Quarq Calibration

I trust my PowerTap.  I’ve ridden it in pouring rain, and even in inches of snow.  I’ve ridden it when temperatures have dipped into the single digits, and I could barely read the display on the computer.  I’ve left it baking in the sun for hours while I’ve been at work. Fact is, I’ve subjected it to all seasons, and all conditions imaginable, and it has always consistently and reliably delivered the data.  In the past three years, I can count on one hand the number of times my PowerTap has dropped a data point.  I never question whether it’s going to work, or the accuracy of the numbers it transmits.

I did not feel the same way about my new Quarq when I first got it.   From the very first ride, I knew I needed to build up my trust in the data it produced.  It’s just how my brain works.  Luckily, I had the means to do so.  For many a ride, I put it up against my old and reliable PowerTap.

Based on feedback from others who have owned both a Quarq and a PowerTap, I was expecting the Quarq numbers to be a bit higher.  After several rides, I found this not to be the case for me.  In fact, the Quarq was consistently lower.  For endurance rides, the Quarq would show 10-20 watts lower, on average.  It would also produce lower numbers for various ranges within a ride (5-min, 10-min, 20-min, etc.).

What bothered me most is that the Quarq was lower than the PT, which doesn’t make sense.  If both power meters were measuring accurately, you would expect the Quarq to read slightly higher due to the dissipation of power through the drive train from pedal to hub.  I mean, physics always wins, right?

I zeroed the torque on both power meters before every ride.  Same result, again and again.  So I perused the Quarq manual, which states that the crank bolts should all be tightened to the same torque, and even provides a recommended torque value of 10 N-m.

Out came the torque wrench.

I was a little surprised when I was able to turn three of the bolts with the torque wrench set to just 7 N-m.  The other two bolts were at least 10 N-m.  I seemed to have found my problem.

After tightening the three crank bolts to 10 N-m, I collected more data.  The comparisons I was most interested in used data from intervals.  After all, that’s what I use my power meter for, primarily.  The results did not make me warm and fuzzy.

In this workout, the intervals were high-intensity, 3-minute efforts.  The Quarq was consistently lower than the PowerTap by 20-25 watts.  This is really bad for training.   During this workout, I had the computer dialed in to the Quarq, so I saw it’s numbers.  (The PowerTap computer was in my pocket).  I was supposed to do 5 x 3:00 intervals, but after seeing the dismal 323 watts on the computer for the third and fourth, I threw in the towel.   323 watts was way below my target range of 350-380 watts.  Had I been looking at the PowerTap data, I probably would have finished the workout.

The Quarq was lower by roughly 6%.  This is not acceptable, as explained above.

So I did the static calibration test listed on the FAQ page on the Quarq web site.  I did a calibration with each pedal in the forward position, first with no static load applied, then with a 20 pound weight resting on the pedal.  The instructions say to take the difference in calibration values and divide by 32 to determine the measured change in torque.

Right pedal: (-175) – (-643) = 468/32 = 14.625 N-m

Left pedal: (-166) – (-634) = 468/32 = 14.625 N-m

Next, calculate the torque applied by the static weight:

Weight (lbs) x crank length (mm) x conversion factor = Torque (N-m)

20 lbs x 175 mm x 0.00444822 = 15.56877 N-m

The error is the difference:

15.56877 – 14.625 / 15.56877 = 0.0606

That number looks familar.  6% error.

What do to?  In the back of my mind, I knew that I never torqued the crank bolts equally.  All I did was torque three loose bolts to 10 N-m.  I never knew what the torque was on the other two bolts.  They were already tight.

Out came the torque wrench.

I loosened all five crank bolts and re-torqued them.  I did it the way you’re supposed to, torquing each bolt incrementally in a star pattern.

I then did the static calibration test again.

Right pedal: (105) – (-384) = 489/32 = 15.28125 N-m

Left pedal: (103) – (-393) = 493/32 = 15.40625 N-m

Two things changed: the calibration values were very different from the results taken before re-tightening the crank bolts, and most importantly, the measured torque was more accurate.  Using the average of the left and right pedal measurements to calculate error:

15.56877 – 15.34375 / 15.56877 = 0.0145

1.5%.  Much better.

Up next, more field tests.  Here is a workout with 10-minute intervals.

That’s more like it.  The first interval is a little curious, but the Quarq was much closer to the PowerTap for the other three.  I wonder if this is because it was a chilly morning at 6:00 AM when I did these intervals.  I’ve seen reports that temperature can affect the Quarq.  Maybe it was just getting warmed up

Here are the ranges from a longer endurance ride.

The numbers for both the Quarq and PowerTap are nearly identical.  Very nice.  Here is a 10 x 30 second sprint workout.

Very close once again for all but three of the sprints.  To tell you the truth, it’s pretty tough to capture the matching 30 second sprints from two power meter files.  Being off by just 1 second can have a big impact on the average power for the interval, so these numbers are more prone to human error than mechanical error.

The Quarq comes calibrated from the factory, but as you can see, stuff happens.  My advice to anyone who thinks their Quarq is behaving badly is to check the crank bolts.  Re-tightening all five crank bolts, progressively, in a star pattern seems to have resolved my issue.  Now my Quarq is in sync with my PowerTap.

So after all this, I can say that I’m much more confident in the Quarq than I was in the beginning.  But I gotta tell you, if I’m testing, or doing intervals where I want to be damn sure of the power coming out of my legs, I still reach for the PowerTap.