What velocity in a pipe actually means
Velocity is flow rate divided by flow area. Nothing more. But because area goes as the square of the bore, velocity is exquisitely sensitive to pipe size: the same 10 gpm moves at 4.1 ft/s in a 1-inch bore, at 16.3 ft/s in a half-inch bore, and at 1.0 ft/s in a 2-inch bore. One trade size makes the difference between a quiet system and one you can hear from the next room.
The number this calculator returns is the mean velocity — total flow divided by total area. Real velocity varies across the bore, from zero at the wall to a peak on the centreline that is roughly 1.2 times the mean in turbulent flow. Every design rule, every erosion limit and every published table is written in terms of the mean, so that is what you compare.
Three separate failure modes are all governed by this one number. Erosion-corrosion strips the protective oxide film off the inside of copper tube, preferentially on the outside of elbows and just downstream of poorly reamed cuts, and it accelerates sharply above the material's limit. Noise comes from turbulence at fittings and rises steeply with velocity. Surge — water hammer — is proportional to the velocity that a closing valve brings to a stop, so a 10 ft/s line generates twice the pressure spike of a 5 ft/s one for the same valve.
Where 0.4085 comes from
Velocity is Q/A in consistent units, and the constant simply collects the conversions so you can work in gallons per minute and inches.
Start with the flow: one US gallon per minute is 231 in³/min, which is 231/1,728 = 0.13368 ft³/min, or 0.0022280 ft³/s. Divide by 448.831 to go from gpm to ft³/s. Then the area of a bore of d inches is π/4 × (d/12)² = 0.0054542d² ft². Dividing gives V = (Q/448.831) ÷ (0.0054542d²) = 0.40850 Q/d².
Turn the equation around and you get the more useful design form: the maximum flow a given pipe can carry at a chosen velocity is Q = V·d²/0.4085. For 8 ft/s that is 19.6d² gpm — a half-inch Type L copper bore of 0.545 in carries 5.8 gpm, a 3/4-inch bore of 0.785 in carries 12.1 gpm, and a 1-inch bore of 1.025 in carries 20.6 gpm. Those three numbers alone will settle most residential branch-sizing arguments.
The velocity head, V²/2g, is worth reading too. It is the kinetic energy of the stream expressed as a height of water, and it is the unit in which fitting losses are catalogued. At 4 ft/s it is 0.249 ft, so an elbow with a loss coefficient of 0.9 costs about 0.22 ft of head — and at 8 ft/s the same elbow costs four times that.
Worked example: is 3/4-inch copper enough for two showers?
A branch feeds two showers, each drawing 2.5 gpm, plus a lavatory at 1.5 gpm, and you expect all three at once in the morning. That is 6.5 gpm through nominal 3/4-inch Type L copper, bore 0.785 in, carrying 120 °F water.
- Flow area. π/4 × 0.785² = 0.785398 × 0.616225 = 0.48398 in².
- Velocity. V = 0.4085 × 6.5 ÷ 0.785² = 2.65525 ÷ 0.616225 = 4.309 ft/s.
- Metric. 4.309 × 0.3048 = 1.313 m/s.
- Against the limit. Hot water at 120 °F is below 140 °F, so the applicable copper limit is 5 ft/s. 4.309 ÷ 5 = 86% of the limit. It passes, with 14% margin.
- Maximum flow at the limit. Q = 5 × 0.616225 ÷ 0.4085 = 7.54 gpm. So this branch has 1.04 gpm of headroom before it reaches 5 ft/s.
- Velocity head. 4.309² ÷ 64.348 = 18.567 ÷ 64.348 = 0.2885 ft.
Now change one assumption. Suppose the recirculation loop pushes the branch to 140 °F or above, so the limit drops to 3 ft/s. The same 4.309 ft/s is now 144% of the limit, and the maximum flow falls to 3 × 0.616225 ÷ 0.4085 = 4.53 gpm, well under the 6.5 gpm demand. The pipe did not change and the flow did not change; the temperature changed the answer from pass to fail. That is why the calculator asks for service temperature rather than just material.
Maximum flow in Type L copper at common velocity limits
| Nominal size | Bore (in) | gpm at 3 ft/s | gpm at 5 ft/s | gpm at 8 ft/s | gpm at 10 ft/s |
|---|---|---|---|---|---|
| 1/2 in | 0.545 | 2.18 | 3.64 | 5.82 | 7.27 |
| 3/4 in | 0.785 | 4.53 | 7.54 | 12.07 | 15.09 |
| 1 in | 1.025 | 7.71 | 12.86 | 20.57 | 25.72 |
| 1-1/4 in | 1.265 | 11.75 | 19.59 | 31.34 | 39.18 |
| 1-1/2 in | 1.505 | 16.63 | 27.72 | 44.35 | 55.44 |
| 2 in | 1.985 | 28.94 | 48.23 | 77.17 | 96.46 |
| 2-1/2 in | 2.465 | 44.62 | 74.37 | 119.00 | 148.74 |
| 3 in | 2.945 | 63.69 | 106.16 | 169.85 | 212.31 |
Bores are ASTM B88 Type L. The 8 ft/s column is the usual cold-water design limit, 5 ft/s applies to hot water up to 140 °F and 3 ft/s above it, per Copper Development Association guidance.
Why hot water gets a lower limit than cold
Erosion-corrosion in copper is the mechanical removal of the protective cuprous oxide film by the flowing water, and the film re-forms more slowly and less coherently as temperature rises. Above about 140 °F the Copper Development Association recommends dropping to 2–3 ft/s, against 5 ft/s for hot water below that threshold and 8 ft/s for cold. Recirculating hot-water loops are the classic casualty: they run continuously at elevated temperature, so a velocity that would be unremarkable in a cold branch pits the return line within a few years.
Note that these are design recommendations from the material's trade association, not code requirements. Your local plumbing code may impose its own limit, and the manufacturer's installation instructions for PEX, CPVC or PVC govern for those materials.
Getting the inputs right
- Use peak simultaneous flow, not total fixture flow. Ten fixtures do not all run at once. Convert a fixture schedule to a probable peak with the water supply fixture unit calculator before you size anything.
- Use the bore, not the trade size. Nominal 3/4-inch copper Type L is 0.785 in, Type M is 0.811 in, Schedule 40 steel is 0.824 in and PEX is close to 0.671 in. Because velocity goes as d⁻², copper and PEX at the same nominal size differ by 37% in velocity at the same flow.
- Check the hottest service the pipe will see. A branch that is cold today may be a recirculation return after a remodel, and the limit halves.
- Do not treat the limit as a target. Designing every branch to exactly 8 ft/s leaves nothing for a flow estimate that turns out low, for scale build-up, or for a future fixture.
- Watch the low end too. Below about 2 ft/s a horizontal run will not carry sediment or purge air reliably, which matters for flushing and commissioning even though it damages nothing.
- Check friction separately. Passing the velocity limit does not mean the pressure drop is acceptable. Run the same pipe through the Hazen-Williams calculator or the Darcy-Weisbach calculator.
Velocity, water hammer and the rest of the sizing job
The surge pressure a stopped column of water generates is given by the Joukowsky equation, Δp = ρ·a·ΔV, where a is the pressure wave speed in the pipe — roughly 4,000 ft/s in copper and 1,300 to 1,600 ft/s in PEX and PVC, which are far more elastic. In copper, that works out to about 55 psi of surge for every 1 ft/s of velocity abruptly stopped. An 8 ft/s branch closed by a quick-acting solenoid valve therefore sees a spike of several hundred psi on top of the working pressure. This is the real reason velocity limits exist in fast-acting appliance branches, and why arrestors are specified at washing machines and dishwashers.
Velocity is one of two independent checks on a pipe size. The other is friction loss against your available pressure, which the water service pipe size calculator works through segment by segment. A pipe must pass both. On short runs with generous pressure the velocity limit usually governs; on long runs the friction budget usually governs; sizing to only one of them is how undersized branches get installed in buildings that have plenty of street pressure.
None of this applies to gravity drainage, where the pipe is not full and there is no pressure to spend. There the relevant velocity check is the opposite one — a minimum of about 2 ft/s to keep solids moving — and it is computed from slope with the Manning's equation calculator and set by the pipe slope calculator.
Why a line that passed on day one can fail this check later
The velocity this calculator reports is only ever as good as the bore you enter, and the bore of an installed pipe is not fixed for its service life. Scale, corrosion products and biofilm all reduce the effective inside diameter over years of service, and because velocity is inversely proportional to the square of the diameter, a small reduction in bore produces a disproportionately larger rise in velocity at the same flow rate.
A pipe that scales down by 10% in diameter carries the same flow at roughly 1 ÷ 0.9² = 1.23 times its original velocity — a 23% increase from a 10% narrowing. A line commissioned at 90% of its erosion limit has no margin left to absorb that; a line commissioned at 70% of the limit does. This is the practical case for the warning this calculator raises inside 10% of the limit: it is not only a caution about the accuracy of today's flow estimate, it is margin against a bore that will not stay at today's value.
Two materials handle this differently. Copper's own erosion-corrosion mechanism removes material rather than depositing it, so its bore can grow slightly over time instead of shrinking, right up to the point where the protective oxide film stops re-forming and pitting begins — which is the mechanism the velocity limit exists to prevent in the first place. Hard-water scaling, by contrast, narrows the bore in any material and is a separate, water-chemistry-driven process that the velocity limit does not address at all; that risk is managed by water treatment, not by pipe sizing.