A control valve is not chosen from a catalog. It is computed from the duty. Seven numbers on the line list fix every downstream decision: the body style, the flow characteristic, the required Cv or Kv, the actuator and its fail-safe action, and the seat leakage class. The number most inquiries omit is the MINIMUM flow, and it is not decorative. Minimum flow sets rangeability, and rangeability decides whether the valve modulates in its usable travel band or spends its life throttling near the seat. A correctly sized valve at maximum flow can still be the wrong valve if that second question is never asked. That gap is common in inquiries. It is the cheapest defect to prevent.
The guide is written for procurement and project engineers holding a line list or a datasheet, often reading English as a second or third language. It covers the seven duty inputs, the characteristic decision between equal-percentage, linear and quick-opening, the Cv and rangeability arithmetic with one worked liquid case, the leakage-class ladder from FCI 70-2 and IEC 60534-4 and what Class IV actually permits, the actuation and fail-safe choice across pneumatic, electric, hydraulic and self-acting families, and the three destructive regimes (choked flow, cavitation and flashing, and aerodynamic noise) that wreck a correctly sized control valve. Every number carries its unit and its test condition, and every claim rests on a named standard rather than on a vendor page. Standards named by reference: IEC 60534-2-1, IEC 60534-4, FCI 70-2, ANSI/ISA-75.11.01 and ASME B16.34.
A control valve selection is computed, not chosen. Seven line-list numbers set the body style, flow characteristic, Cv, actuator and seat class: maximum flow, minimum flow, inlet pressure P1, allowable pressure drop at each of those flows, fluid temperature, a fluid-property pair, and the required shutoff (Bahadori, 2014).
| Input | Symbol or unit | Why it decides something | What goes wrong when it is missing |
|---|---|---|---|
| Maximum flow | Qmax (m³/h, kg/h, Nm³/h) | Sets the top of the Cv envelope and the body size | Loop saturates at high demand |
| Minimum flow | Qmin (m³/h, kg/h, Nm³/h) | Sets required rangeability and the low-travel Cv | Rangeability unverifiable; valve throttles on the seat edge |
| Inlet pressure | P1 (bara or psia) | Sizing datum for the flow equation | ΔP alone cannot size a gas or a cavitating liquid |
| Allowable pressure drop | ΔP at Qmax and at Qmin (bar or psi) | Fixes required Cv at each end and installed characteristic | Wrong characteristic; valve hunts |
| Fluid temperature | T (°C or °F) | Sets material rating, SG and Pv, expansion allowance | Body or trim pressure-temperature rating undersized |
| Fluid-property pair | Liquid: SG, Pv (bara); Gas or vapor: M (or relative density) and Z | Closes the IEC 60534-2-1 sizing equation | Choked-flow and cavitation checks impossible |
| Required shutoff | Class per FCI 70-2 or IEC 60534-4 (III to VI) | Selects seat construction and material | "Tight" reads as bubble-tight; wrong seat quoted |
Qmax and Qmin bracket the duty. P1 is the sizing datum. Allowable ΔP is stated at both flows because it is rarely the same. Minimum flow is the input most often omitted from an inquiry, and omitting it makes rangeability unverifiable: the valve modulates near maximum and sits on the seat at minimum. Rangeability is a Cv-arithmetic question, not a duty-inputs one.
For a liquid, the IEC 60534-2-1 sizing equations need specific gravity SG and vapor pressure Pv at the working temperature; Pv fixes where cavitation begins. For a gas or vapor the pair is molecular weight M (or relative density to air) with the compressibility factor Z at P1 and T. Without Z the compressible equation drifts near critical conditions, and any control valve selection based on that shortcut fails in service (IEC 60534-2-1, 2011).
The shutoff is stated as a class from FCI 70-2 or IEC 60534-4, one of III, IV, V or VI, not as the word "tight". Each class is a measured leakage rate on a new, clean seat at a stated test differential and ambient temperature. Specify the class the seat must still meet after the first service interval, not on the day of shipment.
Body style follows from how much pressure drop the valve absorbs and how tight the shutoff must be, not from what is cheapest per inch.
Three control valve body families cover nearly every modulating duty: globe for severe throttling and tight shutoff, rotary (segmented ball or eccentric plug) for high rangeability and solids tolerance, and butterfly for high-capacity low-differential service. ASME B16.34 sets the body and end-connection pressure-temperature rating across all three (ASME B16.34, 2020).
The globe body is the default when the valve must absorb a large fraction of the loop's pressure drop. Flow turns through the trim at a right angle, the plug loads squarely against a machined seat, and the assembly is accessible from the top through a bolted bonnet. A single-seated globe carries the full differential across one seat, needs the largest actuator thrust to close, and delivers the tightest metal-seated shutoff the geometry allows. A double-seated (balanced-plug) globe splits the pressure across two seats and cuts the actuator size, but matching two seats perfectly at once is not achievable in practice; double seating is used where flow control matters and bubble-tight shutoff does not.
Rotary control valve bodies deliver high rangeability and better tolerance of solids, slurries and viscous fluids than a globe, because the closure member sweeps past the seat rather than stalling into it. A segmented ball uses a shaped, characterized ball segment that shears through the seat as it rotates, producing an inherent characteristic close to equal-percentage over a wide travel band. An eccentric-plug design pivots the plug on an offset shaft so it lifts clear of the seat before rotating, which holds shutoff longer under abrasive service (Bahadori, 2014). Both cost less than a globe of the same Cv and take smaller actuators.
A butterfly is a control valve in the middle of its travel and an isolation valve everywhere else. Even triple-offset designs lose throttling authority at low openings and near full open, where gain becomes non-linear or disappears. Butterfly is the right family for high-capacity, low-differential air, water and cooling duty; the wrong family where the valve must absorb a large pressure drop across a small port. Shinjo supplies pneumatic, electric and hydraulic actuation across the three body families; the in-house workshop machines DIN and API gate, globe and check valve bodies only.
The inherent flow characteristic of a control valve is chosen from the installed pressure-drop ratio, defined as valve pressure drop at design flow divided by total system pressure drop (ANSI/ISA-75.11.01, 2013). Equal-percentage suits ratios below about 0.7, where system losses dominate at high flow; linear suits ratios above 0.7, where the valve carries most of the loop drop. Habit picks equal-percentage on every line; the number picks the right one for the control valve on that line.
The inherent characteristic is a property of the trim, measured on a test rig at a constant pressure drop across the valve: for a given percent of travel the trim passes a defined percent of rated Cv, and that curve is what a datasheet plots (ANSI/ISA-75.11.01, 2013). The installed characteristic is what the control loop actually sees once the pipe, fittings, exchangers and other equipment take their share of the total drop. As flow rises those system losses rise with roughly the square of velocity, the drop available to the valve falls, and the installed curve bends away from the inherent one. An equal-percentage trim installed on a friction-heavy line reads back to the controller as approximately linear; a linear trim installed on that same line reads back as quick-opening, with high gain near the seat and no authority at the top of travel. The terminology and the reference test condition for a control valve are set by the IEC 60534 series and by ISA-75.11.01, with terms such as inherent flow characteristic and rated travel defined in Part 1 (DIN EN IEC 60534-1, 2024; ANSI/ISA-75.11.01, 2013).
Compute the ratio at design flow, then read the recommended inherent characteristic from the band (ANSI/ISA-75.11.01, 2013). Every band below is the ratio at design flow, not at any other operating point.
| Ratio band | Inherent characteristic | Why it works | Typical duty |
|---|---|---|---|
| More than 0.7 | Linear | Valve absorbs most of the loop drop, so the installed curve tracks the inherent curve. | Level and pressure loops on short, low-loss lines |
| 0.25 to 0.7 | Equal-percentage | System share rises with flow, and the equal-percentage inherent curve linearizes the installed one. | Flow and temperature loops on long or friction-heavy lines |
| Less than 0.25 | Equal-percentage | Loop drop falls sharply with flow, so only equal-percentage keeps installed gain usable. | Bypass and recirculation service |
| On-off transitions only | Quick-opening | Not a modulating characteristic; used where the valve moves between fully open and fully closed. | Emergency vent and blowdown |
The two lower bands look adjacent and behave differently in practice: at a ratio of 0.7 the valve dominates the loop and tolerates a linear trim, while below 0.25 only the equal-percentage curve does the corrective work the pipe cannot.
A quick-opening trim reaches near-full Cv in the first quarter of travel and asymptotes after that. The shape is useful when the valve's job is to move between fully closed and fully open on demand, an emergency vent, a blowdown or an interlock trip, and it is the wrong shape for a modulating control valve at any pressure-drop ratio. Above a few percent of travel the loop gain drops toward zero, the controller reads no plant response to further output change, and integral action winds up. Modulating service picks between linear and equal-percentage on the ratio; on-off service picks quick-opening on the duty.
Rangeability follows from the characteristic choice and from the sizing arithmetic.
A valve type is not the input a workable control valve selection starts from. The seven numbers on the line list, maximum flow, minimum flow, inlet pressure, allowable pressure drop at each of those flows, fluid temperature, the fluid-property pair and the required shutoff class, are what let a recommendation fit the actual duty instead of a guess. Send those seven numbers with your inquiry and talk to a specialist about the configuration your duty actually calls for.
Size the control valve across its whole duty range, not at maximum flow. Compute the required Cv at maximum AND minimum flow before comparing anything to a datasheet, because the two together fix the trim's minimum rangeability and its normal-flow travel position, and both are what a datasheet hides behind a single rated figure. Both coefficients are defined by the flow-capacity standard IEC 60534-2 (IEC 60534-2-1, 2011).
Kv is defined as Kv = Q / √ΔP, with Q in m³/h and ΔP in bar, so its units are m³/h·√bar for a water reference; other liquids carry a specific-gravity correction inside the same standard. Cv is the same coefficient expressed in US customary units (gpm, psi, water at 60°F). The conversion is Kv ≈ 0.865 × Cv, and both coefficients belong to the flow-capacity standard IEC 60534-2 (IEC 60534-2-1, 2011). A datasheet publishes rated Cv at full travel, but the number that matters for control quality is the Cv the trim delivers at the travel position the loop actually sits at. Read that off the vendor's inherent Cv-versus-travel curve, not off the rated figure on the cover.
Take a cooling-water duty on a globe-body control valve. At the design point Qmax = 20.0 m³/h with ΔP across the valve = 0.30 bar, because the loop is long and piping friction dominates. At turndown Qmin = 1.0 m³/h with ΔP across the valve = 1.20 bar, because piping friction drops steeply at low flow, so more of the total system ΔP lands on the valve. Compute the required Kv at both ends.
| Duty point | Flow Q | ΔP across valve | Required Kv = Q / √ΔP | Required Cv (Kv / 0.865) |
|---|---|---|---|---|
| Maximum | 20.0 m³/h | 0.30 bar | 36.5 m³/h·√bar | 42.2 |
| Minimum | 1.0 m³/h | 1.20 bar | 0.913 m³/h·√bar | 1.06 |
The rangeability the duty demands falls out of the arithmetic: required Cv spans 42.2 down to 1.06, a ratio of about 40:1. A trim's published rangeability is a hard cap (ANSI/ISA-75.11.01, 2013); a duty demanding more will lose control at the low end no matter how carefully the maximum-flow Cv was chosen. A compressible service needs one further step, a choked-flow check comparing the pressure-drop ratio x against Fk·xT; that check must clear before the sizing result is trusted.
Common engineering practice places the preferred operating band at rated Cv for a control valve at about 30% to 80% of rated travel, derived from the process-gain window of 0.5 to 2.0 in which loop equipment should stay (Emerson, 2023); outside it, control quality collapses. Below 30% the plug throttles on the seat edge and gain goes non-linear; above 80% the valve is approaching wide open, with no headroom left for a flow increase. An oversized valve, chosen on peak flow with a comfortable margin added, spends its working life below 30% at normal flow: the loop hunts, the positioner output never settles, and the seat land has to be reground at the first outage. This is the strongest argument for the two-point sizing: it exposes the oversized-valve trap before an order is placed. Imperial-unit arithmetic step by step is on the site's companion Cv calculation article.
Beyond a certain pressure drop the control valve stops obeying the sizing equation, and the failure that follows is mechanical rather than hydraulic.
The required Cv at maximum flow may still be correct on paper, but the flow the valve actually passes is capped and the trim it passes through is being eroded, imploded or acoustically hammered. The onset is found with the pressure-drop-ratio test in IEC 60534-2-1 (IEC 60534-2-1, 2011), not by inspection.
The check is a single comparison. Compute the pressure-drop ratio x = ΔP / P1 and compare it against Fk × xT, where xT is the terminal pressure-drop ratio for the trim and Fk is the specific-heat-ratio correction from IEC 60534-2-1 (IEC 60534-2-1, 2011). A pinned case of P1 = 101 bara with ΔP = 90 bara gives x of about 0.89 against a standard-globe Fk × xT of about 0.68, so the flow is choked. The equation is then capped at ΔP = Fk × xT × P1: any pressure drop taken above that produces no additional flow, only vibration and acoustic power. Staged trim is needed to bring the working ratio back below the threshold.
For liquids the analogous number is FL, the liquid pressure recovery factor (IEC 60534-2-1, 2011). When the vena contracta pressure inside the trim drops below the fluid's vapor pressure but the downstream pressure recovers above it, vapor bubbles form and collapse inside the valve; metal is removed from the trim and body in the cavitation pattern. When downstream pressure itself stays below the vapor pressure, the fluid remains two-phase into the pipe and the erosion moves out of the valve and into the piping downstream. That is flashing. A high-recovery body such as many butterfly and segmented-ball designs cavitates earlier at the same nominal duty; a globe body with a low-recovery cage tolerates more.
Sound-power prediction is standardized. IEC 60534-8-3 covers aerodynamic noise from gas and vapor service (IEC 60534-8-3, 2010), and IEC 60534-8-4 covers hydrodynamic noise from liquid service (IEC 60534-8-4, 2015). A prediction is used to specify trim geometry and pipe lagging on the control valve against a site noise limit and to demonstrate compliance before the valve is ordered, not to promise a decibel figure in isolation. Independent CFD work has validated the 8-3 method against measurement on diaphragm elements (Fenini et al., 2021), which is one reason it remains the specification default.
Remedies escalate. Replace the trim first with a multi-stage or drilled-cage design that takes the pressure drop in stages, so no single stage crosses the threshold. Change the body style next, to a low-recovery globe or a purpose-built anti-cavitation body when trim alone is not enough. Finally, move the pressure drop out of the valve entirely, splitting it across a restriction orifice or a second valve in series so the control valve sees only what it can survive.
A leakage class is a test result on a new, clean control valve at a stated differential pressure and ambient temperature, not a property it keeps for life, and no class is zero. FCI 70-2 and IEC 60534-4 (ANSI/FCI 70-2, 2013; DIN EN IEC 60534-4, 2023) define the ladder, and per-class figures on supplier pages differ enough between sources that the standard, not a vendor summary, is the reference a specification should point at.
Warning: A seat leakage class is what the valve delivered the day the test report was signed. It is not a lifetime property, so specify the class the seat must still meet after the first service interval, not on the day of shipment.
| Class | Allowable leakage | Test basis |
|---|---|---|
| III | About 0.1% of rated capacity | Water or air at service delta P, ambient temperature |
| IV | About 0.01% of rated capacity | Water or air at service delta P, ambient temperature |
| VI | About 11 bubbles per minute on a 4-inch port, roughly 1.7 ml per minute; the standards Table I scales from about 1 bubble per minute on a 1-inch port to about 45 bubbles per minute on an 8-inch port | Air at 50 psig; the bubble count depends on port diameter per FCI 70-2 Table I |
Every figure in the table belongs to a specific test medium and test pressure, and dropping the test basis is how "Class IV" turns into an empty label on a datasheet. The Class VI figure is not a percentage but a bubble count tied to the seat port, so a larger body carries a proportionally larger allowance at the same class (ANSI/FCI 70-2, 2013). "Bubble-tight" is a colloquialism for the top of the soft-seated ladder, and even that top is defined by a measured test allowance rather than by zero flow.
The class is what the valve delivered the day the test report was signed. Service changes it. A soft seat, usually PTFE or a filled elastomer, will meet Class VI when new, then degrades with temperature cycling, with particulates that score the sealing face, and with compression set from long dwells against the plug (DIN EN IEC 60534-4, 2023). A metal seat starts one or two classes looser and closes that gap only after lapping; it then wears through cycles, and galling on a hard-against-hard pair is the failure mode that ends its shutoff life. Neither surface stays where the test report put it.
Specify Class IV when a small measured bypass is tolerable, and Class VI when the process cannot see gas leakage past a closed valve. Record the test medium and test pressure on the same line as the class, so the number on the datasheet carries its own basis. Then ask what the seat will be doing after a year in the actual medium, at the actual temperature, and choose the seat material against that answer rather than against a new-valve number.
The actuator step in control valve selection turns on required thrust, stroking speed, the fail-safe requirement, and the utilities that actually exist at the valve, not on preference and not on what the supplier stocks. Thrust or torque is specified 25% to 40% above the calculated requirement (Bahadori, 2014), which is why the utilities available at the valve decide the family before any sizing sheet is opened.
A spring-return pneumatic diaphragm is the cheapest fail-safe answer where clean dry instrument air arrives at the mounting flange; on a remote wellhead with no air, the family moves to electric with a battery back-up or to a self-operated design before the sizing sheet is opened.
| Family | Typical duty | What it needs on site | Fail-safe behavior | Watch out for |
|---|---|---|---|---|
| Pneumatic (diaphragm and piston) | Fast modulating across most process lines | Clean dry instrument air at a specified supply pressure | Inherently fail-safe with a spring; direction set by orientation | Loss of air strokes the valve to its spring position |
| Electric | Precise positioning where no air is available | Reliable power and cabling to the required area rating | Fail-in-place by default; a spring return or battery back-up must be specified | Slower stroking; fail-safe is an added option, not a property |
| Hydraulic | Very high thrust on large or high-pressure valves | A hydraulic power unit with its own maintenance | Fail direction set by accumulator design | Extra rotating equipment on the plant to keep |
| Self-acting | Simple pressure or temperature duties without a loop | No external power; the process itself does the work | Set by the sensing element and its spring | Limited range; no remote signal |
Shinjo supplies pneumatic, electric and hydraulic actuation against the specification the duty produces; for actuator geometry and the ISA symbol set, the pneumatic control valve pillar has the depth an EPC drawing office needs.
Actuator thrust or torque is specified 25% to 40% above the calculated requirement (Bahadori, 2014), with that requirement summing unbalanced pressure force across the plug, seat load fixed by the leakage class, packing friction, and any dynamic thrust in high-recovery service. The margin covers instrument-air pressure droop, ambient temperature variation, and packing wear. Stroking speed is a separate calculation: a pneumatic diaphragm can stroke in under one second where a comparable electric linear actuator needs tens of seconds, so a trip demand or a loop-tuning target fixes the family before the datasheet. A positioner does not add thrust and does not replace the margin; it closes the position loop.
In control valve selection, fail-open, fail-closed and fail-in-place are outputs of the process hazard analysis and the SIL determination, not properties of the actuator that happens to be on the shelf (Bahadori, 2014). The question a HAZOP answers is which valve position is safe if instrument air, electric power or the control signal is lost. Cooling water to an exothermic reactor typically fails open; fuel gas to a burner typically fails closed; a level valve on a slug catcher may fail in place so the vessel neither drains nor overfills. The catalog default is not an answer to a hazard question, and a datasheet that leaves fail-safe action to the vendor is a specification defect a reviewer should return.
Warning: Fail-safe action comes from the HAZOP, not from the actuator on the shelf. A datasheet that leaves it to the vendor is a specification defect.
A control valve specification is one line item that collects the duty, the characteristic, the sizing, the leakage class and the actuator into a single instruction.
Illustrative values, internally consistent but not a recommendation for any real service, make that shape concrete. The sizing line carries three numbers rather than one: rated Cv 160, required Cv 110 at maximum flow, required Cv 14 at minimum flow, with the body and end-connection rating per ASME B16.34 (ASME B16.34, 2020).
The corresponding inquiry does not lead with a valve type or a body size. It leads with the seven numbers from the line list: maximum flow, minimum flow, inlet pressure, allowable pressure drop at each of those flows, fluid temperature, the fluid-property pair, and the required shutoff. Add the medium by name and its concentration where the fluid is a mixture or a solution, because material compatibility and vapor-pressure behavior depend on it. Shinjo supplies against those seven numbers. A valve type sent without them commits the sizing before the duty has been read, and the wrong control valve is as easy to quote as the right one.
A control valve modulates the flow of a process fluid to hold a controlled variable (flow, pressure, temperature or level) at a setpoint. The controller issues a signal, the actuator moves the trim, and the trim varies the flow area. It is the final control element of the loop.
No standard fixes the count at three. The phrase is used in three different framings: by actuation (pneumatic, electric, hydraulic, self-acting), by body style (globe, rotary, butterfly), and by duty (throttling versus on-off). Ask which framing the inquiry means before answering, because the three lists do not overlap.
Common alternatives are final control element, modulating valve, throttling valve and regulating valve. Datasheets often name the specific type (globe control valve, segmented ball control valve) in place of the generic term. In loop diagrams the abbreviation FCE is standard shorthand for final control element.
Symptoms include hunting around setpoint, a signal-to-position stroke that no longer covers the full travel, audible flow noise at partial opening, external leakage at the packing, and a downstream temperature or pressure drift that a stroking test reproduces. A positioner diagnostic captures most of these before the loop notices.
No. Class IV under FCI 70-2 and IEC 60534-4 permits leakage of approximately 0.01% of rated valve capacity, measured on a new clean valve at the standard test differential pressure and ambient temperature. Bubble-tight in ordinary use refers to Class VI, a soft-seated ladder step defined by a low bubble count.
It sits close to the seat for most of the duty. Below about 30% of rated travel the trim throttles on the seat edge, loop gain becomes non-linear, small signal changes drive large flow changes, and seat wear accelerates. The preferred operating band at rated Cv is 30-80% of travel.
Specify from the installed pressure-drop ratio, not from the loop variable. A heat exchanger loop in which the valve carries a small share of the total system drop, and that drop falls as flow rises, calls for equal-percentage. Linear applies where the valve carries most of the drop and it stays roughly constant.
Every modulating pneumatic control valve needs one. It closes the position loop, corrects for actuator hysteresis, stem friction and packing load, and delivers the diagnostics that catch a degrading valve early. On-off valves and self-acting regulators do not. Electric actuators integrate the equivalent function inside the drive.
Control valve selection resolves in one direction. The seven duty numbers on the line list fix everything downstream: maximum and minimum flow, inlet pressure, allowable pressure drop at each of those flows, temperature, the fluid property pair, and the required shutoff. The body style follows from how much pressure drop the trim must absorb and how tight the shutoff must be. The inherent characteristic follows from the installed pressure-drop ratio, not from the habit of specifying equal-percentage. The sizing arithmetic is then run across the whole duty range, so that the required Cv at minimum flow lands the trim inside its usable travel band rather than on the seat edge. The leakage class is chosen with its test basis attached, and read as a starting condition rather than a lifetime guarantee. The actuator is picked from thrust, stroke and site utilities, and the fail-safe action is derived from the process hazard rather than from what the actuator happens to do.
The number most often missing from an inquiry is minimum flow. Without it, rangeability cannot be verified, and a valve sized only at maximum flow is a valve waiting to fail on the seat.
Last reviewed and updated: September 2026. Author: Shinjo Valve engineering team. For valve selection support or a quotation against your service conditions, contact our team.