How this instrument works
The valve flow coefficient Cv answers one practical question: how much water would this valve pass, in US gallons per minute, if you pushed it across a 1 psi pressure drop? That figure is measured once at the factory, wide open, on a calibrated water rig, and it becomes the yardstick manufacturers stamp on every model so an engineer can compare a 2-inch globe valve against a 3-inch butterfly without inspecting either one's internal port geometry. Change the pressure drop or swap in a denser fluid and the coefficient itself does not move — only the flow it predicts does, through that square root.
The square root comes from the same physics that governs any flow forced through a restriction: kinetic energy trades against pressure, so velocity, and therefore flow rate, scales with the square root of the pressure difference driving it, never the difference itself. Doubling the pressure drop raises flow by only about 41 percent, not 100. Specific gravity rides inside that same root because a denser fluid resists acceleration more; a given pressure drop pushes fewer kilograms of a heavy fluid through in a minute than it does water, so SG sits under the radical to correct the water-based rating for anything heavier, or lighter, than water itself.
This particular equation covers ordinary turbulent, non-flashing liquid flow, at a single travel position, usually wide open. A valve's installed Cv actually climbs from zero at the closed seat to this rated maximum as the stem travels, following whichever characteristic — linear, equal-percentage, quick-opening — the trim was cut to produce, so one wide-open figure never describes behaviour at 30 percent open. Push the pressure drop high enough relative to the fluid's vapor pressure and it flashes to vapor inside the valve; flow chokes, stops climbing with further pressure drop, and this formula quietly stops applying. Full sizing standards handle that case with a separate recovery-factor correction.
- Enter Flow rate — the duty, in US gal/min, the valve must actually pass at the operating point you are sizing for, not the pipe's rated capacity.
- Enter Specific gravity of the fluid — 1 for water at 60 degrees F, lower for light hydrocarbons, higher for brine or seawater.
- Enter Pressure drop across the valve — the psi budgeted for the valve alone, separate from friction losses in the rest of the line.
- Read Valve flow coefficient (Cv) and check it against a manufacturer's Cv-versus-travel table to pick a size that opens in a usable range.
Worked example — sizing a valve for 100 gpm at a 4 psi drop
A control valve on a water line must pass 100 US gal/min of ordinary water, specific gravity 1, while the process allows it only a 4 psi pressure drop across the trim. Plugging straight into the formula: Cv = 100 × √(1 ⁄ 4) = 100 × 0.5 = 50. That is the number stamped on the datasheet of whichever valve ends up specified — a Cv of 50, wide open, at those conditions.
Fifty is a realistic catalog figure: a 2-inch globe valve commonly rates somewhere near Cv 45 to 60 wide open, so this duty falls squarely inside a standard size rather than forcing an oversized or undersized choice. Had the fluid instead been a light hydrocarbon at SG 0.8, the same 100 gpm and 4 psi drop would need only Cv = 100 × √(0.8 ⁄ 4) = 100 × 0.447 = 44.7 — a smaller, cheaper valve, because the lighter fluid accelerates through the same restriction more easily.
Questions
What does a Cv of 50 actually mean?
It means this valve passes 50 US gallons per minute of 60 degree F water when the pressure drop across it is exactly 1 psi. Cv is a capacity rating measured on a test bench, not a direct description of pipe size or port area — two valves sharing a nominal size can carry very different Cv depending on trim design, which is the entire point of publishing the number: it lets you compare capacity across manufacturers without inspecting internal geometry.
Why does specific gravity sit under a square root instead of being multiplied directly?
Because Cv is defined against a flow rate, and flow rate itself only scales as the square root of the driving pressure difference — doubling the pressure drop lifts velocity by about 41 percent, not 100. Specific gravity enters the same expression because it changes how much pressure a given mass flow needs to accelerate through the restriction, so it has to sit inside that same radical to correctly rescale the rating for fluids denser or lighter than water.
Is Cv the same thing as Kv, the metric flow coefficient?
No, though they describe the same physical property. Kv is defined in cubic metres per hour of water at a 1 bar pressure drop, the convention most European and Asian manufacturers publish. The two convert as Cv is about 1.156 times Kv, or Kv about 0.865 times Cv. Reading a Kv figure off a foreign datasheet as though it were Cv understates true capacity by roughly 15 percent, a costly mistake when specifying imported valves.
Does this formula work for steam or compressed gas service?
Not directly. This is the liquid form of the sizing equation, valid for incompressible, non-flashing flow. Gas and steam expand as they cross the valve, so their version of Cv needs an added expansion factor and is worked in mass or standard volumetric flow rather than a simple gallons-per-minute figure. Sizing standards publish the compressible-flow form separately; feeding steam data into this liquid formula gives an untrustworthy answer.
What if I use the pressure drop across the whole pipeline instead of just the valve?
The result comes out wrong, almost always too small a valve. Pressure drop here is strictly what the valve itself consumes — friction in straight pipe, elbows, and fittings belongs to the rest of the system's pressure budget. Feeding in the total line loss inflates the pressure drop term, understates the required Cv, and leaves a valve that throttles too aggressively once the rest of the system's own losses are accounted for.
Why might two valves rated at the same wide-open Cv still behave differently in service?
Because a wide-open Cv only fixes the endpoint of a curve, not its shape. How Cv rises from zero at the closed position to that rated maximum, the inherent flow characteristic, differs by trim: linear trim raises Cv in step with travel, equal-percentage trim raises it slowly then steeply, and quick-opening trim delivers most of its capacity in the first part of the stroke. Matching the wide-open number without matching that characteristic can still leave a control loop hunting.