Digital manometer: picking a range, holding zero

Three decimal places on a screen are not proof of accuracy. That is the most expensive misconception in the trade: a digital manometer always displays a number, even when it is wrong, and it does so with a confidence no U-tube ever had. The glass tube at least showed a crooked meniscus when it was set up badly. The digital instrument does not — it reads 3.47 in. w.c. whether it was zeroed five seconds ago or the day before yesterday, in a truck at -18 °C. What follows is not about how to take the reading, but about the instrument itself: range, unit, zero, and the limits it gets pushed past without anyone noticing.

What a digital manometer actually measures

A service digital manometer does not measure a pressure. It measures a difference between two ports. With only one port connected, the second one breathes room air, and that local atmospheric pressure becomes the reference. Anything that moves the reference moves the reading: a mechanical room door slamming shut, an exhaust fan cycling on.

That distinction is what drives the range decision. One instrument gets used for three measurements separated by two orders of magnitude: gas pressure at the manifold (a few inches of water column), draft at the flue (a few hundredths, negative), and pressure drop across a filter. No single sensor resolves all three equally well.

What range should a digital manometer have for natural gas?

For gas work, the reference range is ±60 inches of water column, roughly ±150 hPa. That is what the common service instruments in Montreal toolbags cover: the Testo 510i is specified from -60 to +60 in. w.c. at 0.01 resolution, and the Fieldpiece SDMN6 reads to 60 in. w.c. at 0.1 resolution, with a high-resolution mode below 2 in. w.c.

Why 60 and not 10? Because the range has to cover the supply, not just the manifold. A typical nameplate target sits around 3.5 in. w.c. on natural gas, but a building fed by a 2 psi service sees roughly 55 in. w.c. upstream of the regulator. A ±1 in. w.c. instrument — ideal for draft — is fifty-five times off scale there.

The opposite trap is just as real. A ±200 in. w.c. manometer whose accuracy is quoted as a percentage of full scale is useless for a -0.04 in. w.c. draft: its absolute uncertainty is larger than the value you are chasing. The same accuracy percentage means very different things depending on the scale it applies to.

In. w.c., kPa, hPa: the unit that derails a setup

Most modern instruments offer a dozen units. That convenience turns into a hazard the moment a technician sets the display to kPa and then reads a rating plate printed in inches of water column.

The conversion is easy: 1 in. w.c. ≈ 0.249 kPa ≈ 2.49 hPa, so a 3.5 in. w.c. target is about 0.87 kPa. The arithmetic is not the problem — the problem is that 0.87 on a screen looks plausible in almost any unit. Nobody questions it.

Hence the rule: the instrument displays the unit on the rating plate, full stop. In Quebec, gas appliances are labelled in in. w.c. because that is the unit of the installation code. The edition of CSA B149.1 that applies here is the 2020 one, in force since July 31, 2020, referenced by Chapter II, Gas, of the Construction Code and by the Safety Code. CSA Group published a 17th edition in 2025, but until the RBQ adopts it, the 2020 edition is what governs in Quebec.

Zero and cold: the two minutes that decide the reading

A pressure sensor’s zero drifts with temperature. That is physics, not a defect. What makes it bite in Montreal is the swing we put instruments through ten months a year: a bag going from a truck at -18 °C to a 24 °C mechanical room crosses more than forty degrees in minutes.

The datasheets say so, if you read to the end. Dwyer specifies its 475 Mark III series at ±0.5 % of full scale between 15.6 and 25.6 °C, but ±1.5 % in the wider bands on either side — a factor of three, driven by ambient temperature alone. The Testo 510i is rated for operation from -20 to +50 °C, which means it works in the cold, not that it reads the same there as at 20 °C.

Three habits handle it: take the instrument out of the bag on arrival, let it reach room temperature while you walk the equipment, then re-zero in that room, both ports open to air. On a multi-stage appliance, re-check zero between stages — the sensor has warmed up.

Overpressure: the failure that never announces itself

A digital manometer rarely fails outright. It degrades. The usual cause is overpressure: the low-range instrument connected out of habit upstream of a regulator, or left on the line during a piping pressure test. CSA B149.1 sets those tests from a table where pressure and duration vary with service pressure, diameter and length — values that have nothing to do with a service manometer’s scale.

Manufacturers publish two numbers that get conflated: measuring range and maximum allowable pressure. The ±1 in. w.c. model in the Dwyer 475 series tolerates up to 5 psig before damage, a hundred times its range. That is not permission. Between “survives” and “measures” lies the whole zone where the sensor comes back with a shifted zero nobody notices, because the display keeps showing three decimals. Hence the habit that protects the instrument as much as the installation: shut the gas, connect, reopen — never a hot connection, never an instrument left in place during a leak test.

Field case: two manometers, 0.6 in. w.c. apart in Saint-Laurent

Mechanical room of an office building in Saint-Laurent, two natural gas boilers, start-up visit on a January morning. Two technicians, two instruments, same appliance: 3.4 in. w.c. on one, 4.0 on the other. A 0.6 in. w.c. spread against a 3.5 nameplate target — enough for one of them to call for a regulator adjustment.

Sorting it out took ten minutes. The first instrument had been zeroed in the truck before walking into a much warmer room; the second had been at room temperature since the previous day. Both re-zeroed in the mechanical room after ten minutes of acclimation: 3.5 and 3.6 in. w.c., a residual spread consistent with their combined uncertainty.

The regulator was fine. What was not showed up on the next reading: supply pressure dropped below the nameplate minimum whenever the second boiler and the water heater fired together. A real problem that would have stayed hidden another season if the crew had stopped at reconciling an instrument discrepancy — and it was that record of readings, not an opinion, that justified asking the distributor to verify the service connection.

Field verdict: three instrument profiles

There is no single best digital manometer, only three profiles — and a technician working in gas heating and combustion usually carries two of them.

The wireless probe (Testo 510i class) is unbeatable for documentation: the reading goes straight to the phone with no transcription step. Its limit is real-world range inside a steel mechanical room — wireless pressure probes earn their place, but not as the only instrument in the bag.

The rigid-port service manometer (Fieldpiece SDMN6 class) is still the first tool out for gas work: rugged, readable at a glance, often paired with a pressure switch tester.

The low-range instrument is not a gas manometer. It is a draft and pressure-drop instrument — fragile zero, and overpressure that ruins its accuracy long before it hurts the housing.

What to check on yours before the next call

Take the instrument out of the bag now, not on the next service call. Three checks, ten minutes: re-zero at room temperature and write down the displayed value before correction — if it has moved since the last reading, the drift is real on your bench too; confirm the display is set to in. w.c.; then put two instruments side by side on the same pressure tap, and let the spread decide which one goes out for calibration. Log the result with the date, in the same file as your service reports.

For the building manager reading that report, it comes down to this: a pressure taken with an instrument whose range, unit and last zero date are known is usable data, while the same number taken blind is just a number. At Montréal Combustion, that is the line between a reading that goes into the equipment file and a reading that has to be taken again. The procedure itself — where to tap in, when to measure under load, how to set manifold gas pressure — takes over from there.