Stuart, this explanation is wrong.
Vapour pressure, while describing a smooth curve as you say, does not quite act like that at the boiling point of materials, nor, the freezing point. This is why ice is found in a vacume, yet we know that it can melt to water and boil off as steam. But that is a single component system.
The boiling point of the n-butane is higher than for propane. This is where the actual problem lies. On Bob's stove, he boiled off the propane in the fuel mix and never recovered enough heat to boil off the n-butane. He should have applied more heat to the system, somehow. Inverted canisters simply supply liquid fuel to a stove. They provide "heat" to the canister by increasing the pressure locally, in the preheat tube. This pressure cannot be supported in a cold canister so it "condenses" in the fuel, causing the temperature to raise. Not really by conduction. But a remote canister will get warm, anyway. Like making capuchino, steam is used to cook the milk by raising the temp of the milk. Boiled fuel heats the entire canister, not just the preheat tube.
What you describe is incorrect in that you say the ratio of components in the gas remains the same. Not true. The difference in ratios for boiling off liquids changes quite a lot. A good example is distilling alcohol. In a water/alcohol mix, it would be impossible to distill alcohol acording to your explanation. Pressure and temperature are RELATED, indeed, numbers are usually given for STP, standard temperature and pressure. In a closed system, such as a canister, it is no more complicated than making moonshine. About the best that can be done is 95% alcohol and 5% water even though you never boil the water. The vapour pressure increases enough to carry over in the distalate at about 5%. The same for gas canisters, propane boils off but the vapour pressure of the n-butane is high enough to burn off some. So, you might burn off 95% (sorry, I never looked up the vapour pressures for n-butane) propane. Roger said "mainly the 30% propane". Quite correct.
It is possible to have one component liquid but not boiling and the second component at, or slightly above, it's boiling point. This will change the ratio of the gas mix a LOT. Conversly, by dropping the temperature to below the boiling point of one component (n-butane in this instance,) it is possible to boil off the other (propane.) Of course, they will form a euctectic at some point, so you will likely not be able to boil ALL the propane off. Also, by reducing the temperature below the boiling point of n-butane, you CAN force some of the gas back into the base solution, again changing the ratio of the gas mixture compared to the base mixture.