Finding the proper circumference for a cycle computer isn’t very hard, or is it? I guess I’ve heard more than one discussion about mileage on rides, with quiet mumbling up the sleeve to the effect of “you silly fool, your computer is wrong.”

Some manufacturer’s are the silly fools, because their instructions are silly. Shimano’s instructions for their Flight Deck tells users to roll the tire one revolution and measure the distance. They add insult to injury by then giving a table for various wheel sizes and tire diameters. Campagnolo is silent on the method, simply instructing the user to input the circumference — a cop out maybe, but at least it doesn’t confuse people.

A moments though should be enough to realize that the circumference of a wheel is smaller on a loaded bicycle. Simply rolling a wheel along the ground doesn’t cut it. Somehow the measurement must be made while riding, and I will now tell you how I do it.

First of all, I pick a reasonable distance to ride. Not so long as to be unwieldy for accurate measurement, but long enough to represent real riding. I choose 25 meters, since I can mount my bike and ride in a normal fashion in this distance, and I have a tape measure that stretches this far.

The trick is to count the number of revolutions of the wheels. If I know this count, I can simply divide it into 25 meters to get the circumference of the wheels as ridden. People who measure track for field events use a bicycle wheel with a counter attached. That isn’t necessary here, since all that is needed is to observe that there must be about 12 revolutions for a 700 C wheel. This corresponds to a 2083 mm circumference. If my wheel is larger, then I will end up past the finish mark after 12 revolutions, and contrariwise if shorter. All I have to do is adjust the 12 revolutions for the error distance.

I start with the valve down, ride the distance and note where the valve comes down near the finish. I then use the spoke spacing to measure the error. Assume that I have a 36 spoke wheel and suppose that the valve comes down 3 spokes after the finish line. then there have been 12 minus 3/36 revolutions or 11.92 revolutions in my ride. Since the distance is 25 m, this means that the circumference is 25000/11.92 mm or 2097 mm.

I have checked the accuracy of this procedure on a measured 2 mile track, and have checked the way points on a 30 mile rude using two different bicycles and two different cycling computers (Flight Deck and ErgoBrain). For most way points, they agreed to one hundredth of a mile. For a few they differed by at most four hundredths of a mile, which was likely due to minor variations in the course, since the differences occurred at way points and did not cumulate for an error in the 30 mile total.

On one bicycle, the circumference by the above method was 2085 mm, but the measurement of one revolution of an unloaded wheel was 2115 mm. This means that over a 30 mile distance, the second measurement method will,result in an error of almost half a mile. (Interestingly, Shimano’s table gives 2115 for this wheel.)