Every industrial valve ships with a factory test report — but the acceptance criteria behind that report depend on which test standard your project references. The two you will meet most often are API 598 (North American practice) and EN 12266 (European practice, with ISO 5208 behind it). They do not accept the same leakage.
What Both Standards Cover
Both define the shell test (body integrity at 1.5× rated pressure) and the seat closure test (leakage across the closed valve, typically at 1.1× rating). The differences live in test durations, allowable leakage rates, and how they are expressed.
API 598 in Practice
- Referenced by API 600, 602, 623 and most ASME-world purchase specs.
- Resilient-seated valves: zero visible leakage during seat test.
- Metal-seated valves: allowable drop/bubble counts scale with valve size (e.g., a defined number of drops per minute per NPS).
- Test durations are short and size-dependent — from 15 seconds on small bore up to a few minutes on large valves.
EN 12266 in Practice
- Referenced on PED / EN 13480 piping projects and most European datasheets.
- Uses leakage Rates A through G (from ISO 5208): Rate A = no visible leakage; each step up multiplies the allowable rate.
- The spec sheet states the rate explicitly, e.g. “seat test per EN 12266-1, Rate A”.
How to Compare Them on a Datasheet
| Requirement | API world | EN world |
|---|---|---|
| Soft-seated ball valve | API 598 — zero leakage | EN 12266-1 Rate A |
| Metal-seated gate valve | API 598 drops/bubbles by size | Rate C–D typical |
| Shell test | 1.5× CWP | 1.5× PS |
When a Canadian or Turkish EPC orders from a European mill — or vice versa — the mismatch between “API 598” and “Rate A” wording is one of the most common documentation review comments. The fix is simple: state both acceptable standards in the RFQ line.
The Three Tests Both Standards Require
Despite different numbering, both standards test the same three things, and a datasheet that omits any of them is incomplete:
- Shell test. Pressure containment of the body and bonnet with the valve partially open. API 598 applies 1.5 times the 38 °C pressure rating; EN 12266-1 designates this test P10. No visible leakage is permitted through the pressure boundary under either standard — this is the one criterion with no negotiation.
- Backseat test. Applicable to valves with a backseat, verifying the stem seal against the backseat with the packing relieved (P11 in EN 12266-1).
- Seat closure test. Performed at low pressure with air or gas (roughly 4.5–7 bar) and, where required, at high pressure with liquid at about 1.1 times the 38 °C rating. This is the test where the two standards genuinely diverge, because they express acceptance differently (P12 in EN 12266-1).
How the Leakage Criteria Compare
API 598 sets acceptance by seat construction: resilient-seated valves must show no detectable leakage, while metal-seated valves are permitted a size-dependent allowance expressed in drops or bubbles per minute. EN 12266-1 instead uses lettered leakage rates, where Rate A is a zero-leakage acceptance and the subsequent rates permit progressively larger, DN-proportional allowances.
The practical consequence: “API 598” on its own is not a leakage specification for a metal-seated valve, and “EN 12266” on its own is meaningless without the rate letter. A complete requirement reads like “tested to EN 12266-1, seat leakage Rate A” or “tested to API 598, resilient seat, zero leakage”. The two systems are not interchangeable and should not be mixed within one specification.
What the Test Certificate Must Show
- The standard and its edition — acceptance criteria have changed between editions.
- Test pressures actually applied for shell, backseat and closure tests.
- Test medium for each test (air, nitrogen, water) and hold duration.
- The acceptance criterion used — leakage rate letter or the API allowance applied.
- Serial or heat number tying the certificate to the delivered valve.
- Signature of an inspection function independent of production.
Where a witness test is specified, the inspector must be scheduled against the test bench date. Adding a witness requirement after production is complete means re-testing, since a witness cannot certify a test that has already happened.
Frequently Asked Questions
Is EN 12266 Rate A stricter than API 598?
For soft-seated valves they are effectively equivalent (no visible leakage). For metal-seated valves the comparison depends on size and the rate class specified — check the numbers, not the names.
Do I need a witness test?
Standard practice is a manufacturer test with certificates. Third-party witnessed testing (per your inspection level) is specified separately — common on critical service and large-bore orders.
What documents should I receive?
Test certificates referencing the standard and acceptance class, plus EN 10204 3.1 material certs. See our guide to API 600 gate valves for typical documentation packages.
Need valves tested and certified to your project standard? Send ValveAtlas your datasheet.
