The Propane Conversion That Sooted Every Pan

Why this matters

A sooting gas oven is a carbon monoxide source before it is a repair. This case is a range moved from a natural gas house to a propane house and converted by someone who did half the job: the cooktop was right, the oven and the regulator were not. The interesting part is that the converted half looked better than a correct conversion would have, which is why nobody went looking at the rest. It is told in the order it had to happen: stop the hazard, then read the evidence, then correct it against 'How to Convert a Gas Range from Natural Gas to Propane'.

The call

The office booked it as "oven smells, pans black." The tech called ahead from the van.

 TECH:     Is the oven on now?
 CUSTOMER: No, we stopped using it last week. Every
           sheet pan and the roaster came out black
           underneath, and there's a smell when it
           runs, like a car in a garage.
 TECH:     Any headaches or anyone feeling sick when
           it was on?
 CUSTOMER: My husband said he had a headache Sunday.
 TECH:     Keep the oven off and leave the cooktop off
           too until I'm there. Do you have a carbon
           monoxide alarm?
 CUSTOMER: One upstairs. Not in the kitchen.
 TECH:     If it sounds, or anyone feels dizzy, get
           everyone outside and call 911 from outside.
           I'll be there in forty minutes.
 CUSTOMER: The cooktop works fine, though. My brother
           converted it when we moved here.

That last line is the case. The range had come from a natural gas house, the new house runs on propane, and a relative had converted it from the kit. Nobody had connected "converted" with "smell."

Stopping it before diagnosing

On arrival, monitor first. The tech walked in with a personal CO monitor running and read 0 ppm in the kitchen with the range off for a week. No gas odor. Windows opened and the hood fan turned on before anything was lit.

One controlled firing, then off. The oven was lit with the door open, the hood running, and the tech standing at the range with the monitor on. The burner fired with tall, soft yellow flames rolling up past the baffle, and the monitor climbed past 35 ppm in under four minutes. The tech turned the oven off at the control, closed the range shutoff, and let the kitchen air out until the monitor read 0 again. That one look was enough to say the oven was burning rich and had to stay off; there was no reason to run it longer to see how bad it got.

The headache gets a straight answer. The tech told the customer that a headache while a sooting oven runs is a recognized sign of carbon monoxide exposure, and that anyone with symptoms that persist should be seen by a doctor and say the oven was sooting. That is the whole of the medical side; the tech does not judge exposure.

Contain the rest of the house. The tech told the customer the range would not be used again until the visit was finished, and recommended a CO alarm on the main floor near the kitchen and the bedrooms, checking what the local code requires.

Only then did diagnosis start, with the range shutoff closed and the range unplugged.

Reading the evidence

The soot pattern. Before any panel came off, the tech looked at where the soot was. The oven ceiling, the upper walls and the underside of the customer's pans were black; the broil drawer, the burner grates and the bottoms of the pots on the cooktop were clean. Soot lands above the flame that makes it, so the pattern pointed at the bake burner alone and away from the cooktop and broil, which told the tech the conversion was not wrong everywhere, only somewhere the bake burner depended on.

Rating plate. Natural gas manifold 4 in WC, propane 10 in WC, minimum propane supply 11 in WC. No conversion label anywhere on the range.

Top burners. Lit briefly for the record, the shutoff opened for that minute and closed again: small blue flames with short cones, fully around each burner, slow to boil a pot. The spuds, checked with the range off, carried the propane stamps from the kit chart. They had been done right.

Oven orifice. Behind the oven burner, an adjustable hood. It sat backed out, in its natural gas position, never turned down onto its pin. That is the part of the kit people miss: there is no new orifice to install, so a conversion done by following the bag of spuds never touches it.

Broil. Its replaceable spud carried the kit's propane stamp. Converted.

Regulator. Cap off: the insert faced out "NAT."

So: top and broil spuds converted, oven hood and regulator not. One pair of facts explains the whole house.

The manometer settles the regulator

A visual call on a regulator insert is a strong hint, not proof; some inserts are marked badly and some designs hide the setting. The tech connected a manometer at the regulator's outlet (manifold) test port, with a second reading at a test fitting on the regulator inlet, opened the shutoff, plugged the range in, and fired two top burners.

• Inlet: 11.2 in WC, above the plate's 11 in WC minimum. The propane supplier's regulation was fine; nothing upstream of the shutoff needed anyone.
• Manifold: 4.1 in WC, against the plate's 10 in WC for propane.

The regulator was holding natural gas pressure on a propane supply.

Why the cooktop fooled everyone

Flow through an orifice rises with its area and with the square root of the pressure across it. At 4.1 in WC instead of 10, the correctly sized propane spuds on top passed only about the square root of 0.41 of their rated flow, roughly two thirds. Underfired propane flames are short, blue and tidy. To someone checking their work, they look like a textbook conversion. The slow boil was the only symptom, and the customer had put it down to a new pot.

The oven ran the other way. Its natural gas orifice is a much larger bore than the propane one, sized for natural gas, which carries only about 40 percent of propane's heat per cubic foot. Even at the reduced 4.1 in WC, that large bore passed far more propane than the burner's air shutter could mix air for. The result was the lazy yellow flame, soot on everything above it, and the CO the monitor found. The same regulator mistake underfired one half of the range and the open hood overfired the other.

That is why "the cooktop works fine" was not evidence the conversion was complete. It was evidence that the top spuds were right and the pressure was low.

The corrections, in the HowTo's order

With the shutoff closed and the range unplugged:

• Regulator insert reversed to LP, cap reinstalled with its gasket.
• Oven hood turned down until it seated on its pin, snug and no further, per the kit.
• Leak test after gas and power were restored: noncorrosive solution on the regulator cap and the oven hood, no bubbles over a full minute.
• Manifold with two top burners firing: 10.0 in WC, matching the plate. Inlet 11.1 in WC under that load, still above the minimum.
• Simmer screws. The relative had left them at their natural gas setting. At the old 4.1 in WC that had not mattered; at 10 in WC every low flame was far too large to simmer. Each screw was turned in until the low flame was small and steady all the way around, and each held through a quick turn from high to low.
• Oven air shutter. After the hood was set, the oven fired with short yellow tips; the shutter was opened slightly until the flames were blue with defined cones. Broil checked the same way, blue as set.
• CO during the oven run. Monitor on, hood running, 15 minutes of bake at 350 F: 0 ppm throughout.
• Test points. Shutoff closed, manometer removed, port plug and inlet fitting refitted, gas back on and both leak-tested with solution: no bubbles.

Top burners on high now came to a boil noticeably faster, which the customer remarked on before the tech said anything.

What the missing label cost

The kit box turned up in the garage with its conversion label still on the backing sheet. Had it been filled in and attached, the next person to look at the rating plate would have seen a range claiming to be on propane and could have checked the parts against that claim. Without it, the plate said natural gas, the gas in the house was propane, and nothing on the range told anyone which parts had been changed. The tech filled in the label with the date and the shop's name and fixed it beside the rating plate, and bagged the old natural gas spuds inside the range for any future move back.

The soot came out of the oven with a wipe-down once cool; the customer's pans were left to her. The ticket recorded the before and after manifold readings, the inlet readings, the CO readings, and which positions were found unconverted, per 'The Meter Readings on the Ticket SOP'.

The general lesson, and where it stops

A conversion is proved by the parts that are hardest to see, not the ones the customer looks at. Count every orifice position against the kit chart, read the manifold on a manometer, and treat a cooktop that "looks right" as a single data point.

Two things would have changed the visit. Had the inlet read below the plate's 11 in WC, no correction inside the range would have brought the manifold to 10, and the call would have stopped with the range shut off and the supply routed to the propane supplier per 'The Gas-Side Boundary for an Appliance Tech'. Had the monitor read CO in the kitchen on arrival with the range off, the source would not have been the range, and the house would have been cleared and the gas utility or fire department called from outside before anyone looked at the oven.

References

• See related: How to Convert a Gas Range from Natural Gas to Propane; The Gas-Side Boundary for an Appliance Tech; Gas Range Connection and Leak Check Standard
• See related: The Meter Readings on the Ticket SOP
• Manufacturer installation instructions and conversion kit chart for orifice positions, regulator setting and manifold pressures
• Range rating plate for natural gas and propane manifold pressures and minimum supply pressure