Introduction

This blog will cover topics about data analytics on bicycles. For many of you, this will be your first look at data specific to bikes. Rest assured, these are not abstract concepts and we encountered all of this for the first time when designing our system from scratch. Before we dive into data, we'll explore what your goals should be when tuning your bike. Then we’ll go into how you go about this process using data. 

First off, your bike was designed to go downhill (yes, this is obvious), so getting data on flat riding is not going to yield useful information. You can collect data in a macro way and you will find out useful things about your bike, but in general, you’ll want to isolate your bike data on fast aggressive terrain because that’s where you’ll need the support and predictability of properly tuned suspension. The end goal for your bike is to make it support you. You want to ensure you are using most of your travel AND that the suspension velocities are in the correct range. Every rider handles their bikes differently so you want to ensure your tune provides proper support for the way you ride the bike. 

In general, your bike should feel balanced. The front of the bike should not overpower the rear, and vice versa. You want the suspension to move approximately the same distance, at the same speeds, when hitting the same terrain. You want to ensure your bike is not packing when you get into a situation where you are hitting successive bumps. Slow rebound speeds will prevent the suspension from opening up to absorb the next bump and will cause the bike to squat down and increase the transmission of forces to your body. Lastly, you don’t want the geometry of your bike changing too far from the intended design. So if your front is really stiff, yet the rear is a soft noodle,  your bike will be raked out and will not corner as designed. 

In the preceding diagram, we will be comparing the motion of V (front vertical) to Y(rear vertical). 

For data to be valid, it must analyze your bike, not just a damper in isolation, or a front and rear damper without incorporating bike specific geometry. If you just analyze damper motion without considering the bike, you are missing key information. You could have a perfectly performing shock on bike A, but could move that shock to bike B and it could feel terrible. So analyzing any suspension component without considering what it’s on does not go deep enough. You need a systematic way to see how the bike is interacting with the ground. We define this by comparing front and rear axle vertical motion. For this, we need the head tube angle of the fork and the leverage curve for the shock. We measure the shock, but we are using the leverage curve to calculate rear axle vertical position. With some basic trigonometry, we calculate front vertical motion by measuring the fork and extracting the vertical component. Now we have a systematic way to compare front and rear wheel motion. Now we just need to apply some theory on how this should behave. 

First Step: Preload

Before you jump in and start turning knobs, you need to get your preload correct. We always suggest starting with the bike manufacturer’s recommended static sag. Make sure you have the correct sag before you start testing. The recommended sag may be wrong, but you can figure that out as you test. 

Once you have correct sag, go out and ride your trails at your aggressive speed. When you are done with your first ride, you want to look at the following things:

1. Front and Rear Max Position
2. Front and Rear Average Position
3. Front and Rear Histogram

Ignore all other data in the app. It is not useful information until you get your preload correct for your weight and speed. What are the knobs at your disposal for pre-load?

1. Air Pressure
2. Volume Tokens (air forks and shocks only)
3. Spring Pre-load
4. Spring rate

Springs are not as flexible as air. The spring manufacturer will have guidelines on how much pre-load the spring can take. If you are undersprung, you will ruin the spring by clamping down on it to get the correct preload (sag). If it’s oversprung, you won’t be able to settle into your sag even with minimal preload. You need the right spring for you.

Assuming you were able to get the right pre-load, look at your first run.

Regarding Max Position: Shoot for 85% of max. Were you able to get to 85%? Was it past 85%? Or was it down near 60%. At 85% or so, you should have enough margin left in the shock or fork for the really big impacts. 

Regarding average position, it’s important that you look at this on an aggressive downhill run with aggressive terrain. You want to get a sense for how far apart the fork and shock are. Most people will like their fork riding higher than their shock. This makes sense since most of your weight is toward the rear of the bike anyway. 

Up to this point, all you’ve looked at is 2 data points for the front and rear, max and average position. Now let’s look at the position histogram to see what we can learn. 

First thing, what is a histogram? It is a simple chart that shows you recurring events in separate discrete ranges or buckets. This is a really powerful way to get a macro view of a large data set. 

In the preceding diagram, you will see a signal in the time domain. Samples are taken in a periodic time interval. On the y axis, 3 discrete position buckets were defined. For now, let’s call them bottom, middle, and top. On the graph to the right, we’ve taken those buckets and placed them into a histogram. On the y axis, this is simply the number of samples taken per bucket. On the x axis, each bucket has a starting and ending point in mm. So a histogram is simply a chart organized with a series of buckets. Each bucket contains a number of samples that were recorded within that bucket range. Going forward, we’ll refer to a histogram as buckets and samples. The samples could be position, stroke velocity, or other metrics we track. 

For analyzing suspension and bike dynamics, the shape of these histograms will yield a lot of information. It’s one of the best macro charts to convey information of large data sets.

In the preceding picture of an axle position histogram, you can see that about 80% of the travel was being used. Also, note the shape of the curve. This shape indicates a fairly linear use of travel, meaning as you get deeper in the travel, you are using less. The curve is fairly flat which indicates that the air spring is not too progressive. The key takeaway from the above histogram is that less time is spent in deeper locations of the fork or shock. This graph visualizes where your fork or shock is spending its time. 

There are 3 knobs at your disposal to change the shape of this histogram: 

1) To shift the whole curve to the left, add air 

2) To shift the whole curve to the right, remove air 

3) To change the shape of the ramp, add tokens or remove tokens. 

Removing tokens will allow movement deeper into the travel with less progression. Adding tokens will restrict movement deeper into the travel by adding progression to the air spring.

Note: For coil shocks, typically you have progression in the leverage ratio to help with added resistance (progression) deeper in the stroke. There are also progressive springs that are now on the market. Progressive springs mean the spring rate, k, increases as you compress the spring. 

Once you have adequate range of motion of your front and rear suspension, regardless of how the bike feels, you are now ready to start looking at balance. If you short circuit to this point, you will waste a lot of time turning knobs,  when you should have figured out spring rate and sag first. 

Bike Balance

What is bike balance? How do you quantify it? This is something we’ve heard everyone talk about but we’ve seen no data on how to quantify bike balance. Even if you could quantify balance, on what trails should you measure it? We spent a lot of time working with many riders on the quest to quantify this mysterious metric.

The first thing to think about is that your bike is interacting with the ground and there is a balance of forces that are happening in real time. A simple view of your balance of forces can be summarized in a simple free body diagram:

The rider (when standing up) has a down force centered on the bottom bracket, there are equal and opposite forces the ground exerts to the wheels. Between the ground and free body forces, there is a front and rear damper that is dispersing energy from these forces in the form of heat. When a damper compresses or rebounds, energy is distributed in time to reduce the sudden impact (defined as jerk, or acceleration of acceleration) of incoming forces. 

We need to quantify the front and rear damping characteristics to see if they balance out. There are several things we want to know when comparing the front and rear motion:

• Average position, or dynamic sag
• High and low compression speeds of the front and rear axle
• High and low rebound speeds of the front and rear axle
• Total up/down distance traveled from front and rear. 

Conceptually, if your rear shock preload was so high that the shock barely moved, you can visualize that all of the force that the earth puts on the rear wheel is going to get transferred to the front of the bike, and vice versa if the fork was locked out. This example can be carried on with damping settings, IE: if your front compression is super damped, but the rear is not, a lot of forces are going to get transferred to the rear of the bike. On rebound, the damper is expanding and putting exerting a force back to the rider, and earth. So if one of your dampers was opened all the way, but the other was not, then the open damper is going to transfer energy to the other end of the bike. The rider will feel all of these forces on their feet, hands, and possibly saddle too. The bike will feel rocky, unstable, and will be difficult to ride. 

First and foremost, your bike was designed to handle terrain on downhill. Your bike’s head tube angle, shock leverage curve, etc. was optimized to keep the rider center of gravity in a way to ensure the bike is stable. You don’t want the rider feeling like they are going to go over the bars on a steep downhill for example. This is all part of the bike geometry. What is much more difficult to understand is how to set the bike up for the terrain you typically ride. This is where data is crucial. 

Regarding bike geometry and stability, going back to the unbalanced preload example, if your fork was over inflated (air) and your shock was under inflated, then your bike geometry is going to change significantly, IE: the head tube angle is going to decrease because the front of your bike is riding high or “Raked Out”. In the reverse case, your head tube angle will increase making your bike feel really unstable going downhill.

Front and Rear Dynamic Balance

Force = mass x acceleration. By measuring the position of the fork or shock, we can simply calculate acceleration from the collected position data. The instantaneous slope at any point along the position curve is velocity. Acceleration is just the instantaneous slope at any point along the velocity curve. The difficulty is knowing instantaneous mass. Your body weight will change as it carries momentum from free fall, cornering, etc.  If the rider stays in a fixed position on the bike, then it’s easier to know the instantaneous mass, but we know riders are all over the bike. So it’s nearly impossible to directly measure the front and rear forces. Our goal, however, is to balance the damper forces from both sides of the bike. By doing this, the rider will get a consistent, stable feel. 

Since we can't measure force directly, we need another method. A visualization of how the front and rear wheel is interacting with the earth will give us a really good indication of balance. Since the front and rear wheel are mostly hitting the same objects and terrain, they should be moving with the same speed for the same wheel displacement. Again, taking prior examples of unbalanced pre-load, the same theory applies for dynamic wheel movement. If your front wheel moves out of the way quickly when hitting a rock, but the back end does not, then your rear will apply a reactive force back to the front, and vice versa. Velocity and displacement are two important metrics when looking at balance. 

In the diagram above, we have plotted every compression stroke for a short ride. Each dot represents a compression event. The dot is placed on a 2 dimensional graph based on the maximum velocity of the stroke (mm/s) and the displacement distance in mm. Notice the fork compression strokes are moving much faster than the rear. Also notice that the rear is getting much more travel. This does not look like a balanced bike. The front is faster than the rear and not getting much travel. 

This conceptual scatter plot is the underlying thesis of our bike balance. We’ve taken an additional step and have performed a linear regression analysis for each set of dots, front and rear. This regression analysis basically creates a best fit line through the data. 

Notice after calculating the regression lines, how much different they are. If you were to try and balance this out, what would you attempt to change in your setup? If you were to remove some preload on the front fork, you may get more travel. This would have the effect of spreading out the dots across the graph. On the rear, you may be happy with the use of travel, but maybe you’d want to speed things up by opening up the compression a bit. 

After making the changes and re-running the test, we’ve been able to get the bike more balanced. The fork is using more travel, the shock is compressing faster. The regression lines are now pretty balanced. 

With any algorithm, you need to have context into the analysis. Bike balance is just a description of balancing forces on both ends of a bike. We employ some simple concepts to describe balance. Just because your bike is balanced, doesn’t mean it is set up correctly. You may have balanced the bike, but at ½ the necessary speed. So the bike is going to feel very over damped. 

Also, there may be situations where you need to imbalance the bike for a certain race. IE, a super steep sustained downhill may require a slacker front end to make the bike feel stable. The downhill guys will accommodate this by slackening the front end with adjustable headsets, or by raising or lowering the handlebars. 

It’s important to correlate this data along with the feeling of the bike to get the correct setup. This will take trial and error, but once you have this experience under your belt, you will be very confident in your setup which will yield dividends at the race. 

Conclusion

First and foremost, it’s a waste of time dialing in a suspension that hasn’t been serviced in a while. So do that first before jumping headfirst into data analytics.

The goal of this document was to give you an overview of how to set up a bike using data. Modern bikes with top line suspension have billions of settings. Also, each adjustment is not made in isolation. Without being able to see the interaction of the front and rear wheels together, it will be difficult to find the right balance on a bike. We caution against skipping what may be broken first which is proper sag and preload.

In general, we’ve seen some really great bikes set up very poorly. In many cases, once the bike has been setup properly, the riders have claimed it’s a completely different bike. The data highlights this as well. In some cases, bikes were set up way over damped on compression and rebound. We’ve seen cases where the bike was balanced perfectly, but at really slow velocities (over damped). For many, having a properly tuned bike was transformational. This is especially true for aggressive light women riders. For light aggressive riders, the stock OEM suspension tunes are just too over damped to provide the right range of motion, speed, and support. But with data, it was obvious what needed to change and we could measure it before and after a re-valve. 

We encourage you to consult with an expert when tuning your bike. If you want to tackle this on your own, Motion Instruments has a great solution, also there are other players in the market. What we can guarantee is that without using data, you’re leaving a lot of your bike’s potential on the table. By unlocking the true potential of the bike through proper setup, you’ll be able to reach your full potential as a rider.