Pipe sizing is a pressure budget, not a flow chart
The question people ask is "what size water line do I need?" The question the calculation answers is "how much pressure do I have left, and which pipe fits inside it?" Those are the same question, but framing it the second way makes the answer obvious and shows you exactly which assumption to change when the answer is unwelcome.
You start with the static pressure at the source. Four things take a bite before any water moves through pipe. The fixture at the end needs a minimum flowing pressure, commonly 15 psi for a tank-type fixture and 25 psi for a flushometer valve. Elevation takes 0.433 psi for every foot the fixture sits above the source, which the psi to feet of head converter makes explicit. The water meter takes its share, read from the manufacturer's curve at your peak flow. And every softener, filter and backflow preventer takes its own.
What survives all four deductions is the friction budget. Divide it by the equivalent length of the run and you have an allowable loss per 100 feet — the single number that selects the pipe. Every published friction table is written in those units precisely so this comparison is a lookup.
Then there is a second, independent test. Velocity has to stay under the limit for the material, regardless of what the friction budget allows. A short run with generous pressure will often pass the friction test at a bore that runs at 15 ft/s, and that pipe will erode and howl. Both tests must pass, and the calculator reports which one fails on every size.
Getting each input right
Peak demand is the probable simultaneous flow, not the sum of the fixtures. A house with 25 gpm of installed fixture capacity might peak at 12 gpm, because nobody runs everything at once. Convert a fixture schedule properly with the water supply fixture unit calculator, which applies the Hunter probability curve.
Static pressure should be the utility's guaranteed minimum, not the reading you took on a Tuesday afternoon. Street pressure swings seasonally and drops during peak-hour demand on the main. Sizing on a summer-morning high reading is how a system that works at commissioning fails in August.
Developed length is measured along the pipe, following every rise, drop and offset, from the source to the most remote fixture — the one with the worst combination of distance and elevation, which is not always the farthest one horizontally.
The fittings allowance converts elbows, tees and valves into extra pipe. Adding 50% of the developed length is the standard first-pass assumption and is usually conservative on a straight service run and optimistic in a fitting-dense mechanical room. If you have the equivalent-length table for your fittings, count them and use the exact figure instead.
Meter loss is frequently the largest single deduction and the one most often left out. A 3/4-inch displacement meter at 20 gpm can take 10 psi or more. Read it from the manufacturer's curve at your actual peak flow, because the loss rises roughly with the square of the flow.
Worked example: a two-storey house on 65 psi street pressure
A house draws a peak of 11 gpm. Street pressure is 65 psi. The highest fixture is 20 ft above the meter. Developed length to it is 120 ft. There is a 3/4-inch meter costing 8 psi at that flow, no treatment devices, Type L copper, and the fixture needs 15 psi residual.
- Elevation cost. 20 ft × 0.4331 psi/ft = 8.66 psi.
- Budget. 65 − 15 − 8.66 − 8 − 0 = 33.34 psi available for friction.
- Equivalent length. 120 ft × 1.50 = 180 ft.
- Allowable loss. 33.34 ÷ 180 × 100 = 18.52 psi per 100 ft.
- Test 1/2-inch (0.545 in bore). Hazen-Williams at C = 140: 4.52 × 111.852 ÷ (1401.852 × 0.5454.87) = 4.52 × 84.85 ÷ (9,433 × 0.05203) = 383.5 ÷ 490.8 = 0.7813 psi/ft, or 78.1 psi per 100 ft. That is four times the budget. Velocity is 0.4085 × 11 ÷ 0.545² = 15.1 ft/s, nearly twice the limit. It fails both tests.
- Test 3/4-inch (0.785 in bore). 0.7854.87 = 0.30764, so the loss is 383.5 ÷ (9,433 × 0.30764) = 383.5 ÷ 2,902 = 0.13215 psi/ft, or 13.21 psi per 100 ft — inside the 18.52 allowance. Velocity is 0.4085 × 11 ÷ 0.616 = 7.29 ft/s, under the 8 ft/s limit. It passes both.
- Check the delivered result. Total friction is 0.13215 × 180 = 23.79 psi. Residual at the fixture is 65 − 8.66 − 8 − 23.79 = 24.6 psi, comfortably above the 15 psi required.
So a 3/4-inch service is correct — but with only 9.55 psi of the 33.34 psi friction budget left unspent, and 7.29 ft/s already at 91% of the velocity limit. If the peak demand estimate is low by 2 gpm, that margin disappears. This is the arithmetic that makes 1-inch services so common on new construction.
Typical residual pressure requirements
| Fixture or device | Typical minimum residual (psi) | Typical flow (gpm) |
|---|---|---|
| Lavatory faucet | 8 | 1.5–2.0 |
| Kitchen sink faucet | 8 | 1.8–2.2 |
| Shower head | 8 | 1.8–2.5 |
| Bathtub filler | 8 | 4.0 |
| Water closet, flush tank | 15 | 3.0 |
| Water closet, flushometer valve | 25 | 25–35 |
| Urinal, flushometer valve | 15 | 15 |
| Hose bibb | 8 | 5.0 |
| Dishwasher | 8 | 2.75 |
| Clothes washer | 8 | 4.0 |
Size the whole system on the most demanding branch, which is normally a flushometer if one is present — its 25 psi residual and high instantaneous flow set the budget for everything upstream of it.
Above 80 psi you must reduce, and that changes the sizing
Plumbing codes in the IPC tradition cap static pressure in a building water distribution system at 80 psi and require a pressure-reducing valve where the supply exceeds it. That matters here because the budget must then start from the reduced pressure, not the street pressure. A 120 psi street reduced to 60 psi gives you a 60 psi budget, and the 60 psi you gave away is not recoverable by pipe sizing.
The reverse trap catches well systems. A pressure tank cycles between a cut-in and a cut-out — 40/60 psi is a common setting — and the system must work at cut-in, the lowest pressure it ever sees. Size on 40 psi, not on 60.
Assumptions and limits of this method
- It sizes one segment, not a whole system. Real distribution piping steps down: the service is sized for full demand, a branch feeding two fixtures for far less. Run the calculator once per segment with that segment's own demand and length.
- It uses Hazen-Williams. That is the standard method in plumbing sizing tables, but it is a water-only empirical fit — see the Hazen-Williams calculator for its range of validity, and Darcy-Weisbach for anything else.
- The fittings allowance is an estimate. A percentage of length is a stand-in for counting fittings. On a complex layout, count them properly.
- It does not check the code's own tables. Some jurisdictions size by a prescriptive table rather than by calculation, and where they do, the table governs even if your calculation says a smaller pipe would work.
- It assumes the peak demand you enter is right. Everything downstream of that number inherits its error, and demand is the input people estimate most loosely.
- It does not size the meter. Meter sizing has its own rules and its own loss curve, and a meter one size up often buys back more pressure than a pipe one size up.
What to do when nothing passes
When the budget goes negative, no pipe size can help — the pressure simply is not there. The levers, roughly in order of cost, are: reduce the residual requirement if the fixture selection allows it; replace a high-loss meter or treatment device; reduce the developed length by rerouting; and finally add a booster pump. Upsizing pipe does nothing at all in this case, which is why the calculator says so explicitly instead of returning the largest size in the list.
When the budget is positive but every size fails on velocity, the demand is simply too high for the pipe family you selected, and the answer is a larger diameter than the list covers or a second parallel run. Check the velocity figure against the material limits in the pipe water velocity calculator before deciding the limit itself is negotiable — for hot water it drops from 8 ft/s to 5, and that alone will change the answer.
When several sizes pass, take the smallest one but look at the margin. A size that lands at 95% of the friction allowance and 91% of the velocity limit has nothing left for a demand estimate that turns out low, for scale build-up over twenty years, or for the extra bathroom someone adds later. One trade size up is cheap during rough-in and expensive afterwards.
The method's basis, and where a code review can diverge from it
This calculator implements the friction-loss segment method set out in IPC Appendix E: build an explicit pressure budget by subtracting residual, elevation, meter and device losses from static pressure, then divide what is left across the developed length and pick a bore that fits inside both the resulting friction allowance and a velocity limit. That is a calculation route, and it is the one a plan reviewer expects to see justified step by step when a design departs from a prescriptive minimum.
Other codes in the plumbing tradition, including the Uniform Plumbing Code's Appendix A, offer a second route: a prescriptive table that maps a fixture-unit or flow total directly to a pipe size for stated pressure ranges, without requiring the designer to build a friction budget by hand. Both routes ultimately rest on the same starting point — converting a fixture count to a probable simultaneous demand, the step Roy Hunter's 1940 National Bureau of Standards work established and that the water supply fixture unit calculator carries out. This calculator picks up from the gpm figure that step produces.
Where the two routes matter is which one governs on a given job. A calculated result smaller than the prescriptive table's answer is not automatically acceptable; it is acceptable only where the adopted code explicitly permits an engineered or calculated method as an alternative to the table, and the reviewing authority may still ask to see every deduction in the budget rather than accept the final size on its own. Check which route your adopted code and jurisdiction expect before treating the output above as the submittal figure.