Sunday, August 2, 2009

PowerCranks – The Final Verdict

My year with PowerCranks (PCs) has come and gone. I received them back in July 2008 and immediately installed them and took them for a spin. Instantly I knew these were something different as I was working muscles not previously stressed in the past despite the tens of thousands of cycling miles on my legs. But would that new pain I was experiencing with these new mechanisms translate into more power on the bike or merely a device for torture?

To answer that question I embarked on a disciplined training regimen of significant riding on the PCs. While I had initially agreed to ride the cranks exclusively, I quickly realized that I wanted to get on the regular cranks at least once a week. This was for several reasons: 1) I could not ride my usual Saturday hard group ride with them and keep up (in other words, those rides would quickly turn to solo endurance rides rather than provide a training stimulus for my VO2 and anaerobic zones), 2) I could not safely ride my group ride with the PCs due to the way they affect both bike handling and remounting (as someone focused on safety this was a big issue for me), and 3) I needed a little variety during the week. So while I did not use them as expected, I feel as though I've been able to assess the cranks and their potential to help riders.

First the facts: I felt I had stagnated in terms of the power I am able to produce. I've got power data going back to before 2003, and over the years that data suggests that I'm topped out in terms of performance. As a time trialist and road racer, my interest lies in aerobic power. As PCs are targeted towards triathletes which are clearly aerobic engines, I thought using them could provide just the right kind of training shakeup to perhaps give me 5 percent or more boost in performance. My workouts were geared towards enhancing aerobic power – a steady mix of tempo, threshold, and VO2 riding.

My best performances of the year were actually in the fall of 2008, within the first few months of using PCs. The graph below shows my normalized power for my time with PCs (red line) and compares against a few key time periods. The green line represents the season prior to using PCs. Do to travel and illness, I just couldn't get very good power down. The black line is an envelope of my all-time best numbers prior to using PCs. A few things to note about this chart: 1) even with a subpar year (the green line) prior to PCs, my power was only 10-15 W lower in the 30-60 minute range, about 4%, 2) my time with PCs shows only a couple of instances where I achieved higher power, and those differences were in the single watt range (i.e., 1-5 W differences), and 3) this graph represents absolute best performances (i.e., cherry picking) rather than repeatable power. This last fact is important. A one-off performance could be due to things like instrumentation error, a new chain, or a newly cleaned chain. More important to me is repeatable power. Nonetheless, the data is what the data is, and while I did set some personal records during my time with PCs, they were not significant in the slightest (1-2% at most). All plotted is my average power. That graph tells much the same story as normalized power.

Looking at a broader set of data, the graphs below show my full set of data for 2003 to present. You may notice that 2003 and 2004 were clearly the low points in my cycling. Those were the initial years of my serious cycling and I would expect those points to be the low hanging fruit. If you neglect those first 2 years you can see how tightly bundled all these data are.

But like I said, I care more about repeatable power rather than one-off performances. I want to know what I can expect to bring come race day. In that vain, I took all my daily power data and found the best performances. I averaged these top performances to come up with power numbers I could bank on. The plot below takes the average of my top 20 performances for times from 10 minutes to 60 minutes. What we see is the 2008 was clearly an aberration. It sticks out as being a below par year, yet even 2008 was not that bad. Power at any duration was no worse than 7-8 W. In other words, even though it was lousy in terms of absolute power, my repeatable power was actually quite good.

Of course, the careful reader may wonder about taking 20 rides and averaging them. That's a legitimate concern, and to mitigate that concern somewhat, we can compare 20 rides to 10 rides and also the peak values. The plot below for 2009 shows only a 1-2 W difference between 10 and 20 rides, and only a 5-10 W difference between absolute peak and a 10 ride average. The same can be seen for the 2007 data.

So if I look at all this data objectively, I would conclude that for me, PowerCranks did not positively affect my long-term power (that being 20 minutes or longer). The data collected indicates that my power numbers were within the norms of prior years, and any perceived increase in power was within the stated error of the measurement device itself (a Saris Powertap).

Now the manufacturer of PCs will counter that an inability to show an increase could be due to a few things: 1) I did not ride the PCs exclusively or 2) my stroke was already good and PCs would not be able to help further.

The first point is factually accurate. As already mentioned I did not use the PCs exclusively. However, PC use did account for over 50% of my training time over the last year (which in total was nearly 500 hours). I would have used them more but I experienced some injuries associated with the comination of the PCs and the Time ATAC pedals. I was getting tendonitis-like pain in my left ankle only when using PCs. I stopped using them for several weeks to heal. Upon returning to them, the ankle pain came back, though my hip flexors were ready for long rides with the PCs. Subjectively, I felt I was adapted to them within 2 months of use. I got to the point were I could ride for 3-4 hours with no pain and while sustaining my overall power. Since the manufacturer of PCs can not provide an objective way to determine adaptation, all we are left with is subjective reporting, and my subjective interpretation is that I was indeed adapted.

Regarding the second point, that is something we may never know. However, I do feel my stroke is pretty good. I live in the Houston, Texas area, which is pancake flat. My specific area has long, uninterrupted flat roads with a coastal breeze in the warmer months and consistent cold north wind in the winter. The terrain does not allow for much variable power. You just sit on the bike and grind away. I've got well over 70,000 miles of this riding in my legs.

So perhaps I'm just the exception and a genetic freak who has a perfect spin already. Or perhaps too much is made of unweighting on the upstroke. To this end, I sought out some actually pedal force data. Unfortunately there isn't very much pedal force data out there, but the exception is the work of Kautz, et. al. In “The Pedaling Technique of Elite Endurance Cyclists: Changes with Increasing Workload at Constant Cadence” Kautz and his colleagues measured pedal force data of category 1 and 2 riders. The study was performed many year before things like PCs, Rotorcranks, or elliptical chainrings hit their stride, and on a positive note the raw data is available on the internet. An example of some of this data is shown below, which is for the rider with the “worst” pedal stroke. The blue line shows the force tangential to the crank. You can see that after the rider passes the 6 o'clock position his stroke becomes “inefficient” as negative torque is applied, as indicated by the greenish line.

We can examine the normalized pedal force for all subjects in the plot below. In this graph, tangential pedal forces for each rider are divided by the peak tangential force for the rider. The dashed red lines indicate the absolute min and max values. Error bars in blue somewhat mask the average for the entire data set, but it can be seen that the overall average of the many subjects is positive through the entire stroke. In other words, riders tend to naturally generate a good stroke.

But what if the rider with the “lousy” stroke used PCs to get rid of that negative torque. The manufacturer has made claims on the internet that PCs force the rider to “unweight and nothing more”. So if we analytically remove negative torque from the riders we get an estimate of the potential increase in power due to that “perfect” pedal stroke. The results are rather surprising. The objective raw data would indicate the following improvements: 8.08%, 4.28%, 4.13%, 4%, 3.64%, 2.61%, 2.15%, 2%, 1.84%, 1.29%, 1.07%, 0.99%, 0.47%, 0.32%, 0.17%, 0.07%, 0.01%, and 4 riders at 0%. The average improvement for this set of riders is a whopping 1.77%. To put this in perspective, that's about 5 W on a 300 W threshold.

So clearly there's the potential (albeit not much based on the Kautz data) for improved pedal technique, but can PCs actually deliver that improvement? The figure below is from the manufacturer and shows a single user on PCs and regular cranks. The user switches between the cranks and force data is recorded. What we see is the PC data shows no negative pedal forces (as it shouldn't), yet this PC rider produces negative force when using regular cranks. I contend that if this PC rider trains from birth to death with PCs, he will likely never generate a force distribution on regular cranks that mimics PCs. Why? One can never develop perfectly coordinate pedal motion. Because fixed cranks are, well, fixed, all the pistons would have to be firing EXACTLY 180 degrees out of phase. I feel I'm pretty coordinated, yet even I know I can't get 180 degrees exactly. The slight phase shift means at some point my downward moving leg will be assisting my upward moving leg. This assistance will always cause the force distribution to deviate away from the PC force distribution. That's simple mechanics.

So is it all bad news for PCs? Are they nothing but a bunch of hogwash? Well, in my opinion yes and no. They are not some sort of magic bullet which allows you to take something from nothing. If you are at your physiological limit on regular cranks PCs, in my opinion, will not improve your power further.

But there's the rub – how to do you know when you are at your physiological limit? Unless you have many years of power data it's hard to know. After using PCs, I do feel they can aid a rider in developing their power. For riders that are still improving, PCs provide additional difficulty that could help them raise their pain thresholds and improve their power. Thus, I see them as more a mental tool rather than a physical one. Do I think you can get the same improvements without them? You bet. But for those riders that need a little nudging, they make work.

And so ends the PC experiment. My n=1 experience tells me they are as effective as a placebo solution at raising my power. They did nothing for my power. The interesting thing will be this fall when I get back in the full swing of more intense training. Due to a late May crash, my summer season was a washout, and I've had difficulty getting my power back in the oppressive heat (worst summer in decades). So the question will be can I return to the same form using regular cranks that I had last fall on PowerCranks? My bet is yes...