How to Measure Flow With a Bucket and a Stopwatch
Why this matters
A flow figure feeds a surprising number of field calls: how long a pool fill will take, whether a showerhead is the one stamped on it, whether a low-flow complaint is the fixture or the supply, whether a hose can keep up with a pressure washer's inlet, whether a filter has started to clog. Most of those calls get made on a remembered number ("a hose is about 10 GPM") or the figure printed on a part, and both can be off by half. A bucket, a mark and a phone stopwatch give a real number in under five minutes, and this card owns that measurement: how to set it up so it is honest, how many trials to take, and how to tell a bad trial from a real reading. Every card that needs a measured flow, including How to Work Out a Fill or Drain Time, points here.
The formula
GPM = gallons in the container x 60 / seconds to fill it.
The units close on paper: gallons divided by seconds is gallons per second, and multiplying by 60 seconds per minute leaves gallons per minute. A 5 gallon container that fills in 40 seconds is 5 x 60 / 40 = 7.5 GPM. For gallons per hour, multiply GPM by 60 again: 7.5 GPM is 450 GPH.
The slip to watch for is inverting it. Seconds / gallons x 60 gives 480 for the same bucket, a number that looks like a flow and is nonsense. A quick sense check catches it: a 5 gallon pail filling in well under a minute is a strong flow, so the answer should be several GPM, not hundreds.
The record
FLOW RECORD
--------------------------------------------------
Outlet: ______________ Valve: fully open? Y / N
Measured at: discharge end of ____ ft hose / fixture
Container: ______ marked at ____ gal (how marked)
Static pressure if gauged: ____ psi
Water: cold / mixed Ran ____ s before timing
Trial Seconds Within 5% of median?
1 _____ ____
2 _____ ____
3 _____ ____
Median seconds: _____
Mean of accepted seconds: _____
GPM = gal x 60 / mean seconds = _____
--------------------------------------------------
The steps
1. Mark the container with a known volume, not its label. A pail sold as 5 gallons usually holds more than that to the brim, and molded marks on the side may be approximate. Pour in a known measure (a 1 gallon jug five times, or a graduated container) with the pail on a level surface, and mark the water line on the inside with a marker or tape. That mark is your gallons. Skip this and every reading carries whatever error the brim has, in the same direction every time, so averaging trials does nothing to remove it.
2. Pick a container that takes at least 20 seconds to fill. Your thumb on the stopwatch is slower than the water. Each start and each stop carries a few tenths of a second of reaction error. On a 40 second fill that is around 1 percent; on an 8 second fill it is around 5 percent, and on a 4 second fill the timing is mostly noise. For a strong flow, use a bigger container (a large trash can marked the same way), not a faster thumb. For a weak flow, a smaller marked container is fine as long as it still takes 20 seconds. Twenty seconds is a shop starting point; longer is always better.
3. Open the valve fully and let it run before timing. Flow through a partly open valve measures the valve, not the system, unless the partial setting is what you are testing. Let water run 10 to 15 seconds first so air clears from the hose and the flow settles. Write "full open" on the record, because the next person to repeat the test needs the same condition to compare against.
4. Measure where the water will actually be used. For a fill, at the end of the real hose. For a fixture, at its outlet with its aerator or showerhead in place. Flow at a bare bib is higher than through a long hose, and a reading taken at the wrong point answers the wrong question.
Hazard at this step: time cold water, or mixed water you have tested by hand away from the full stream. Water from a heater can scald faster than a hand can be pulled away. Keep the bucket and the stream away from receptacles and cords, and on a finished floor put the bucket in a tub, a basin or on a towel, because a missed stop floods the floor in seconds.
5. Use the same start and stop events every trial. Start the clock the instant you swing the empty container fully under the stream; stop it the instant the water surface touches the mark at eye level. Do not start with the container already under the stream, and do not stop on "about there." The point is that every trial has the same two errors in it, so they compare.
6. Take three timings, and test each against the median. Find the middle value of the three. Any trial more than 5 percent from the median gets thrown out and re-run. Then average the accepted times and compute the flow from the average. With equal volumes, averaging the times and converting once is the right way round; averaging three separately computed GPM figures lands very close on readings this tight, but it is the times that were measured. See related: Averaging Readings: Mean, Median and Weighted.
7. Dump the water deliberately. A full 5 gallon pail is about 42 lb (5 x 8.34 = 41.7). Lift it with your legs, close to your body, and pour it where it can drain, not on a customer's lawn bed or down a basement stair.
Worked example 1: a hose bib for a fill
The tech needs a flow for a fill-time estimate and sets up at the end of the 75 ft hose that will do the job. The pail is marked at 5.00 gallons from a jug. Valve fully open, run 15 seconds, cold water.
Trial Seconds Diff from median 39.9
1 40.6 +1.8%
2 39.5 -1.0%
3 39.9 0.0%
Median 39.9 s
Mean accepted (40.6 + 39.5 + 39.9) / 3 = 40.0 s
GPM 5.00 x 60 / 40.0 = 7.50 GPM
Check: per-trial GPM 7.39, 7.59, 7.52
mean of those = 7.50
All three trials sit inside 5 percent of the median (the widest, trial 1, is 0.7 s over, 1.8 percent). The flow is 7.5 GPM. Averaging the three per-trial flows gives 7.50 as well, which confirms nothing odd is hiding in one reading. That 7.5 GPM, written with "75 ft hose, full open," is the number the fill-time card divides into the volume.
What the label would have said. If the tech had used "about 10 GPM for a garden hose" instead, a 12,000 gallon fill would have been planned at 1,200 minutes (20 hours) when the measured flow gives 1,600 minutes (26.7 hours). The plan would have been more than six hours short.
Worked example 2: a showerhead, and a trial that does not belong
A customer says the new showerhead "uses more water than the old one." The head is stamped 1.8 GPM. The tech holds a bucket marked at 2.00 gallons (marked from a jug) under the running shower, mixed to lukewarm and checked by hand first.
Trial Seconds Diff from median 51.4
1 52.1 +1.4%
2 51.4 0.0%
3 44.0 -14.4% REJECT, re-run
3b 51.8 +0.8% (new median 51.8)
Mean accepted (52.1 + 51.4 + 51.8) / 3 = 51.77 s
GPM 2.00 x 60 / 51.77 = 2.32 GPM
Trial 3 was 7.4 seconds short of the median, 14.4 percent. A fill that fast would mean a flow of 2.73 GPM, and a showerhead does not jump about 18 percent between one minute and the next. The tech recognized the cause: on that trial the clock started late, after water was already landing in the bucket, which shortens the time and inflates the flow. A splash that throws water outside the bucket does the opposite, lengthening the time. Either way the remedy is the same: discard and re-run. The re-run at 51.8 seconds sits with the others.
What 2.32 GPM against a 1.8 stamp means. A showerhead's rated flow is stated at a test pressure (federal rules state the showerhead maximum at a test pressure of 80 psi), and a head on a house running below that pressure usually flows at or under its stamp, not over. Reading 2.32 GPM, 29 percent over the 1.8 stamp (2.32 / 1.8 = 1.29), points at the head, not the house: a missing or removed flow restrictor, or a head that is not the one the box described. If the tech had accepted trial 3, the report would have said 2.44 GPM (2.00 x 60 / the three-trial mean of 49.17 s) and the conclusion would have been the same, which is worth noticing. A bad trial does not always change the answer, and that is exactly why it has to be caught by the rule rather than by whether the answer looks wrong.
When a bucket is the wrong tool
- The flow is too strong to time. Above roughly 20 GPM even a large can fills too fast for a fair reading, and the stream throws water. Use the water meter instead: with every other fixture off, time one full sweep of the meter's dial or count the register over a measured minute.
- The flow changes while you watch. A well pump cycling on a pressure switch, a pump drawing a tank down, or a supply shared with an irrigation zone all give a flow that moves. Time it at several points and write down when, rather than averaging a moving target into one number.
- You need flow at a pressure. If the question is whether a fixture meets its rating or whether a supply can deliver at a given pressure, gauge the pressure at the same time and write both. A flow without its pressure cannot be compared to a rating. See related: Pressure and Head: The 2.31 and 0.433 Pair.
How to verify you got this right
- The container was marked from a known measure, and the record says how.
- Every accepted trial is within 5 percent of the median, and any rejected trial is shown with its re-run.
- The fill took at least 20 seconds on every trial.
- The flow passes the sense check: several GPM for a pail filling in under a minute, not hundreds.
- The record names the measuring point and the valve setting, so a repeat test next month can be compared to this one.
References
- See related: How to Work Out a Fill or Drain Time
- See related: Averaging Readings: Mean, Median and Weighted
- See related: Pressure and Head: The 2.31 and 0.433 Pair
- Federal plumbing product efficiency standards for showerhead maximum flow and test pressure (US Department of Energy); some states set lower maximums