LP Conversion With the Regulator Changed but the Orifices Left Alone
A technician converts a 76,000 BTU/h natural gas furnace to LP. The manufacturer's kit contains four LP orifices, an LP regulator spring, and a conversion label. The technician installs the spring, sets manifold pressure to 10.0 in. w.c. with the burner door in place and the unit running, and applies the label - but leaves the original natural gas orifices installed, reasoning that manifold pressure is what controls input and the spring is the only part that actually needs changing. On start-up the flame is long, yellow, and roaring, lifting off the ports, with black soot forming on the burners and in the vestibule. A combustion test shows CO of 900 ppm air-free and O2 of 2.0%. The flame sensor drops out after about 4 minutes and the board locks out. LP has roughly 2,516 BTU/ft³; natural gas is about 1,030 BTU/ft³.
1. Explain quantitatively what happens to input when 10 in. w.c. of LP is delivered through orifices sized for 3.5 in. w.c. of natural gas. Why is this the real cause of the symptoms?
Read the worked answer
Orifice flow depends on the square root of pressure and the inverse square root of specific gravity, while heat input also depends on the energy content per cubic foot.
Flow factor from pressure: √(10.0 ÷ 3.5) = 1.69 Gas density factor: √(0.60 ÷ 1.53) ≈ 0.63 (natural gas SG ≈ 0.60, LP vapor ≈ 1.53) Volume increase ≈ 1.69 × 0.63 ≈ 1.06× Energy per cubic foot: 2,516 ÷ 1,030 ≈ 2.44×
Combined, input is roughly 1.06 × 2.44 ≈ 2.6×, or about 195,000 BTU/h into a heat exchanger rated for 76,000. That is the diagnosis: the unit is massively overfired.
The symptoms follow directly. O2 of 2.0% means almost no excess air for the gas being burned, so combustion is incomplete - hence 900 ppm air-free CO, yellow lifting flame, and soot. The flame sensor drops out because soot insulates the sensor and the flame pattern leaves the sensing zone. The 10 in. w.c. manifold reading is correct for LP and proves nothing about orifice sizing; manifold pressure and orifice size are two separate variables.
2. The technician insists the input is correct because the manifold pressure matches the plate. How would you prove otherwise with a meter clock, and what should the numbers be?
Read the worked answer
Clocking the meter is the objective test: input = gas rate × heating value.
Correct LP input: 76,000 ÷ 2,516 ≈ 30.2 ft³/h ≈ 0.50 ft³/min. Two cubic feet of test dial takes 4.0 minutes.
Actual, at ~2.6× firing: 195,000 ÷ 2,516 ≈ 77.5 ft³/h ≈ 1.29 ft³/min. Two cubic feet takes about 1.6 minutes.
That is a decisive difference - a dial that moves 2 ft³ in roughly 1.6 minutes instead of 4 minutes cannot be explained by minor gauge error. Procedure: shut off all other gas appliances, turn the furnace on, watch the smallest test dial for a timed interval, and use the LP heating value from the supplier (LP blends vary; use the actual delivered value, not a generic number). Some LP installations use a tank gauge or a flow meter rather than a dial - use whatever provides a rate.
Note also that the technician's premise is backwards. On an LP conversion the regulator spring change sets the pressure, but the orifices set the area; the kit supplies both because both are required. Clocking also catches the opposite error - LP orifices with a natural gas spring, which underfires and produces sooting and pilot/ignition problems.
3. State the corrective actions, the safety communication to the customer, and what additional inspection this event makes necessary.
Read the worked answer
Corrections, in order:
- Shut the unit down and lock out gas before any further firing. It has been operating far outside its listing.
- Install the LP orifices from the kit, all four, verified against the kit list and the appliance's LP rating plate.
- Re-confirm manifold pressure at 10.0 in. w.c. (or the plate value) with the burner door on and the unit at steady state.
- Clock the meter to confirm ≈30 ft³/h, then run a combustion test: target air-free CO well below the 400 ppm ceiling and O2 in the plate's band, with a sharp blue flame.
- Verify temperature rise against the rating plate and check that the limit switch is not cycling.
Additional inspection this event makes necessary: the heat exchanger, burners, and vestibule must be inspected and cleaned of soot, and the flue checked for soot accumulation. Sustained overfiring at 2.6× rating thermally stresses the heat exchanger and can crack it; a combustion test in the supply airstream before and after the blower starts is a reasonable screen, with visual or camera inspection of the primary and secondary surfaces. Replace the conversion label with a correctly completed one and leave the kit documentation with the appliance.
Customer communication: state plainly that the furnace was firing at roughly two and a half times its rating and producing high carbon monoxide, that it must not be operated until corrected, and document the notification. Also advise that any CO alarms in the home be checked and that the appliance be retested after the repair.