Power2Max review by Steve Irwin

v1.2 15th Feb 2012

Notes on this version

I originally wrote this review in late May 2011, shortly after purchasing a Power2Max. I have since purchased a second Power2Max, and also my Quarq Cinqo failed in late 2011 and was replaced under warranty. I've re-written this review fairly extensively to cover what I have learned between May 2011 and Feb 2012.

I'll refer to my first and second Power2Max and Cinqo in this review. First Power2Max refers to the one that I previously covered in my review, second Power2Max refers to the additional one I purchased a few months later. First Cinqo refers to the one that I previously talked about in my review, and second Cinqo refers to the warranty replacement after the first Cinqo failed.

I want to say up front that it seems likely that some of the issues with my first Cinqo that I mentioned in my previous review were due to it being faulty. It didn't exhibit problems consistently enough for me think that I ought to return it, but the second Cinqo is definitely much better. Also, it's worth mentioning that when the Cinqo did fail completely, Bob Tobin at CyclePowerMeters provided excellent service, sending me a replacement the same day he received the faulty one.

What is a Power2Max?

Power2Max appeared on the market in early 2011 selling a crank based power meter, very similar in concept to the Cinqo and SRM. They arrived on the scene with none of the usual prolonged anticipation and endless waiting, with most people only becoming aware of their existence a relatively short period of time before the product became available. Their pricing is very good compared to a Cinqo or SRM, and I was particularly interested in the fact that the Power2Max doesn't use a magnet / reed switch combination, as I had seen issues related to this with my first Cinqo.

As soon as it became clear that some users were indeed receiving their Power2Max power meters, and they seemed to be working okay, I was willing to order one, as I was having quite a few problems with my first Cinqo at the time.

I had used two PowerTaps and a Cinqo prior to using the Power2Max. I have never used an SRM, so will not be able to draw any comparisons with that. I have only ever used Garmin head units with my power meters.

I wrote this detailed review because at the original time of writing there wasn't a lot of information out there about Power2Max, but at the same time there was a lot of interest from people who were keen to buy one if their confidence level was increased.

Buying a Power2Max

The Power2Max website is http://www.power2max.de/ I ordered a Rotor 3D 130mm version for my first Power2Max, choosing the option to include the crank and tool. The tool is needed to assemble and disassemble the spider and drive-side crank arm.

My order was placed on 30th March, and I received an email from Power2Max confirming my order and requesting payment of 1007.90 Euros (including shipping to the UK) by bank transfer. This method of payment may surprise some customers, but it is apparently common practice in Germany. I had until 25th April to pay, so didn't need to send the money straight away. The email also gave an expected delivery timescale of week 19. My UK bank charged 9 GBP to send the payment via a method that took a few working days. After currency conversion the total cost was 920.30 GBP.

By my calculation, the middle of week 19 is 11th May. My Power2Max was despatched to me on 12th May via DHL, and was delivered to me in the UK on 18th May.

When I ordered my second Power2Max, the timescale was very similar to the first time, and the process was mostly the same. The only difference was that the second time, I had bought the crankset myself from Rotor USA as I wanted to get 165mm cranks, and Power2Max required me to post the driveside crank to them for assembly testing. This was returned with the Power2Max power meter. Power2Max now sell 165mm cranks themselves, so there would be no need to go through this more complex process.

What is in the box

Included in the box is:

Note that chainrings are not included.

Weight

The original Rotor spider weighs 58g. The Power2Max spider weighs 264g, so the extra weight is 206g (130mm version).

Using with TT rings

There are a couple of "gotchas" waiting for the unwary user who wants to fit solid TT rings to a Power2Max.

The first "gotcha" is that if you assemble the spider and drive side crank arm then try to put a large SRAM TT ring onto it, you will find that you can't do it, the hole in the chainring is too small to be able to get the battery cover through. You have to first put the large chainring in place, then attach the crank arm. This means that if you ever want to change the chainring, you'll need to disassemble the crank arm and spider to remove it. The small chainring is no problem as it goes on the other side so the battery cover doesn't have to fit through it.


Figure 1

The second "gotcha" is that the battery cover has a chain catcher protrusion on it, and this will touch a solid chainring. Power2Max say that you should sand this down if using a solid chainring, to leave a 1mm gap between it and the chainring. You aren't damaging the casing at all by doing this, it's purely a protrusion beyond the outside of the casing, and is easily filed or sanded down.

The chain catcher protrusion can be seen at the bottom of the image in figure 1.

Battery compartment

The Power2Max takes a single CR2450 battery, the same as a Cinqo, and specified battery life is the same as a Cinqo at 400 hours. I had to replace the battery in my first Power2Max after 8 months, so the figure of 400 hours seems plausible, but I haven't tracked how many hours I've used it for.

The battery compartment is secured by 3 small crosshead screws. This is no problem at home, but if your battery dies while out on a ride, it would be quite a bit more fiddly to change compared to a Cinqo's screw off cap that can be removed and replaced by hand, whereas with the Power2Max you'd need to open up your toolkit to get an appropriate screwdriver out.

In use

The Power2Max is easily detected with a Garmin head unit just like any other ANT+ power meter. It transmits both power and cadence, and doesn't require a magnet to be attached to the bike.

Zeroing

I have found that many people do not understand the concepts of slope and zero offset as they relate to power meters, so some explanation is merited here.

Internally, power meters measure force with strain gauges. The electrical signals must be translated into torque measurements in known units. You may recall from maths lessons at school that the equation of a straight line is: y = mx + c At least, that is the form that we were taught at my school. m is the slope or gradient, and c is the intercept or zero offset, i.e. where the line crosses the y-axis when x=0.

In the context of a power meter, we can think of x as representing the raw electrical signal from the strain gauges, and y as the measurement in known units that we are interested in. c is referred to as the zero offset, because it is needed to correct for the fact that even when no force is being applied by the rider there will still be an electrical signal generated by the strain gauges.

The zero offset of all power meters tends to change with temperature, and also with changing mechanical characteristics, e.g. as the interface between the chainrings and the spider "beds in" with use. When you use a Garmin head unit to issue the "calibrate" command, what you are actually doing is setting the zero offset.

The slope of a power meter tends to be much more stable than the zero offset, and it needs to be, as it effectively cannot be changed while riding. A true calibration of a power meter entails determining the correct slope using a static torque test, where a known weight is hung from the pedals, and the power meter's torque measurements are used to determine the correct slope. Power2Max claim that their power meter's slope will always remain valid, and thus it is only the zero offset that the user need be concerned with. There is no way to change the slope of a Power2Max (the user can change the slope of a Cinqo or SRM), but the user is able to check the slope, a process that I will cover later in this review.

The Power2Max manual states that it will automatically set the zero offset whenever you stop pedalling for 2 seconds. There is no way for the user to know that this zeroing has taken place, or how much of a change has occurred, nor is there any way to turn this feature off. Some users may not like this lack of control.

You can still manually set the zero offset using a Garmin head unit in exactly the same way as you would with other power meters, using the "calibrate" command when there is no force being applied to the pedals. When you do this, the new zero offset is displayed on the screen, so if you remember what it is each time, you know how much it has changed and you will have an idea of how often and under what circumstances you need to zero it again.

"Null" watts vs zero watts


Figure 2

One source of annoyance for me with the Cinqo is what happens when you stop pedalling. Because it only knows how fast the cranks are moving once per pedal revolution, it simply continues to transmit the same power value when you stop pedalling. A Garmin head unit will display the same power value for 3 seconds, then either go to 0W if you're moving, or "null" watts if you're stationary. This can really mess up your mean max power stats, e.g. look at this section from a ride with a Cinqo in figure 2. Note how the yellow line (power) goes to zero while the blue line (speed) is non-zero, but as soon as I stop, the yellow line disappears. It says my average power was 254W, but 50kJ of work over 3:38 is actually an average power of 229W. The difference arises from the period of "null" power, which WKO+ effectively treats as not having happened. With a Power2Max, the yellow line would have stayed at 0W and the correct average power of 229W would have resulted. In this respect the behaviour of a Power2Max is like that of a PowerTap rather than a Cinqo.

Cadence range


Figure 3

The specification says that the Power2Max works from 20 to 200rpm. I can't pedal fast enough to verify the 200rpm figure, but I can confirm that it works perfectly down to 20rpm. Even though it is taking 3 seconds to complete each pedal revolution at 20rpm, the power figure still updates once per second. Below is an example from when I was testing the low cadence behaviour:

Dropouts

My first Cinqo suffered from an extremely variable level of dropouts. Most of the time I didn't get any dropouts, but on some occasions the problem was bad enough to make it almost useless. I haven't had the same problems with my second Cinqo. I've now been using a Power2Max for 9 months, and have never had any problems with dropouts.

Power spikes


Figure 4

What I did almost always see to some extent in files from my first Cinqo is power spikes, where the cadence and power change together to an incorrect value. This can happen both upwards and downwards. Figure 4. is an example from a Cinqo file.

The changes in power and cadence represent incorrect data, speed remains more or less constant. I haven't seen any examples of such problems with the Power2Max. My second Cinqo is much better in this respect, but there are still some small spikes in the recorded data, though nothing like as big as the ones in the graph above.

Micro-variations in data


Figure 5

The human body naturally produces slightly varying power each pedal stroke, but power meters also differ in how their measurements vary on a second by second basis. For example, the PowerTap and Cinqo work in fundamentally different ways, meaning that, at least in theory, a Cinqo should exhibit less second to second variation than a PowerTap, however I have seen it said by others that this is not the case in practice.

So I have dug out some old PowerTap files, as well as looked at Cinqo files and my recent Power2Max files to try to find examples of where my power was very constant over a one minute period, to see what differences can be seen. The power axis is scaled to show 200W to 300W in all cases (I haven't scaled the cadence and speed axes to match).

My subjective verdict is that the Power2Max has the lowest level of micro-variation, followed by the Cinqo, then the PowerTap.

Accuracy

Comparison of speed on a Kurt Kinetic turbo trainer

The Kurt Kinetic is generally very consistent in terms of the power vs speed relationship. I have now used my first Power2Max and second Quarq for many turbo sessions, and my conclusion is that they give extremely similar power readings at the same Kurt Kinetic speed, approximately within 1% of each other.

Static Torque Test

All of the Cinqo, PowerTap and Power2Max support checking of the calibration using a Garmin and hanging a known weight from the pedal with the crank arm horizontal.

I have done this with my first Cinqo in the past and found it to read around 2% too low. I have never seen much variation in this figure over time, i.e. the Cinqo's slope appears to be quite stable.

How you perform a static torque test with the Power2Max is as follows:

  1. With the bike held stable in an upright position, and the crank arm horizontal, use the "Calibrate" feature on the Garmin and note down the number displayed (-806 in my case)
  2. With the rear brake applied, attach a known weight to the pedal spindle and repeat (gives a number of -728 in my case)
  3. Calculate the actual torque being applied as W * 9.80665 * C / 1000 where W = weight in kg, and C = crank length in mm (actual torque = 13.337044 in my case with an 8kg weight and 170mm cranks)
  4. The Power2Max appears to use some sort of arbitrary units of "ppm" for the number displayed on the Garmin. It doesn't really matter what they represent, as all we need to do is calculate the slope as T / (Wppm - Uppm) where T is our calculated actual torque, Wppm is the number displayed by the Garmin with the weight attached, and Uppm is the unweighted number. In my case, this works out as 13.337044 / (-728 -(-806)) = 0.1709877436
  5. The Power2Max includes a calibration certificate in the box that you can compare this number against. In my case, it specifies 0.170035 Nm/ppm, so the error in my slope measurement is 0.6% with the large chainring. With the small chainring the error worked out at 3.1%.
  6. However, a problem with this is sensitivity to changes in the numbers. If my 2nd number had changed by 1 digit, to -729, that would change the calculated error with the large chainring to 1.9%.

    The Cinqo is better in this respect, because the same weight changes the number it returns by ~420 rather than ~80 for the Power2Max.

    If you want good accuracy in performing a static torque test with a Power2Max then you're going to need a very heavy weight to mitigate this issue.

    Zero offset drift during a turbo trainer session

    If we manually zero ("calibrate") the power meter at the start and finish of a turbo trainer session, we can see how much the zero offset has moved by.

    Based on the static torque test, the Power2Max's power readings will change by approximately 1.5W for each unit of movement in the zero offset. From what I have seen with my two Power2Maxs, and information from other owners, it seems that this figure is roughly the same for all Power2Maxs.

    For a Cinqo, the power reading will change by approximately 0.25W per unit of zero offset.

    My first Power2Max tends to drift by around 9W during a 40-60 minute session.

    My second Cinqo tends to drift by around 7W.

    My second Power2Max tends to drift by around 45W.

    Based on my two Power2Maxs, and information from other Power2Max owners, it seems there is a large amount of variation in how much they drift during turbo trainer sessions. 9W is at the low end of the range of variation, and 45W is at the high end, but certainly doesn't seem to be especially unusual. My second Power2Max was returned to Power2Max in Germany for them to look at, but they have told me that the behaviour is normal, and they expect the issue to be compensated for by the auto-zero when the rider freewheels.

    Accuracy conclusion

    When properly zeroed, I have no reason to doubt the accuracy of any of the power meters I have owned. They all give a very similar relationship between speed and power on the Kurt Kinetic, and the static torque tests I have done suggest they are accurate.

    The major issue with the Power2Max is that some units exhibit a lot of movement in the zero offset, and it isn't always convenient to freewheel to make the auto-zero take effect, or to manually zero. For example, suppose you are riding up a mountain, where the temperature at the top is much lower than at the bottom. If it has a relentless steep gradient, you will not be able to re-zero the power meter, and the readings will become increasingly inaccurate. Or, on the turbo trainer, it can take a long time for the offset to stabilise. I wanted to use my second Power2Max for a ramp test, during which it would not be acceptable to freewheel. So I had to warm up first, manually zeroing the power meter every two minutes, waiting for the zero offset values to stop changing. It was only after 22 minutes that the zero offset had become sufficiently stable that I could start the ramp test. This makes my second Power2Max pretty much unusable on the turbo trainer, as I'm not willing to spend over 20 minutes of each session waiting for the offset to stabilise. On the other hand, my first Power2Max and second Cinqo are fine, as was my first Cinqo (in terms of zero offset drift).

    However, for normal outdoor riding, the Power2Max is fine. The flow of air past it when riding means that it reaches a stable offset much faster than it does on a turbo trainer - approximately 10 minutes. And also, outdoors, road features such as roundabouts, traffic lights, other vehicles, tend to force the rider to freewheel, at which point it auto-zeroes.

    It is also fine for many turbo sessions where you have the chance to re-zero, e.g. if doing an interval session, you can re-zero after you warm up, and then also in the recovery period between each interval.

    Whether or not the zero offset drift will be a problem will depend where you live and the type of riding you do. It appears to be a lottery just how much the individual unit you receive will drift.

    If you use the turbo trainer a lot for sessions where it is important to ride continuously without the interruption of the 5-10 seconds it takes to stop pedalling and re-zero, then it may not be worth the risk of receiving a Power2Max unit that drifts by tens of Watts. If you live in an area where you have big climbs with a lot of temperature difference between the bottom and top, where you wouldn't be able to re-zero during the climb, the same applies.

    Responsiveness


    Figure 6

    10 mile TT start

    The following image shows the start of three different 10 mile TTs on the same course using each of the three power meters. For the two with a grey section on the left, the change to black indicates when I pressed the lap button just before setting off. For the one without a grey section I started a new file by pressing the start button just before setting off.

    Power2Max and PowerTap both have a similar delay before the power starts to increase. Power2Max shows the cadence and speed increasing at the same time as the power increases. PowerTap has a slight delay between power increasing and speed and cadence increasing. The Cinqo has the worst behaviour with power and cadence not starting to be recorded until several seconds after the speed starts to increase.

    The graph below (Figure 6a) shows an example from another TT, where the Power2Max didn't respond as well at the start as it did in the above example. This time, using the Duotrap for both speed and cadence, two non-zero cadence samples were recorded while the recorded power was still 0W.


    Figure 6a

    Turbo trainer responsiveness test


    Figure 7

    The following test was carried out:

    The same gear was used throughout the test, so cadence and speed should remain directly linked throughout the test.

    The Cinqo and Power2Max tests were carried out on a Kurt Kinetic with a pro flywheel fitted. As I no longer own a PowerTap, Jim Ley kindly carried out that test for me, and his Kurt Kinetic didn't have the pro flywheel fitted, hence the more rapid acceleration and deceleration, but this doesn't affect what we're looking at in the data.

    There are some important differences between the behaviour of the 3 setups as can be seen in figure 7:

    Power vs speed

    The Power2Max and Cinqo both have the increase in power lagging very slightly behind the increase in speed. The amount of lag appears identical to me. The PowerTap has the increase in power occurring in advance of the increase in speed, it is hard to say whether or not it is incorrect in the opposite direction to the Power2Max and Cinqo, as obviously the rider has to initiate the increase in speed by applying more power. Given the results above from the 10 mile TT start, I would say that the PowerTap has probably exhibited a slight lag again in the speed data, and the timing difference between power and speed is at least to some extent incorrect.

    Speed vs cadence

    As the gear was constant throughout the test, speed and cadence should be linked by one being a constant multiple of the other.

    The Power2Max exhibits a clear lag in the cadence data relative to speed. There is a 3 second timing difference between peak speed and peak cadence.

    The Cinqo looks to be very good in this respect, apart from a slight lag in the cadence changing at the start, but that could easily be due to the effort starting from ~50rpm. There will be a random element at this rpm as it will depend on how the start of the effort matches up with the 1 second interval ANT+ updates. Overall, cadence and speed are linked as we would expect.

    The PowerTap also looks pretty good, but not quite perfect. The cadence line is noticeably below the speed line at peak speed, but this may be correct, as it is also below the speed line during the low power periods, and the size of the gap should increase with speed. However it doesn't remain below the speed line during the periods of speed increasing and decreasing.

    The Power2Max behaviour in this test is disappointing. It is difficult to see how it can be an inherent hardware problem, especially given the results of the 10 mile TT start, but the lag in the turbo trainer scenario is 100% reproducible, and I also noticed the lag on the road. Perhaps they will be able to fix it with a firmware update in the future.


    Figure 7a

    This problem can be worked around by using a 3rd party cadence sensor. I attached a crank magnet to get cadence data as well as speed data from the Duotrap, and repeated the test. The Garmin automatically used the Duotrap's cadence in preference to the Power2Max cadence, with better results as in figure 7a.

    Summary

    The Power2Max appears to essentially work fine for many people's needs, and hence is good value for money for anyone who wants a crank based power meter with the freedom to use different wheels.

    Pros

    Cons