Trek Madone Gen 8 aerodynamics

A full year before the Madone Gen 7 and IsoFlow were unveiled, our engineers were beginning conceptual design work on the next-generation lightweight Émonda. We knew that the Madone was our most aerodynamic road race bike yet, and we were excited to see how we could add some of our aerodynamic research to the Émonda.  

Pretty quickly after creating our first Émonda concepts, which were fairly radical, we realised that the aerodynamic performance gap between Madone and Émonda could be significantly reduced. That’s when the possibility of a single performance race bike became seriously discussed. But first, we needed to prove to ourselves that our athletes and customers wouldn’t be making compromises with this change.  

Weight vs aerodynamics

Our first concept bikes on the next-generation Émonda project were “A1” and “A2”. Our wind tunnel and computational fluid dynamics (CFD) testing showed some promising results that narrowed the gap between previous-generation Madone and Émonda. However, it would still lead to an unacceptable aerodynamic penalty if we only wanted to produce one performance road race bike.  

Our next step was to fully explore the design space. We created more prototypes in order of increasing aerodynamic performance to add to the lightweight A-series, all the way up to the mostly aero E-series. After hundreds of CFD and structural analysis iterations, we brought the A-, C- and E-series bikes back to the wind tunnel.  

With wind tunnel results and weight predictions for all three prototypes, we could virtually test how the different options would perform over a variety of race conditions. With the same wheels and tyres, we wanted a frame design that would outperform both Madone Gen 7 and Émonda on any slope.

The plot below shows the performance of Madone Gen 7 and Gen 8 prototypes compared to Émonda on slopes ranging from flat (0% grade) to quite steep (12% grade). The aerodynamic Madone Gen 7 (in the dashed white line) is faster than the lighter-weight but less aerodynamic Émonda from 0% to just over 3% gradient, as the weight difference has a small effect on flats and mild climbs and the rider is moving at a faster speed, giving aerodynamics an increased impact.  

When we look at the prototypes, they all make improvements over the Madone Gen 7 and Émonda at certain slopes, but only prototype C3 is faster than both bikes over all the conditions. The lightweight A1 option climbs well, but wouldn’t be a good option for fast sprints on level ground. The more aerodynamic E2 is best at 0% grade, but would require riders to sacrifice the climbing performance of Émonda.  

With these results, we saw that prototype C3 had the potential to combine the weight of an Émonda with the aerodynamics of a Madone, which (after much more optimisation) paved the way for a single performance road race bike – the Madone Gen 8. 

New Full System Foil aero shapes

How do you create concepts like A1 to E2, spanning from super-light to super-aero, without years of trial and error? The answer lies in our new system of tube cross-sections – Full System Foil.  

Before Full System Foil aero shapes, we primarily used Kammtail Virtual Foil (KVF) tube shapes. These shapes were revolutionary at the time, meant to maximise aero performance with the now-defunct UCI 3:1 length-to-width tube aspect ratio rule. Kammtail shapes are very aerodynamic, but performance degrades as you depart from the 3:1 aspect ratio of more traditional aero bikes.

To improve on KVF shapes, we devised a more flexible cross-section generator that could permit a huge variety of shapes by allowing software to “push” and “pull” on the cross-section walls. Then, we coupled this base shape with an optimisation algorithm that would push/pull to generate a new shape and virtually test its aerodynamics and structural efficiency. The optimisation learns from previous design iterations, with enough time and computations, arrives at a Pareto front which is the set of cross-sections that represent the best combinations of weight and aerodynamics.   

This animation shows a small subset of some of the shapes the optimisation algorithm tried out. As you can see, the flexible nature of the push/pull shape modification creates some crazy shapes that would never be practical, but this allows the whole design space to be explored – even the bad parts. This wide exploration is important, since it can generate shapes that might not occur to a human designer. 

As we monitored the progress of our optimisation, we noticed that, in general, the shapes that were being generated had rounder back ends than KVF shapes and showed better aerodynamics in real-world unsteady aerodynamics situations where the wind direction is rapidly changing. The front ends of many structurally efficient shapes were relatively rectangular, which made sense intuitively.  

Where aero isn’t everything 

Armed with a collection of shapes ranging from structurally efficient (lightweight) to super aerodynamic, the next step was to identify where on the frame to slot in the different options. This is always an element of our bike frame design, but we took it to the next level for Madone Gen 8, using thousands of CFD and Finite Element Analysis (FEA) simulations to identify the best spots on the frame for each cross-section option.  

For example, the downtube shape favours structural efficiency over aerodynamics, because the slow-moving wake behind the front wheel lessens the importance of low-drag shapes in that area. In the opposite direction, the upper seat tube, IsoFlow, and seat post are all designed to be very aerodynamic shapes, because the airflow is accelerated between the legs of the rider and magnifies the drag in that region. Essentially, each shape is designed with consideration as to how the air moves over the shapes in front and behind it.   

Cylinders are slow

We’ve known for a long time that cylinders are not a very good shape aerodynamically. All the way back in 1953, NASA’s predecessor, the National Advisory Committee for Aeronautics (NACA), showed that simply changing a cylinder to a 2:1 ellipse reduces drag by 40%.

Long before then in 1912, Gustave Eiffel (yes, that Eiffel) was discovering how much drag shapes like cylinders generate in his drop tests from the Eiffel Tower.

So why do we put cylindrical water bottles on otherwise highly optimised aerodynamic race bikes? Aero bottles are not a new concept but not often raced due to usability and practicality. We set out to design a bottle and cage that would be practical in race situations for the Lidl-Trek team. The team riders and staff set two requirements: the aero bottle cages also needed to be compatible with standard bottles, and the downtube and seat tube bottles needed to be identical/swappable.  

These requirements prohibited the use of a super-streamlined bottle like the Speed Concept downtube bottle. To make up for this, we designed the cross-sections of the downtube and seat tube bottles to work in conjunction with the frame and wheel cross-sections and create a virtual airfoil. In the image below from our CFD simulations, slow-moving air is shown in grey. Fast-moving air sees these slow-moving wakes much like solid objects and flows smoothly around them, reducing drag.  

Our RSL Aero Bottles save 1.8 Watts at 35 km/h when compared to standard 620 ml bottles and are faster than no bottles at all.

While these bottles are optimised in conjunction with the Madone Gen 8, they were also tested in CFD on a range of bike frames and they reduced drag relative to standard bottles in every case.  

Designing around the rider system 

Bikes don’t ride themselves, and a rider creates a massive impact on the flow field around a bike. That’s why we test around the whole rider system (bike, rider, components, bottles/cages) and include a rider in our aerodynamic design from the first computer simulations to our wind tunnel tests using Manny the pedalling mannequin. This emphasis on rider aerodynamics is what led to developments like IsoFlow on Madone Gen 7. IsoFlow returns with some subtle optimisation on the Madone Gen 8, increasing rider comfort while providing structural and aerodynamic benefits.  

On the Madone Gen 8, the handlebar may jump out to an observer as looking less aerodynamic than the previous generation, and that’s because it is! At least in isolation.  

The cross-section of the Gen 8 handlebar tops got thicker and rounder than that of the Madone Gen 7. If you put the bike alone in the wind tunnel, that handlebar will increase drag. However, when a rider pedals behind it, the wake of the thicker handlebar slightly reduces drag on their pedalling legs by slowing down the air in front of them. The effect is small, but the pedalling legs are the biggest contributor to total system drag, so small changes to their airflow can be impactful. Just like our Full System Foil shapes, the handlebar cross-section was optimised using the same process, but including the effect of the legs behind the bar. 

The results

When all was said and done, the wind tunnel results show that the Madone Gen 8 is in an aerodynamic tier above the Émonda. Compared to the Madone Gen 7, the Gen 8 sees the majority of aerodynamic improvements at the low yaw angles, which are the yaw angles most commonly encountered by riders.

We tested the bikes at a range of wind tunnel speeds to support simulations of a wide variety of race scenarios. Shown here are the results at a 22 mph tunnel speed, which is a lower speed that makes testing a little more challenging, but is more attainable than pro peloton pace. We also tested at speeds up to 40 mph in order to feed simulations for our Lidl-Trek pro athletes (more on that below).  

The simulations made it clear to Lidl-Trek that the Madone Gen 8 would require no sacrifices in crucial scenarios when compared to the Madone Gen 7 or Émonda. Catching the attack on the steep 10% grade is traditionally where the riders would like to have an Émonda for as little weight as possible, but the Gen 8 Madone beats the Émonda slightly in that scenario. Both generations of Madone are in a class of their own in the sprint scenarios, which place utmost importance on aerodynamics, but Gen 8 edges out Gen 7 in both cases – a little more comfortably on the uphill finish.

In all these scenarios, the effects of acceleration are simulated. Though a slight effect, it is easier to accelerate the lighter weight frame of the Madone Gen 8 up to speed when compared to the Madone Gen 7, for example. 

In the end, what started as an update to the Émonda became the ultimate climbing and sprinting bike that excels in the demanding use case of our Lidl-Trek riders while bringing the exhilarating experience of a lightweight performance race bike to those who can only dream of sprinting at 1,500 Watts (like me).