Valve Leakage Classes Explained: Matching the Test Rate to the Service

The Short Answer

Valve leakage classes are the acceptance levels that define how much a valve may leak across the closed seat during a seat test, and they are the single most-cited specification line in valve procurement. The two systems in widest use are API 598 (and MSS SP-61), which expresses acceptance mostly as “no visible leakage” for resilient-seated valves and a measurable rate for metal-seated valves, and ISO 5208, which defines lettered rates (A through F) with specific allowable leak rates in bubbles or volume per time. The class you specify changes the design: a rate that demands a bubble-tight seal usually requires a soft seat or a lapped metal seat, while a looser class allows standard metal-to-metal seating. Specifying the right class is a cost decision — every step tighter costs more in trim, testing, and rejection — so the class should match the service consequence of leakage, not a habit of always specifying the tightest.

The Two Test Systems

API 598, “Valve Inspection and Testing,” is the standard test for industrial valves, and its seat-leakage acceptance is expressed in two ways: for resilient-seated valves, no visible leakage through the seat during the test; for metal-seated valves, a measured leakage rate (commonly expressed in drops per minute or a volumetric rate) with acceptance limits that depend on valve size. MSS SP-61 is the companion standard with similar leakage requirements. ISO 5208, “Industrial valves — Pressure testing of metallic valves,” defines leakage rates A through F: Rate A means no visible leakage, Rate B allows a small number of bubbles, and Rates C through F allow progressively larger measurable leak rates, usually stated as a volume per unit time for the test. In practice, “bubble-tight” in American practice corresponds roughly to ISO 5208 Rate A (or, for some services, Rate B with a defined bubble count) — but the rates are not directly interchangeable, and the exact allowance must be read from the standard’s tables for the valve size.

Leakage Classes at a Glance

System	Class / Rate	Allowance (typical)	Typical Seat Design
API 598	Resilient seat	No visible leakage	Soft seat (elastomer/PTFE)
API 598	Metal seat	Measured rate by size	Metal-to-metal seating
ISO 5208	Rate A	No visible leakage	Soft seat or lapped metal seat
ISO 5208	Rate B	Small defined bubble count	Lapped or precision metal seat
ISO 5208	Rates C-F	Progressive measurable leak rates	Standard metal seat

The table is a practical guide, not a substitute for the standard’s tables — the exact allowable values are size-dependent and edition-dependent (data needs further verification for your specific standard edition and valve size).

How Leakage Is Tested

A seat test pressurizes the closed valve with the test medium (air, gas, or water) on one side at the specified test pressure and measures what passes through the seat. For gas tests, the accepted method is bubble counting: the downstream side is submerged or connected to a bubble tube, and the technician counts bubbles over a defined period against the class allowance. For water tests, leakage is collected and measured by volume or drops over time. The test pressure, medium, and hold time are defined by the standard and often modified by the purchaser’s specification — a valve that passes a water seat test does not automatically pass a gas test at the same class, because gas finds smaller leak paths. Class acceptance is always tied to the test conditions: the same valve can pass Rate B under one test medium and fail it under another.

Application Example: A Purchasing Habit That Cost a Project

A project specification called for “bubble-tight shutoff” on all isolation valves, and the purchasing team applied the same requirement to a battery of metal-seated gate valves in a utility service. The vendor quoted lapped metal seats and extended testing, and the cost premium was substantial — all because the specification confused a resilient-seat capability with a metal-seat requirement. The service review showed the utility lines tolerated a measurable leakage rate: a Rate D or a standard metal-seat allowance was fully acceptable, and the premium was unnecessary. The specification was corrected to class the valves by service consequence, and the project saved the premium without changing any safety requirement. The same class-vs-service thinking applies to how leakage is documented on certificates — see valve leakage class calculation, which covers converting measured leakage into the class figures.

Common Misconceptions

The first misconception is that “bubble-tight” is one universal level — it means different allowances in different standards and for different valve sizes. The second is that the tightest class is always better — tighter classes cost more in trim and testing and can force a soft seat where a metal seat is more appropriate for the temperature. The third is that a valve that passes its seat test at the factory stays at that class forever — service wear degrades the seat, and the class is a new-valve test result, not a lifetime guarantee. The fourth is that water-test and gas-test results are equivalent — gas finds smaller leak paths, so a gas test at the same class is stricter. The fifth is that the class alone defines the valve — the test pressure, medium, and duration are part of the acceptance, and a class without its test conditions is an incomplete specification.

Practical Recommendation

State the leakage class together with the test standard, test medium, test pressure, and valve size on every datasheet. Match the class to the consequence of leakage: critical safety and high-value services justify Rate A or a resilient seat; utility and general services can accept a measured metal-seat rate. Verify that the chosen class is compatible with the seat material and the service temperature — a soft seat gives tight shutoff but limits temperature. And confirm the class on the factory test certificate against the standard’s table for the actual valve size before acceptance, rather than trusting the class name alone.

Conclusion

Valve leakage classes translate “how tight is tight enough” into a testable number — no-visible-leakage for soft seats, measured rates for metal seats, and lettered rates under ISO 5208. Specify the class with its test conditions, match it to the service consequence, and the cost of tight shutoff goes exactly where it is needed — no more, no less.

This article provides general educational knowledge only and is not professional engineering advice.

Sources & Further Reading

API, MSS, ISO

Tags: Valve Standards