Behind the Scenes.
How we put e-bike drives through their paces.
Welcome to the engine room of our data analysis. Anyone reading our test reports will inevitably stumble across diagrams, performance curves, and thermal analyses that go far beyond what is typically found in a ride report. But how do these data come about? How do we know that Motor A runs more efficiently at a cadence of 80 rpm than Motor B?
The answer lies in Zeil am Main, near Frankfurt. There, on the PTlabs test bench, we conduct all our measurements. But data is only as good as the method used to collect it. Transparency is our currency. That's why we are disclosing here how we work, why we do it that way, and the philosophy behind our dyno runs.
1. Basics
The heart of our analyses is not just any treadmill, but one of the most modern bicycle dynamometers in Europe. Developed and built by Veit Müller, the laboratory manager of PT Labs. Veit is not just an engineer with a penchant for precision, but someone who understands that an e-bike is more than the sum of its parts.
The Complete System Counts
A crucial difference to many pure industry tests is our approach to the test object: We test the entire bicycle.
Manufacturers often test "naked" motors on motor test benches. This provides isolated values for the unit but does not always reflect reality. A motor sits in a frame, it drives a chain (or belt) via a chainring, the power runs through the cassette into the hub, through the spokes to the tire, which finally makes contact with the roller. Furthermore, in our tests, the motors are always powered by the batteries actually used in the bike. In recent years, also in exchange with various manufacturers, we have learned that a motor often behaves differently in the complete "bicycle" system than in a laboratory vacuum. Resonances, frame stiffness, and peripherals play a role.
Why a Test Bench at All?
Now one might ask: "Why don't you just ride in the woods?" We do. Extensively. Nothing replaces the feeling of a drive on the trail, the response behavior on a wet root, or the modulation in a hairpin bend. That is the subjective side of the coin, which we illuminate in our ride reports and YouTube videos.
But feelings are hard to compare. The test bench, on the other hand, provides us with the objective, reproducible truth. It creates framework conditions that remain exactly the same for every drive. No daily form of the rider, no headwind, no different ground surfaces. Only here can we let drives, motors, and systems compete fairly against each other. We "scientify" the riding feel a bit to understand why a bike feels the way it feels.
The context is important to us: The dyno values are extremely valuable and helpful, but they are not the sole, absolute truth. A motor can seem weaker on paper but feel stronger in the terrain thanks to ingenious sensor technology. Both belong together: The data from the lab and the sweat from the trail.
2. Power Measurement
When we speak of power, we mean the pure motor power in our data sheets. But how do we determine that if we measure at the rear wheel? Here we go to enormous effort together with PT Labs.
The Principle of Loss Calculation
We measure the power at the rear wheel via a high-precision measuring drum. That is the output of the complete system. To know what the motor actually delivers, we must know what gets "lost" on the way from the motor to the roller.
For every single measuring point, we therefore carry out a strict protocol:
1. Measurement without support: We drive the system only via the crank, without the motor helping. We input a defined power/cadence and measure what arrives at the roller. 2. The Baseline: The difference between input and output is our power loss (tire flex, chain slip, bearing friction, frame twisting, etc.). 3. The Calculation: We add this calculated power loss to the measurement with motor support later.
Why is this so important? Because losses are not linear. A tire has a different rolling resistance at 50 watts input and 70 cadence than at 200 watts and 100 cadence. In addition, chain skew has an influence. Since we pre-calibrate every measuring point practically "naked", we get a motor power output in the end that is as close to reality as physically possible – cleaned of factors that have nothing to do with the quality of the drive.
Full Steam Ahead: The Battery Factor
A banal but critical factor is the energy content. We conduct all power-relevant measurements (especially maximum power) with a battery as full as possible (90–100 % charge level).
The reason lies in cell chemistry: If the charge level drops, the voltage in the battery cells drops (Voltage Sag). Many motors can only call up their absolute peak power if the voltage is high enough. If we were to test a motor with 30 % battery, the results would simply be wrong or at least not comparable with a competitor starting fully charged. For us, frequent recharging or the use of many spare batteries is mandatory to ensure equal opportunities.
Maximum Support Mode
We measure almost exclusively in the highest support mode (Turbo, Boost, Race, etc.).
Critics might object: "But I mostly ride in Eco or Tour mode!" That is correct. But these modes are not standardized between manufacturers. What is "Tour" at Bosch might be "Trail" at Shimano. These modes are software applications and often adjustable by the user.
The maximum mode, however, shows us the physical limit of the system. Here manufacturers usually "show their hand". It is the only common denominator to make the hardware's potential comparable. If we know what the motor can do, we can better estimate how it works in the medium modes.
3. General Test Parameters
To make results comparable, we must eliminate variables. The biggest uncertainty factor on the bicycle is usually the human on top of it.
The Pneumatic Rider
No human sits on our test bench who gets tired or pedals unevenly. Instead, we simulate a load. The bicycle is loaded with a pneumatic system that simulates a rider weight of approx. 80 kilograms.
This is important for the contact pressure of the tire on the roller and thus for rolling resistance. Also thermally, it makes a difference whether a "dummy" sits on the saddle or not, as this influences the airflow. We try here too to stay as close to practice as possible. An 80kg rider is a realistic average value for our target group.
Wind from the Industrial Plant
An e-bike motor is an electric machine that produces waste heat. Without cooling, every modern motor would overheat on the test bench within minutes and throttle performance (Derating). In reality, the airstream provides cooling.
In the lab, a massive industrial fan takes over this task. It is placed directly under the motor and provides plenty of fresh air.
Admittedly: The wind speed we generate here is probably slightly above the average of what one has on a steep uphill passage in the Alps. But our goal is comparability. We want to know what the motor delivers when it is kept thermally stable.
The Art of Crank Modulation
Here it gets special and this is a point that caused us a lot of headaches at PT Labs, but is essential: Crank modulation.
We drive the system mechanically, but not bluntly linear. A human pedal stroke is never perfectly round. Even more importantly: Different motors prefer different pedaling styles.
Some drives need a "stomping" kick with high torque peaks to wake up. Other systems prefer a very round, even pull on the chain.
If we were to drive all motors with an identical, perfectly round sine wave, we would disadvantage some systems. A motor programmed to react to torque peaks (the "kick" into the pedal) might release less power with a perfectly round pedal stroke than it could.
Therefore, we try to find out for every new drive: How must it be pedaled to perform maximally? We adapt the modulation of our drive unit to the "sweet spot" of the motor. This is complex, but it is the only way we treat every system fairly and elicit from it what it would perform under an experienced rider.
4. Special Test Parameters
We run various protocols to illuminate different aspects of the drives. Here are the key data points:
Power Measurement: The Feel-Good Cadence
When we determine maximum power, we usually do so in a cadence range of 75 to 80 revolutions per minute (rpm).
Why here of all places? Because it is the "feel-good cadence" for sporty e-bikers. Those who ride ambitiously rarely pedal below 60 and rarely permanently above 90. In this window of 75–80 rpm, most electric motors have their best efficiency and highest power output.
There are exceptions – some motors like to rev higher, others are torque monsters in the basement. If this discrepancy is huge, we point it out separately. But for the comparability of our standard charts, we pin the cadence in this area.
Cadence Measurement: Cruising instead of Sprinting
In our diagrams, you often find curves showing how the motor supports across the entire rev range (from very slow pedaling to the "hamster wheel").
We do not conduct this measurement at maximum rider effort. Instead, we fix the input power at 130 watts.
130 watts correspond to comfortable but brisk riding – a classic everyday situation, no sprint. We want to see: How does the motor behave when the rider just pedals "normally"? Does it push hard already at low cadence? Does it collapse at the top? If we were to go in here with 400 watts rider power, we would constantly run into limiters and mask the nuances of the motor characteristics.
Thermal Measurement: With or without Cover?
A hot topic (pun intended). How stable is the motor on the mountain?
We thought long and tested: Should we remove the motor covers to ensure better cooling? The answer is: No, we usually leave them on.
Our tests have shown that the plastic cover itself has a surprisingly small influence on the core temperature. Much more important is the heat dissipation via the mounting points on the frame (the "sides" of the motor). Since manufacturers design the integration including the cover, we test that way too.
An interesting detail on the side: If you see a slight drop in performance (approx. 1–2 % per minute) in our 15- or 20-minute thermal stability curves, this is almost never thermal throttling. It is simply the dropping battery level. As mentioned above: Less charge = less voltage = minimally less power. Real thermal problems (Derating) are recognized by massive drops, not by a gentle glide downwards.
For comparability, we also use – wherever possible – carbon frames. The material conducts heat differently than aluminum, and to keep the variable "frame material" as constant as possible here too, carbon is our standard.
5. Dialogue with Drive Manufacturers
Finally, a word on cooperation. We test independently, but not in isolation. Before and after the tests, we are often in dialogue with the engineers of the drive manufacturers.
This has nothing to do with influence – our measured values are sacred and are not embellished. But it is about technical understanding and fairness.
- Software Status: We ensure that we do not test beta firmware containing bugs that have long been fixed in the series.
- Understanding Anomalies: If a motor shows strange behavior on the test bench (e.g., an abrupt shutdown at a certain speed), we ask why. Often there are technical reasons (component protection, gear preservation) that one must know to interpret the result correctly.
- Integration: Sometimes problems lie not with the motor, but for example with the pre-series integration in the test bike.
This exchange helps us not only to deliver naked numbers but also the explanation for why these numbers are the way they are. We want to understand how the developers defined the "sweet spot" of their system, and then check whether we can validate this on the test bench.
We hope this insight into the "torture chamber" of PT Labs helps you understand our data better. At the end of the day, all these watt numbers, Newton meters, and temperature curves serve only one goal: To help you find the right e-bike.