I’ve looked again at the various graphs I developed back in 2006. I found various Excel programs which I used to generate the graphs.
The graphs are:
Constant vs Temperature. This shows that as temperature increases, the constant decreases – the tire diameter increases.
Constant vs date. This shows that as time between measurement dates increases, the constant decreases.
Precal constant vs date. As measurement dates go by, using the same initial pump-up pressure, the constant decreases.
Precal constant vs temperature. This shows that as temperature increases the constant decreases.
Constant vs temperature change. I am not sure what this shows – too much time has gone by since I did the work.
In all of the above it is seen that there is a decline in constant – the tire gets bigger. This is the normal experience of measurement. The postcal is almost always smaller than the precal. My data were derived solely from daytime rides, in which the temperature at postcal was higher than at precal.
There is considerable scatter to the data in all cases, but the trends are clear.
It has been suggested that it is possible to separate the effect of temperature and pressure. I’ve not had much luck with this. The only pressures I recorded were precal pressures, measured using the pressure gauge on my tire pump. I tried a few times to use the pump to record postcal pressures, but a bit of air escaped each time I tried, and I was not satisfied that the readings were solid data.
In answer to Mark's question asking how much calibration changes with temperature, my eyeballing of the graph of precal counts vs temperature at constant pumpup pressure leads to my estimate of a calibration decrease of 2 counts per kilometer for every 10 degrees of temperature increase. With so much scatter in the data a guess is about as good as relying on the linear trend line.
Now I just noticed the graph in my post of 27 January 2006 in which I estimated calibration drop of 5 counts per km for a 10 degree temperature increase. Which guesstimate is right? Reader's choice.
Most importantly, my data are not universally applicable. I used only Armadillo tires, which may or may not behave as other kinds of pneumatic tires.
There are too many variables for my thinker to cope with. What you see is what you get.
I made an attempt to factor in the expansion of the tire rim with temperature, but it led me nowhere.
However – I found that the data was of use to me, as I could use it to predict how my tire would behave under various conditions. Anyone wishing to expand the study will have to begin taking data and seeing where the exercise takes them. The task could become monumental depending on how many tires are investigated.