Hydrostatic pressure test field guide for plumbers

The plumbing pressure test, and what it is

A plumbing pressure test fills the piping, or part of it, with water or air, raises it above the pressure it will see in service, and holds it while you watch for any leak or any drop. On the drain side that means a column of water standing in the stack. On the supply side it means the lines pressurized to their working pressure or above. Either way the test proves the same thing. The joints hold before anything covers them.

The reason the test matters is timing, not paperwork. A leak found at rough-in is a fitting you cut out and redo on dry, open framing with nothing finished around it. The same leak found after the drywall is up, the tile is set, and the ceiling below is painted is a demolition job and a water-damage claim. You test before close-in because that is the last moment every joint is in front of you.

Two systems, two tests, and people blur them. The DWV piping, the gravity drain-waste-vent network, gets a water or an air test that proves it is watertight and gastight at low pressure. The water supply piping, which runs under real pressure in service, gets a hydrostatic or an air test at or above its working pressure. The numbers and the methods differ, and the code that governs both is whatever edition of the IPC or UPC the jurisdiction adopted, with local amendments.

What does a plumbing pressure test prove?

A plumbing pressure test proves the piping is leak-free at the test pressure and, on the supply side, that it has the strength to hold above its working pressure. It is a tightness check first. Every solvent weld, every soldered or pressed copper joint, every threaded fitting and crimped PEX connection in the tested section either holds the head or the gauge, or it shows itself.

What the test does not prove is worth naming so nobody oversells it. It does not prove the drains are sloped right or the vents are placed to protect the trap seals, which is a separate inspection covered in the DWV venting and sizing guide. It does not prove the system is clean or the fixtures are set. And it does not prove a joint that holds today will survive twenty years of thermal movement and water hammer. The test is a snapshot. It says that, as it sits at rough-in, the piping holds.

The value is in catching the bad joint while it is cheap. A weeping solvent weld behind a finished wall is found by a stain on the ceiling months later, after someone else owns the leak. The same weld at rough-in is found by your eye while you walk the stack, and fixed before lunch.

How do you test a DWV system with water?

You test a DWV system with water by capping the openings, filling the piping until at least a 10 ft head of water stands above the highest joint in the section under test, and holding it for 15 minutes with no drop and no visible leak. That 10 ft column is the IPC figure, and it puts a few psi of head on the lowest joints, which is plenty to find a bad solvent weld or a joint someone dry-fit and forgot to glue. Confirm the head and the hold against the adopted code, because the IPC and UPC both run a water test but the wording and the section numbering differ by edition.

On anything taller than a single story you test by section, not all at once, because a full-height column would put far more head on the bottom-floor joints than the pipe and the test plugs want to see. You plug the base of the section, fill the stack for that floor to 10 ft above its highest joint, walk it, then drop down and test the next section. The rule is that the upper 10 ft of the section below also gets re-tested, so no joint except the very top of the system ever sees less than a 10 ft head.

Water is the honest test on DWV because the system runs on gravity, is meant to be watertight, and a water failure weeps where you can see it instead of letting go. Fill from the bottom, bleed the air, and walk every joint with a light. A drop in the visible level inside 15 minutes, with the temperature steady, is a leak, and the only fix is to find it, cut it out, and retest the section.

Can you air test a DWV system?

You can air test a DWV system where the code and the pipe material allow it, holding a gauge pressure of 5 psi, or about 10 inches of mercury, for 15 minutes with no drop. The IPC permits that air test for piping other than plastic. For plastic DWV, which is most of what gets installed now, the code direction is a vacuum test instead, pulling the section down to about negative 5 psi or negative 10 inches of mercury and holding 15 minutes, or a water test. Verify the method and the value against the adopted code and the pipe manufacturer before you pressurize anything.

The reason plastic gets a vacuum or a water test instead of a positive air test is safety and the manufacturer's own warning. PVC and ABS makers caution against testing their pipe with compressed air or gas, because a solvent weld that lets go under stored air energy does not weep, it ruptures, and a blown cap or a shattered fitting under air is a projectile. Water stores almost no energy. Air stores a great deal, and that difference is the whole argument.

Air has its place on a long horizontal run where filling with water is impractical, or in freezing weather where a water column would burst the pipe. When you do air test, gauge it, hold it, and brush soapy water on the joints to find the leak by the bubbles. But on plastic the default is water or vacuum, not a pressurized air test, and an inspector in many jurisdictions will stop you if you reach for the air on a glued system.

What pressure do you test water supply piping to?

You test water supply piping hydrostatically at not less than the working pressure of the system, held 15 minutes with no drop, and for piping other than plastic the code also allows an air test of not less than 50 psi. That is the IPC basis. Many specs and many inspectors push higher than the bare working pressure, commonly to 100 psi or to 1.5 times the working pressure, to give margin and to stress the joints, so confirm the number against the project specification and the adopted code rather than from habit.

The material drives how you test the supply lines. Copper, soldered or pressed, takes a hydrostatic test fine once the solder has cooled or the press is made. CPVC needs its solvent welds fully cured before any pressure goes on them, the same as plastic DWV. PEX is the one with a real exception: a positive air test on PEX is allowed only where the pipe and fitting manufacturer specifically authorizes it in their instructions, and otherwise PEX gets a water test like everything else. Do not assume you can air test PEX because it is plastic and forgiving.

Run the supply test on potable water, because the lines are about to carry drinking water and a dirty fill seeds the system. Pressurize, hold 15 minutes, and walk every fitting, every manifold connection, and every crimp or clamp ring. The supply side fails at the connection far more often than the tube, so the joints get the close walk, not the middle of a stick of pipe.

The final test: smoke, peppermint, and odor

The rough-in test proves the joints. The final test, run after the fixtures are set and the traps are filled, proves the finished DWV system is gastight and that every trap seal and connection holds sewer gas out of the building. The code version is a water or a low air test through the finished system. But the two final tests a pro reaches for when chasing a complaint are the smoke test and the peppermint test.

The smoke test pushes a dense, visible smoke into the DWV system, often after plugging the vents, and brings it to a low pressure on the order of an inch of water column. Then you walk the building and the roof looking for smoke escaping at a bad joint, a dry trap, or a cracked fitting. The peppermint test is the old low-tech version. Pour a strong oil of peppermint into the top of the stack with hot water, plug the vent, and have someone with a clean nose, who has not handled the peppermint, walk the floors to smell where it leaks out.

These are the tools for the call where a finished building smells like sewer and the rough-in passed clean. Verify the final-test method and any pressure against the adopted code, because the code spells out the acceptable final tests, and the smoke and peppermint methods are traditional ones the inspector may or may not require by name.

The building sewer test

The building sewer, the gravity line from the building drain out to the public sewer or the septic tank, gets its own water test. You plug the end at the point of connection to the public sewer, fill the sewer with water, and bring it to not less than a 10 ft head, held 15 minutes with no drop. That is the IPC gravity sewer test, and it proves the buried line is watertight before it is backfilled.

The timing is the whole point on a sewer, because it is going under grade. A leak in a buried sewer is found by a wet spot in the yard, a sinkhole over the trench, or roots in the line years later, and the repair is an excavator and a permit. Test it before the trench is closed, while the pipe is exposed and a bad joint is a fitting you can still reach.

Watch the head and the plug. A 10 ft head on a sewer plug is real force, and a test ball or a mechanical plug that is undersized or under-inflated will blow out and dump the test, sometimes into the trench you are standing in. Size and seat the plug for the head, and keep clear of the line of a plug under pressure.

How do you fill and vent the system before a water test?

Fill from the low point and bleed the air from every high point, because trapped air ruins a water test. Air in a water-filled line is a spring. It compresses as you build the head, so the level wanders on its own, and an air pocket at a high point cushions the very spot a leak would show, hiding the weep. The goal is a water-solid section with the air pushed out ahead of the fill.

So you open the high points, fill slowly from the bottom, and let the rising water drive the air up and out until each vent or opening runs solid water with no spitting. On a supply manifold that means cracking the farthest and highest fixture stop. On a DWV stack it means filling until the column stands clean at the top with no gurgle. Fill too fast and you trap air the openings cannot catch.

Air makes the test spongy, and spongy reads as both a false leak and a hidden one. A section that will not settle to a steady level usually has air, a passing plug, or a real leak, and the cheap way to tell which is to bleed it fully first, then watch it again.

Isolating the test

Half of a clean pressure test is deciding what not to pressurize. You cap and plug the open ends with test plugs, test balls, or threaded caps rated for the head or the pressure, and you isolate every device in the section that is not rated to take the test. The water heater is the first one people forget. Its tank and its controls are not built to sit at a 100 psi or a 1.5 times working-pressure test, so you isolate it at its stops and test up to it, not through it.

The relief valve is the item that bites both ways. A temperature-and-pressure relief valve on a water heater, or any relief valve in the section, will lift the moment the test pressure crosses its setting and dump your test. Isolate it out of the section, or remove it and plug the connection, and put it back the instant the test is done. A relief valve left isolated and forgotten is a removed safety on a live system, which is a far worse problem than a failed test.

The backflow preventer, the pressure-reducing valve, control valves, gauges, and any low-rated fixture or appliance come out of the test section the same way. The backflow assembly in particular has its own test discipline and its own failure modes, covered in the backflow failed-test repair guide, and you do not prove it with a system pressure test. Mark on the drawing or the plan what you capped, what you isolated, and what you removed, so the reinstatement puts the system back exactly the way it has to be.

The gauge

The gauge is the witness, and a wrong-range or uncalibrated gauge throws out the whole test. Size it so the test pressure lands near the middle of the scale, because a gauge reads most accurately in its mid-range and worst at the ends. A 100 psi test on a 0 to 160 psi gauge sits where you can read it. The same test on a 0 to 600 psi gauge buries the reading in the bottom of the scale, where a small drop disappears in the width of the needle.

Calibration is what an inspector or a witness checks first. A gauge with a current calibration sticker, dated inside its interval, is the one whose reading means something, and the cal record belongs with the test paperwork. An expired or missing calibration is a finding on its own, no matter what the needle showed during the hold.

On a long supply hold or a high-value test, a calibrated pressure recorder or a datalogger trends the pressure and timestamps it, which beats a hand-logged reading for proving a flat hold and for catching a slow drift the eye misses between checks. On a 15-minute DWV water test you are mostly reading the visible water level and walking joints. The supply test is where you want a real gauge you can trust.

Thermal change and the false drop

On any hold longer than a few minutes, temperature moves the gauge, and the move looks exactly like a leak. Water is nearly incompressible and a filled line is nearly rigid, so a small change in water temperature makes a surprisingly large change in pressure. Fill a supply line with cold water on a warm slab and the water warms, the pressure climbs, and the gauge reads high with nothing wrong anywhere. Pressurize in a warm afternoon and let the building cool overnight, and the morning gauge reads low with every joint tight.

The fix is to record the water temperature and the ambient temperature alongside the pressure, so when the gauge moves you can ask the first question: did the temperature move with it. A pressure change that tracks a temperature change, and recovers when the temperature recovers, is thermal, not a leak. A change with steady temperature is a real leak. You cannot tell the two apart from the gauge alone, which is why the thermometer is test equipment, not an afterthought.

The classic trap is the overnight supply hold. The crew pressurizes before they leave for the night, the building cools, the gauge is down in the morning, and someone condemns a joint that is perfectly tight. Run long holds when the temperature is stable, fill near the ambient temperature so the water is not chasing equilibrium for hours, and read any drop against the temperature log before you blame the pipe.

Finding the leak

When the level falls or the gauge drops and the temperature is steady, you walk the section and find the leak by hand and eye. On DWV you are looking for water, so you walk every joint with a light: the solvent welds first, since a dry-fit or under-glued PVC joint is the common one, then the hubs and the gaskets, then the threaded adapters and the cleanout plugs. A bead at a fitting, a damp spot under a hub, a drip building on the underside of a horizontal run. That is your leak.

On a supply test the joints to walk are the soldered and pressed copper, the threaded brass, and the PEX crimp or clamp rings. PEX fails at the fitting, not the tube, and usually because the ring was not square on the fitting or the crimp tool was out of adjustment, so the rings get the close look and a go/no-go gauge. Run a dry hand or a paper towel around the underside of a suspect joint, where a drip you cannot see collects.

On an air test you cannot see water, so you brush soapy water on the joints and watch for bubbles, the same as gas work. A high-pressure pinhole on a supply line can spray fine enough to feel as a mist before you see it, so listen at the quiet joints. Mark what you find, drop the pressure, fix it dry, and retest the section from the start. A joint nipped up under pressure may seal for the test and weep in service. The retest is not optional.

Materials and the joint method

The joint method sets when you can test, and rushing it is the most common self-inflicted failure. A solvent-welded joint in PVC, ABS, or CPVC is not a glued joint, it is a chemical weld that has to cure, and it has almost no strength for the first few minutes and only reaches full strength after a set cure time that grows in cold weather and on larger pipe. Test a green solvent weld too soon and you pull the joint apart or it weeps, and you have failed a test on a joint that would have held an hour later. The manufacturer's cure chart, against pipe size and temperature, is the number that governs, not the clock in your head.

Copper is ready when it is cool. A soldered joint holds as soon as the solder solidifies and the joint comes down to where you can handle it, and a pressed joint holds immediately once it is crimped with the right jaw. The catch on copper is the joint you thought you soldered and only tacked, which the test finds for you.

PEX is mechanical and ready at once, but only if the connection was made right. A crimp ring has to be square and pass the go/no-go gauge, and an expansion fitting has to be given its few seconds to shrink back tight on the fitting, longer in the cold, before any pressure goes on it. The table lays out when each material is ready to test.

Freeze and the cold-weather test

Cold weather changes a water test in two directions, and both bite. A solvent weld cures slower as the temperature drops, so the cure time before you can pressurize a plastic joint stretches well past the warm-weather figure, and a joint that would be ready in an hour at 70 F may need several at 40 F. Check the manufacturer's cold-weather cure, not the standard one.

The bigger hazard is freezing the test water. Water left standing in a tested section that drops below freezing will split a fitting or burst a pipe, and you have created a leak after the test passed. In freezing conditions you either keep the section warm, test with air where the code and the material allow it, or, the usual answer, drain the section completely and promptly the moment the test is accepted.

Draining fully is harder than it sounds. The low points, the trap arms, and the dead legs hold water that a quick blow-down misses, and a missed pocket freezes and splits just like a full line. Plan the drain points before you fill, not after the forecast turns. A burst from trapped water is a callback you handed yourself.

The witnessed test and the inspection

A pressure test that nobody witnessed is a test you may get to run again. The rough-in inspection is usually the AHJ inspector signing off the DWV and supply tests before close-in, and the inspector wants to see the system actually holding, not a note on a form that says it did. Schedule the inspection for the hold, with the water standing or the gauge up, so the inspector reads the level or the gauge for themselves.

The rough-in sign-off is a hold point. You do not cover, insulate, or backfill until the test passes and the inspector releases the work, because covering an unproven joint is the exact failure the test exists to prevent. On commercial work a commissioning agent or the owner's representative may witness alongside or instead of the AHJ, and the same rule holds. The witness sees the hold, signs the record, and the record gates the next trade.

Record the test the inspector witnessed: the section, the medium, the head or pressure, the hold time, the result, and who signed. A test with no recorded head, pressure, or duration is a story, not a record. A year later, when a joint weeps in a wall, the question is whether the section was ever proven, and the record is the only thing that answers it.

Medical gas and special systems

Medical gas, fuel gas, fire sprinkler, and other special systems are not tested like ordinary plumbing, and treating them like it is a serious mistake. Medical gas piping carries its own far stricter test and certification regime, with an initial pressure test, a standing pressure test, and a cross-connection and purity verification, performed and documented to the medical gas standard, commonly NFPA 99, by certified installers and verifiers. Do not improvise it from a plumbing water test.

Fuel gas piping gets a pressure test to the fuel gas code, at a pressure and a hold set by that code and the local amendments, with the appliances and regulators isolated. Fire sprinkler piping gets a hydrostatic test, commonly at 200 psi or 50 psi above working pressure for a longer hold, to the sprinkler standard, commonly NFPA 13, and it is witnessed and certified separately. Each of these has its own governing standard, its own pressure, and its own paperwork.

The point is to know what you are testing before you set a number. The DWV and water-supply tests in this guide are for ordinary sanitary and potable plumbing. The moment the system is medical gas, fuel gas, sprinkler, or a process or specialty piping system, stop and pull its own standard, because the test that proves a drain line says nothing about a system that has to hold a life-safety or a clinical duty.

Field example: a two-story house at rough-in

A two-story house was tested at rough-in before close-in, and the plumber ran the DWV stack by section. The upper section was capped at the second-floor base, filled until the water stood 10 ft above the highest joint, and held. At 15 minutes the level was steady and a walk of the solvent welds came up dry, so that section passed. The lower section was then plugged at the building drain and filled so the upper 10 ft of it carried head along with the floor below, and it held the same 15 minutes.

The supply test ran next. The water heater was isolated at its stops and its temperature-and-pressure relief valve was left out of the section, the hose bibbs were capped, and the manifold was filled from the low point with the highest fixture stop cracked to bleed the air. The lines came up to 100 psi, the local practice above the working pressure, on a 0 to 160 psi calibrated gauge, with the water and the ambient both near 65 F at the start.

At 15 minutes the gauge read 96 psi. Before anyone called a leak, the temperature log answered it. The slab had warmed the cold fill and then the afternoon cooled, the water had dropped a few degrees, and the small drop tracked the temperature. The crew topped it back to 100 psi, let it settle, and it held flat. Logged as a pass with the temperature noted, so the inspector and the next reader could see the drop was thermal, not a weep.

What to document

A test you cannot reconstruct from the record is a test that did not happen, as far as the next plumber and the inspector are concerned. The record is what proves the piping was proven, and it is what answers the callback a year out when a joint weeps inside a wall and the question is whether that section ever held.

Capture the system and section tested, the test medium, the head or pressure with its basis, the hold time, the start and end readings, the water and ambient temperature, who witnessed it, and the pass or fail. If a drop was thermal, note the temperature that explains it. If a joint was repaired, note it and that the retest ran from the start. Tie it to the code and edition you tested to, because a head or a pressure only means something against the code that called for it.

Common mistakes

• Testing a green solvent weld before the manufacturer's cure time, so the joint pulls or weeps on a weld that would have held an hour later.
• Filling a water test without bleeding the high points, so trapped air makes the level spongy and hides a leak.
• Leaving the water heater, the relief valve, the backflow preventer, or another low-rated device in the test section.
• Reading a pressure or level drop as a leak when the water temperature changed, with no temperature log to tell the difference.
• Using an uncalibrated gauge, or a range that puts the test pressure at the bottom or top of the scale.
• Air-testing plastic DWV under positive pressure, against the manufacturer's caution and the stored-energy hazard.
• Covering, insulating, or backfilling before the witnessed test passes, hiding the joints the test exists to inspect.
• Nipping up a weeping joint under pressure instead of depressurizing, repairing dry, and retesting from the start.
• Leaving water in a section through freezing weather with no heat or drain, splitting a fitting after the test passed.
• Treating a closed stop or isolation valve as a test boundary when a passing seat confuses the hold.

Field checklist

Standards and references

The framework lives in the two model plumbing codes. The IPC, published by the ICC, and the UPC, published by IAPMO, both put the tests in their general regulations chapter: the DWV water test at a 10 ft head held 15 minutes, the DWV air test at 5 psi or 10 inches of mercury for 15 minutes for piping other than plastic, the vacuum option for plastic, the water-supply test at not less than the working pressure or a 50 psi air test for non-plastic, and the gravity building sewer test at a 10 ft head. The section numbering, commonly in the IPC's Section 312 testing provisions, shifts between editions, so confirm the section, the value, and the hold against the edition the jurisdiction adopted and any local amendments before you cite one on a permit set.

The pipe and fitting manufacturer governs two things the code defers to them: the solvent-weld cure time before testing, and whether PEX may be tested with air. Their printed instructions control, and a positive air test on plastic against their caution can void the warranty and create a hazard. Backflow and cross-connection assemblies run on their own ASSE test standards and their own field test, covered in the backflow failed-test repair guide, and are not proven by a system pressure test.

Special systems carry their own stricter standards. Medical gas piping is tested and certified to the medical gas standard, commonly NFPA 99, by certified installers and verifiers. Fuel gas piping is tested to the fuel gas code, and fire sprinkler piping is hydrostatically tested and certified to the sprinkler standard, commonly NFPA 13. The standard that controls any given test is the one the AHJ has adopted and enforces. The model code is the starting point; the adopted edition, the local amendments, the project spec, and the manufacturer's instructions are what govern the work.

Units and terms

Pressure testing carries vocabulary from the code, from pipe fabrication, and from the field, and the same idea reads differently across a test record, a code section, and an inspector's note. The terms below travel across the whole test.

Two notes on the numbers. A head of water is a pressure expressed as the height of the column, and 10 ft of head is roughly 4.3 psi at the bottom, which is why the DWV water test is written as a height, not a gauge pressure. Developed length, the measured run a vent or drain is sized by, belongs to the DWV sizing guide and has nothing to do with the pressure test. Do not carry it into a test record.

Hydrostatic test

A pressure test using water, the safe default because water stores almost no energy and a failure weeps instead of bursting

Head

Pressure expressed as the height of a water column; 10 ft of head is about 4.3 psi at the bottom

psi / psig

Pounds per square inch; psig is gauge pressure above atmospheric, which is what the test gauge reads

DWV

Drain-waste-vent, the gravity sanitary and vent piping, tested watertight and gastight

Rough-in test

The test of the bare piping before fixtures and before close-in, proving the joints hold

Final test

The test after fixtures are set, proving the finished DWV system gastight, including the smoke and peppermint methods

Vacuum test

A negative-pressure test, the code option for plastic DWV instead of a positive air test

Test ball / plug

An inflatable or mechanical plug that caps an opening to hold the test head or pressure

Thermal drift

A pressure change on a long hold caused by the test water warming or cooling, read against the temperature log so it is not mistaken for a leak

AHJ

Authority having jurisdiction, the inspector or office that adopts and enforces the code and witnesses the test

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What slope does a drain need?
• Why did my RPZ fail the test?
• What is a pressure reducing valve?
• What is a trap primer?
• How do you size a grease interceptor?