CO-202 explains the variation in temperature gauge reading with increasing altitude.

At 07:56 PM 7/18/02 -0600, James Nazarian wrote:
>"Does altitude affect the readings/accuracy of a mechanical temp gauge?"

Yes.

>"and by how much?"

In very round numbers, about 0.3% for each 1000 feet of altitude. Sorry to disappoint you if you were expecting more.

>"I took a number of observations but I am going to withhold them until the initial opinions are in. I am speaking of a VDO gauge specifically, but I'll take input about other manufacturers as well."

I dunno nuttin about no VDO gauge, but I do know about the original MG "safety" gauge. I refer to the MGA and early MGB with the combination oil pressure and temperature gauge (and maybe also some later MGB after switching from electrical back to mechanical temp gauge). The analog temperature gauge is internally nearly identical to the oil pressure gauge, which is why the "safety" gauge was built into a common case by a single assembly line. The temperature gauge is driven by vapor pressure created by boiling a volitile fluid in the sensor bulb attached to the cylinder head, in fact the temperature gauge is actually a pressure gauge internally. It has a curved flattened tube inside that tends to uncurl with increasing internal pressure, mechanically driving the indicator needle (same as the oil pressure gauge).

Now making that rash assumption that the temperature gauge uses the same pressure range as the oil pressure gauge, mid scale (180dF) is about 50 psi sensor pressure. Increasing your altitude from sea level to 10,000 feet drops the ambient air pressure by about 30% from 15 psi to about 10.5 psi (round numbers). The pressure gauge works with the difference between line pressure and atmospheric pressure. At sea level it may have 50 psi inside and 15 psi outside (35 psi difference), which drives it to half scale (either 50 psi or 180dF reading). At 10,000 feet it will have 50 psi inside and 10.5 psi outside (39.5 psi difference), which will drive it relatively higher (about 13% higher), to about 56.5 psi or (estimating) 186dF on the temperature scale. Note that the temperature scale on the gauge face does not start at zero, and it is non-linear, so this estimate is based on the amount of physical motion of the needle when compared to the oil pressure gauge.

So you have a change in gauge reading of 13% on the oil pressure scale but only about 3.3% on the temperature scale (because it's non-linear) due to the altitude change, or just about 0.33% increase in indicated coolant temperature for each 1000 feet of change in altitude. Note that the change in atmospheric air pressure with altitude is also not linear, so all of this is a sort of hokey round numbers calculation.

Now if you were expecting or observing a larger increase in temperature reading (or even a decrease in reading) with increasing altitude, bear in mind that there are other factors that have a much larger effect on your coolant temperature. On the bad side, there is much less mass of air passing through the radiator core at higher altitude, so the cooling is less effective. But to offset that you also have less fuel and air going into the engine through the carburetors, so you are processing less power and less heat, similar to running it at only 70% maximum throttle (10,000 feet full throttle vs sea level full throttle). If you were to run it at similar throttle settings (say 50% throttle) at sea level and high altitude, you would have less power at altitude, and the actual coolant temperature might be about the same in either case, or even cooler at high altitude. But if you're going to make the same road speed, then you need more throttle at higher altitude to process a similar amount of power. In that case you are making about the same amount of heat in the engine, but the radiator will have a much harder time removing it, so the coolant temperature may be considerably higher. And if you happen to be going up hill at full throttle for prolonged periods of time at high altitude (mountain climbing), all bets are off, take along extra water jugs or give it an occasional rest stop or ease up on the throttle for a while.

In any case, the accuracy of the analog temperature gauge does not change much with altitude. If you are encountering a substantial change in coolant temperature at higher altitude, then trust the gauge, because it's real.

And then at 12:15 PM 7/19/02 -0400, Bob Howard wrote:
>"The internal pressure is forcing the curved tube to straighten. .... If the force were working against a diaphram, I would agree that altitude might have barely measureable effect, but in this instance seems to me that the effort is expended against the metal, not atmosphere."

The key is that the curved (flattened) metal tube is a diaphram.

When this tube is formed into a flat and curved part the inside circumference is forced to a smaller length than the outside circumference (an arc with thickness). As the pressure differential increases, the tube expands in thickness, tending to reform towards being round in cross section instead of flat. This wants to make the inside radius smaller and the outside radius larger, but the length of these surfaces stays about the same as before, so the tube uncurls. The amount of "bulge" in the flat tube is a function of the pressure differential from inside to outside of the tube.

For the oil pressure gauge, if you were to run the signal tube through a rubber stopper, and put the gauge in a bell jar with the open signal tube exposed to atmospheric pressure, and then draw a vacuum in the bell jar around the gauge, the gauge would read about 15 psi from the atmospheric pressure signal. The temperature gauge is of similar construction.

Barney Gaylord
1958 MGA with an attitude (and often plenty of altitude)