Power measurement on a road bike: 9 power meters in the test

Topics in this article

• Powermeter: Short & sweet
• What the measured values bring
• How TOUR tests power meters
• The power metres in the individual test

The first world-class cyclist to recognise the value of the special measuring cranks was Greg LeMond. The American, who won the Tour three times between 1986 and 1990, was electrified by the possibilities of these cranks, which the inventor Ulrich Schoberer completed in his basement workshop in Jülich. LeMond was able to objectively record his performance while riding, and he quickly realised that he no longer had to spend more and more hours in the saddle in order to improve - instead, he could use intensive training sessions much more efficiently with the values in front of him. Class instead of mass. That was 1991.

Powermeter: Short & sweet

Powermeters are measuring devices that record the rider's pedalling power while riding. Prices start at 599 euros. The recorded power data (given in watts) is transmitted to the computer on the handlebars or to a mobile phone. When analysed, they are primarily used to control the load during training and races, but can also be used to optimise equipment.

Schoberer's company SRM is still the main supplier of power meters for the professionals: eight teams rode in the Tour de France 2016 with the product from Germany. However, a thriving scene of numerous other suppliers has grown alongside it. Pedalling performance is measured pretty much everywhere on the bike: on the crank spider, crank, bottom bracket shaft, pedal, rear hub and shoe sole.

The measuring principle of most devices is based on converting the elastic deformation of a mechanical component under the force of the footfall into electrical signals. Every electronic scale works in this way. However, the bicycle has the handicap of being moved in a harsh environment: It is exposed to lashing water, rough impacts and changing temperatures. In addition, the devices must be able to cope with the incredibly dynamic muscle power: A power meter for cyclists must produce clean values in the range of 100 to 2000 watts and have reserves in order to map even higher load peaks. A horror scenario for accurate measurement. For hobby cyclists, 170 instead of 150 watts of continuous power is already worlds apart. Professionals, on the other hand, want interval power outputs of 480 or 500 watts to be displayed accurately; or to know whether the top power output in a sprint is 1450 or 1490 watts.

Measured values show the progress

This brings us to the core of what it is all about: recording performance in exact figures. Have I got better? Accurate measurements provide a satisfactory answer to this question. If there is any doubt as to whether the figure is correct, the entire measurement effort is worth nothing, as this renders the objectification of feelings and rankings invalid, and then you can do without them.

Professional sports are dominated by power meters that measure at or around the bottom bracket. The SRM principle of measuring in the crank spider has been copied since the patent expired, which puts pressure on prices. SRAM (or Quarq) and Power2max offer significantly cheaper devices with a comparable measuring principle; they can also be found in professional teams that are not officially sponsored. In addition to the eight SRM teams, a further nine rode with crank power meters from other manufacturers at the 2016 Tour. No team used measuring pedals.

It is astonishing that the highly professional and financially strong Team Sky, of all teams, uses the measurement technology from Stages, which only records the power of the left leg. Although the technology is cheaper, it is also less accurate because the power of one leg is only multiplied by two. As both legs are rarely equally strong and the proportion of the total power output accounted for by the right and left leg therefore fluctuates, this measurement clearly lags behind the current status of the competition. However, Stages has been working on two-leg measurement for some time, and prototypes were used at the Tour de France last year.

Power meter on the road bike: the trainer's cult of numbers

More than 25 years of experience with performance data have shaped and changed cycling. Their analysis has been constantly refined, and today cyclists are among the best-studied endurance athletes. This is because pedalling performance directly reflects what a rider can do in a given situation. It's that simple and at the same time complex. It is perhaps best compared to the telemetry data from racing cars, which shows the engineers on the trackside what is happening in the car every second.

Today, professional coaches use a wealth of indicators to assess the strengths and weaknesses of riders. They can use them to decide which rider suits which route, who needs to ride in which position and when in order to make the team as fast as possible in a team time trial, when the training load is too high and so on... Power meters have created a cult of numbers in the training industry and continue to fuel it.

What the measured values from the road bike power meter tell us

Professionals use power metres in training to implement their coach's instructions, but also in races to manage their power. The data is also used indirectly to improve aerodynamics, even without measurements in the wind tunnel. This is because a more streamlined seating position, for example, normally means that the racer can ride faster with the same power output or has to produce fewer watts for a certain speed. Amateur athletes can Powermeter in the same way: If you think your way into the figures, you can control your training yourself, or you can give your data to a trainer who can use it to plan your training more precisely and control it better. The greatest and most immediate benefit is being able to precisely record your own training and intensity ranges and thus apply them precisely. In addition, progress in form can be objectively recognised. This is described in detail in the TOUR training planner, which is included as an extra with the February issue every year.

Data analysis programmes, such as the open source software Golden Cheetah, are now so comprehensive and detailed that you can analyse your training and competition records in every conceivable way. However, you can also get lost in it...

Powermeter: A test field from 600 to 2600 euros

We invited all the relevant companies to take part in our test. Eight manufacturers sent nine power meters into the race, including five new developments: SRM Origin, Rotor 2Inpower, Power2max NG, SRAM Quarq Red DZero and Stages Carbon. We were also able to test the Infocrank cranks from Verve for the first time. We also tested the revised Vector 2 pedal from Garmin and the P1 pedal from PowerTap. The classic G3 hub from PowerTap rounds off the test field with a price range of € 600 to € 2600. The new Shimano power meter, introduced at the beginning of the year, was unfortunately not available in time. We were unsuccessful with Pioneer, who no longer have a European distributor, with 4i, who did not respond to our invitation, and with BePro, who did not want to send us a device.

For the first time, we were able to directly compare crank power meters under laboratory conditions on a newly built test stand. This also allowed us to analyse how the devices react to temperature jumps, for example when moving from indoors to outdoors or when driving up high mountains. If the system is sensitive to temperature, the measured values can be falsified. We also undertook numerous comparison rides on the roller and on the road, with two to four power metres on the bike, in order to be able to directly compare measured values in many situations. We examined a power spectrum from 150 to 1200 watts and not only looked at average values, but also analysed raw data sets in detail to see if there were any outliers. This also shows whether relative deviations are constant - i.e. caused by the appliance - or dependent on the power.

Some devices deliver very similar values

The results of the six crank powermeters are within a range of plus/minus two per cent in the laboratory test. Manufacturers had long specified this resolution as the standard. Some now promise an accuracy of plus/minus one per cent; Rotor even states plus/minus one watt. We do not know which of the devices comes closest to the actual power output with its measured values: the measuring effort required for this with a calibrated sensor with very high resolution would be extremely high and expensive. For similar reasons, we were also unable to show whether all devices from one manufacturer deliver identical measured values. To do this, we would have to test a large number of devices against a reference. However, we had two models from Infocrank at our disposal - so we used one as a reference to directly compare the measured values of different devices. We noticed that the values from SRM, Quarq and Rotor were particularly close to each other.

The inexpensive Powertap hub works well apart from a few outliers and proved to be stable over thousands of kilometres in the long-term test. However, the raw values of the hub fluctuate more than with all other devices. The P1 pedals from Powertap measure around ten watts too little across the entire power band. This has a relatively greater effect at low power levels than at high ones; here the pedals are exemplary compared to crank power meters within the specification - and unbeatable when switching from bike to bike. The experience with Garmin's Vector 2 pedals is less convincing. The left pedal of the test sample repeatedly showed clearly too little. It required resets several times, even while riding, and was not a reliable source of data overall. Users describe similar experiences in forums.

Long-term purchase

Our test shows that you don't have to go for the most expensive device to get usable data. But it also shows that quality assurance does not work for all companies. In our opinion, the correct measurement of performance should take centre stage. Features such as the separate measurement of the left and right leg or pedalling analysis are interesting, but not always reliable; when we rode two devices on one bike in parallel, there was rarely a consensus as to which leg was stronger and pedalled "rounder".

Our tip: Anyone considering buying a power meter should see it as a long-term investment that can be used for around ten years. A device that works according to the SRM principle, with interchangeable cranks (modular design), seems to us to be the best long-term option for combining accuracy, durability and frame compatibility. Anyone looking for detailed information beyond watt values should take a look at the Infocranks or the new Rotor cranks.

How TOUR tests road bike powermeters

The absolute accuracy of the devices cannot be determined with reasonable effort. However, being able to directly compare the measured values is also very informative. For the first time, we tested crank power meters in direct comparison on a new test bench. We used one of the Infocrank models as a reference and compared the measured values of all other crank-based systems - including a second Infocrank model. The result shows how much the systems differ from each other and how the deviation of each device changes with different pedalling power. The setup also makes it possible to cool the test system and check how a temperature jump of 15 degrees affects the measurement result. In addition, we compared all the power meters with each other on numerous test rides (roller and road). The power range was between 150 and 1200 watts. For the measured values, we analysed both the smoothed average values and the raw data sets of the devices. We used Garmin Edge 520 computers with identical settings to collect the data. The data was analysed in Excel and Golden Cheetah.

The power metres in the individual test

Powertap P1

A pedal power meter for real left-right measurement. The P1s are quick and easy to adjust and always delivered slightly too low but stable values in the test. The pedalling surface is higher above the pivot axis than with normal pedals. This has a negative effect on pedalling, but you can get used to it. The cleats are not officially compatible with Look, but they worked perfectly for us. Real plug & play: If you ride several bikes and always want to measure your performance, the P1 is the perfect choice.

• Price: 1299 Euro
• Weight: 439 grams/pair without plates
• Software: PowerTap app, firmware update possible
• Connection: ANT+, Bluetooth LE
• Battery*: 1 x AAA per pedal, 60 hours
• Accuracy*: not specified
• Cadence: from pedal rotation

+ Easy to change, consistent data, temperature stable
- Short battery life

Garmin Vector 2

Garmin also offers a pedal power meter for true left-right measurement, which must be mounted with a torque spanner. The transmitters are attached to the pedal axles and connected with plugs in the pedal axles. Even if the wearing parts, including the pedal body, are replaceable: It all seems quite delicate. A long-term test would have to clarify whether five-digit kilometre figures can be achieved with it. Theoretically, the pedal can record many key figures relating to pedalling, but in practice it failed to determine reasonable watt values; we already had to make this experience in 2014 when testing the Vector 1. Not ideal for frequent changes from bike to bike.

• Price: 1149 Euro
• Weight: 359 grams/pair without plates
• Software: Data analysis, firmware update via ANT+ dongle
• Connection: ANT+
• Battery*: 1 x CR 2032 per pedal, 175 hours
• Accuracy*: not specified
• Cadence: from pedal rotation

+ Lightweight, many measured variables
- unreliable

Powertap G3

Proven classic for years with very light measurement technology. The hub is a solid part that has already completed several thousand test kilometres without any problems. The measured values usually differ from the SRM level by one to two per cent. The raw data tends to be erratic, which is bad for pacing training. The hub also automatically determines the correct speed. In addition to the hub, ready-built wheels are also available. The Powertap hub is a very favourable alternative to pedals for measuring power on several bikes and is much more accurate than any single-leg measurement.

• Price: 599 Euro
• Weight: 325 grammes
• Software: Firmware update via USB
• Connection: ANT+, Bluetooth LE
• Battery*: 1 x CR 2032, around 60 hours in test mode
• Accuracy*: +/- 1.5 per cent
• Cadence: calculated from torque peaks

+ relatively cheap, robust
- short battery life, annoying display

Verve Infocrank

The Verve cranks are designed as cradle bars and can therefore measure the left and right legs separately. The technology requires comparatively massive and heavy forged cranks, available with BB30 shaft. The construction is uncompromisingly designed for the measuring job, with separate magnets for left/right, which are well positioned on the bearing caps; with the latest firmware, the magnets can be omitted. Verve is an official supplier to British Cycling and the UCI. Very short and extra-long cranks are in preparation, as is a torque analysis. Robust professional system with left-right resolution.

• Price: 1599 Euro
• Weight: 865 grams (175 mm, 52/36)
• Software: in preparation, firmware update via ANT+ dongle
• Connection: ANT+
• Battery*: each crank 2 x SR 44, 500 hrs.
• Accuracy*: +/- 1 per cent
• Cadence: magnet (62 grams) or mathematically from torque

+ Left-right measurement, easy to handle, temperature stable
- Relatively heavy, no modular design

Rotor 2Inpower

The Spaniards combine two fundamentally different measuring principles to create a power meter: Strain gauges in the right crank and in the bottom bracket shaft record the power of the left and right leg separately. This is particularly helpful for users of the oval Q-ring chainrings, as it allows them to adjust their chainrings better (under "Optimum Chainring Angle", OCA). The internal battery is powered by a magnetic plug. We rode the crank for around a thousand kilometres and it worked flawlessly. An interesting addition for Q-ring fans.

• Price: 1499 Euro (without blades)
• Weight: 851 grams (175 mm, with 53/38 aero blade)
• Software: Rotor evaluation software, firmware update
• Connection: ANT+, Bluetooth LE
• Battery*: Internal rechargeable battery, 300 hours
• Accuracy*: +/- 1 watt
• Cadence: Acceleration sensor

+ Left-right measurement, temperature-stable
- Relatively heavy, no modular design

Stages Carbon GXP

The Stages crank replaces the standard left crank; it only measures the power of the left leg and multiplies it by two, which limits the measuring accuracy in principle. The sensor is located on the inside of the crank and is extremely light: it weighs just eight grams more than the SRAM Red crank. However, the wireless connection via ANT+ is as full of holes as a Swiss cheese and the data quality varies greatly. Under laboratory conditions, a measurement is successful and shows that the measuring principle can work in principle. However, it is hardly usable in this form.

• Price: 799 Euro
• Weight: 175 gram (175 mm)
• Software: StagesPower app, firmware update possible
• Connection: ANT+, Bluetooth LE
• Battery*: CR 2032, 200 hours
• Accuracy*: +/- 2 per cent
• Cadence: Acceleration sensor

+ light and invisible, price
- Unstable wireless connection, poor data quality

SRM Origin

New top model from the inventor of the Powermeter, in co-operation with Look. Super light, with exchangeable BB30 shaft. Modular design (interchangeable cranks) and therefore future-proof. Rotating pedal eyes for quick adjustment of the Look carbon cranks to different lengths (170/172.5/175 mm). The focus is entirely on clean data acquisition with 16 strain gauges in the spider and a magnet to record the cadence without interference. The crank works with any ANT+ device, but the optional SRM PowerControl VIII computer allows better data handling. High-end system for professionals.

• Price: 2618 Euro
• Weight: 658 grams (52/36 Shimano)
• Software: Data evaluation, firmware update via PC-VIII computer
• Connection: ANT+
• Battery*: installed, 1,400 hours
• Accuracy*: +/- 1 per cent
• Cadence: Magnet

+ light, data transparency, modular design, software
- Price, no left-right resolution

Sram Quarq Red DZero

Blends seamlessly into the Red group with carbon crank arms. Similar to SRM, the device measures the torque between the crank and blade and calculates the cadence with acceleration sensors as well as the left-right balance from the torque. Optionally available with BB30 or -24 mm shaft. Good app. The zero point shifts when the temperature jumps, which resulted in a deviation of up to 20 watts in our test (15 degree difference). SRAM claims to compensate for the temperature by calculation. The best price-performance ratio among the crank sensors in the test.

• Price: 1075 Euro
• Weight: 751 grams (175 mm, 52/36)
• Software: Qalvin Bluetooth LE app, firmware update possible
• Connection: ANT+
• Battery*: CR 2032, 200 hours
• Accuracy*: +/- 1.5 per cent
• Cadence: arithmetically

+ Price, modular design
- Temperature-sensitive

Power2max NG Rotor 3 D 24

A crank sensor based on the SRM principle that fits a whole range of cranks; it is charged via mini-USB, a rubber plug seals the socket. The raw measurement signal is more unstable than comparable power meters, which is probably due to the calculation of the cadence using an acceleration sensor. Temperature compensation worked perfectly in the test. Also available as a BB30 version. The predecessor model Type-S is significantly cheaper: It runs on batteries, is specified to an accuracy of +/- 2%, but otherwise offers comparable performance features (sensor with aluminium crank from 690 euros). Carefree system for cost-conscious riders.

• Price: 1257 Euro
• Weight: 795 grams (175 mm, 52/36)
• Software: Firmware update possible
• Connection: ANT+, Bluetooth LE
• Battery*: Internal battery, 150 hours
• Accuracy*: +/- 1 per cent
• Cadence: Acceleration sensor

+ temperature stable
- nervous display

Crank powermeter in comparison

The Graphic shows how widely the measured values of the crank power meters vary. The values of an Infocrank model serve as a reference (horizontal line at zero) - the coloured dots show the deviations of the measured values of the other devices (and the second Infocrank) at 100, 200 and 300 watts (single-leg drive). At 300 watts, the values deviate between +2.3 and -1.5 % - not bad, but a little more than some people claim. SRM, Quarq and Rotor (pink, partly covered) are extremely close to each other. The second Infocrank deviates the most (light blue). Who is measuring correctly? There is a lot to suggest that the correct values are to be found where the coloured dots are clustered.