# While You Pressure Test LP, Read the Drop Not Just the Gauge > A timed LP test either holds or it does not, but the rate and shape of the drop tell you whether you are chasing a fitting, a valve seat or a temperature swing. Source: https://ocrv.info/blog/lp-pressure-test-read-the-drop Published 2026-07-19. Category: How-To Guides. About 6 minutes. A timed pressure test measures whether a closed LP system holds a set pressure over a fixed interval. Pass or fail is only half the information. A fast initial drop suggests an open fitting, a slow steady drop suggests a seat or thread, and a drop that stops usually means temperature, not a leak. ## The short answer Set the system to test pressure, close the supply, let it settle, then watch the manometer over a timed interval. Whether it holds matters, but so does how it falls. Rate, shape and whether the fall stops all point at different faults. Temperature change during the test will move the reading on its own, so control for it before you declare anything. ## What a timed pressure test actually measures The test isolates the piping downstream of the supply valve and pressurizes it to a known value, then observes whether that value is maintained over a set period with nothing consuming gas. Two things are worth understanding about what this does and does not prove. It proves the **piping and connections** are gas tight at test pressure, across every joint, every flare, every threaded fitting, every flex connector and every appliance shutoff valve that is closed. That is a meaningful statement about the whole distribution system at once, which is why the timed test is the standard first move rather than hunting joint by joint. It does not prove the appliances are safe, it does not evaluate the regulator, and it does not tell you anything about combustion or venting. Those are separate tests. It also does not locate the leak. It tells you a leak exists and roughly how large it is, and then the work of isolating it begins. The instrument matters. A water column manometer, digital or the traditional U tube, resolves small pressure changes that a mechanical gauge on a tank simply cannot show. A tank gauge is not a test instrument. ## Setting up the test correctly Most bad test results come from bad setup rather than from bad piping. The sequence: 1. **Turn off every appliance** and every appliance shutoff valve. Pilot lights out. Refrigerator off, not on automatic, because an automatic changeover can open a valve mid test. 2. **Connect the manometer** at a proper test port. Many systems have one at the regulator or a tee downstream. If none exists, a fitting has to be introduced, and that fitting becomes part of what is being tested. 3. **Open the supply** briefly and bring the system to test pressure, then close the supply valve. 4. **Let it settle.** This is the step everyone rushes. Pressure needs to equalize through the whole system and temperature needs to stop moving from the act of pressurizing. Give it several minutes before you start the clock. 5. **Record the starting value**, then time the interval without touching anything. 6. **Record the ending value** and the ambient temperature at both ends. Two setup faults account for most false failures: a test port fitting that itself leaks, and an appliance valve that was assumed closed and was not. Soap the test connection before timing, and physically verify every appliance valve rather than trusting the handle position. ## Reading the drop: what different rates suggest Here is where the useful information lives. **An immediate, rapid fall to zero or near zero.** Something is fully open. An appliance valve left on, a disconnected line, a missing cap on a capped stub, or a fitting that was never tightened. Look for the obvious before you look for the subtle. **A steady, linear fall over the interval.** A genuine leak of consistent size. Threaded joints, flare fittings that were over-tightened and distorted, a cracked flare, a flex connector at the end of its service, or a valve seat passing internally. The steeper the line, the larger the passage. **A fall that slows and then stops.** Frequently temperature. Gas cooling after the pressurizing event contracts, pressure falls, and once thermal equilibrium is reached the reading stabilizes. It can also be a valve that was slightly unseated and then seated under pressure. Retest after the system has been sitting at ambient for a while. **A reading that rises.** Almost always temperature going the other way, a bay warming through the morning. Occasionally it is a supply valve that is not fully closed. **A fall that varies in rate.** Look for something mechanical that moves: a fitting under strain from a component that shifts, or a valve seat that seals intermittently. Record the numbers rather than the impression. "Held for the interval" and "fell most of a column inch in the interval" are different findings that lead to different work. ## Temperature is the variable that fools people Gas in a fixed volume changes pressure with temperature. That is not a leak, and it will produce a drop or a rise large enough to change a pass into a fail. Practical consequences: - A test started in a cool morning bay and finished after the roll-up door has been open in the sun is not a valid test. - A system just filled or just pressurized is still equalizing thermally. Wait. - A unit brought in from a hot drive has piping at a different temperature than the surrounding air, and it will be moving for a while. - Long copper runs through an underbelly respond to temperature more than short interior runs. The correction is procedural rather than mathematical. Run the test in a stable environment, note the temperature at the start and end, and if it moved meaningfully, repeat the test once conditions settle. A test that fails once and passes twice under stable conditions was probably measuring the weather. ## Isolating the leak once the drop is confirmed You now know there is a leak and roughly how big. The next step is finding it, and the efficient method is halving rather than searching. Work in sections. Close a valve that splits the system, retest, and see which half holds. Repeat within the failing half. On most units the natural division points are the main branch to the interior appliances, the branch to any exterior connection, and each appliance shutoff. Once the section is small enough, go to a leak detection solution on every joint in that section. Apply generously, wait, and look for growing bubbles rather than an initial foam. Check flares, threaded joints, the regulator connections, flex line ends, and any place a line passes through a bulkhead where chafe is possible. Two frequent findings on older units: a flex connector at an appliance that has hardened and cracked at the ferrule, and a flare that was reused too many times and no longer seats. Both are replacements rather than tightening jobs. Over-tightening a distorted flare makes it worse. If the leak is inside an appliance rather than in the piping, the work moves to that appliance. On a furnace that means the gas valve and manifold, covered under [furnace and heating service](/services/plumbing-appliances-and-hvac/furnace-and-heating-repair). On an absorption refrigerator it usually means the burner assembly or its supply line, which falls under [refrigerator service](/services/plumbing-appliances-and-hvac/refrigerator-service). ## The regulator test that belongs in the same visit While the manometer is connected, test the regulator. It is the same instrument and a few extra minutes. Two measurements: **Delivery pressure.** With the supply open and an appliance drawing, read the manometer at the test port. It should sit within the range specified for the system, and it should not sag substantially when a second appliance fires. A regulator that cannot hold delivery pressure under combined load starves the appliances, and the symptom the owner reports is usually a furnace or water heater that will not stay lit when something else is running. **Lockup.** Open the supply, let the system come up, then close every appliance valve. Delivery pressure will rise slightly and should then stop and hold. A regulator that keeps creeping upward is not locking up, and that is a replacement. Regulators are consumable. They have a service life measured in years, and the vent can also become obstructed by insects or debris, which changes behavior in ways that look like a piping fault. Checking it in the same visit as the pressure test avoids a second trip, which matters for owners driving the twenty or so miles from Eastvale to Yorba Linda by SR-71 to SR-91. All of this work is performed in shop, and the rate structure is published on the [labor rates page](/prices/labor-rates). --- OCRV Center. 23281 La Palma Ave, Yorba Linda, CA 92887. (949) 799-3387. info@ocrvcenter.com. Serving Eastvale, California in Riverside County, about 20 miles from the shop. All work is performed in shop at the Yorba Linda facility. No mobile, roadside or on-site service is offered. Scope is body, paint, structural, fiberglass, interior and vehicle systems work. Engine rebuilds, transmission rebuilds, drivetrain work, DOT inspections and emission testing are not performed.