I have bought the Campangolo Khamsin wheelset, and now I would like to set the right wheel size in my Trek Incist 8i cycle computer. Before I do the custom wheel setting, I would like to know what would be the best choice?
The wheel has ETRTO 622x15c printed on it.
The diameter of the wheel is 622 mm. The circumference of the wheel is 1954 mm. Thus, if you ride on wheels, you have to enter 1954 mm on the cycle computer.
However, most cyclists choose not to ride on wheels but rather ride on tires. This is when things become complicated. Most cyclists measure the weight of the products and get very angry if the manufacturer stated the weight of some component incorrectly in marketing. Most cyclists do not measure the width of a tyre, and even if they did, the width of an inflated tyre depends among other things on rim width. So if you buy a 28mm tire, and measure its width as 25mm, the manufacturer can claim that the stated 28mm width was on some other rim size than the one you're using.
Because of this, if you buy a 28mm tire you can actually get a 25mm tire, as the manufacturer has an incentive to sell lightweight products, and the easiest way to sell lightweight tires is to sell a 25mm tire as a 28mm tire.
Not only this, but the tire's rolling diameter and circumference are affected by the weight placed on the bike, whether the cycle computer uses front or rear wheel for measurement as the weight distribution is uneven (most e-bikes use rear wheel, most separate cycle computers use front wheel) and the inflation pressure.
A just-inflated 28mm front tire (7 bar) has about 0.3% larger rolling circumference than a tire inflated two weeks ago (5 bar) 1.
The good news is that if you have some map service that is able to measure paths very accurately, and if you have a cycle computer with 0.01 km resolution, you can enter some guessed value as the rolling circumference and then ride a 30-50 km path. Use the map service to determine the precise length of the 30-50 km path. Then ride the path, taking no shortcuts and not stopping on any nearby shops. In principle, the longer the path the better, but on longer routes you need to stop somewhere to drink unless you carry a drinking bottle, and on even longer routes you need to stop somewhere to eat.
The true rolling circumference is
true_path_length / cyclometer_trip_distance * guessed_value.
So, for example I rode a 34.92 km path where the cyclometer gave 35.28 km. The guessed rolling circumference was 2149 mm. Calculating
34.92 / 35.28 * 2149 gives 2127 mm as the true rolling circumference of these tires (Continental Grand Prix 5000 tube-type, 28mm stated width, 622mm bead seat diameter, on 622x19C rims, inflated to 7 bar and reinflated after two weeks when at 5 bar, at the measurement time pressure was approximately 6 bar, combined weight of bike and rider is 130 kg, this was the front wheel).
If you want to be very precise, you need to ride the path halfway between your regular tire inflation interval. That way, on recently inflated tires the error is in one direction, and on tires you have to inflate very soon the error is equal but in the other direction.
The largest problem of this is to find a good map service. Google Maps is not good in areas where bicyclists use bike paths and car drivers use roads. I use a regional service that gives the path length with 0.1 km precision and also shows markers between 0.2 km. I look at the path, and carefully measure with a ruler the distance between the last 0.2 km marker and the endpoint. I also carefully measure with a ruler the distance between two 0.2 km markers. Then with some math I can get the true length of the path to 0.01 km precision even though the service has only 0.1 km precision in its output.
If you do this procedure carefully, including doing the measurement halfway in tire inflation interval, the good news is that you can be accurate to within about ±0.2% as the error is in one direction with recently inflated tires and in the other direction with tires needing reinflation in a short time.
(1): I measured this on my old cycle computer that had 0.1 km resolution. It was based on the same route where a just inflated tire gave 34.3 km and two weeks ago inflated tire gave 34.4 km. Thus, the measurement is very imprecise due to the very low resolution. Also, the old cycle computer was an analog wireless one prone to misreceive counts, whereas the new cycle computer is a digital wireless one not prone to such misreception. I should re-do the measurement on my new cycle computer that has 0.01 km resolution.