A bucket test estimates average flow for one defined collection
Calculate faucet flow as collected US gallons × 60 ÷ elapsed seconds. To make the result useful, collect the entire stream over the same interval you time, identify the control setting and retain the uncertainty in both volume and duration. A household collection does not reproduce a certification test or reveal water pressure.
This article focuses on whether a small difference between two calculated flow values is meaningful. The inputs are constructed examples, not measurements of a faucet. It is separate from choosing an aerator thread or interpreting a pressure-compensating mechanism.
Define the collection before using the formula
EPA's 2012 inspection guidance, printed page 6, describes a ten-second faucet collection using a bucket or flow-measuring bag. That historical procedure supports the collection concept; it is not presented here as the current home-certification protocol. Our uncertainty calculations below are independent arithmetic, not EPA tolerances.
For an ordinary observation, record the faucet and accessory identity, handle setting, volume unit, container marks, elapsed time and concurrent water uses. Use a safely manageable container and safe water temperature; do not remove safety controls or change heater settings for a measurement. Record spills or incomplete capture as invalidating that collection, rather than estimating them away.
Quantify a small timing difference
Suppose exactly 0.25 US gallon is collected. If the interval is ten seconds, average flow is 1.50 GPM. If the true interval was nine seconds, it is about 1.67 GPM; if eleven seconds, about 1.36 GPM. These values describe the same assumed volume with different denominators. They are not three different faucet ratings.
| Assumed collected volume | Elapsed seconds | Calculated GPM, rounded |
|---|---|---|
| 0.25 gal | 9 | 1.67 |
| 0.25 gal | 10 | 1.50 |
| 0.25 gal | 11 | 1.36 |
The interval from about 1.36 to 1.67 GPM follows only from an assumed ±1-second timing bound. It does not establish that a person actually has that timing accuracy. To compare two household readings, first ask whether the apparent difference is smaller than plausible collection and timing effects.
Include volume uncertainty without counting it as a product defect
Now suppose the volume estimate is 0.25 ±0.01 gallon and the interval is 10 ±1 seconds. Treat these as independent allowed bounds, not statistical confidence intervals. The lowest possible ratio in that rectangle is 60 × 0.24 ÷ 11 = approximately 1.31 GPM. The highest is 60 × 0.26 ÷ 9 = approximately 1.73 GPM.
A written result of “1.500 GPM” would hide those input limits. A clearer record is “nominal estimate 1.5 GPM; illustrative input bounds allow approximately 1.31–1.73 GPM.” For a real test, substitute justified bounds for the actual tools and method, or state that the uncertainty was not established.
A longer collection can reduce the proportional effect of a fixed timing bound, if flow and operating conditions remain comparable. With an assumed exact 0.75 gallon collected over 30 ±1 seconds, the ratio spans approximately 1.45–1.55 GPM. A longer interval does not correct a misgraduated container, a spill or changing flow.
Keep repeated readings and conclusions separate
Record each valid collection individually before calculating any pooled value. Note changes in control setting, other running fixtures or serviced parts. Do not discard a reading only because it disagrees with the marking. Investigate whether the conditions or measurement explain the difference before attributing it to the product.
The pressure-compensation guide explains why equal measured flow does not prove a regulating mechanism. The faucet research page explains rating context. A volume-and-time observation can supply an average-flow input for your use estimate; it cannot establish inlet pressure, a full performance curve or WaterSense certification.