Pressure Test on a Repair Before Close

Purpose

A repair is not finished when the joint stops dripping, it is finished when a gauge with a stated resolution held a stated pressure for a stated time with the water temperature stable. "No visible drip" is not a test, it is an observation made by someone who wants to go home. This procedure gives the shop a repeatable close that catches the joint that will weep at two in the morning, and it produces a record that answers the only question that matters when a customer calls back: was it tested, at what pressure, for how long, and did anything move. The second failure it prevents is a tech who tests a repair and never restores the relief valve or expansion tank he isolated to do it.

Scope

Covers verification of a completed repair or alteration on a pressurized potable water line in a house, small office, small retail space or multifamily unit, before the access is closed and the ticket is signed.

Does not cover the fuel gas test, which is a different medium, a different pressure and a different set of consequences, and which the gas line pressure test SOP owns. Does not cover the rough-in test an inspector witnesses on new work, which runs to the code test pressure and duration rather than this shop standard. Does not cover drain, waste and vent testing. Does not cover locating an unfound leak, which is the acoustic and tracer gas leak location SOP.

Roles and the handoff

Procedure

1. Choose the medium, and choose water unless something makes it impossible. Acceptance: the medium recorded, with a written reason where air was used. Stop rule: air is not used on plastic piping without the pipe manufacturer's written permission, and many prohibit it outright. Hazard: compressed air stores far more energy than water at the same pressure, so a failure under an air test throws fittings rather than spraying, which is why nobody stands in line with a cap or an end and the area is cleared before the compressor starts.

2. Define the test boundary and write down everything you isolated. Acceptance: a list naming each isolation point and each device valved out or removed, including the expansion tank, the water heater, the pressure reducing valve, any backflow assembly, filters and any existing gauge. Stop rule: a device that cannot be isolated sets the ceiling for the test pressure, so the test runs at or below that device's rating or does not run. Hazard: closing a valve upstream of a water heater that is still energized can leave it heating into a closed volume, so the heater is switched off at its disconnect or gas control before the boundary closes.

3. Fill and purge until air is gone, then purge again. Acceptance: solid water with no spitting at the highest and furthest point on the isolated segment. Stop rule: a segment that will not stop spitting has a high point with no vent, so the test does not start; trapped air compresses and then redistributes, and the gauge falls for reasons that have nothing to do with a leak. Hazard: purging a hot-side segment sends water at 120 to 140 F out of a fixture at hand height, and 140 F water scalds in a couple of seconds, so purge cold first and glove up.

4. Pick a gauge whose resolution is finer than the acceptance. Acceptance: gauge increments no coarser than half the drop you would call a failure, so a 1 psi acceptance needs a gauge marked in 1 psi or finer. Stop rule: a gauge marked in 5 psi increments cannot resolve a 1 psi acceptance, so the test does not run on it and the job does not close; get the right gauge. Hazard: a gauge threaded onto a hose bib is a projectile if the threads are corroded, so back the bib body with a second wrench and never stand square in front of a gauge while bringing pressure up.

5. Record water temperature at the start, at the same point every time. Acceptance: a temperature reading with its location, taken at the segment rather than at the room. Stop rule: a segment just filled with cold water in a warm building is still equilibrating, so wait until two readings five minutes apart agree within 1 F before the hold starts. Hazard: none here, a thermometer against a pipe; record it as checked rather than blank.

6. Bring pressure up slowly to the stated test pressure and log the start. Acceptance: the test pressure reached, with the clock time and the gauge reading written down at that moment. Stop rule: never exceed the lowest rated component still inside the boundary, and for plastic tube remember the rating falls as temperature rises, which for PEX is published per ASTM F876. Hazard: bringing pressure up puts a tech's face at the fresh joint, so stand off its plane, come up over a minute rather than in a second, and keep everyone else out of the room.

7. Hold, then walk every joint you disturbed, not just the new one. Acceptance: no measurable drop on the chosen gauge over a hold of at least 15 minutes, with end temperature within 1 F of the start, and every joint in the boundary wiped and inspected dry. Stop rule: any measurable drop with temperature stable means a leak, so find it and re-test from step 6 rather than extending the hold hoping it settles. Hazard: reaching over and behind piping under test puts hands next to a fitting that may let go, so walk with a light and a rag rather than a bare hand on the joint.

8. Interpret the pair, never the pressure alone. Acceptance: a written verdict pairing the pressure change with the temperature change, in the order pass, thermal or leak. Stop rule: a drop with a falling temperature is ambiguous and does not pass; it is re-run with temperature stable, because in a closed rigid system a small temperature change moves pressure far more than a small leak does over the same interval. Hazard: none here, a judgement at the gauge; discharge it rather than leaving it blank.

9. Put every isolated device back and prove each still does its job. Acceptance: expansion tank reinstalled and its pre-charge confirmed against static, relief valve back in service with its discharge piping reconnected and terminating as it was, pressure reducing valve back in service with static re-read at the same gauge point, filters reinstalled, heater re-energized, and static within about 5 psi of the arrival reading. Stop rule: any device on the step 2 list you cannot restore is called out in writing before you leave, and the ticket does not close. Hazard: this is the step that puts pressure, heat and fuel back into a building with a family in it, so the relief valve is verified free and unplugged before the heater is fired, and no plug, cap or test fitting leaves the site still installed.

The record this produces

Seven fields: the medium with its reason where air; the isolation list from step 2; the gauge resolution; start pressure with clock time; start and end water temperature at the same point; end pressure with clock time and the hold duration; and the restore confirmation with static at the same gauge point as arrival.

That line is what a callback is argued from. A shop that can say the segment held 65 psi for 15 minutes on a gauge marked in 1 psi increments with water at 58 F throughout is in a different position from one that says the tech did not see any water.

One worked pass, including a step that fails

A 3/4 in copper repair at a laundry wall, one coupling and one compression stop replaced, wall open.

Step 1: water. No reason for air on this repair, recorded.

Step 2: boundary is the laundry branch, isolated at the branch valve and at the fixture stop. Nothing else is inside the boundary, so no expansion tank, relief valve or regulator is isolated on this job and the list says so explicitly rather than being left blank.

Step 3: filled and purged at the laundry bib until solid, then purged again at the high point of the branch.

Step 4 FAILS. The gauge in the van is marked in 5 psi increments, and the shop's acceptance is no measurable drop against a 1 psi resolution. Five psi increments cannot resolve that. Stop rule taken: the test does not run and the wall stays open until the 1 psi gauge arrives from the shop. The delay is written on the ticket rather than absorbed by lowering the standard.

Step 5: with the correct gauge on, water at the branch reads 58 F, and again 58 F five minutes later, so it is equilibrated and the hold can start.

Step 6: pressure brought up over about a minute to 65 psi, which is normal static here, logged at 14:02.

Step 7 FAILS on the first hold. At 14:17 the gauge reads 64 psi and water reads 58 F. That is a 65 minus 64, or 1 psi drop with temperature unchanged, which meets the acceptance's definition of a failure exactly. Walking the boundary finds a damp ring under the new compression nut. Stop rule taken: the nut is snugged, the joint wiped dry, and the test restarts from step 6 rather than the hold being extended.

Second hold: 65 psi at 14:31, 65 psi at 14:46, water 58 F at the start and 59 F at the end. Zero measurable drop over a 15 minute hold, with a 1 F rise that is inside the step 7 allowance and would if anything push pressure up rather than down.

Step 8: verdict written as pass, pairing zero psi change with a 1 F temperature rise, at 65 psi over 15 minutes on a 1 psi gauge.

Step 9: nothing was isolated beyond the two valves, both reopened. The list from step 2 is re-read to confirm it, rather than trusted from memory. Static reads 65 psi at the laundry bib against 66 psi on arrival, a 1 psi difference and inside the 5 psi acceptance. The test gauge comes off the bib and the bib is proved to shut and hold.

When the site does not match this procedure

The customer wants the wall closed now. The gauge, the hold and the temperature are the product. Close the wall untested and the shop owns whatever happens next with no record to stand on. Offer to return with the gauge the same day.

A device inside the boundary cannot be isolated. The test pressure drops to that device's rating, or the boundary moves. Never test above the lowest rated component still in the loop, and write which component set the ceiling.

Air is genuinely the only option, such as on a line that cannot be filled in freezing conditions. Get the pipe manufacturer's position first, clear the area, and record the medium and the reason. On plastic, the default answer stays no.

Permitted new work. The inspector's test governs, at the pressure and duration in the adopted code, commonly IPC Section 312.5 or UPC Section 609.4. This shop test is in addition, not instead, and the ticket names both.

References

• IPC Section 312.5 and UPC Section 609.4 for the code water supply test, in the edition your authority having jurisdiction has adopted, which governs permitted new work rather than this repair-verification hold.
• ASTM F876 for PEX pressure ratings and their reduction with temperature; the pipe manufacturer's published position governs any air test on plastic.
• See related: the gas line pressure test SOP for fuel gas, and the expansion tank verification on a closed system SOP for the pre-charge check step 9 refers to.