AERO COMPONENTS

FROM CYCLINGPOWERLAB

More and more cyclists are realising that the real elephant in the room when it comes to the forces that slow a rider down is aerodynamic drag, not weight. In fact, no matter how much power a rider CAN produce, minimising the power he HAS TO produce for a certain speed will always tend to improve results. To understand the relative importance of aerodynamic drag in this objective consider that as a rule of thumb an average rider will be using the majority of his power output to overcome aerodynamic drag on any road gradient upto around 3% when the battle against gravity takes over. Then consider how much of your typical training or racing routes feature a gradient of 3% or more...suprisingly little in many cases. 

Aerodynamics haven't always enjoyed such importance. For many years we have seen riders obsessing over weight, particularly component weight, to a point that component weight databases such as the one at weightweenies  sprung up. Of course there is nothing wrong with obsessing over weight, all marginal gains aggregate to help the rider, but this is best done with an equal or greater obsession and due prioritisation for aerodynamic drag minimisation. The reader who wants to learn a bit more about aerodynamic drag and in particular how it can be measured and optimised using a power meter may wish to review our aerodynamics primer.

Component Drag Data

An increasing amount of component drag data is making its way into the public domain via manufacturers or third parties who have run independent wind tunnel tests. It wasn’t always the manufacturer’s intention to put this on general release, but nowadays the consumer demands it. Frankly we are past an era when serious cyclists would buy components on a manufacturers promise that “these are aerodynamic”. It is with all of this in mind that we have compiled a component drag database which forms part of and sits behind our component drag evaluator. In compiling this database we relied on public-domain data but also made enquiries with several manufacturers. We would like to acknowledge the quality of the data &/or responses from Bontrager, HED, Trek, Specialized & Zipp, while we scored a disappointing zero of manufacturer data from Campagnolo/Fulcrum, Shimano/Pro and Mavic. We would of course welcome being provided with further data or details of the precise testing protocols used to collect it.

Interpreting Drag Data

Mount a time trial or tri bike in a wind tunnel sited somewhere near sea level – San Diego perhaps - mount a rider or rider shaped manikin, crank up the airflow to 30mph, spin the wheels at 30mph, and the unit as a whole may register about 3000 grams of drag. An object with 3000 grams of drag moving through sea level air at 30mph, which implies a CdA of around 0.267m^2, would require about 394 watts of power to keep it moving at that speed if there were no other forces slowing it down. Understand however that manufacturers dont test their components at 30mph to somehow exaggerate aerodynamic benefits given the convenient assumption that everybody can crank 394 watts to the finish line, this is simply a speed that keeps things sensitive yet provides results that scale well to the range of real-world cycling speeds. 

Take the rider or manikin off and a slippery bike may measure about 600 grams – hence the oft quoted claim that "the rider makes up 80% of aerodynamic drag". Swap various front wheels in and out of something that looks like a frame, subtract the drag reading of the wheel-less setup (a process known as taring), and you have drag numbers for various wheels. These may be in the range 100-200 grams while the same wheels, if sited in the rear of the bike, would account for a lot less drag - between 30% and 60% of that attributed to the front wheel. The percentage here depends on the size of the rider and the aerodynamic attributes of the frame, how these disturb the airflow towards the rear of the bike and how the frame itself keeps the rear weel faired from airflow. 

Component drag tends to be tested at a range of yaw angles because, in the real world, meteorological wind ensures that airflow seldom hits a bike head-on. For an introductory discussion of yaw have a look at our article of a while ago on yaw, drag & component choice. Aero components tend to have lower drag at non-zero yaw angles and so, in order to truly evaluate the effectiveness of any aerodynamic component, we need to consider the full spectrum of drag data - typically offered between 0 and 20 degrees of yaw. This range of data, compared with some knowledge of the yaw angles experienced by a cyclist either generally or on a specific course, is what we really need to interpret the aerodynamic promise of components in a meaningful way.

Component Drag Evaluation Methodology

Our database contains component drag data through the range of yaw angles from 0 – 20 degrees in 2.5 degree increments. In cases where data was not available with this 2.5 degree resolution we filled in the gaps using linear interpolation and in cases where a 20 degree value was missing (certain frame data) we assumed the 17.5 or 15 degree number. Frame data is usually provided at both negative (drive side) and positive (non drive-side) yaw angles. As a simpification we store only the latter and apply it equally to negative yaw situations. In cases where this data came from charts (most) we settled for interpreting the charted values to the nearest 5 grams. 

We assumed all of this data was tested at sea level, where “ISO Standard Atmosphere” air density is 1.225kg/m^3. We also assumed testing at 30mph, or that the data was normalised to 30mph of airflow along the axis of bicycle motion, because 30mph appears something of a bicycle industry standard for wind tunnel testing. Given these assumptions drag numbers can be converted to equivalent CdA’s (or component contribution to system CdA) which have usability at any ride speed. Mouse over the drag number in our model to see these equivalent CdA’s. 

Given raw CdA numbers through a range of yaw angles we can realistically evaluate any component at any speed on any course by calculating apparent wind yaw angles on each sector of the course. We can calculate these yaw angles using a power model to define rider speed given average power output and baseline CdA, wind speed and direction, and the interaction between rider speed, the compass bearing of course sectors, wind speed and wind bearing. Given these sector by sector yaw angles we can compare component contribution to CdA, using interpolation where necessary, subtract this from the contribution of whatever baseline component (you specify this) and then it is a relatively simple matter to revaluate the relationship between a riders power and speed on that course. This revaluation leads to figures of “Time Dif.” (holding baseline power constant) and “Power Dif.” (holding baseline average speed or time constant).

Inputs - Rider & Bike

• Baseline CdA – Input your baseline CdA assuming riding with the reference component (as specified in the "Comparing To") dropdown. Consider reading our aerodynamics primer to help you estimate this.

• Evaluate Component Type – Select “Wheel” or “Frame”. This defines the type of the components that will be evaluated in the result matrix.

• Comparing To (Data Source) - Select a reference component and data source against which all other components of your type of interest will be compared

• Rear Wheel Drag Scaling – Leave this as 0% to evaluate the aerodynamic impact of changing only a front wheel. Select a value between 30 and 70% to evaluate a pair of wheels where the rear wheel accounts for 30 to 60% of the front wheel drag.

Inputs - Course Characteristics

• Course Layout (Bearings) – Select from a list of course “shapes” that may best represent the type of course on which you want to evaluate component benefit. A range of polygon shaped courses is available (having equal length sectors) - this is important to simulate a range of yaw angles.

• Wind Bearing – Adjust wind bearing (the compass bearing the wind is blowing FROM) to define how wind conditions would interact with the course sectors and their bearings. For example, a wind bearing of 0 degrees simulates a tailwind out and headwind back on the “Out and back (180,0)” course.

• Wind Speed (KPH) / Gradient. – Enter meteorological wind speed (the type that is reported by a weather station) . Check the Gradient box to adjust this wind read at “weather station height” (10 metres as a convention) to wind at bike height. This is important because wind slows down as it gets closer to the ground due to friction effects. For more explanation of this phenomenon see our article on wind gradients & correction factors.

• Course Distance (KM) – Enter a course distance in kilometres. In conjunction with the average speed you enter this will define “Baseline Time”.

• Average Speed (KPH) – Enter an average speed in KPH. Wind effects may determine that the ride is modelled faster or slower on different sectors. In conjunction with the average speed you enter this will define “Baseline Time”.

Outputs

• Sector Yaw Angles – The yaw angles determined for each of the course sectors together with the percentage of total time riding that sector at that yaw angle.

• Component Evaluation - An alphabetic list of all the components of your type of interest (wheel or frame) from our database including the associated yaw data, sometimes from more than one source. Given this data and our evaluation methodology the matrix shows “Time Dif.” and “Power Dif.”. The row highlighted in blue represents the “baseline” component against which all others are being compared. The row highlighted in red represents the component offering the greatest time (and by implication, wattage) saving given the course and conditions you have specified. Keep in mind that this in no way implies this is the outright “best” component – simply that it might be expected to offer the greatest advantage given the ride scenario you have specified.