API 598 vs ISO 5208: Valve Pressure Testing Explained

Every industrial valve is supposed to leave the factory pressure-tested, yet the two documents that govern that testing are frequently confused in procurement documents. API 598, Valve Inspection and Testing, dominates North American and oil-and-gas specifications, while ISO 5208, Industrial valves — Pressure testing, is the reference for European and international projects. The two standards are close cousins — same test philosophies, nearly identical pressure multipliers — but their leakage acceptance criteria are structured differently, and mixing them up on a purchase order is one of the most common causes of inspection disputes at the vendor's test bench.

What API 598 Requires

API 598 covers the inspection, examination and pressure testing of gate, globe, check, ball, butterfly and plug valves supplied against API product standards. Its core test program is built on three elements:

• Shell (hydrostatic) test — the valve, partially open, is pressurized to 1.5 times its rated pressure to prove the integrity of the body, bonnet and pressure-boundary joints. Acceptance is straightforward: no visible leakage through the shell, and no weeping at gasketed or packed joints.
• Seat leakage tests — with the disc or closure member fully closed, a high-pressure liquid test at 1.1 times rated pressure checks sealing under maximum differential. For resilient-seated and soft-seated valves, a low-pressure air test (commonly specified at about 0.6 MPa, roughly 80 psi) verifies tight sealing at realistic service differentials; air leakage is judged by counting bubbles, liquid leakage by counting drops.
• Backseat test — where a valve has a backseat (typical for gate and globe valves), it is pressurized fully open to confirm the stem backseat seals and protects the packing.

API 598 expresses its acceptance criteria as tabulated allowable leakage: the permitted number of drops per minute or bubbles per minute depends on the valve type and its nominal size. Resilient-seated valves are held to zero visible leakage, while metal-seated valves are allowed a small, size-scaled leakage rate — a practical recognition that metal-to-metal sealing cannot be bubble-tight at high differential.

How ISO 5208 Structures the Same Tests

ISO 5208 applies to metallic industrial valves of all common types and defines the same two fundamental tests: a shell test at 1.5 times the maximum allowable pressure, and closure (seat) tests at 1.1 times that pressure. A low-pressure gas closure test at approximately 6 bar is the pneumatic counterpart used to verify seat tightness at low differential.

The signature feature of ISO 5208 is its graduated leakage-rate system. Instead of a single pass/fail table, the standard defines a ladder of leakage rates — from Rate A (no visible leakage, effectively zero) up through intermediate rates such as AA and B, to the more permissive rates used for non-critical service. The allowance scales with the valve's nominal size, so a DN 200 valve is permitted proportionally more leakage than a DN 50 valve at the same rate. Test durations also scale with size — short holds for small bore valves, several minutes for large ones — and the buyer selects the rate.

If the purchase order does not specify a rate, the applicable valve product standard's default applies. In broad terms, soft-seated valves default to Rate A, while conventional metal-seated valves default to the middle of the scale. Buyers who never state a rate are leaving the tightness decision to whoever fills in the paperwork.

The Practical Differences

Neither standard is "stricter" in absolute terms. An API 598 resilient-seat requirement and an ISO 5208 Rate A designation both demand zero visible leakage; a metal-seated gate valve under either document will pass a small, defined leak rate. The systems differ in how the acceptance number is expressed — fixed table values versus size-normalized rates — which is precisely why a specification must never quote a leakage number from one standard while citing the other. "Rate D" printed next to a reference to API 598 is a contradiction an inspector is entitled to reject.

Procurement Takeaways

• Name one standard and stay inside it. State API 598 or ISO 5208 in the purchase order, and take all acceptance criteria from that document alone.
• If you use ISO 5208, specify the leakage rate. Rate A for isolation duty on gas, toxic or hydrocarbon service; the product-standard default is acceptable for general metal-seated utility service, but say so explicitly rather than by omission.
• Match the rate to the seat material. Resilient seats should be Rate A; demanding Rate A from a metal-seated trim inflates cost without a defined engineering benefit, while accepting a permissive rate on a soft seat defeats the purpose of the soft seal.
• Check test duration and medium. Both standards scale hold times with valve size; large-bore valves need long holds to reveal slow leakage, and gas service valves should receive a gas closure test even when the shell test is hydrostatic.
• Make the certificate carry the numbers. The test certificate should state the standard, the test pressures, the holding times and — for ISO 5208 — the leakage rate achieved. A certificate that says only "tested and passed" is not evidence of tightness class.