When a control valve is fully closed, it is rarely perfectly tight, and how tight it actually is has a name: its seat leakage class. The ANSI/FCI 70-2 standard, mirrored internationally by IEC 60534-4, sorts valves into leakage classes from I through VI, each specifying how much fluid a shut valve is allowed to pass. This is the tight-shutoff rating you see on a valve data sheet, and it drives a practical decision that trips up a lot of specifications: how tight does this valve really need to be, and is a control valve even the right device to isolate with? This page lays out the classes and the decision that follows from them.
Valve Seat Leakage Class (ANSI/FCI 70-2) in one line: Valve seat leakage class, defined by ANSI/FCI 70-2 and IEC 60534-4, rates how much a closed valve leaks, from Class I with no specified limit up to Class VI, the tightest, a soft-seat bubble-tight standard. Class IV is a typical metal-seat rating, Class V allows a small measured liquid leak rate, and Class VI is verified by counting bubbles on a soft seat. Isolation and shutoff valves need Class V or VI; throttling valves usually tolerate Class IV.
The classes run from loosest to tightest. Class I is essentially an agreement between buyer and manufacturer with no defined leakage limit, used where shutoff simply is not a requirement. Classes II and III allow progressively less leakage as a percentage of the valve's rated capacity, and Class IV tightens that further; Class IV is the common benchmark for a standard metal-seated control valve, allowing a small leakage expressed as a fraction of rated flow. These lower classes reflect that a metal plug closing against a metal seat, ground to fit but not gasketed, will always pass a little fluid.
Class V is a step tighter and is defined differently, as a specific allowable leak rate measured under a defined test with water, scaled to the seat size and test pressure, rather than as a percentage of capacity. It represents a high-performance metal seat lapped for tight shutoff, and it is a common requirement where a control valve is also expected to isolate reasonably well. Because Class V is a measured rate rather than a bubble count, it can be verified quantitatively, which is part of why it is specified for services that need dependable, repeatable tightness from a metal seat.
Class VI is the tightest and is the province of soft-seated valves, where an elastomer or PTFE seat deforms against the plug to seal. It is verified by a bubble test: the shut valve is pressurized with air or nitrogen and the leakage is measured as a permitted number of bubbles per minute, scaled to the port size, with the allowance being very small. Class VI is what you specify when you need near-bubble-tight shutoff, and achieving it generally means accepting a soft seat, with its own temperature and service limits, rather than a metal one. Together the six classes give a common language so a data sheet number means the same thing to everyone.
The practical decision is matching the class to the job the valve does. A valve whose whole purpose is to throttle, sitting mid-travel most of its life and only occasionally closed, rarely needs better than Class IV, because a little leakage when shut does not matter to a service that is about modulating flow, not sealing it off. Paying for Class V or VI on a pure throttling duty adds cost and often a softer seat with tighter limits for no real benefit. So throttling valves commonly land at Class IV, and that is the right answer for them.
The moment the valve is expected to isolate, though, the requirement climbs. An isolation valve, or a control valve doubling as an emergency shutdown element, has to genuinely stop the flow, and that pushes you to Class V for dependable metal-seat tightness or Class VI where you need near-bubble-tight shutoff, such as for a hazardous fluid that must not weep past a closed valve. The class you specify follows directly from the consequence of leakage: nuisance leakage on a throttling service tolerates Class IV, while a safety or environmental consequence demands Class V or VI.
Even so, there is a hard limit worth stating plainly: a control valve is not a substitute for a dedicated block valve. Control valves are built to modulate, and their seats wear as they throttle, so even a valve that leaves the factory at Class VI can degrade in service as the trim erodes. Where positive isolation is genuinely required, for maintenance lockout, for a tight battery limit, or for safety isolation, the correct practice is a dedicated on-off block valve designed for tight shutoff and not asked to throttle. The leakage class tells you how tight a control valve is when shut; it does not turn that control valve into an isolation device.
The leakage class you choose ripples straight into trim material and cost. A metal-seat Class IV valve is comparatively simple and durable, tolerant of temperature, dirt, and erosion, which is why it is the workhorse for throttling. Reaching Class V demands a carefully lapped, harder metal seat and tighter machining, and reaching Class VI generally means a soft seat, an elastomer or polymer element that seals beautifully but brings temperature ceilings, chemical-compatibility constraints, and a wear part that will not last as long in abrasive or hot service. Higher shutoff class, in short, usually means more cost and often a narrower operating envelope, so over-specifying tightness has real downsides beyond price.
Once a valve is in service, the concern shifts to whether it still holds its class, because a valve that is passing, leaking through when it should be shut, wastes product, defeats isolation, and can be a safety or emissions issue. That passing is often invisible at the valve but visible in the process data. A block or shutoff valve that is supposed to be closed but shows a downstream temperature creeping up, a pressure that will not fall, or a flow that should be zero but is not, is telling you it is leaking through, and those are exactly the signals a monitoring system already carries.
A cloud SCADA platform such as Merobix historizes the pressures, temperatures, and flows around valves across a facility, so a passing valve can be caught as a trend rather than discovered during a shutdown that will not hold pressure. For remote and unmanned sites this matters because there is nobody to feel a valve body warming from throughflow; the evidence has to arrive as data. Trending a supposedly closed valve's downstream conditions turns seat leakage from an invisible loss into an alarmable, reviewable condition, and it helps an operation know when a valve's real-world shutoff has drifted away from the class it was bought at.
Class IV is a typical metal-seat rating allowing a small leakage as a fraction of the valve's rated flow. Class V is tighter, defined as a specific measured liquid leak rate scaled to seat size and test pressure, representing a high-performance lapped metal seat. Class VI is the tightest, a soft-seat standard verified by a bubble test with a very small permitted leakage. Tightness increases from IV to VI, and Class VI generally requires a soft seat rather than metal.
Not reliably. A Class VI rating means the valve is near-bubble-tight when new, but control valves are built to throttle, and their seats wear as they modulate, so the tight shutoff degrades over time. Where positive isolation is genuinely required, for maintenance lockout or safety isolation, the correct choice is a dedicated on-off block valve designed for shutoff and not asked to throttle. The leakage class describes tightness; it does not make a control valve an isolation device.
An emergency shutdown or isolation valve needs a tight class, typically Class V for dependable metal-seat tightness or Class VI where near-bubble-tight shutoff is required, because its whole job is to genuinely stop the flow when it closes. A throttling control valve, by contrast, usually tolerates Class IV since a little leakage when shut does not undermine its modulating duty. The class follows the consequence of leakage: safety and environmental consequences push you to V or VI.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.