three quiet open jet wind tunnels. With careful attention to sound and vibration damping, they are remarkably quiet for the volume of air moved. We have constructed "test heads" (Custom Acoustic Test Simulators - CATS) that are equipped with realistic silicon ears and simulated ear canals. Miniature condenser microphones are placed in the ear canals - at the location of the ear drum. This is critical for correct real ear WNR measurements.
We test our products across multiple helmet brands (Bell, Giro, Bontrager, Rudy, etc.) and retail price points (<$50 to ~$250). We also test with different cycling glasses. Helmets and eye wear can impact the amount of wind noise.
Two Custom Acoustic Test Simulators (CATS) and Three Quiet Open Jet Wind Tunnels (small, medium and large)
allow us to perform controlled independent testing variations in both Colorado and New Jersey.
Our quiet open jet wind tunnels allow us to measure sound pressure levels while carefully controlling wind speed, direction, and head orientation. One limitation of our tunnels is that we can't make them free of acoustic reflections (anechoic). Another limitation is that we can't see precisely what is happening with the airflow.
Road to Laboratory Coherence
Investigating and understanding possible differences between road and laboratory testing is important. Accordingly, we perform sound vs. wind noise mapping - to understand the threshold of hearing in different wind velocity / turbulence environments. This allows us to identify / compensate for microphone vs. human ear differences (ATF). Understanding perceived wind noise helps ensure that our performance data and conclusions are more accurate.
Since wind noise is created by air flow, we need to understand the aerodynamics of air flowing around a cyclist's head, and the impact on that air flow when using Cat-Ears. Visualizing air flow, especially when looking for very localized turbulence, is very difficult. However, air and water are both fluids and at low speeds their dynamic behaviors are essentially the same. Thus it is possible to do "aerodynamic testing" using water, and observe those flows in great detail.
We designed and built a water tunnel - otherwise known as a hydrodynamic flume - to meet Cat-Ears unique test requirements. The test chamber is large enough to hold a "slice" from one of our simulated heads. We can pump water through it to reach an air speed equivalent of 15 MPH. We inject fluorescein dye into the flow and illuminate it with LED and laser ultra-violet lamps. We then capture high resolution videos and photographs, allowing us to see differences in turbulent flow patterns as we make design changes to our products.
The physics of turbulent fluid flow is very complex and not fully understood. Equations exist that seem to describe this ubiquitous natural phenomenon quite well, but they are for most real world problems beyond present human ability to solve directly. Over many decades, techniques have evolved to reach approximate solutions to these equations and over the past twenty years, with the widespread availability of powerful computers, the technique of computational fluid dynamics, or CFD, has gone from academic research to engineering use. We've taken some baby steps with CFD to see if it may be useful for design of future products. We have explored the use of SU2, which is an open source CFD system, based on mathematics originally developed at Princeton and now supported by Stanford. We did get it to run - although on our modest computers some runs took 19 hours! As we continue to refine our understanding of the turbulent air flow underlying wind noise, we will consider further use of CFD.
Laminar flow around a blunt object and cylinder. SU2 (Stanford University open source CFD) output into Paraview (open source data visualization). Source: Cat-Ears
Unmatched Performance Testing Capabilities = Unmatched Product Performance