An API 598 pressure test is the only objective evidence that a valve as built will hold the line. Everything else in a purchase — the casting certificate, the machining tolerances, the coating thickness — describes the intent. The test describes the result.
And yet the API 598 pressure test report is often the document buyers review most casually. It arrives as a signed sheet, it says “passed”, and it goes into the project file. That sheet is also the only record that will exist if the valve leaks during commissioning two months later.
The API 598 pressure test, step by step
Shell test (body test). The valve is opened and the body cavity filled and pressurised to 1.5 times the pressure rating at 38 °C, with the ends closed or blanked. It proves the pressure-retaining boundary — body casting, bonnet joint, gland — has no through-wall defect or leak path. Any visible leakage fails the test; there is no acceptance band.
Seat test (closure test). With the valve closed, pressure is applied to one side and leakage past the seat is measured. It proves the seating arrangement is capable of performing its function. This is the test that has an acceptance band, and where most misunderstandings occur.
Backseat test. For valves with a backseat feature, pressure applied to the back of the stem proves the backseat seals, allowing stem packing to be replaced with the valve in service. Not all designs include this; smaller and lower-pressure valves frequently do not.
Functional and operating test. Where the specification requires it — commonly in API 6D pipeline orders — the valve is operated under differential pressure and the operating torque recorded. This is what catches an undersized actuator or a gearbox that cannot move the valve under real load.
Low-pressure closure test. A separate low-pressure check is specified in some standards to confirm seating behaviour under conditions closer to those that prevail when a valve is closed against a low-pressure system. Not every API 598 pressure test specification calls for it.
Test pressures and durations
The usual basis for an API 598 pressure test is: shell test at 1.5 × the 38 °C rating, seat test at 1.1 × the 38 °C rating, with the rating taken from the applicable standard such as ASME B16.34.
Minimum holding durations under API 598 follow the valve size. Widely used values are:
| Valve size | Shell test duration | Seat test duration |
| NPS 2 and smaller | 15 s | 15 s |
| NPS 2 to 6 | 60 s | 60 s |
| NPS 8 to 12 | 120 s | 120 s |
| NPS 14 and larger | 180 s | 120 s |
Verify these against the current edition of the standard and the purchaser’s specification — they are a practical reference, not a substitute for the document. Higher pressure classes and different valve types can carry different requirements.
Two details that cause avoidable failures in the test bay:
- Test pressure is measured at the valve, not at the pump.A gauge on the pump discharge with 15 metres of hose between it and the valve is not evidence of the pressure at the valve.
- Gauges must be calibrated and of a suitable range.A gauge reading in the bottom fifth of its scale is not accurate enough to certify a test pressure. Use a range of roughly 1.5 to 2 times the test pressure, with a calibration certificate within 12 months.
Understanding ISO 5208 leakage rates
ISO 5208 defines leakage acceptance in rates designated A through H, with A the most stringent. In practice the distinction that matters most is:
- Rate Ameans no visible leakage. This is the requirement for soft-seated valves and where the specification says “zero leakage”.
- Rates B through Hpermit progressively larger measured leakage, expressed as a function of nominal size and time. Metal-seated valves are commonly accepted at Rate D or better, because a metal-to-metal seat at ambient temperature cannot be expected to produce an absolutely dry seal.
This is where project disputes begin. A buyer who writes “zero leakage” for a metal-seated gate valve on a 600 °C steam line has specified something the design cannot guarantee at ambient test conditions — and a supplier who accepts that order without a written exception has created a problem for both parties.
The correct approach is to state the acceptance rate and the standard it comes from in the specification: “Seat leakage not to exceed ISO 5208 Rate D, tested per API 598.” An API 598 pressure test is only as meaningful as the criterion written against it.
Eleven things to check on a test certificate
- Certificate number, and a unique identifier tying the report to the specific valve or batch.
- Valve tag numbers, matching the packing list.
- Heat numbers for body, bonnet, disc and stem, matching the material certificates.
- Test medium — water, kerosene or other, as specified.
- Shell test pressure and duration, and the pressure used to derive it.
- Seat test pressure and duration, for both directions if bidirectional sealing is required.
- Observed leakage, expressed against the acceptance criterion — not just “OK”.
- Test water temperature, with the chloride content for austenitic bodies (many project specifications limit chlorides to around 30 ppm to avoid pitting).
- Tester name and signature, plus the date.
- Witness signature and stamp if third-party inspection was required.
- Date of the test relative to the shipment date. A test certificate dated four months before shipment describes a valve that has since been drained, painted, packed, moved and possibly disassembled.
Documentation levels: 3.1 versus 3.2
EN 10204 3.1 is a certificate issued by the manufacturer’s inspection department, which must be independent of the production department. It is the standard requirement for most commercial valve orders.
EN 10204 3.2 is a certificate issued by an independent third party — a notified body or an inspection agency. It is required where the end user or the applicable regulation demands independent verification, and it needs to be arranged before production, because the inspector has to be present for defined hold points.
If your project requires witnessed testing, notify the manufacturer early. Inspection agencies typically need two to four weeks’ notice, and a valve that is already painted and packed cannot be re-tested for a witness without additional cost.
Frequently asked questions
Does a passed factory test mean the valve will not leak on site? It means the valve met the acceptance criteria of the API 598 pressure test at the time of test, under test conditions. Site conditions differ — temperature, differential pressure, line strain and contamination are all outside the scope of a factory test.
Can valves be tested pneumatically instead of hydrostatically? Pneumatic testing is permitted in defined circumstances but stores far more energy and carries a real safety risk. Where it is required, agree the medium and pressure in writing and ensure the test bay is properly secured.
How long should a shell test be held? The minimum duration in an API 598 pressure test is set by the standard and varies with size; see the table above. A shorter hold proves less and is not compliant.
Why is the test pressure based on 38 °C ratings? Because the standard ratings behind an API 598 pressure test are tabulated at that reference temperature. Testing at a higher metal temperature requires a reduced pressure, calculated from the rating tables. Do not apply ambient-temperature test pressures to a hot valve.
What happens if a valve fails the seat test? It is reworked and re-tested, and the original failed test should remain in the record. A factory that discards the failed report rather than documenting the rework is telling you something about how it handles quality records generally.
API https://www.api.org/products-and-services/standards
ISO https://www.iso.org/standards.html
EN https://www.en-standard.eu/
We test every valve before shipment and issue a report showing the actual pressures, durations and observed leakage rather than a single word of approval. Third-party inspection can be arranged with the agency of your choice — tell us the requirement at enquiry stage.
