Quick Summary:
- Valve seat leakage classes — from Class I (no test required) to Class VI (bubble-tight) in ANSI/FCI 70-2, or Class A through D in ISO 5208 — define the maximum permissible leakage rate through a closed valve, and specifying the correct class for the application avoids both the quality shortfall of under-specifying and the cost premium of over-specifying bubble-tight seats in services that don’t require them.
- Class VI (bubble-tight) shutoff using resilient soft seats is achievable and standard for ball and butterfly valves in clean service, but it is not achievable from metal-seated valves under the same test conditions — requiring Class VI from a metal-seated gate valve is a specification error that cannot be met and causes procurement disputes that delay projects.
- The test conditions under which leakage class is measured — test pressure, test fluid, test duration and measurement method — are as important as the class designation itself: a valve that meets Class IV at 6 bar may not meet the same class at 100 bar, and the standard test conditions do not replicate worst-case in-service conditions in many applications.
Why Leakage Classification Exists
Not all valves in a process plant need to provide bubble-tight shutoff. A block valve on a drain line, a gate valve on a utility water header, or an isolation valve on a non-critical bypass — none of these need to prevent every last molecule from passing through when closed. Requiring bubble-tight shutoff from every valve adds cost (soft seats, precise machining, tighter quality control), and in many valve types (metal-seated gate valves, large butterfly valves in throttling service) it is physically unachievable. The leakage classification system provides a structured way to specify realistic shutoff requirements for each application, avoiding both inadequate shutoff (which causes safety or operational problems) and over-specification (which adds unnecessary cost).
ANSI/FCI 70-2 Leakage Classes
ANSI/FCI 70-2 (Control Valve Seat Leakage) defines six leakage classes specifically for control valves. Class I — no seat leakage test required; for valves where shutoff is not important. Class II — 0.5% of rated Cv as maximum leakage rate; very generous, for simple on-off service. Class III — 0.1% of rated Cv; still generous, for most general control service. Class IV — 0.01% of rated Cv; the standard class for most globe control valves and the default in most specifications. Class V — 0.0005 ml/min per psi differential per inch of port diameter; essentially zero for liquid service, standard for tight shutoff control valves. Class VI — bubble-tight in air at 50 psi (3.45 bar); zero measurable leakage using the bubble test method, the highest class available and standard for critical isolation using soft-seated valves.
The critical thing to understand is that Classes I through V are measured with liquid (water or equivalent), while Class VI is specifically a pneumatic test with air. This asymmetry means that a valve certified to Class VI is bubble-tight to air at the test conditions, but its actual liquid leakage rate is not directly covered by the Class VI test. In practice, a valve that achieves Class VI in the pneumatic test will generally also achieve Class V or better in liquid service, but this is not formally guaranteed by the standard.
API 598 Leakage Classes
API 598 (Valve Inspection and Testing) defines leakage allowances for general industrial valves — primarily gate, globe, check, ball, plug and butterfly valves — rather than specifically for control valves. API 598 test classes are defined differently from ANSI/FCI 70-2, with specific volumetric allowances for different valve types and sizes, tested at defined pressure conditions (typically the valve’s full rated pressure for the relevant class).
For metal-seated gate valves, API 598 allows a leakage rate of (DN in mm × 0.1 mL/min for liquid, or DN × 1.0 mm³/s for gas) — a practical acknowledgment that metal-seated valves cannot achieve bubble-tight shutoff and that a defined low leakage rate is acceptable. For resilient-seated ball valves and soft-seated plug valves, API 598 requires zero measurable leakage under the liquid seat test conditions. This is why specifying “API 598 compliant” for a metal-seated valve gives a different shutoff requirement than for a soft-seated valve — the standard applies different criteria depending on the valve design.
ISO 5208 Classification
ISO 5208 (Industrial Valves — Pressure Testing of Metallic Valves) defines leakage classes A through G, where Class A is no leakage (zero measurable) and higher letter classes permit progressively more leakage. Class A is equivalent to ANSI/FCI 70-2 Class VI for liquid service. Class D and Class F are commonly specified for metal-seated valves where some leakage is accepted. The ISO 5208 standard is widely used on European projects and on international projects where ISO rather than API standards are specified. The class designations are not numerically the same as ANSI/FCI 70-2 — always specify the applicable standard explicitly rather than just a class number to avoid ambiguity.
Choosing the Right Class
For control valves in throttling service: Class IV is the standard default; Class V for tight shutoff service (critical isolation where the control valve doubles as a block valve); Class VI for safety-critical service where zero leakage is non-negotiable (toxic fluid control valves, SIS final elements). For isolation valves: zero leakage (API 598, Class A/ISO 5208 Class A) for soft-seated ball, butterfly and plug valves in clean service; API 598 allowance for metal-seated gate valves where bubble-tight shutoff is physically impossible. For check valves: Class IV through VI depending on the consequences of backflow — critical reverse flow prevention requires a tighter class than a drainage check valve.
Summary
Valve seat leakage class specification requires matching the class standard and test conditions to the valve type and the process consequences of leakage. ANSI/FCI 70-2 for control valves; API 598 for general industrial valves; ISO 5208 for international projects. Specify realistically — don’t require Class VI from metal-seated valves, but don’t accept Class IV where the service demands bubble-tight shutoff. Always specify the test standard explicitly, not just the class number. And remember that test conditions (pressure, fluid, temperature) define what the class actually means in practice — a leakage class at 3.45 bar test pressure tells you less than you need to know for a valve operating at 100 bar in service.