A smell is not a test. A gauge is a test.

The leak-down test

The system is brought up to pressure, isolated at the tank, and watched for a fixed period on a gauge. Either the pressure holds or it does not. That is the test, that is a written result, and it is the only thing that reliably finds a slow leak.

Smell is not the test. The odorant added to propane is there to warn you about a substantial leak, and it does that job well. It will not reliably tell you about a slow one at a fitting behind a panel, and people go nose-blind to a faint smell inside an enclosed rig within minutes of walking in. If a shop tells you your gas system is fine because they had a sniff around, you have not had a gas test.

The other reason we test after any work: fittings loosen with vibration. A rig that has been towed a thousand miles over Florida expansion joints has had every flare fitting on it worked back and forth. That is why the test happens at the end of a job as well as the start.

The regulator, and the number nobody quotes

Outlet pressure is measured in inches of water column, and it should sit around eleven. That is a genuinely small pressure, which is the point — appliances are designed around it, and a regulator drifting out of spec produces symptoms that get blamed on everything else: a furnace that will not stay lit, a fridge that will not get properly cold on gas, a hob that burns yellow and sooty. A regulator can pass a leak test perfectly and still be feeding your appliances wrong, so it is a separate measurement.

Two-stage regulators are also wear items. They have a diaphragm and a vent, and both age. If yours has never been measured, it is a five-minute check that answers three complaints at once.

Absorption refrigerators

The fridge in most rigs has no compressor. It works by heating an ammonia solution — on gas or on an element — and letting it circulate by gravity through a sealed cooling unit. Two things follow from that, and both catch people out.

The first is that it must be reasonably level to work at all. Run badly out of level for long enough and the solution can settle and crystallize inside the cooling unit, and that is not a repair — it is a new cooling unit, which is most of the cost of a new fridge. The rule of thumb the trade uses is that if the slope is uncomfortable to sleep on, it is too much for the fridge.

The second is that it is slow, and it needs airflow. A fridge that is not getting cold often has a blocked flue, a dirty burner, a wasp nest in the outside vent, or no airflow behind it at all. We clean the flue and the burner, check the vent path, and then actually run it — on gas and on shore power — for long enough to prove it gets cold, rather than declaring it fixed because the flame is blue.

Furnaces

The classic RV furnace fault is not the burner. It is the sail switch: a flap in the airflow that has to prove the blower is really moving air before the gas valve is allowed to open. Dust, a lazy blower motor, a restricted duct or a blocked return all stop that switch making, and the symptom is a furnace that runs the fan and never lights. The limit switch is the other one, and it does the opposite job — it shuts things down when the box gets too hot, usually because airflow is restricted somewhere. Both are safety devices doing exactly what they were designed to do, which is why replacing them blind is the wrong instinct.

Water heaters

Two families, and which one you have decides the maintenance. A steel tank has a sacrificial anode rod that is meant to corrode instead of the tank, and it is a consumable — it should be inspected annually and replaced when it is eaten away. An aluminum tank has no anode rod, and fitting one to it is actively wrong. Getting these two confused is common. Either way the tank should be drained and flushed once a season, because sediment sits in the bottom and insulates the burner from the water it is meant to be heating.