How Altitude Changes Combustion

Why this matters

The same appliance that ran clean on a bench at sea level can make carbon monoxide at 6,000 feet with nothing wrong inside it. Nothing broke. The air got thinner and the burner never found out. Shops that work a mountain territory learn this the expensive way: a run of new installs that all pass a quick startup, then a winter of callbacks for soot, rollout, and cracked exchangers, because the crew commissioned to a sea-level habit. The correction is small, it is written on the rating plate, and it takes ten minutes if you do it at startup.

Before you stand in front of an operating burner

Carbon monoxide is the hazard here and it is invisible. Wear a personal CO monitor and turn it on before you enter the mechanical space, not after you light the appliance. If it alarms or you read measurable CO in the room with occupants present, get everyone outside into fresh air, shut the fuel off at the appliance shutoff valve, open the space to outdoors, and do not relight until you have found and corrected the cause. Anyone with headache, nausea, or confusion goes to emergency medical care, not to the couch.

Before you touch the gas train, leak-check every joint you break with a bubble solution or an electronic detector, and never with a flame. If you smell gas at any point, leave the building with the occupants, touch no switches, use no phone inside, and call the gas supplier from outside.

If your work includes metering an energized control circuit, the exposed-live work is permitted only under the troubleshooting exception at 29 CFR 1910.333(a)(1), which allows energized work where de-energizing introduces additional hazards or is infeasible, and testing is the named example. Use a meter and leads rated for the category of the circuit you are on, CAT III for distribution and branch-circuit work, and inspect the leads for cracked insulation before each use. An under-rated meter across a line-voltage fault is the failure that kills.

What altitude actually changes, and what it does not

Oxygen is still 20.9 percent of the air by volume at 10,000 feet. Anyone who tells you there is less oxygen in the mountains is describing the effect, not the mechanism. What changes is density: fewer molecules of everything in each cubic foot, because the column of atmosphere above you weighs less.

At about 6,000 feet, atmospheric pressure runs near 80 percent of the sea-level value. At the burner, combustion air is drawn at room temperature in both places, so the density that matters to you is essentially that pressure ratio: each cubic foot of air the blower moves carries about 80 percent of the oxygen mass it carried at sea level.

Nothing tells the burner. The blower turns at the same speed and sweeps the same volume. The orifice is the same hole.

Why a fixed orifice drifts rich, and by how much

Two effects run in opposite directions and they do not cancel.

Gas flow through a fixed orifice at a fixed manifold pressure varies with the square root of the gas density, so a thinner gas gives you more cubic feet per hour, roughly 1 divided by the square root of the density ratio. At a density ratio of 0.80 that is about 12 percent more volume. But each of those cubic feet now carries only about 80 percent of the sea-level heat content. Net delivered input lands near the square root of the density ratio, about 89 percent of nameplate.

Combustion air, meanwhile, fell to 80 percent.

So the air-to-fuel ratio at altitude is roughly 0.80 divided by 0.89, about 0.90 of its sea-level value: the burner is running about 10 percent leaner in air than it was designed to. Whether that matters depends entirely on how much excess air the appliance had to give away.

• An appliance commissioned at 30 percent excess air (air-to-fuel 1.30 times stoichiometric) lands at 1.30 x 0.90, about 1.17, so roughly 17 percent excess air. Thin, but still oxidizing.
• An appliance commissioned at 15 percent excess air lands at 1.15 x 0.90, about 1.03. Three percent margin, and any soot, a partly blocked filter, or a windy day pushes it under stoichiometric, which is where carbon monoxide is made.

That is the whole story of why altitude bites the marginal appliance and leaves the well-set one alone. This article states the arithmetic of the shift; the sibling article on excess air owns the rule for where to set it in the first place.

The gate: what the rating plate decides

Read the rating plate before you compute anything. Manufacturers certify appliances to an elevation range under the ANSI Z21 and Z83 product standards, and the plate states it.

• Plate shows a sea-level-to-2,000-foot rating with no high-altitude provision. The appliance needs input correction. The widely applied United States convention, carried in the National Fuel Gas Code, is to reduce input 4 percent for each 1,000 feet of elevation above 2,000 feet, applied in whole 1,000-foot steps, and the manufacturer's own instructions govern where they differ. Many manufacturers meet this with a listed orifice kit rather than a pressure change.
• Plate lists the appliance for elevations to 10,000 feet. No derate. The correction is already inside the appliance, either as a factory orifice selection or as a listed conversion kit installed at the factory. Deratng it again makes it under-fired.

Local amendments exist and some western jurisdictions publish their own derate tables. Check the adopting jurisdiction before you set anything, because the enforceable number is theirs.

Case A: the plate stops at 2,000 feet

Install elevation 6,000 feet. Nameplate input 100,000 Btu/hr, illustrative but a common size.

Steps above the threshold: (6,000 - 2,000) / 1,000 = 4. Derate: 4 x 4 percent = 16 percent. Target input: 84 percent of 100,000, or 84,000 Btu/hr.

Clock the gas meter with every other gas appliance off. Say the 1-cubic-foot test dial takes 34 seconds for one revolution. Flow is 3,600 / 34 = 105.9 cubic feet per hour. Get the local heating value from the gas supplier for that delivery area rather than assuming a textbook 1,000 Btu per cubic foot, because at altitude the delivered energy per actual cubic foot is lower and the supplier's stated value is the one your calculation has to use. Say they give 830 Btu per cubic foot.

Input: 105.9 x 830 = about 87,900 Btu/hr.

Against the 84,000 target that is 4.6 percent over. Not a catastrophe, and exactly the kind of miss a tech eyeballing manifold pressure would never see, because the manifold pressure is correct. The correction is the manufacturer's high-altitude orifice, sized in their table for this elevation, followed by a re-clock. Manifold pressure adjustment alone is a last resort and only within the range the instructions permit, because the regulator setting is also what holds the burner stable through supply-pressure swings.

Re-clock after the orifice change and confirm you land within a few percent of 84,000, then run a flue gas analysis and record it.

Case B: the same site, the plate lists to 10,000 feet

Same elevation, same clocked input, same 4.6 percent variance from a sea-level nameplate figure - and the call is the opposite. This appliance is supposed to fire near its plate rating here, so a rich flue reading is not an altitude problem and derating it will only make combustion worse by dropping the burner below its stable firing range.

Go looking for the things altitude gets blamed for: a combustion air opening sized on paper but blocked by storage, a vent that condensed and partly closed, a filter loaded past the blower's ability to move design airflow, a heat exchanger crack pulling circulating air into the burner path, or a room that goes negative when the dryer and the range hood run together.

One gate, two opposite conclusions, and the only thing that separates them is a plate you can read in five seconds.

What your analyzer will and will not show you up here

Oxygen and carbon dioxide come off the analyzer as percentages of the sample, and a percentage does not know what altitude it was taken at. A burner running 4 percent oxygen reads 4 percent oxygen at any elevation. That is useful, because it means your excess-air judgment travels unchanged.

What does not travel:

• Anything absolute. Delivered input, total heat output, and airflow in cubic feet per minute are all density-dependent, and the standard-air constants baked into field formulas assume 0.075 pounds per cubic foot. At 80 percent density those constants are 20 percent optimistic unless you correct them.
• Carbon monoxide as measured. Convert to an air-free basis using the measured oxygen before you judge it, exactly as you would at sea level, because a diluted sample reads low. The conversion is where an altitude reading and a sea-level reading become comparable.
• Net stack temperature. Subtract the temperature of the air the burner is actually breathing, not the room you are standing in, which is usually a different temperature in a mountain mechanical room with an outdoor air duct.

How to verify you got this right

Clock the meter twice, once before and once after the orifice change, with all other gas appliances off both times. Two clocks that disagree by more than a couple of percent mean another appliance fired during one of them.

Run the appliance to steady state, take flue gas readings at the same probe depth on both runs, and record oxygen, air-free carbon monoxide, net stack temperature, elevation, and the supplier's heating value on the ticket. The elevation and heating value are the two fields that let next year's tech compare their reading to yours instead of guessing.

Then run the appliance under the condition that breaks it: interior doors closed, clothes dryer and every exhaust fan on, and watch for spillage at a draft hood and for carbon monoxide to climb. An appliance that is correct on input and still spills has a building problem, not a fuel problem.

References

• NFPA 54 / ANSI Z223.1, National Fuel Gas Code, for high-altitude input reduction and appliance installation requirements
• ANSI Z21 and Z83 appliance product standards, as reflected on the appliance rating plate's certified elevation range
• 29 CFR 1910.333(a)(1), OSHA general industry, for the energized-work troubleshooting exception applied to live control-circuit testing
• Manufacturer installation instructions and high-altitude orifice tables, which govern where they differ from the general convention
• See related: What Excess Air Does to Efficiency; How Combustion Actually Works; How to Read a Flue Gas Result