Downamics, Andover (2025)

25/11/2025

How does the roll rate impact displacement?

Using IMU sensors we can measure the bike’s angular velocity — in this case, the X-Axis roll rate (how fast the bike is leaning left or right).
We can’t measure absolute lean angles with IMU sensors, but the rate of lean change tells us a lot about how the shock is being loaded dynamically.

The heatmaps (slides 1 & 3) show us the roll rate (degrees per second) and direction on the X-Axis and rear wheel displacement (mm) on the Y-Axis. The two heatmaps are from different locations and different bikes. Slide 1 shows us that at around ±2°/s the shocks displacement is most active. Once we go outside of that, there is little change in the colouring. This also explains the different width of the lower roll rate bins (blue) in the histogram in slide 2.

The histograms (slides 2 & 4) show us how the displacement is used in the slower roll rates (bottom 25% of the range) and in the higher roll rates (top 25% of the range) as a sample count on the Y-Axis and rear wheel displacement on the X-Axis.
This helps us to see how differing roll rates can contribute to changes in the usage of the displacement. For example, in the first histogram, we can see that during the higher 25% roll rates, the shock is loaded more in mid third of the stroke, in comparison to the last third where the difference becomes less. This is a different story in the 2nd histogram, but using the heatmap in slide 4, we can see that this track has a much narrower range of roll rates in comparison to slide 1.

To give some context, we also need to understand that the sample population for a bike in lower rate rolls is typically larger than the population size of higher rate rolls, therefore there is a bias in the data. We can adjust for this by using probability instead of sample count in the histograms.

It's not the perfect analysis, but it does provide me with a valuable insight into something that I previously hadn't had eyes on.

24/11/2025

Bike stability isn’t just about chainstay length or wheelbase!
The real stabiliser is the rear mechanical trail — and it’s almost never talked about.

When we speak about “stability”, we usually point directly at chainstay length or wheelbase. All too often we hear about how longer chainstays or rear centre can be more predictable and a longer wheelbase is more stable. Shorter bikes are often referred to as being more playful, agile or poppy.
This is only partially true, and not the complete picture.

A much bigger factor in stability is the rear mechanical trail — the distance between where the rear tyre contacts the ground and where the steering axis of the rear end would meet the ground (if it had one).
Chainstay length/rear centre influence this number, as does wheelbase, but neither actually helps you to really understand the reason why.

So when a bike feels “stable”, it isn’t just long chainstays or wheelbase at work — it’s the rear mechanical trail guiding the rear wheel into line, just like the fork’s trail can correct the front wheel.
This is why two bikes with similar chainstay lengths can feel completely different on track.

This is covered in more detail in the Data Driven Descents book, where I cover how mechanical trail works and what affects it, especially as it’s a dynamic metric.

23/11/2025

I found this piece of analysis from some in-depth testing of the back in 2023. This was part of a complete case study I did on pedal kick back and how the O-Chain reduces the effects.

It emphasises how important video and data can be when combined. If you've seen this done before it's typically suspension data overlaid. You can also create other overlays such as accelerations and rotations, braking and GPS positions. In this example I had a system monitoring the suspension and braking and a separate IMU sensor mounted to crank. This data was then synchronised with the data from the BYB in order to understand rear suspension behaviour and any correlation with crank accelerations.

I always run video when testing, but I learned the hard way when I've not run any camera and missed some opportunities to learn more. If any action camera brands want to support then let's talk 😉

22/11/2025

How do forces travel through a mountain bike?

It starts with the front wheel, and every part of the bike acts as a filter on the signal on the way to the rider.

These plots show the vertical accelerations measured at the front axle, rear axle, frame (bottom bracket) and bars. The front axle, rear axle and frame are what I refer to as the “three axles”, which combine to make that bikes specific fingerprint. The bars are our metric for adding context to rider feedback.

1. Front axle → bars & frame
The signal spreads out and becomes more chaotic — this is what the rider feels most clearly and will likely be how they communicate their sensations.

2. Rear axle → bars & frame
The force path is narrower and more predictable.
The rear wheel is doing a different job: stability & drive.

3. Front axle vs rear axle
Even on the same bumps, the wheels experience different loads, different timing, and different frequencies.
The front wheel sees the terrain with less filtering, whereas the rear sees the result of:
– pitch
– weight transfer
– wheel path differences
– suspension movement
– braking load

The important bit to note - Impact forces don’t move straight up the bike.
They are subject to transformation, where each stage of the bike filters, delays, shapes and/or amplifies what the rider feels.

This is why front and rear setup have different requirements, as a result of chassis dynamics, timing, and the live load on top (the rider).

These plots show a variety of trails from a world cup track to a blue flow trail, a world cup rider to capable amateur and a carbon chassis to an aluminium one. There are about as many variables in these three examples as possible. Usually we aim to reduce the variables but for this case study, a wide range makes for better examples.

19/11/2025

This is getting on for nearly a year's worth of work now and we're on the final stretch. Although it technically started 5 years ago when I took the leap into this field.
It started off as a book to showcase how using data for bike set up can be incredibly beneficial, but has expanded into a pretty hefty book about all aspects of bike performance with understanding this using data at its core.

There are so many people to thank for helping with this and taking an interest in it. I'm super keen to get it wrapped up and out there now. It currently sits at 370 pages but could be closer to 400 by completion and around 80,000 words.

It's due for release early 2026 but you can preorder now using the link in the bio.

16/11/2025

The devil is in the details!

This is a project I'm deep into now and the sheer amount of variables and combinations is not only daunting but also exciting.
To go to a granular level on parts of the bike which are often over looked or taken for granted usually ends up with the rabbit hole getting deeper but also splitting 😂

Something I noticed last year was how if any of the components that make up a pivot assembly get damaged, aren't installed correctly or are manufactured to unrealistic tolerances, it can cause not only a singular issue at that location, but a chain reaction of issues across multiple pivots.

This is most likely only applicable to a race bike, and one that gets regular and thorough maintenance, but realistically, those are the bikes that are going to benefit most from marginal gains like this.

So with a host of different parts, preparation techniques and assembly procedures to try, I think winter is sorted 😂

15/11/2025

This was the last time I used the Stendec system. World champs 2022 in Les Gets with
Installing the system was a big job, especially with brake pressure sensors requiring multiple bleeds and then new lines after. However, the data was gold, especially when combined with the video footage from the onboard camera, all synchronised with the data.

With modern systems, the sample rates are higher, sensors are better quality and more appropriate and installation is much more efficient, but I feel we still need brake pressure data and onboard video as standard channels.

I am looking at quick connectors used in Motorsport hydraulic lines and it's not too much work to get POV footage with data overlay these days, but it goes to show how advanced Dave Garland's thinking was back in 2017 when he started building the Stendec system.

📷

13/11/2025

Small bump sensitivity - what the data shows!

These slides show front displacement (blue trace), rear displacement (red trace) and vertical acceleration in green, so we can see position and the results of the forces the axle is subjected to.

Slide 1, 2 and 3 are the same but with different annotations.

In slide 1 the purple box shows the initial small bump displacement and the axle accelerations that are the result of the bump force. We can see a series of three consecutive displacements, ranging from ~38mm - 48mm. As dynamic sag is at 29% typically (58mm) this would suggest the front wheel is significantly unloaded at this point. We can also see the accelerations at the axle which cause this displacement. The accelerations at the axle then consequently cause displacement to change fractions of a second later.

A similar chain of events can be seen in slide 2 where the suspension is extending, to the right of the red line (rebound) but during this, the axle is subjected to a couple more vertical accelerations which change the shape of the rebound event, shown in the red rectangles.

Interestingly though, the bigger compression event which teaches ~180mm displacement, is the result of an axle acceleration which gets clipped as it's out of the sensor's range of measurement, 32g in this case. The axle then continues to oscillate as the displacement reaches its peak and beyond. This is a result of the wheel assembly (Inc spoke tension) tyre construction and tyre pressure. Change some of these components and this oscillation pattern will change.

Slide 3 is the same as the first 2 slides but with no annotations and slide 4 is an example of this at the rear end.

This is small bump sensitivity in the real world — not just a feeling, but a measurable sequence of forces, accelerations and movements.

09/11/2025

I was just tidying up a chapter in the book where I cover the topic of using distance to compare runs rather than time.

Distance as a metric allows you to compare 2 runs and see how the bike is behaving on the same point on track, as opposed to time, where the rider may (and usually does) arrive at that point on the track at slightly different times, offsetting the events.

Using distance does rely on GPS data, which is prone to wander and inaccuracies, but it's usually accurate enough to allow this to work. The bonus is that the events are usually offset by a fraction which allows differences to be easier to see....there's always a positive.

The slides in this post show:

📊 Slide 1: distance Vs displacement front and rear with GPS speed.
📈 Slide 2: the same section of track using time instead of distance.
📊 Slide 3: front and rear braking data with GPS speed plotted against distance.
📊 Slide 4: front and rear brake data with GPS speed plotted against time.

This type of analysis was done externally as proprietary software typically uses time, so exporting the raw data and doing some processing allows us to get a different insight.

07/11/2025

This was an idea I was pretty stoked to put into action!

Each mechanics station would have a USB cable which dropped down, enabling the .bike loggers to be plugged in by the mechanic.

The USB cables then went to a hub, which was then fed back to the data station. This made it easier to get data from any bike with minimal interference to the mechanics, and an efficient workflow.

It saved walking around with the laptop and getting in peoples way, and it meant that the files could be downloaded as soon as the bike was in the stand, all while the mechanic could carry on with routine (and sometimes not-so-routine) work.

I have developed this further with automated file retrieval and management, and I'd really like to improve the hardware for a cleaner install and cable management, including retractable cable reels to keep things tidier above the mechanics.

Slides 1-3
Slide 4

06/11/2025

Video is also data and I use it a lot to analyse a bike's performance and a riders.

Having clips like this is a great way to help a rider remotely as it gives you valuable context to their description of what's happening.

I'll get a clip like this one from along with his thoughts, then I'll analyse the video closely and breakdown what I observe and recommendations for set up changes to try.

You do need to bear in mind that a clip such as this one only represents a very small percentage of a run and track, so you need to take that into consideration too.
📷 Video is data, 💬 rider feedback is data... Don't just focus on the numbers and graphs 📊

04/11/2025

Following on from yesterday's Engineering Speed post, we take a look at what braking data looks like from the magnetic sensors.

As mentioned these sensors can either deliver an analogue signal, which is what we see in the first slide, or a digital signal as shown in slide 2.

The analogue signal provides us with more information regarding how the brakes are applied, how the brakes are released and what's happening in-between.
This kind of trace has a lot more in common with hydraulic brake pressure, however unlike brake pressure, the magnetic sensors require a bit more calibration in order to understand actual brake usage and not just lever displacement.

The digital signal is purely on or off, as can be seen in the 2nd slide, where the braking events are just an up and down signal, where we can't see any information about how the brakes are actually used. But even this data is better than no braking data as we can get an idea of braking time and bias, and even just seeing where the brakes are applied on track.

I've learned a lot about the braking habits of different types of riders too, for example, Enduro riders use their brakes in a different way to downhill riders. This is partly down to a substantial difference in track time, but also how they use the brakes to rotate the bike. Braking data is some of my favourite.

Downamics

25/11/2025

24/11/2025

23/11/2025

22/11/2025

19/11/2025

16/11/2025

15/11/2025

13/11/2025

09/11/2025

07/11/2025

06/11/2025

04/11/2025

Address

Website

Alerts

Contact The Practice

Shortcuts

Share