Carbonfibreboy's formula is based on LaPlace's law, which pops up whenever pressure vessels (the heart, arteries, steam boilers, and tires) are analyzed.

For cylinders, T = p * r / t,

where T is tension developed in the wall of the cylinder, p is the indicated gauge pressure, r is the radius of the cylinder, and t is the thickness of the cylinder wall. This "thin-walled" case holds when the radius of the cylinder is at least 10 times the thickness of the wall, so for most tires (and large arteries) we're OK. A moment of algebra will show that for constant wall thickness and a maximum safe design tension, the product of pressure and radius is a constant, the "3000" used in Carbonfibreboy's rule of thumb. It's messier in real life because on a wheel, the pressure vessel consists of the tire carcass, the rim, and the rim-tire interface where the two beads mate. Also, wider tires might (I don't know, never measured) be thicker than narrow ones; so to be more accurate you should include the tension-lowering effect of thicker sidewalls, which would allow a wider tire, if thicker, to hold a pressure higher than the rule-of-thumb value might suggest and still remain at a safe tension. But because over-inflating a tire blows the bead off the trim -- it doesn't split the tire carcass -- it is probably prudent not to view thickness as a safeguard against failure from over-inflation, especially since to my casual "thumb-and-middle-finger" testing, most bike tires are about the same thickness anyway. In other words, Carbonfibreboy's omission of tire thickness in his rule of thumb is not a fault but rather a feature.


So a rule of thumb that says divide 3000 by the tire diameter to get the max. safe inflation pressure seems a sensible place to start and is supported by physics. It's the same principle that shipbuilders use to figure out how thick the steam boiler in a nuclear submarine has to be. (The magic number for a steel boiler is of course many many thousands: hundreds of psi * several feet of radius.) The caution for us is that the weakest point is the bead interface. Some tires, regardless of safe inflation pressure, are difficult to get to seat reliably on some rims -- we've all been there. I think most of us, sensibly, would avoid those combinations, and not underinflate such a tire merely to keep it on the rim.


Joining the trend to wider tires, we've been experimenting with lowering the pressure to below the sidewall max, since they seem to be immune from pinch flats. In the past we used 25 mm tires inflated to 120 psi "because that's what tandemists do." We currently use 30 mm, the biggest that will fit, inflated to just 80 and Ms. C. loves the ride on our chip-seal. We're no slower, either, just as Jan Heine would have predicted all along. Our magic number might now be just 2400. (Crew + vehicle weight ~325 lb also. Friends of ours who are much taller and heavier (and faster) have profited greatly by modifying their bike to take 43 mm tires.)