API 527 Seat Tightness Testing: Pressure, Hold Time, and Pass Criteria

The API 527 seat tightness test checks whether a pressure relief valve’s seating surfaces hold a seal at pressures just below the valve’s opening point. You pressurize the valve inlet to 90 percent of set pressure (or 5 psi below set pressure if the valve is set at 50 psig or lower), hold that pressure for a defined time, and watch for leakage using a method matched to the service medium: bubble count for gas or vapor, visual observation against a dark background for steam, and droplet observation for liquid.¹ The valve passes or fails based on criteria specific to its seat construction and service. The standard covers conventional, bellows, and pilot-operated designs with set pressures from 15 psig up to 6,000 psig.

What the Test Proves and What It Does Not

API 527 is a seat tightness procedure, nothing more. It does not verify set pressure, and it is not a design standard. You are only answering one question: at pressure just below the opening point, does the seat leak more than the standard allows? If a purchaser needs tighter seat performance than API 527 requires, stricter criteria can be written into the purchase order, but the baseline test is the one described here.¹

Test Pressure

For most valves, hold the inlet at 90 percent of set pressure.¹ This puts the valve at the high end of normal operating conditions, just under lift.

The exception is low-pressure valves. When the set pressure is 50 psig or less, test at 5 psi below set pressure instead.¹ Ninety percent of a low set point leaves too little margin and can cause the valve to begin lifting during the test. A valve set at 40 psig gets tested at 35 psig.

Setup by Service Type

Match the testing medium to the service the valve was built for. Using the wrong medium produces unreliable data because gas, liquid, and steam behave differently against a seating surface at high pressure.¹

Gas or vapor service: clean, dry air or nitrogen, supplied through a regulated source that holds inlet pressure steady throughout the hold period.
Liquid service: water. Fill the valve’s inlet body bowl and let it stabilize until there is no visible flow from the outlet before starting observation.
Steam service: saturated steam. After removing condensate, raise inlet pressure to the test level and hold it at least three minutes to heat the valve body before checking for leakage. Skipping the heat soak means the valve has not reached thermal equilibrium and the results will not reflect operating conditions.

For gas or vapor tests, leakage is counted as bubbles through a specific tube arrangement. The tube has a 5/16-inch outside diameter and is submerged vertically 12.7 mm (0.5 in) into a water reservoir.¹ The exact depth creates consistent back-pressure so bubble counts are comparable between tests. Cut the tube end cleanly. A burr causes erratic bubble formation and throws off the count.

Pressure gauges must be accurate to within 1 percent, with the expected test pressure sitting in the middle third of the dial range.¹ A reading near the top or bottom of the scale loses sensitivity where you need it most.

Ambient temperature during the test should stay between 5°C (40°F) and 50°C (122°F). Outside that range the medium behaves unpredictably. Keep the test area free of excessive vibration, which can create phantom bubbles or mask real leakage.

Hold Time and Observation

Once the test pressure stabilizes, hold it before beginning the leakage check. For valves with a nominal inlet pipe size of 2 inches or smaller, the hold is at least one minute. Larger valves need longer for a leak path to develop and for the medium to reach steady-state conditions through the valve body.

Watch for leakage using the method that fits the service:

Gas or vapor: count bubbles forming at the submerged tube over a timed interval. Each bubble represents a measurable volume of gas past the seat. The count needs to be steady and repeatable, not a single burst from residual outlet pressure.
Steam: observe the outlet against a dark background so escaping wisps become visible. Steam is nearly invisible against a light or cluttered backdrop, so the dark background is part of the measurement, not staging.¹
Liquid: watch the outlet for droplets. A single drop forming and falling is a failure.¹

Inlet pressure must stay at the test level for the full observation period. If it drifts down because supply cannot keep up with leakage, you are measuring a moving target and the data is not usable. Continuous pressure monitoring is what makes the leakage reading meaningful.

Pass and Fail Criteria

Metal-Seated Valves

Metal-seated valves get a limited bubble allowance because metal-to-metal contact permits microscopic gas passage even when properly finished. API 527 sets maximum allowable bubble rates based on the valve’s effective orifice size and set pressure. Larger orifices and lower set pressures get higher allowances: larger seating surfaces have more circumference for leakage, and lower pressures produce less force compressing the seat.

As one specific point on the table, an F-orifice valve set below 1,000 psig is allowed up to 20 bubbles per minute.¹ Smaller orifices carry proportionally lower limits, and for valves with nominal inlet pipe size under 1 inch, the allowable rate is capped at a very low volumetric flow. Read the full leakage table for the orifice designation and pressure range of the valve you are testing. Generalizing from one row to another can lead to accepting a valve that actually fails, or rejecting one that passes.

Soft-Seated Valves

Soft-seated valves, which use an elastomer or polymer insert against the nozzle instead of metal-to-metal contact, are held to zero leakage. No bubbles are permitted during the observation period. A resilient seat should conform completely to the nozzle. Any bubbles at the submerged tube point to a defect, misalignment, or damage in the seat that has to be corrected before service.

Steam Service

Steam valves must show no visible leakage against a dark background for the entire observation period.¹ Any wisp is a failure. There is no allowable quantity. The reason the criterion is binary: even a small steam leak at high temperature erodes the seating surfaces, and the erosion accelerates.

Liquid Service

Liquid valves follow the same zero-leakage rule. A drop forming and falling from the outlet during observation is a failure.¹ A wet film or slight moisture at the outlet that does not form into a drop is generally not treated as a failure, but the moment a distinct droplet breaks free, the seat is not tight enough.

After a Failure

A failed test does not mean the valve is scrapped. The usual corrective step for metal seats is lapping, which uses an abrasive compound to grind the nozzle and disc contact surfaces back to a flat, smooth finish. For soft seats, the elastomer insert is typically replaced. After any repair, the valve goes through the full API 527 test again from the beginning. There is no shortened retest. The repaired valve has to meet the same criteria as a new one.

1
American Petroleum Institute. API Standard 527 – Seat Tightness of Pressure Relief Valves