Water meter leak test: the 60-minute method
A water meter leak test tells you how much a commercial building loses per hour: overnight flow, riser isolation and the real cost of a leak in Montreal.
At a Glance
A water meter leak test needs no special equipment: an empty building, two readings and a subtraction. In a commercial building the hard part isn't measuring — it's telling a real leak apart from the automatic loads that run all night.
Seven in the morning, in the basement of a strip mall on the North Shore. Nobody has been in the building since eleven the night before. The water meter dial, though, is turning — not fast, but without stopping, as if a tap had been left open somewhere across three floors. No one has seen water anywhere, and last year’s volumetric bill landed without raising an eyebrow.
Hold on to that picture. A water meter leak test needs no thermal camera and no specialist contractor, just an empty building, two readings and a subtraction. The difficulty lies elsewhere: a commercial building is never truly empty. A hot water loop circulating, cooling tower make-up topping off, a softener regenerating at three in the morning — on a dial, each one looks exactly like a leak.
Water meter leak test: what the reading actually settles
Shut every fixture, read the meter including its decimals, wait sixty minutes with no water use, read again. Any movement is either a leak or an automatic load you haven’t accounted for. The test doesn’t tell you where; it tells you how much you lose per hour — and that number is what justifies a search.
Separating leaks from automatic loads is the whole job. A building losing 20 litres an hour and a building whose softener regenerates 20 litres an hour show the identical dial. The method below sorts the two out before you spend a dollar on leak location.
Tools required: what the test really takes
- Direct access to the meter dial, with a flashlight. City-installed meters often sit in a locked room — settle access ahead of test night.
- A phone. A timestamped photo of the dial beats a number scribbled on a clipboard.
- A hose-bib pressure gauge with a maximum-reading pointer, for pressure-hold tests on isolated sections.
- The plumbing drawings, or an hour spent mapping risers and isolation valves.
- An ultrasonic leak detector if the leak is confirmed but invisible: it locates what the meter only quantified.
⚠ Safety: what you don’t shut off in an occupied building
A sprinkler supply is not closed because someone is hunting a leak: any interruption of fire protection follows the building’s own safety procedure. Same logic for a commercial kitchen in service. Notify occupants, lock out the valves you close, and reopen slowly — a sudden repressurization stirs up deposits in an older riser and sets off water hammer. Reading and isolating are operations work, but the repair takes a licence in Québec: that belongs to a mechanical plumbing contractor.
1 · Log the overnight flow before closing anything
Read the meter at closing and again at opening, every business day, for two or three weeks. Overnight volume divided by hours gives the building’s hydraulic signature.
One detail matters here. Montreal’s reading guide tells you to ignore the decimals when you report a reading — fine for billing, useless for diagnosis, because decimals are the only part of the dial that moves in an hour. Don’t count on the City’s consumption portal either: it compares annual consumption and groups several buildings together, which helps for a portfolio, but a leak that starts in November shows up there the following spring.
2 · The 60-minute test, with the building genuinely shut
Park everything that consumes water without a human touching it, and write it down: cooling tower make-up, humidifiers, ice machines, automatic urinal flushing, irrigation, and above all softener regeneration — reschedule it outside the test window. A domestic hot water recirculation loop is closed and should consume nothing. If it consumes during the test, you already have your branch.
Read, wait a full hour, read. If the dial hasn’t moved a single decimal, the building is tight at the flow the meter can see: every meter has a minimum flow below which it doesn’t register reliably, and an oversized one will pass a weep unnoticed. A negative test on a building whose annual consumption keeps climbing deserves a second look at meter sizing.
3 · Isolate by riser, not by floor
Montreal’s building stock resists floor-by-floor isolation. In a 1960s-to-1980s building, plumbing reads vertically — risers feeding the same rooms from one floor to the next. Close a riser, repeat the one-hour reading, log it: each attempt halves the suspect zone.
Two findings show up along the way, as useful as the leak itself: isolation valves that no longer seal, and abandoned branches left pressurized after a tenant fit-out, with nobody watching them.
4 · What overnight flow is abnormal, and what does it really cost?
No universal threshold, only a line of reasoning. A building using 6,000 m³ a year uses roughly 16 on a business day; if an eight-hour unoccupied reading accounts for more than a third of that, something runs non-stop.
Putting the flow into dollars holds a surprise. A steady 1 litre per minute is about 525 m³ a year. At the $0.110 per cubic metre tier that applies between 1,000 and 10,000 m³ in Montreal’s non-residential buildings, that’s roughly sixty dollars. Lost water is close to free here. What costs money is what comes with it.
| Where that litre per minute leaks | What it actually costs |
|---|---|
| Buried service entrance | Backfill washout, slab damage, trench repair |
| Interior cold water to drain | Crossing into a higher rate tier, with no alarm |
| Domestic hot water | ≈ 114 GJ of heat a year, before appliance efficiency |
| Recirculation loop | Fresh make-up water, dissolved oxygen, scaling |
That 114 GJ checks out in one line: 525,000 litres raised from 8 to 60 °C at 4.19 kJ per litre per degree. A gas load comparable to a small building’s entire heating bill, for a leak nobody sees — which is why anything on the hot water side jumps the queue regardless of flow rate.
5 · Confirm before you open a wall or a slab
The meter quantified, the isolation narrowed a zone. Confirm with a second method: a pressure-hold test on the isolated section, acoustic listening on piping and valves, or a thermal scan over a buried hot water line. And before suspecting a pipe, walk the usual suspects — all of them drain away, so all of them are invisible:
- A urinal or closet flushometer with a diaphragm stuck partly open. First suspect by a wide margin in low-traffic washrooms.
- The relief port of a backflow preventer passing a trickle. In Montreal the device has to sit on the same pipe segment as the meter and be maintained regularly, so it’s right there, easy to check during the test.
- A relief valve on a water heater or tank, piped to a drain.
- Cooling tower make-up with a float stuck open, or a continuous overflow.
- A softener stuck in regeneration, which on its own explains a steady overnight flow.
6 · Close the loop with a number, not an impression
Once repaired, rerun the sixty-minute test under the same conditions, then add one line to the monthly operations log: the meter volume, read on the same day each month. That line is what turns the next leak into a visible variance instead of an unexplained annual bill — the same reflex Québec’s drinking water strategy asks of municipalities, which are pushing their network leakage index below 4 while the provincial average sits around 6.
What we measured in Laval: 11 m³ a night and one urinal
Light industrial building with an office section in Laval, annual consumption up from roughly 5,000 to more than 13,000 m³ in two years with no change in activity. No trace of water anywhere, no complaints.
Overnight reading across nine hours: 11 m³, a little over 20 litres per minute continuous. With the valve downstream of the meter closed, the dial stopped — the leak was inside. Riser-by-riser isolation on a Saturday morning: the warehouse washroom wing carried the entire flow.
Two causes stacked. A urinal flushometer, in a washroom used a few times a week, with its diaphragm stuck partly open, running to the drain silently in a room nobody enters. And the relief port of the backflow preventer, two metres from the meter, trickling into the floor drain.
Fix: a rebuild kit on the flushometer, repairs to the backflow preventer. Overnight reading the following week: 0.4 m³. Over a year the building dropped back below the 10,000 m³ tier — and the whole search cost less than a single month of the overconsumption.
The Friday-night and Monday-morning reading
This week’s job is two photos per building: the meter dial at Friday closing, the same dial at Monday opening. Divide the difference by the hours the building was shut. Any building whose result isn’t near zero goes on your work list, ranked by flow, and you built that list without picking up a tool.
That weekend reading is the first thing Montréal Combustion asks for when a water bill has doubled: two minutes to know whether you’re facing a leak or an operating habit, and the assurance you won’t pay for leak detection on a badly programmed softener.
Frequently Asked Questions
How do I know if I have a water leak without seeing any water?
How can I tell if a water leak is before or after the meter?
What is a normal overnight water flow for a commercial building?
Does a water leak increase a commercial building's water bill in Montreal?
Sources
- Installer un compteur d'eau dans un immeuble non résidentiel — Ville de Montréal
- Lire le compteur d'eau d'un immeuble non résidentiel — Ville de Montréal
- Immeubles non résidentiels : tarification volumétrique de l'eau — Ville de Montréal
- Consulter la consommation d'eau de son organisation — Ville de Montréal
- About the Québec Water Strategy (Stratégie québécoise d'économie d'eau potable) — Gouvernement du Québec