Testing the Power2Max and Look KeO Power pedals

How many power meters is too many?

I'm currently suffering from a surfeit of bike related gadgets, the latest being a Power2Max crank-based power meter and a set of Look KeO Power pedals. Both devices measure your power output and, being a bit of a power meter geek, I've been asked to test them.

It's going to take some time to complete the full reviews so I thought I'd blog about how I'm going about it and any technical problems that I encounter. If anyone has any tips or experience with these power meters, I'd welcome your comments. I'll update the blog as I go.

The kit

We were sent a Power2Max with a set of BB30-compatible Rotor 3D Plus cranks – no chainrings unfortunately, but I was able to scrounge some from another bike. And no head unit – you'll need an ANT+ compatible one, such as a Garmin Edge 500, 705 or 800, Bontrager Node 1 or 2, or an O-synce Macro X. I wouldn't recommend using a CycleOps Joule, which is ANT+ enabled but doesn't have a function to check the calibration of the Power2Max. The Power2Max kit weighs 776g (power meter, chainring spider and bolts, crank arms and bottom bracket) and costs €1,165 (£1,000) direct from Germany. Cheaper options are available.

Power2Max

The Look KeO Power pedals came with pedals, cleats, sensors and a Polar CS600X computer with Protrainer 5 software – £1,699.99 for all that or £1,499.99 without the computer. But don't think you can get away with using a Garmin with these pedals – they're not ANT+ compatible, so you'll need a Polar CS500 or CS600X to read the data via Polar's proprietary WIND protocol. The whole Look-Polar setup is seriously light, with the pedals weighing 344g, the sensors 36g and the computer just 40g.

Look KeO Power pedals

Test rig and installation

My regular road bike didn't have a BB30 bottom bracket so I commandeered a Cannondale SuperSix 105 that we tested recently. I swapped the back wheel out for one of the PowerTap SL+ wheels that I use in training to act as the control. I own three PowerTaps and have found them to be reliable and stable under changing temperature conditions. As far as I can tell, they agree with other to within a couple of watts. The one I'm using in this test was about 5-8W lower than a Quarq power meter that I tested earlier in the year, which is entirely consistent with losses you'd expect through the drivetrain.

I'm not the world's best mechanic so I sent the two power meters over to our workshop, where our hard working wrench George Ramelkamp installed them. He's experienced in dealing with new-fangled bits and bobs so it didn't take him too long to work everything out. The Power2Max was the easiest to install, as it's just a crank and it comes with the right tool.

The KeO Power pedals were a bit more fiddly, as they require you to be quite precise with the alignment of the sensors (you get a tool to do this too). Once installed, you need to tell the Look sensors what length cranks you use. This is quite quick and is done via a button on the left-hand sensor, although you're limited to just four lengths: 170, 172.5, 175 and 177.5mm.

First ride

Initially I opted to pair my PowerTap with my Garmin 705, the Power2Max with the Joule (oops) and obviously the Looks had to go with the Polar CS600X. This doesn't take long, although when you've got two ANT+ power meters running at once you need to be careful with which one you 'wake up' to pair.

I performed initial zero offset checks indoors, where it was about 20°C, then set off for my first ride, where it was about 2°. The first problem became apparent immediately: the Joule would display the power of the Power2Max but I couldn't record with it because it needed a heart rate or a speed signal. I was using a non-ANT+-compatible Polar heart rate strap so that was out. And my Garmin, which does record speed, was talking to the PowerTap, so that was out.

No matter, this was only a first ride – I'd wake up the Joule from time to time to eyeball the power compared to what I was seeing on the PowerTap/Garmin. It looked okay right near the start but then drifted downwards to about 30W below my PowerTap.

Second problem: the Polar was set to record every five seconds, so I wasn't seeing much in the way of meaningful power figures. I also couldn't see average power on any one of the screens so I just amused myself by checking my left/right balance (58:42% unless I was going hard) and my heart rate. From what I could tell, the Look KeOs were reading lower than my PowerTap too.

Three hours of solid riding done and I downloaded the data I had. The PowerTap numbers were perfectly consistent with my effort level and my recent training. The Power2Max didn't record so I couldn't compare, while the Look KeOs gave me an average power about 25W lower than my PowerTap. Given that I didn't do any mid-ride manual zero offsets, that was okay. At least they worked.

Second ride

Lesson learned from the first outing, I swapped the head unit pairings around so I had PowerTap/Joule and Power2Max/Garmin 705. This meant I could record off both head units as well as being able to stop and zero the torque on the PowerTap and manually zero the Power2Max during my next ride (the Looks are zeroed just by turning the transmitter off and on and the Power2Max should zero whenever you stop pedalling).

I made a mental note to do this a couple of times to try to iron out drift caused by temperature variation. I also set the Polar CS600X to record every second rather than every five. This gives a total recording time of six hours – not great considering you can get double that on a Garmin, but still well within my needs. Four-and-a-half hours of battling a typical December day later and I had some data. And was quite hungry.

Data

PowerTap/Joule Power2Max/Garmin 705 Look KeO-Polar CX600S
Overall avg (W) 243 216 216
Distance (km) 137.3 137.3 137.2
Total time, inc stops 4:32:04 4:32:05 4:29:00 (not inc all stops)
Avg moving speed (km/h) 30.9 30.9 30.9
Total climbing (m) 1372 1423 1275
Avg cadence (rpm) 87 87 88
Work done 3876kJ 3465kJ 3958kcal (burned)
Selected segments
Pre-ride zero 525 -335 OK
Hill 1 287 264 266
Hill 2 295 272 285
Hill 3 289 257 273
Hill 4 288 260 268
Stop to zero 524 -277 OK
Tempo 1 298 270 271
Brief stop off the bike
Tempo 2 282 258 259
2min section of last climb 294 272 268
Post ride zero 524 -298 OK

Conclusions so far

I'm getting data and it's fairly consistent from all three power meters. This is a good thing but there are some anomalies. At first glance it might appear that either the Power2Max and Look KeOs are reading low or my PowerTap is reading high. Anything is possible but I'm disinclined to believe the latter as, based on my race data, it would imply that I'm more aero than just about anyone in the world.

The Power2Max appears to read lower on climbs than the other two. I'm also curious about the 58:42 L/R balance measured by the Looks, which evens out to 53:47 or even closer to parity when I'm going hard. Is this because I favour my left leg that much on a long ride, or is the right pedal under-reading because it's slightly out of alignment? Something to check in the next set of testing.

Here's something a bit clever. The following graphic shows how the three power meters stack up against each other using virtual elevation modelling, a method invented by Robert Chung (who kindly did this graph for me) for estimating CdA (coefficient of drag x frontal area) and Crr (coefficient of rolling resistance) using a power meter. You can do it the other way round, holding CdA and Crr fixed and plotting the elevation you'd expect from the power and speed data. Or as Robert did in this case in order to show the variance between power meters, hold Crr fixed, choose a CdA for each power meter and match up the shape of the curves.

Virtual elevation profiles of Look KeO Power pedals (Blue), Power2Max (Green), PowerTap (Red) and measured elevation (Black dotted line). CdA different for each power meter. Courtesy: Robert Chung

The virtual elevation won't, in general, match up with the actual measured elevation because wind, braking, changing position and road surfaces are all factored in as elevation. Also, measured power was different across all three power meters. What we can do is get an idea of how the power meters track against each other over time. The fourth section in the graph above shows this quite well: the blue Look and red PowerTap lines are nearly superimposed, whereas the green Power2Max line drifts around.

Fixing CdA at 0.404 (PowerTap value, which we think is realistic for all the gear I was wearing and the bike I was riding), the Look and Power2Max appear to be under reporting as they quickly drop off the virtual elevation profile. The alternative is to fix CdA at 0.35 (which we think is unrealistic) and the PowerTap will appear to over report. Courtesy: Robert Chung

After seeing this I'm more satisfied that the Look and PowerTap are giving consistent results, with the Looks possibly under reporting. But I'm not happy with the Power2Max at the moment. In ride number 3 today, I performed a zero offset at the start and got -329 (similar to Sunday) then again at the finish after two hours and it was -262. During this time, the Power2Max started off reading 40W higher than the PowerTap but finished 25W lower. Upon further inspection of the file it took about 20 minutes to stabilise to the lower value.

That's a big swing, but Power2Max say this can happen in the first part of the ride until the unit adapts to the temperature (2-4W/degree). A similar problem can arise if you're climbing a long hill where the temperature changes from bottom to top. Alpine climb anyone?

What next? Julius Jennings from Koolstof Coaching has offered to bring his calibrated Computrainer up to Bath so we can independently test each power meter against a standard. I will also do a torque calibration based on known weights. Polar have checked my pedal setup and have confirmed it's OK but they are sending me a new computer as the battery in the CS600X looked rather lacklustre. They also tell me that they're going to be some major updates to their head units and web software next year which should make data collection and transfer even easier.

To be continued...

Related Articles

Comments

Back to top