Views: 0 Author: J-VALVES Publish Time: 2026-10-26 Origin: Site
In almost every enquiry that lands on a valve desk, the same two candidates are on the table: a butterfly valve and a gate valve. Both isolate flow, both are quarter-century-old designs, and both are quoted in the same pressure classes. That is where the similarity ends. A gate valve is a full-bore, multi-turn, metal-seated wedge that stops flow with almost no added resistance. A butterfly valve is a short, quarter-turn disc that sits permanently in the flow path and stops flow by pressing an elastomer or a metal seat against the disc edge. Choosing between them changes the pump head, the weight hung on the pipe, the torque the actuator must develop, the energy bill and the maintenance plan for the next thirty years.
This article is deliberately quantitative. It works through the flow capacity numbers, then the volumetric cost of the residual pressure drop, then the structural weight and space savings, then the torque and actuator consequences, and finally the failure modes that make one valve wrong for a duty even when the numbers look favourable. The aim is a set of rules a project engineer can apply in a design review, not a list of adjectives.
A gate valve operates by linear motion. The stem lifts a wedge or slab out of the flow path, and at full open the bore is essentially a straight pipe with a small seat recess. Closure is multi-turn — typically 12 to 20 handwheel turns on a DN150 valve, more on large sizes — and sealing is metal-to-metal on two inclined faces, driven by the wedging action described in our guide to solid, flexible and split wedges.
A butterfly valve operates by rotation. The disc turns 90 degrees inside a short body, and sealing is achieved by interference between the disc edge and a resilient seat (concentric design) or by a metal seat contacted only in the final few degrees of rotation (double- and triple-offset designs). Closure is quarter-turn, a few seconds by hand or actuator, and the seat is replaceable on most body styles.
Those two mechanisms produce all the downstream differences: flow capacity, face-to-face length, weight, torque, sealing tightness, and what happens when the valve is left part-open for a year. For the wider product families behind both designs, see the butterfly valve range and the gate valve range on the same site.
The honest starting point is the valve loss coefficient zeta, defined by Δp = zeta · ρv²/2, using the mean pipe velocity. Published ranges for fully open valves, based on Crane TP-410 and manufacturer catalogues, are tight enough to design with:
Valve type (fully open) | Typical zeta | Equivalent length | Behaviour |
|---|---|---|---|
Gate valve | 0.10 – 0.20 | 5 – 10 D | Full bore, straight-through; loss is almost entirely seat recess and body transition |
Ball valve, full port | 0.05 – 0.10 | 3 – 5 D | Lowest of any isolation valve |
Butterfly valve, DN50–DN200 | 0.25 – 0.50 | 12 – 20 D | Disc thickness and shaft dominate; small sizes are proportionally worse |
Butterfly valve, DN250–DN600 | 0.15 – 0.35 | 8 – 12 D | Improves with size because disc thickness scales less than diameter |
Double/triple-offset butterfly | 0.15 – 0.30 | 8 – 12 D | Slimmer, profiled disc; metal seat adds little at full open |
Globe valve | 4 – 10 | 300 – 500 D | Included only for scale |
Converted to flow coefficient on a DN200 (8 in) line at Class 150, a full-port gate valve sits near Cv 5,100, while a typical high-performance butterfly disc at 90 degrees sits near Cv 2,200 — roughly 43 percent of the gate valve's capacity for the same bore. Two useful rules follow:
At equal size, a butterfly valve costs you about 0.1 to 0.3 bar of head at 2–3 m/s water velocity. That is small in absolute terms and large in annual pumping energy, which is quantified below.
A butterfly valve is not "one size fits all" for capacity. Because the disc occupies the bore, the published Cv at 90 degrees varies by more than 30 percent between a thin profiled disc and a thick rubber-lined one. A data sheet that gives only the pressure class, not the Cv at full open, is not enough to design with — the Cv and Kv sizing calculation method converts the required flow and allowable drop into the coefficient you should be asking for.
This is the calculation that decides most large-diameter water projects, and it is rarely done. The hydraulic power lost in a fully open valve is the product of volumetric flow and pressure loss; the electrical draw is that figure divided by the pump, motor and drive efficiency:
Pshaft = Q · Δp / (ηpump · ηmotor) (Q in m³/s, Δp in Pa, result in W)
Take a DN500 main at 1,500 m³/h (0.417 m³/s) of water, velocity 2.1 m/s, 4,000 operating hours per year, electricity at USD 0.12/kWh, and a combined pump-and-motor efficiency of 0.72. The velocity head is 2,255 Pa, so a ζ of 0.15 gives 338 Pa (about 3.4 mbar) and a ζ of 0.30 gives 677 Pa (6.8 mbar).
Item | Gate valve, ζ = 0.15 | Butterfly valve, ζ = 0.30 | Difference |
|---|---|---|---|
Velocity head ρv²/2 (Pa) | 2,255 | 2,255 | — |
Valve loss at full open (mbar) | 3.4 | 6.8 | 3.4 |
Hydraulic power lost (kW) | 0.14 | 0.28 | 0.14 |
Shaft power at the motor (kW, η = 0.72) | 0.20 | 0.39 | 0.19 |
Annual energy (kWh) | 780 | 1,570 | 790 |
Annual cost (USD) | 94 | 188 | 94 |
20-year cost at 4% discount (USD) | 1,280 | 2,570 | 1,280 |
The honest conclusion is that the energy difference on a single mid-size valve is small — a few hundred dollars a year, roughly USD 1,300 over twenty years — and it will not on its own justify a gate valve that costs several times as much. Two things change the answer. First, valve loss scales with the square of velocity, so raising the same DN500 main from 2.1 m/s to 3.0 m/s multiplies every figure in the table by about two, and running two or three parallel trains multiplies it again by their number. Second, the penalty scales with size: a ζ difference of 0.15 is worth very little on a DN150 line and a good deal on a DN1000 main running fast. Run the arithmetic on the real duty — and note that it disappears entirely on a bypass, a drain, a tank inlet used a few hours a week, or a fire main that is shut for years: there the butterfly valve's cost, mass and speed of operation win outright.
One caution on the gate valve side: the low ζ assumes a full-port, straight-through body. Reduced-port and some forged patterns have significantly higher losses, and a gate valve that is habitually left 70 percent open to "balance" the line loses more head than a butterfly valve at full open. Use the installed position, not the catalogue position.
The difference in mass is not incremental; it changes the structure. Indicative figures for Class 150 cast steel bodies:
Size | Gate valve, face-to-face and mass | Wafer butterfly, face-to-face and mass | What it changes |
|---|---|---|---|
DN150 (6 in) | ~280 mm, 60–90 kg | ~56 mm, 12–18 kg | One-person handling, no crane |
DN300 (12 in) | ~400–500 mm, 200–300 kg | ~78 mm, 35–55 kg | Support bracket versus pipe hanger redesign |
DN600 (24 in) | ~700–900 mm, 900–1,400 kg | ~130 mm, 150–220 kg | Concrete pier versus flange bolt-up |
DN1200 (48 in) | > 3,000 mm face-to-face, 4,000–8,000 kg | ~300 mm, 700–1,200 kg | Gate valve is often not buildable; butterfly is the default |
Three engineering consequences follow, and each has a number attached:
Support steel. A 1,200 kg valve on a DN600 line adds a concentrated dead load that the pipe stress model must carry, often forcing a spring hanger or a pier. A 200 kg valve usually does not. On skid-mounted packages the effect is worse: skid steel is priced by the kilogram, and the valve drives the frame.
Face-to-face length. A gate valve of large size consumes 1.5–3 m of straight pipe that the plot plan may not have. In retrofit and tie-in work this alone can eliminate the gate valve option.
Shipping and lifting. Above DN400 the logistics difference is comparable to the valve price, and the civil works to land a heavy valve can exceed it.
The flip side: gate valves resist pipe loads better. A butterfly valve body is a short ring clamped between flanges; it has little bending stiffness, so flange misalignment, thermal movement and unsupported spans transfer directly into the seat. Gate valves with long bodies tolerate far more. If the piping is not aligned to within the manufacturer's tolerance, a butterfly valve will leak and a gate valve may not. Specify alignment, and use lug or double-flanged bodies where the valve has to carry line load.
Actuator cost is where butterfly-versus-gate comparisons are most often got wrong, in both directions.
A gate valve needs a large thrust over a long stroke. A DN300 Class 150 gate valve typically needs 30–60 kN of stem thrust and perhaps 120–200 N·m at the handwheel, with 15–25 turns for full travel; a gearbox and a thrust-rated actuator follow, and a full-stroke actuation time of 30–90 seconds is normal. A butterfly valve needs perhaps 400–700 N·m on the same bore, but for only 90 degrees of travel, so the actuator is smaller, cheaper and strokes in a few seconds — which is exactly what emergency shut-off and frequent cycling need.
The trap is that butterfly torque is not a single number:
Seating and unseating torque peak at 0 degrees, driven by the interference fit of the disc edge in the seat. This is the number most catalogues quote.
Bearing and packing friction is roughly constant and rises with age, dirt and inadequate lubrication.
Hydrodynamic (dynamic) torque comes from the flowing fluid acting on the disc. It peaks not at full open but typically between 60 and 80 degrees of opening, and it acts to push the disc closed.
An actuator sized only on catalogue seating torque will stall mid-stroke on a high-velocity line, or slam the disc shut when the torque reverses. Size against the worst combination at the maximum differential pressure the valve will actually see — not line pressure, and not the shut-off case alone — then apply a safety factor: about 1.25 to 1.5 for on/off duty per AWWA C504 practice, and higher (1.5 to 2.0) for dead-end service, modulating duty or dirty media where friction drifts upward. The AWWA C504 field relationship T = K · D³ · Δp is a useful cross-check, but K must come from the valve maker's torque curve, not from a generic table.
Gate valve actuation has the opposite profile: high thrust, low torque, long stroke, and a much larger margin between running and breakaway loads. Breakaway thrust after long idle periods — the condition the industry calls stiction or cold welding between wedge and seat — can exceed the design seating thrust, which is why gate valve actuators are often specified on breakaway rather than running torque. Our pneumatic actuator sizing and fail-safe guide covers the thrust-to-torque workflow for both types in detail.
The common claim that "butterfly valves seal better than gate valves" is true only for resilient-seated butterfly valves in clean, low-temperature service. The actual comparison depends on the seat system, not the valve type.
Design | Typical achievable tightness | Limits |
|---|---|---|
Resilient-seated (concentric) butterfly, EPDM/NBR | Bubble-tight at test; ISO 5208 Rate A / FCI 70-2 Class VI equivalent | Temperature-limited (-20 to +120 °C for EPDM, lower for NBR); elastomer swells in hydrocarbons; seat damage from debris |
Double-offset butterfly, PTFE/RPTFE seat | Rate A to ISO 5208 in clean service, bidirectional | Seat extrusion at high differential; fire-safe requires secondary metal seat |
Triple-offset metal-seated butterfly | Rate A achievable, bidirectional, to 425 °C class dependent | Higher torque; needs correct installation direction and clean seat faces; cost approaches a gate valve |
Metal-seated gate valve | API 598 acceptance, not bubble-tight by definition | Leakage permitted is a function of size (drops or bubbles); lapped faces must be protected from debris |
Resilient-wedge gate valve | Bubble-tight, buried water distribution standard | Rubber temperature limits; not for hydrocarbons or steam |
Two practical points. First, a metal-seated gate valve is allowed to leak a small amount under API 598; rejecting it on that basis is a specification error, and the acceptance criteria are set out in our API 598 pressure testing guide. Second, a bubble-tight butterfly seat is only bubble-tight while the elastomer is intact: a single weld bead or a piece of mill scale passing through the valve can nick the seat and turn a Class VI valve into a steady drip. Strainers upstream of resilient-seated butterfly valves are cheap insurance on new lines.
Excluding installation error, the two designs fail in almost completely different ways, which makes the maintenance strategy quite different.
Failure | Butterfly valve | Gate valve |
|---|---|---|
Seat degradation | Elastomer swelling, compression set, tearing at the disc edge; the dominant wear mode | Lapped face erosion, wire drawing from a persistent leak, scratching from debris |
Throttling damage | Disc edge and seat erode; cavitation damage downstream of the disc at small openings | Rapid seat and wedge scoring; gate valve is not a throttling valve at all |
Long idle periods | Generally benign; elastomer takes compression set but recovers on cycling | Stiction between wedge and seat; breakaway thrust can exceed actuator capacity |
Solids and slurry | Disc edge and seat damaged by particles trapped at closure | Wedge pocket and body cavity trap debris; seat faces score |
Stem / shaft failure | Sheared shaft from oversized actuator or a jammed disc | Bent stem from cheater bars or thermal binding |
Repairability in line | Seat replaceable on most bodies; disc and shaft usually workshop-only | Seat rings and wedge renewable; lapping usually requires removal |
Body and pressure boundary | Wafer body has no flange stiffness; gasket joint is the weak point | Robust flanged body tolerates line loads |
There is one additional gate valve failure that deserves its own line because it is a design decision rather than a defect: the wedge can only be trusted to seal if the valve is cycled occasionally. A gate valve held closed for two years at temperature can require breakaway thrust far beyond its design value. If a plant's isolation philosophy is "install it and forget it", a butterfly valve is the more predictable asset.
The envelope determines whether the comparison is even available.
Parameter | Gate valve | Butterfly valve |
|---|---|---|
Pressure class | Class 150 to 2500 (and 4500 in special designs) | Class 150 standard; Class 300 common for high-performance; Class 600 available in reinforced triple-offset |
Temperature | -196 °C (cryogenic) to 650 °C with appropriate trim | -20 to +120 °C resilient-seated; up to about 200 °C with FKM; 425 °C and beyond with triple-offset metal seat |
Size | Economical to DN600–DN900; above DN1200 the mass becomes prohibitive | DN50 to DN3000; the natural choice above DN600 |
Media | Any, including slurry, solids, two-phase and pigging duty | Clean liquids and gases; abrasive and fibrous media attack the disc edge |
Above Class 300 and above roughly 400 °C, the comparison mostly disappears and the gate valve wins on availability alone. Below DN600 in clean water at moderate temperature, the butterfly valve wins on cost, mass and speed of operation. The contested middle — DN300 to DN800, Class 150/300, water and non-aggressive liquids — is where the energy calculation above decides the answer.
Use these in sequence; each one can close the question.
Can the valve be built at all? Pressure class above 600, temperature above 400 °C, or size above DN1200 with high pressure: gate valve or a triple-offset butterfly with a verified class. Cryogenic service: both, with the appropriate extended bonnet.
Is the medium clean? Fibrous, abrasive, slurry, or containing solids that can bridge: gate valve, or better a knife gate, pinch or full-bore ball valve. A butterfly disc edge is a debris trap that damages the seat.
Must the line be pigged, swabbed or rodded? Gate valve (slab or through-conduit) or full-bore ball. A butterfly valve cannot pass a pig.
Is the valve in continuous flow duty? If yes and the line is large or runs at high velocity, run the energy calculation above on the real flow, hours and electricity price. Where the discounted energy penalty of the disc approaches or exceeds the purchase-price saving of the butterfly valve, choose the gate valve — or specify a larger butterfly valve to bring the velocity, and therefore the loss, back down.
Is space or weight constrained? Skids, retrofit tie-ins, elevated pipe racks, offshore modules and plant rooms: butterfly valve, using lug or double-flanged bodies where line load or dead-end duty applies.
Is fast or frequent operation required? Gas isolation, emergency shut-off, automated isolation valves on a batch cycle: butterfly valve, because quarter-turn actuation is cheaper, faster and easier to automate, and because it does not suffer stiction.
Does the duty require bubble-tight shut-off at process temperature? Resilient-seated butterfly for water to about 120 °C; triple-offset metal-seated butterfly for higher temperature with verified Rate A; otherwise a gate valve with a recognised leakage acceptance level.
Two supporting rules cut across the list. Neither valve type should be used to throttle continuously: the butterfly valve erodes its seat and cavitates near closure, and the gate valve scores its faces. If a line needs regulation, put a control valve in it and use an isolation valve of either type for isolation. And where a large butterfly valve must be installed close to a pump discharge or an elbow, check with the maker: asymmetric flow raises dynamic torque and can overload an actuator sized on catalogue data.
Most specification failures are omissions. For whichever type is selected, the enquiry should state:
Standard and design basis: API 600 / API 603 / ASME B16.34 for gate valves; API 609 or EN 593 for butterfly valves, with the body style (wafer, lug, double-flanged) named.
Size, pressure class, face-to-face dimension and the flange standard and facing.
Body, disc/wedge, stem and seat materials, with hardness where NACE or erosive service applies.
Seat type and certified leakage class, referenced to API 598 or ISO 5208, with the test medium.
Full-open Cv or zeta, and the Cv curve if the valve may be used part-open.
Break-to-open, running and seating torque (butterfly) or seating and breakaway thrust (gate), at the maximum differential pressure, cold and hot.
Temperature range with the seat or packing material that achieves it, and the fire-safe requirement (API 607 or API 6FA).
Actuation interface (ISO 5211 flange size), fail position, and required stroking time.
Coating or lining system and its certification for potable water, plus any fugitive-emission class for the stem seal.
At J-VALVES both families are quoted with the flow coefficient, the torque or thrust at the specified differential pressure, and the leakage class stated on the certificate rather than left to the buyer to infer. A comparison between butterfly and gate valve is only meaningful when both quotations carry those numbers.
Which has lower pressure drop, a butterfly valve or a gate valve?
A gate valve. A full-port gate valve at full open has a loss coefficient of about 0.10 to 0.20, close to a straight pipe, while a butterfly valve at full open is about 0.25 to 0.50 in small sizes and 0.15 to 0.35 above DN250, because the disc and shaft stay in the flow path. On a DN200 line that typically means a fully open gate valve passes about twice the flow of a butterfly valve at the same pressure drop. The gap narrows with size but never closes.
At what size does a butterfly valve become clearly cheaper than a gate valve?
The crossover is usually around DN250 to DN300. Below that, the two are close in price and the gate valve's lower pressure drop and robust body often win. Above DN300 the butterfly valve is lighter, shorter and cheaper by a widening margin, and above about DN600 the gate valve becomes difficult to justify on mass and space alone. The exception is continuous high-flow duty, where the pumping energy penalty of the disc can outweigh the purchase saving within a few years.
Can a butterfly valve be used for throttling and flow control?
Only within limits. A butterfly valve has usable control characteristics roughly between 30 and 70 degrees open, but below about 20 degrees the velocity across the seat rises sharply, causing cavitation, noise and seat erosion, and the disc edge wears quickly in continuous throttling. For genuine regulation, use a globe or dedicated control valve, and keep the butterfly valve for isolation or for occasional trimming. A gate valve should never be used for throttling at all: partial opening wire-draws the seats and destroys the sealing faces.
Why does my butterfly valve actuator stall or let the disc slam shut mid-stroke?
Almost always because the actuator was sized on catalogue seating torque only. Total butterfly torque is seating torque plus bearing and packing friction plus hydrodynamic torque, and hydrodynamic torque peaks between roughly 60 and 80 degrees of opening, not at the closed position. On a high-velocity line it can exceed the seating torque and it reverses direction past mid-stroke. Size the actuator against the worst-case torque at the maximum actual differential pressure, then apply a factor of 1.25 to 1.5 for on/off duty and 1.5 to 2.0 for modulating, dead-end or dirty service.
Is a gate valve or a butterfly valve better for a line that stays closed for years?
A butterfly valve, in most cases. A gate valve that sits closed at pressure for a long period develops stiction between the wedge and the seat rings, and the breakaway thrust needed to open it can exceed the original seating thrust, leading to bent stems or stripped yoke bushings when operators use extension bars. A resilient-seated butterfly valve takes some compression set but generally still opens, and its quarter-turn actuation gives operators a much better chance of freeing it. Whichever type is installed on a long-idle line, specify annual partial cycling as part of the maintenance plan.
