Power Games

Ask any follower of cycling what the latest must-have gadget is and they are likely to mention some kind of power meter. As a cyclist who is keen to improve I wanted to understand how they work and how they could possibly help me.

The latter question is simpler to answer, so I shall address it first. The point of a power meter is that it gives you a way to rapidly determine if your training is taking you in the right direction and to quantify to what extent you have progressed. It will not of itself make you faster, unless of course you were planning to carry the cash you have now spent on it while riding.

There are of course cheaper methods to determine if your training is working, such as timing how long it takes you to ride a predetermined route. Unfortunately this (and most similar methods) are very sensitive to the weather and other external factors. To see what I mean, play around on bikecalculator.com and you will find that if you add a very light 5 kph headwind to the initial calculation you have to raise the rider’s power from 150W to 195W to maintain the same speed: a huge difference in power to overcome a minuscule breeze you might not even notice. Sadly you can’t entirely solve this by riding in a loop: the same parameters with a 5kph tailwind require 112W so you would still need to raise your average from 150W to 153.5W; only 2% more, but still not nothing. Ambient temperature, air pressure, tyre condition, tyre inflation pressure, clothing: all these things will likely have a larger effect on your time from one week to the next than changes in your fitness.

One way of riding which does not have many of these problems, because it takes aerodynamics out of the equation entirely, is to time how long it takes to ride a set distance on a static turbo trainer. It is, however, very boring and quite unlike real riding; on the turbo trainer I find it difficult to keep my hands on the handlebars and instead tend to sit more vertically. It is very likely that my performance in these conditions is only vaguely related to my performance when my hands are on the bars.

Of course a significant advantage of the real-world-timing methods is that they inherently take into account the effect of changes in your weight. The turbo-trainer will not do this, and nor will any of the power meters.

Now, let’s survey the various ways to estimate and/or measure power and how they work.

Low-end turbo-trainer

If you have a turbo-trainer already you may be able to find its resistance curve versus speed described online. From this you can calculate a curve mapping indicated speeds to the power outputs required to achieve them. The formula for my Cyclops Fluid2, for speeds in miles per hour, is estimated by someone on the internet as:

Power = 0.0115 * speed3 – 0.0137 * speed2 + 8.9788 * speed

Others on the Internet suggest, with good evidence, that there is a variance of at least 50W depending on how well-inflated your tyre is and how tightly you have pressed the trainer to the tyre, and that the resistance increases after about 5 minutes once the fluid in the trainer has warmed-up. This means the numbers I can calculate are shaky at best, but should be comparable to each other as long as I warm up for 5 minutes first (which I always do), keep the tyre well-inflated and the turbo cam in the same place.

The obvious advantage of this is that you only need your existing turbo trainer, wheel-speed sensor, spreadsheet program, and a formula that with any luck someone on an internet forum has already worked out. The other big advantage is that what you are measuring is analogous to the power delivered to the road: if you improve your pedalling technique so that your upward-moving foot is providing less opposition to your downward-moving foot you will see that in your numbers.

The disadvantages are that it takes an amount of effort to work the number out, you have to do it after riding so you don’t get live data, your numbers should be comparable to each other but they cannot be compared to anyone else’s numbers, and it doesn’t give you a way to ride to a target power at an event. Also riding on the turbo trainer is a bit boring.

Net cost (for me): £0.00 + time.

High-end turbo-trainer

Some more expensive turbo-trainers have built-in power-measurement features which are a lot more accurate than the DIY method above. These generally have an electric brake to supply the counter-force to the wheel rotation and can therefore get a very accurate measure of the power you are outputting by measuring how much power needs to be supplied to the system which counteracts it, then factoring in the power-efficiency of that system. They typically also have links to a PC and provide virtual rides for you to go on which make using the turbo trainer more enjoyable.

This has all the advantages of the previous method, to which you can add live power readout & the VR features.

The main disadvantage carried forward is that it doesn’t give you a way to ride to a target power at an event.

Net cost (for me): £1000.00 or so.

Hub power meter

Hub power meters like the CycleOps PowerTap work by measuring the strain on the rear hub as it conveys torque from the drivetrain to the wheel via the spokes.

One advantage of this is that you can fit your power-sensing wheel to any bike you have, so if you want to ride to a given power output at an event you can do so quite easily. It also removes the tyre condition and its inflation pressure from the variables affecting the measured power: the only things left are losses in the chain links and drivetrain bearings, and these are less likely to vary hugely week-to-week. As with the turbo-trainer methods you will naturally see any improvement in technique as an improvement in power.

The biggest advantage is that you can go and ride anywhere you like, you are not limited to your garage or living room as you are with a turbo trainer.

Net cost (for me): £1000.00 or so.

Crank-spider power meter

Crank-spider power meters like the SRM and Quarq work by measuring the strain on the “spider” which transfers torque to the chainrings.

Compared to the turbo-trainers these are more portable: you can ride them outside. Compared to the hub power meters they are less portable. If you have more than one bike it is a massive pain to move the cranksets between them; and that presumes that they have the same bottom bracket so that it is possible. On the other hand, if you have one bike and multiple sets of wheels for various conditions then the crank-based systems will be considerably more convenient.

The advantage of this style of system is that it limits the potential sources of variable losses to the bearings in the pedals and the bottom bracket, neither of which are likely to change significantly from ride-to-ride when compared to the chain links or the rear-tyre condition.

The disadvantages are that it is much more difficult to move from bike-to-bike than the other systems and it is strongly tied to your choice of chainring tooth-counts; if your power meter is 39/53 and you want to ride a given course with 34/50 you will be doing so with no power meter. They also cost a lot, although if you are building a bike from scratch you can deduct the cost of the non-power-meter chainset you would otherwise have needed.

Net cost (for me): £1500.00 or so.

Crank-arm power meter

Crank arm power meters like the StageONE work by measuring the strain on one or both of the crank arms. The mechanisms we have looked at so far are at points in the drivetrain where the action of both of your legs have been mechanically combined, so if your pedalling technique has your upward-moving foot opposing the force of your driving foot you will see that as reduced power. Unfortunately, if you measure this in a naïve way with a crank-arm strain gauge it will show up as increased power. For this reason I am skeptical about the accuracy, and therefore usefulness, of any of single-crank-arm power meter.

The main advantage of a single-arm system when compared to a crank-spider system is that only the non-drive side crank arm needs to be replaced, which does not affect the rider’s choice of wheels, pedals or chainrings at all.

A two-arm-measuring system gives you the same combined power number as a crank-spider system, although the sides are combined in software not mechanically. The advantage is that pedal technique defects can be directly detected and highlighted: a significant benefit? Perhaps. The disadvantages are the same as for crank-spider strain gauges.

Net cost (for me): £1500.00 or so (two side).

Pedal axle power meter

Pedal axle power meters work by measuring the strain on the pedal axles. The systems I have seen measure both pedal axles; a single-axle system could be made but would obviously have the same problems as a single-crank-arm system.

Provided that a pedal axle power meter for your preferred type of pedals is available, these systems are much more attractive than crank-arm or -spider systems. It is easy to transfer pedals from one bike to another (although not as easy as switching a wheel) and it does not in any way limit your chainring or wheel choices.

The other advantages of the two-crank-arm systems are also available in these systems.

Net cost (for me): £1500.00, plus I have to switch to Look pedals which I don’t want to do.


Right now I would favour getting a rear-hub power meter from eBay, if I were to do anything. Fancy turbo trainer is a strong second because making it less boring to ride in the garage for hours at a time is an excellent side effect. For the future I think pedal axle meters are the most promising.

One Response to “Power Games”

Leave a Reply

  • (will not be published)

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>