A Quart Every Evening

Why the numbers do not clear it: the standard differential tests measure the checks' ability to hold a forward pressure differential. They say almost nothing about whether check No. 2 is drip-tight against reverse pressure. Check No. 2 at 2 psid is a pass, and it can still leak backward.

Root cause, most probable: thermal expansion.

• The PRV at the service and the RPZ downstream of it effectively close the domestic system. There is no expansion tank.
• Heating 500 gallons of cold water expands it. Downstream pressure rises past the supply pressure.
• If check No. 2 is not tight against reverse pressure, that elevated pressure bleeds backward into the zone between the checks. The relief valve sees zone pressure rise and opens near its 2 psi set point - which is exactly its job.
• Evening timing fits: 6-9 pm is peak hot-water draw and recovery, which is when expansion is greatest.
• Onset fits: discharge began when the heater was replaced. A new heater, a new T&P, or a change in settings can change the expansion behavior.

The assembly is the messenger, not the disease. A quart per evening is a small, intermittent, correctly-triggered discharge, not a gross seat failure. That is why the tech's parts-swapping did not fix it.

What tightening actually does:

• It raises the differential at which the relief valve opens, above the design set point. The drip stops because the valve no longer reaches its opening point.
• The relief valve is the assembly's protection against a fouled or failed check No. 1. Deliberately desensitizing it removes that protection while leaving the assembly looking intact.

The risks

1. False security. The device can still pass a field test (the opening point may now read well above 2.0 psi, which looks like a pass) while no longer providing meaningful relief protection.
2. Listing/spec violation. The spring is an OEM part set at the factory. Altering it takes the assembly outside its approval and the tester's certification basis.
3. The underlying problem survives. Thermal expansion is still happening; the pressure now has nowhere to go except the water heater's T&P or the plumbing itself.
4. Diagnostic damage. The tech has now removed the symptom that pointed at the root cause, and future testers will be reading a modified device.

Correct action: restore OEM parts and settings, then fix the system - not the spring.

1. Restore the assembly. Replace the altered spring and any disturbed parts with OEM components. Verify relief opening at 2.0 psid or greater and drip-tight at 1.0 psid or greater.

2. Test check No. 2 for reverse leakage, using the field procedure for backpressure tightness - not the forward-differential test that already passed. If it leaks, clean the seat and replace the disc; rebuild if needed.

3. Fix the system. Install a properly sized ASME thermal expansion tank on the domestic system, and confirm the water heater T&P relief is functional and correctly rated. This is the actual correction; steps 1 and 2 just put the assembly back the way it was.

4. Verify the fix works. Recheck for relief discharge over several evenings of normal hot-water use, not a five-minute observation. Intermittent symptoms need an observation window long enough to include the trigger.

5. Full retest and documentation. Run the complete five-criteria test sequence after all repairs. On the report, record the as-found condition (including the altered spring), the parts replaced, the root cause, and the expansion tank installation. A tester who signs off a repair without noting the system-side cause has left the next person to rediscover it.

6. Tell the owner plainly that an RPZ on a closed domestic system without an expansion tank will keep producing this call, and that the assembly is functioning correctly when it discharges under those conditions.