Flow sensor selection is driven by three engineering gates: the fluid's electrical conductivity and cleanliness, the required turndown between minimum and maximum GPM, and the acceptable permanent pressure loss across the meter body [S3][S4].
Get any of those three wrong and the instrument either reads zero, drifts in calibration, or starves the process of head. The eight ifm flow-sensor families spanning 1/4" to 16" line sizes and -40 °F to 302 °F process temperatures illustrate the full envelope a spec engineer has to filter against [S1].
Match the Measuring Principle to the Fluid
Magnetic-inductive (magmeter) flow meters operate on Faraday's law and require a conductive medium with at least 20 μS/cm; they deliver roughly ±0.5% accuracy on water-based conductive media across 1/2" to 6" lines and process temperatures of -20 to 150 °C (-4 to 302 °F) [S1].
For deionized water, ultrapure water, or non-conductive media, ultrasonic time-of-flight meters (1/4" to 2", ±1%) are the correct family because they do not rely on fluid conductivity; conversely, attempting to use a magmeter on deionized water will produce a near-zero signal. Vortex meters use a bluff body and piezoceramic sensor to count shedding vortices, making them independent of pressure and temperature fluctuations, but they are restricted to water-based liquids in the 1/2" to 2" range at ±2% accuracy (water) and -20 to 100 °C medium temperature [S1].
For compressed air, an SD mechatronic meter covers 1/2" to 3/4" lines at ±2% with 14 to 194 °F media, while an SA thermal-mass meter extends to 1/2" to 1 1/2" lines at ±5% and 14 to 212 °F [S1]. The flow sensor category page consolidates the principle-vs-fluid matrix that any selection should be anchored to.
Verify the Turndown, Not Just the Line Size
The single most common specification error is matching the sensor line size to the pipe, meter, or master-valve size instead of to the operating flow range; a 4-inch T-style impeller sensor, for example, requires roughly 40 GPM before it starts reading accurately, which can mask a 10 GPM drip zone entirely [S4].
Modern hydrometer-style units such as the WeatherTRAK Flow3 cut that minimum to about 21 GPM at 4 inches, while a 4-inch iron ultrasonic sensor can resolve flow below 1 GPM and a 4-inch HD-plastic ultrasonic sensor starts at roughly 4 GPM, giving leak-detection resolution that impeller-based units cannot match [S4]. For irrigation systems running MP Rotator, tree, or bubbler stations under 10 GPM, a 2-inch sensor with a 10 GPM minimum flow threshold will simply not register those zones, and operators will be forced to disable flow monitoring for them, losing all the diagnostic value of the install [S4].
As a working rule, lock the sensor's minimum detectable GPM to at least 50% below the lowest expected zone flow, not to the nominal pipe diameter, before finalizing any bill of materials.
Compare the Main Sensor Families on Decision Criteria

Across the eight families published by ifm, four decision criteria drive the shortlist: fluid compatibility, accuracy class, line-size coverage, and process temperature envelope [S1].
Magmeters (SM sanitary / SM industrial) lead on accuracy at roughly ±0.5% for conductive water and cover the widest 1/2" to 6" range, but they will not work on oils, glycol, or deionized water. Ultrasonic meters (SU sanitary / SU industrial) handle the non-conductive fluids and the largest line sizes (2" to 16") at roughly ±3% accuracy and -20 to 100 °C, suiting building HVAC and large water-distribution mains. Vortex (SV) and mechatronic (SB) units cover water-based process liquids in the 1/2" to 4" range at roughly ±2% to ±6%, while thermal-mass types (SA, SD) are reserved for gases and compressed air at the cost of accuracy (±5% to ±10%) [S1].
For buyers cross-shopping against capacitive sensor or inductive sensor proximity products, note that those are presence/position devices and are not interchangeable with flow meters; they share the same process-connection vocabulary but solve entirely different measurement problems.
Accuracy, Cost, and System Impact: The TRIZ Trade-off
The two most heavily weighted factors in any flow-meter purchase are total installed cost and required accuracy, and they trade against each other: the lowest-priced meters are also the least accurate because they lack automated noise filtering and signal amplification [S5].
Secondary factors are system impact (how much pressure drop the meter introduces and whether it disturbs upstream or downstream flow) and the meter's flow-disturbance signature, both of which scale with internal geometry. Coriolis and ultrasonic designs generally have the lowest permanent pressure loss, while multi-port impeller and orifice-plate meters can impose 1 to 3 psig at design flow in clean water service. For a related decision framework on adjacent motor-driven systems, see the comparison of general-purpose and definite-purpose AC motors.
Specifying a meter whose stated accuracy is tighter than the plant's calibration loop can resolve is wasted spend, so always bound the accuracy requirement to what downstream control or billing actually consumes, not to the best number on the data sheet.
Output, Installation, and Environmental Constraints

Beyond the principle, four installation-side constraints decide whether a chosen sensor will actually work in the field: minimum straight-pipe run upstream and downstream of the meter, orientation (horizontal vs vertical) versus line-fill conditions, output protocol (4-20 mA analog, pulse, HART, IO-Link, or switching), and ambient versus process temperature ratings [S7][S6].
ELF (Enhanced Low Flow) sensors from Creative Sensor Technology target the residential and drip-irrigation segment and are intended for low-flow zones such as green roofs, atriums, and hanging-basket planters, while FSI sensors are specified for higher-flow mainline work, so the ELF/FSI split is the first gating decision on irrigation projects [S2]. Pressure sensors and flow sensors are not substitutes: a pressure transmitter measures the line's static or differential pressure and infers flow only with a known Cv, while a true flow sensor measures the moving stream directly, and they should be selected against measurement objectives and budget rather than cross-specified [S6].
Limitations, Failure Modes, and Who Should NOT Pick the Mainstream Option
Magmeters fail silently on non-conductive fluids; vortex meters fail below their minimum Reynolds number and on multiphase flow; thermal-mass meters drift on contamination build-up because their heat-transfer coefficient changes when the wetted surface fouls; and impeller-based T-style sensors under-read or stop registering entirely below their minimum GPM threshold [S1][S4].
Buyers who should NOT default to a standard magmeter include those measuring hydrocarbons, deionized or ultrapure water, steam, or slurries with high solids loading, and any plant where the line-size and operating flow rate are mismatched by more than 5x. For gases and compressed air, a thermal-mass or mechatronic meter is the correct default, and ultrasonic is the right pick for large-diameter water mains where insertion probes avoid the cost of a full bore cut. Confirm that the selected meter's published minimum and maximum process temperature spans the worst-case ambient plus fluid heating condition, and that the wetted-material compatibility list (often 316L stainless, PVDF, or EPDM) is on the sensor's published datasheet before releasing the purchase order.