T&P Valve and Discharge Piping Verification
Purpose
The temperature and pressure relief valve is the one device on a water heater whose failure mode is a rupture, and the one nobody looks at until an inspector does. Two different things go wrong and they need different habits: the discharge piping fails visibly and gets corrected on any visit, while the valve fails silently and only reveals itself when you disturb it. This procedure verifies the piping every visit and treats a lever lift as a commitment, not a check, because a valve that will not reseat has to be replaced before you leave.
Scope
Covers verification of the relief valve and its discharge piping on a residential or small-commercial storage water heater, gas, electric or hybrid, on a service, install or annual visit.
It does not cover sizing or setting expansion control, which the expansion tank verification SOP owns and this procedure hands off to as soon as thermal expansion is identified as the cause of a drip, or a leaking tank, which goes to the failure investigation SOP.
Roles and the handoff
| Role | Owns | Hands off |
|---|---|---|
| Office | Books a drip complaint as a diagnostic, not a valve swap | Tells the tech if the property has a pressure reducing valve or a recent meter change |
| Service tech | Every step below, including the decision not to lift the lever | Returns the static reading, the recovery peak and the discharge disposition |
| Owner or lead | Any scope the visit uncovers that the customer did not call about | Quotes pressure work as its own line, so a valve swap is never sold as the fix for expansion |
The handoff that breaks: a tech replaces a dripping valve, the drip returns in a week, and the shop eats a second visit. The valve was never the fault, and the reading that would have said so takes two minutes at a hose bib.
Procedure
Read the valve's own markings and the tank's rating plate together before anything else. Acceptance: the valve's marked set pressure at or below the plate maximum working pressure, its marked relieving capacity in BTU per hour at or above the appliance input, and a certification marking to ANSI Z21.22 / CSA 4.4 present. Residential stock is commonly marked 150 psi and 210 F, but read yours. Wrong looks like a relieving capacity under the appliance input, which is undersized however new the valve is: replace it. Hazard: the flue connector on a firing gas appliance burns on contact, so shut it down and let it cool before reaching across it.
Confirm the valve is where the tank's manual puts it, sensing the hot zone. The sensing element has to sit in the top of the tank, in the tapping the manufacturer designated, not on a side port chosen for convenience. Acceptance: valve in the designated tapping with the element in the tank, not in an extension nipple that holds it out of the water. Wrong looks like a nipple long enough that the probe never reaches stored water: it can then only respond to pressure, so the temperature half of the device is decorative. Hazard: none, it is looking, but keep clear of the vent connector.
Walk the discharge piping against the section your jurisdiction adopted, top to termination. Acceptance: full size of the valve outlet with no reduction anywhere, no valve and no trap in the line, continuous downward pitch, material rated for the valve's marked temperature, a plain unthreaded end, and a visible termination discharging to an approved location with an air gap rather than direct-connected to the drainage system. The termination height figures are stated numerically in International Plumbing Code 504.6 and in Uniform Plumbing Code 608.5, they differ between the two and have moved between editions, so pull the adopted section rather than working from memory. Wrong looks like ordinary PVC on a line that must carry 210 F water, a reduction to 1/2 in, or a threaded end a homeowner can cap: stop, the appliance does not go back in service on that discharge. Hazard: a line that has been relieving holds hot water, so open it with the outlet pointed away and your hands off the end.
If anything is wet, measure the system before you blame the valve. Gauge cold static at a hose bib with nothing running, then leave a gauge with a lazy hand on through a full recovery. Acceptance: static in psi recorded, closed or open stated as a yes or no from the presence of a pressure reducing valve, backflow preventer or metered check, and the recovery peak in psi recorded against the valve's marked set. Wrong looks like a peak that reaches the valve's marked set: the valve is not failing, it is relieving thermal expansion, and this goes to the expansion tank verification SOP. A static above 80 psi is separately its own scope under International Plumbing Code 604.8 in your adopted edition. Hazard: thread the gauge fully home before opening the bib, because a gauge blown off a fitting at high static is a projectile.
Decide whether to lift the lever, and be honest that this is a commitment. Lift only where the valve is not currently weeping, the discharge terminates somewhere nobody is standing, occupants are out of the space, and a correctly-rated replacement is on the truck today. Acceptance: all four conditions true and stated, or the test declined with the reason recorded. Wrong looks like a lever lifted on an eight year old valve at the end of the day with no replacement in stock: decline it, because scale under a seat that has never moved routinely stops it reseating. Hazard: none, it is a decision made before you touch anything.
Lift the lever fully and briefly, standing clear of the outlet. Acceptance: full-bore flow at the termination for the two or three seconds the lever is held, then the lever returned to seat. Wrong looks like a trickle or nothing, which means the valve is seized and is replaced. Hazard: discharge can be at stored temperature, which on these systems means 120 to 140 F normally and up to the valve's 210 F rating in a runaway, hot enough to scald bare skin in seconds, so nobody holds a container in the stream, nobody stands at the termination, and the customer is out of the room before the lever moves.
Watch the termination for reseat, and take the stop rule if it does not. Acceptance: flow stops completely inside about a minute of releasing the lever, confirmed by looking at the termination, not by feel. Wrong looks like a continuing dribble: the valve is replaced now, on this visit, and the appliance is shut down and the cold inlet closed until it is. Hazard: do not put a hand under the outlet to check for flow, and do not cap or plug the discharge to stop it, since blocking a relieving valve is how a tank becomes a pressure vessel with no way out.
Replace and return to service with the protective function proved. Shut the appliance down, close the cold inlet, relieve pressure at a low fixture, drain below the valve tapping, fit the new valve with sealant on the threads, refit the discharge to the same rules in step 3, then refill until a solid stream runs at a hot fixture for 60 continuous seconds before energy goes back. Acceptance: no weep at the valve or the tapping at 20 minutes under full static, and a recovery peak measured below the valve's marked set with the gauge still fitted. Wrong looks like restoring energy on a part-filled tank, which destroys an element in seconds and is worse on a gas burner. Hazard: this step puts fuel or electrical energy back with the customer in the house, so on a gas unit run a spillage test at the draft hood about a minute after the burner fires with a personal carbon monoxide monitor on you, and on an electric unit close the breaker from the hinge side.
When the visit cannot end clean
Valve will not reseat and no replacement is on the truck: appliance off, cold inlet closed, customer told the hot water is out until the part arrives, return booked before you leave. Discharge direct-connects to waste: that is a cross-connection concern as well as a code fault, so it gets rebuilt rather than noted. Customer refuses the rebuild: the appliance is not left in service, and the refusal goes in the record in their words. Static above 80 psi with no pressure reducing valve: quote it under IPC 604.8 in your adopted edition, because every downstream fix is temporary until it is done.
The record this produces
- Valve marked set pressure, marked relieving capacity, certification marking, and the plate maximum working pressure and input it was compared against
- Valve location and whether the sensing element reaches stored water
- Discharge piping disposition item by item, and the code section it was measured against
- Cold static in psi, closed or open as a yes or no, and the recovery peak in psi
- Whether the lever was lifted or declined, with the reason, and the reseat result
- Post-work leak check at 20 minutes and the recovery peak after any pressure work
The static and peak pair is what stops the next tech replacing a valve that is behaving.
Worked pass: 8 year old gas tank, customer reports a drip by the heater
Single family house, meter with an integral check, no pressure reducing valve, no expansion tank. Complaint is an intermittent drip, worst in the evening.
- Steps 1 and 2: valve marked 150 psi, 210 F, relieving capacity above the plate input of the 40,000 BTU per hour appliance, certification marking present, plate maximum working pressure 150 psi, so the set pressure sits at the plate limit and is acceptable. Valve in the designated top tapping, element in the tank.
- Step 3: FAILED. Discharge run in ordinary PVC, reduced from the valve's 3/4 in outlet to 1/2 in at the first fitting, and terminating in a threaded end over a floor drain. Three separate faults, any one of them enough. Stop rule taken: the appliance does not go back in service on that discharge, and it was rebuilt in CPVC, full 3/4 in throughout, no reduction, continuous fall, plain end, terminating with a visible air gap over the drain to the adopted IPC 504.6 heights.
- Step 4: cold static 92 psi at the hose bib. Metered check makes the system closed, and there is no expansion control. Gauge left through a recovery: lazy hand peaked at 150 psi, which is exactly the valve's marked set. The valve has been opening because the system reaches its set pressure, so the valve is working and the drip is an expansion problem.
- Steps 5 to 7: lift test declined and the reason recorded. A valve that has been cycling on expansion for years has a seat that has been passing water, so it is replaced as part of the correction rather than tested.
- Step 8: customer approved the pressure work on the spot. Pressure reducing valve fitted at the service and set to 65 psi, which brings static under the 80 psi threshold in IPC 604.8; expansion tank fitted and pre-charged to that same 65 psi under the expansion SOP; new relief valve on the rebuilt discharge. Refilled to a solid stream at 65 seconds, gas relit, spillage test clean a minute after firing, no weep at 20 minutes. Gauge left through a full recovery: peak 72 psi, so 72 minus 65 is a 7 psi rise, inside the expansion SOP's 10 psi gate and far below the valve's 150 psi set.
A tech who swapped the valve and left would have been back inside a fortnight, and the discharge that has to carry the next real relief event would still be half-size PVC with a threaded end.
References
- ANSI Z21.22 / CSA 4.4 for the certification and markings a water heater relief valve carries, including set pressure and relieving capacity
- International Plumbing Code 504.6 or Uniform Plumbing Code 608.5, in the edition your authority having jurisdiction has adopted, for discharge piping and termination requirements, which differ between the two documents
- International Plumbing Code 604.8, in your adopted edition, for the pressure reducing valve requirement above 80 psi static
- Manufacturer installation manual and rating plate for the tank, for the maximum working pressure, input rating and designated relief tapping
- See related: the expansion tank verification SOP, the water heater annual service SOP, and the water heater failure investigation and warranty claim SOP