What an Efficiency Number Actually Compares

Why this matters

A tech measures 82 percent at the flue on an appliance whose nameplate says 80, tells the customer it is running better than rated, and is wrong. Not slightly wrong in the arithmetic, wrong in the comparison: those two numbers count different things going in and different things coming out. Efficiency arguments in this trade are almost never about performance. They are about boundaries, and once you can name the boundary of each number in front of you, most of the disagreement disappears and what is left is a real finding.

Three declarations hide inside every efficiency figure

An efficiency is useful output divided by input, which sounds like one number and is actually three decisions somebody already made for you.

1. The numerator boundary. What counts as useful output, and where is it measured? Heat leaving the appliance? Heat arriving where people are? Work delivered at the shaft or at the load?
2. The denominator boundary. What counts as input? Fuel only? Fuel plus the electrical energy the appliance consumes to run itself? Fuel on which energy basis, since a sibling card covers how the higher and lower heating value bases differ by roughly a tenth for natural gas and shift the figure by around eight points.
3. The condition set. At what load, what entering conditions, what duty cycle, and over what period? A steady-state figure and an annualized figure describe the same machine and cannot be compared.

Two efficiency numbers are comparable when all three declarations match. When any one differs, the difference between the numbers tells you about the declarations, not about the equipment.

Before you take a flue reading to get one

Combustion efficiency is measured in flue gas, so carbon monoxide comes first. Run a personal carbon monoxide monitor in the occupied space before the appliance fires, not after. If ambient carbon monoxide climbs toward the 50 ppm 8-hour time-weighted average of 29 CFR 1910.1000 Table Z-1, or the 200 ppm ceiling NIOSH publishes, everyone leaves the building immediately, no switch is touched on the way out, and the space is ventilated from outside before anyone re-enters. Never let a decision about staying in the building rest on how the air feels or on an averaged readout.

The probe sits in gas hot enough to burn through a light glove: insert and withdraw it by the grip, and set it on a non-combustible surface rather than on the appliance jacket. Do not adjust anything on the burner to chase an efficiency figure until you have confirmed the heat exchanger is sound and the draft is correct, because leaning an appliance with a defect converts a small carbon monoxide problem into a large one.

The four figures you will meet, and what each one counts

The first row is the only one you can measure with a probe and an analyzer in an afternoon. The third row cannot be measured in the field at all, by anyone, with any instrument: a seasonal rating is the output of a prescribed laboratory test procedure, and for residential equipment those procedures live in 10 CFR Part 430. It is a fair basis for comparing two products against each other and it is not a prediction of what one installation will do.

That single distinction resolves the opening scenario. An 82 percent stack-loss reading and an 80 percent seasonal rating are not two measurements of the same thing, so one being higher is not evidence of anything.

Worked example: one appliance, two correct efficiency numbers

Take an appliance measured at 82.0 percent steady-state combustion efficiency at the flue, at a stable firing rate, with the analyzer set to the correct fuel and the sample line confirmed tight. That figure counts fuel energy in against fuel energy not lost up the flue. Nothing else is inside its boundary.

Now walk the same equipment out to the installed-system boundary. The two adjustments below use illustrative fractions to show the method; measure or estimate your own for a specific job rather than borrowing these.

Adjustment one, widen the denominator. The appliance consumes electrical energy to run itself, for the combustion air blower, the circulating device and the controls. Say that electrical input, expressed on the same energy basis as the fuel, amounts to 3 percent of the fuel input. The denominator grows by a factor of 1.03, so the efficiency becomes 82.0 divided by 1.03, which is 79.6 percent.

Adjustment two, shrink the numerator. Some of the heat that leaves the appliance never reaches the space it is meant to serve, because the distribution path runs through an unconditioned area. Say 8 percent of delivered heat is lost that way, so the numerator is multiplied by 0.92: 79.6 times 0.92 is 73.2 percent.

Same appliance, same day, same instant. One figure is 82.0 percent and the other is 73.2 percent, a gap of 8.8 percentage points, and both are correct. Nothing about the equipment changed between them. Only the boundary moved.

Note carefully what this does and does not license you to say. It does not say the appliance is underperforming, and it does not say the manufacturer's rating is dishonest. It says that the number a customer experiences in their fuel use is not the number on the analyzer, and the difference is mostly outside the appliance.

Which figure answers which question

• Setting up a burner, or detecting appliance degradation over time: use steady-state combustion efficiency, always measured the same way at the same firing rate. Its value is that it is comparable to itself. A drop of two or three points on the same appliance measured the same way is a genuine finding; the same drop against a different appliance's figure is not.
• Comparing two products before purchase: use the seasonal rating, because that is exactly what it was built for and because both products were tested under the same procedure.
• Talking to a customer about what they will actually spend to run it: use the installed-system boundary, and say out loud which losses you included. This is the only boundary honest enough for that conversation, and it is the one most likely to reveal that the biggest available improvement is not in the appliance.
• Justifying a repair against a replacement: use the installed-system boundary for both options, or you will compare a measured field number for the existing equipment against a laboratory number for the new one, which flatters the replacement by construction.

That last one is worth stating as a rule. When you correct one side of a comparison to a wider boundary, correct the other side too, or say in the same breath that the comparator is on a narrower boundary. A repair-versus-replace case built on a field stack-loss figure for the old unit and a seasonal rating for the new one is not a comparison at all.

What changes the answer

• Part load versus full load. A steady-state figure is taken at one firing rate. Equipment that modulates or cycles spends most of its life somewhere else, and the efficiency at that operating point can differ materially in either direction depending on the technology.
• Entering conditions. For equipment that can condense the water vapour in its products, the return temperature decides whether condensation happens at all, and therefore whether a large block of recoverable heat is captured or sent up the flue. The same appliance, same setup, different return conditions, gives genuinely different efficiency, and that is a real performance difference rather than a boundary artifact.
• Excess air. The relationship between the surplus air a burner runs and the heat it sends up the stack is owned by a sibling card in this library. What matters here: whatever excess air does to the flue loss shows up in the steady-state figure immediately and in the seasonal rating not at all, because the rating was set in a laboratory on a different appliance of the same model.
• Duty cycle. Standby and off-cycle losses are invisible to a steady-state measurement, which is taken while the appliance is firing. An appliance that short-cycles badly can measure well and perform poorly, and only the wider boundaries catch it.

The failure mode

The version that costs a shop its credibility is quoting a laboratory rating as though it were a measurement. A customer told their new equipment will deliver a specific seasonal efficiency in their house has been given a number that no field instrument can confirm and that their fuel bill will not reproduce, because their distribution path, their duty cycle and their parasitic loads are not in that number's boundary. When the bill comes in higher than the number implied, the shop has no defense, because the claim was never checkable.

The version that costs a tech a diagnosis is the opposite: dismissing a genuine appliance problem because a field reading looks fine against a rating. A stack-loss figure sitting close to a nameplate seasonal rating tells you nothing about the appliance's health, and treating it as reassurance ends the investigation before it starts.

Catch both by writing the boundary next to every efficiency number you record. Three or four words is enough: "stack loss, steady state, high firing rate" or "system, including parasitic and distribution."

How to verify you got this right

Take any efficiency number in front of you, from a nameplate, an analyzer, a proposal or a manufacturer's sheet, and answer the three declarations out loud: what is in the numerator, what is in the denominator, and under what conditions. If you cannot answer all three from the document the number came from, the number cannot be compared against anything and should not appear in a recommendation.

Then check the last comparison you made between two efficiency figures. Did both sides carry the same boundary? If one was measured and the other was rated, redo the comparison or restate it with the mismatch named in the same sentence.

References

• 10 CFR Part 430, federal test procedures for residential appliance efficiency ratings, as the basis for seasonal and annualized figures
• 29 CFR 1910.1000 Table Z-1 carbon monoxide permissible exposure limit; NIOSH ceiling value for carbon monoxide
• Manufacturer documentation for the energy basis, firing rate and test conditions behind a published rating
• See related: What Excess Air Does to Efficiency; The Units That Get Confused and What It Costs; What Your Instrument Is Actually Measuring