The Norcold That Cooled on Shore Power and Quit on Gas

Inside an absorption refrigerator is a sealed steel circuit charged with ammonia, water and hydrogen, plus sodium chromate as a corrosion inhibitor. Heat applied at the boiler drives the cycle. There is no compressor and no moving refrigerant pump. The heat can come from a 120 volt resistance element clamped to the boiler tube or from an LP burner firing up a flue beside it, and the circuit cannot tell the difference.

So a fridge that makes cold on shore power has a functioning cooling unit by definition. The ammonia has not migrated, the tubes have not developed a leak, and the sodium chromate has not crystallized into a blockage. Every one of those failures kills both modes at once. When only the gas side is down, the sealed system is a bystander and the money belongs somewhere else.

The list is short and it is nearly always mechanical rather than electronic. A burner orifice partially blocked by rust scale that fell down the flue. A burner tube occupied by a spider web, which sounds like folklore and is the single most common cause we find on units that live outdoors. A flue baffle that has slipped out of position, so combustion gases race past the boiler without transferring heat. Low regulated LP pressure. A tired thermocouple that no longer holds the gas valve open. A control board that is not seeing flame confirmation and locks out.

• Orifice partially obstructed by rust scale from the flue
• Spider or mud dauber nesting inside the burner tube
• Flue baffle displaced, so heat leaves without reaching the boiler
• Regulated LP pressure below specification at the manifold
• Thermocouple output too weak to hold the safety valve open
• Board lockout after repeated failed light attempts

Cheapest and most likely first, which on this appliance means looking rather than testing. The lower access panel comes off, the burner assembly comes out, and the orifice and tube get inspected under a light. On this Class C the burner tube had a compacted web about two thirds of the way along it, dense enough to hold a fingertip of resistance when probed with a piece of wire. That alone will not stop a burner from lighting. It will starve it enough that the flame no longer reaches the correct height and the boiler never gets to temperature.

Then pressure. A manometer goes on the test port and the regulated pressure is read with a burner and a stove top element running, because static pressure lies. Then the flue baffle position, checked from the top with the cover off and a mirror. Then flame quality with everything reassembled: crisp blue, steady, drawn up into the flue, with no yellow tips and no lifting off the port.

Flame appearance is the most information dense observation available on this appliance and it costs nothing to take. A crisp blue flame that sits on the burner and draws upward is correct. Yellow tips mean incomplete combustion, which points at restricted primary air or a partly blocked orifice. A flame that lifts and dances is usually pressure too high or a burner sitting off centre in the tube.

A short, lazy, mostly blue flame is the one that fools people, because it looks close enough to right. That is the signature of a partial obstruction, and it is what this unit had. The fridge would light, hold flame, satisfy every safety circuit, and simply not make cold, which is why nothing in the control system ever complained.

Because it was on shore power in the driveway. That is the detail that had confused everyone, and it is the detail that solved it. The owner plugged in at home, so the electric element ran and the fridge was cold every time he checked before a trip. On the road the unit switched to gas automatically and the temperature climbed over about eighteen hours, which is slow enough that it reads as a gradual failure rather than a switch flipping.

This pattern is common enough that we ask about it directly at intake. A complaint framed as intermittent cooling that correlates with being plugged in is not intermittent at all. It is two modes, one working and one not, and separating them takes one question rather than an afternoon.

Burner assembly out, tube cleared with compressed air and a nylon brush, orifice inspected and cleaned, flue baffle reseated, manometer check at the manifold, reassembly and a two hour cool down test with a probe thermometer in the box. Just under two hours of bench time plus the cool down watch, at the mechanical and electrical rate.

A cooling unit replacement on the same appliance is a different order of job: the fridge comes out of the cabinet, which on a Class C often means removing trim and sometimes a section of countertop, and the replacement unit itself is the largest line on the invoice. That is why proving which side is at fault matters before anybody orders anything.

Run one controlled test. Unplug from shore power, switch the fridge to gas only, put a thermometer in the fresh food compartment and read it at four hours and again at twelve. Write both numbers down with the ambient temperature beside them. Then repeat on shore power. Two pairs of numbers turn a vague complaint into a defined one and they cost you nothing but a day of patience.

Bring the model and serial from the data plate inside the fridge door, not the year of the coach. Absorption fridges change part numbers across production runs and the serial is what determines which burner assembly and which board fit. Owners who arrive with a photograph of that plate save us a phone call and themselves a parts delay.