Explained

Anemometers, Explained

After reading, you'll understand why an anemometer's CFM reading is only as good as the area you entered, how to tell a useful accuracy spec from a useless one, and which sensor type belongs on which job.

Updated August 23, 2026 · Companion to our Tech & Gadgets review

In this guide

The 30-second version
The products
What actually matters
Specs in plain English
Green flags vs red flags
The brands
How to read a listing
How much to spend
FAQ

Almost everyone buying a first anemometer is buying it to answer a volume question: how much air is coming out of this register, is this fan moving what it claims, is the bathroom exhaust doing anything. The meters are sold on that promise, with CFM printed on the box. But no handheld anemometer measures volume. It measures speed at the sensor, and then multiplies by a number you told it. Once that clicks, the rest of the spec sheet starts to read very differently.

The 30-second version

The meter reads velocity, and every CFM or CMM figure is that velocity multiplied by a cross sectional area you entered yourself.
A fixed error term is the tell, because plus or minus 100 fpm looks small next to a 4,900 fpm range and is a third of a 300 fpm reading.
Hot wire goes in ducts, vanes go on grilles, because a slim probe reaches traverse points and a wide vane averages across an outlet face.
One centre reading is not the duct average, since wall friction slows the edges, which is why the log-Tchebycheff traverse exists.
At an air outlet a capture hood is usually the right tool, though it is not an absolute instrument either and a pitot traverse in straight duct remains the reference.

The products this guide is about

Our verdict on each, with the full reasoning on the companion review.

What actually matters

Air velocity - this is the only thing the sensor physically responds to, whether that is a heated wire losing heat to passing air or a vane being spun by it. Everything else on the display is derived from it.
Volume flow - velocity multiplied by area, in CFM or CMM. Extech's manual prints the formula, Fluke's manual tells you to determine the duct area first and even to ask the duct manufacturer, and Kestrel labels the input as user supplied duct shape and size.
Free area - the actual open area air passes through at a grille, which is smaller than the grille's outside dimensions and smaller than the duct behind it. Fluke's 925 literally labels its area setting Free Area, which tells you the distinction matters.
Minimum measurable velocity - the bottom of the stated range. A vane needs enough force to overcome bearing friction before it turns, so vane meters have a floor while thermal sensors are typically specified from zero.
Accuracy format - percentage of reading, percentage of full scale, or a fixed offset, sometimes all three combined. This one distinction separates a meter you can quote from one you can only wave around.
Traverse - taking multiple readings across a duct or an outlet face at defined positions and averaging them, because the flow profile is never flat. Good meters have multi point and timed averaging built in for exactly this.
Calibration traceability - whether the meter ships with any calibration document, whether an NIST traceable version exists, and whether calibration can be restored in the field. Only the Kestrel among the four here lets you restore it yourself, by swapping the impeller.

The specs, in plain English

fpm and m/sFeet per minute and metres per second, the two velocity units you will see. One metre per second is about 197 feet per minute. HVAC work in the US is nearly all in fpm, so a meter that only resolves to 0.1 m/s is resolving to about 20 fpm.

CFM and CMMCubic feet per minute and cubic metres per minute, the volume units. These are always calculated, never measured. If the meter shows CFM it has already multiplied a velocity by an area you set.

Percent of readingAn error that scales with the number displayed. Three percent of reading is 9 fpm at 300 fpm and 27 fpm at 900 fpm. This is the format you want, because the meter stays proportionally as good at the bottom of its range as at the top.

Percent of full scaleAn error expressed against the top of the range rather than the reading. On a 4,900 fpm meter, two percent of full scale is about 98 fpm no matter what the display says. At a low reading that is a huge fraction of your answer, which is why the format flatters the spec sheet.

Fixed offset termThe constant added to a percentage, written as something like plus or minus 1.5 percent of reading plus 59 fpm. It behaves exactly like a percent of full scale figure: it does not shrink, and at low velocity it is the whole error.

Hot wire or thermal anemometerA heated element whose cooling rate is converted to velocity. Small probe diameters let it fit into ducts and tight gaps, and it responds at velocities too low to turn a vane. It is fragile and has an upper air temperature limit.

Rotating vaneA small windmill whose rotation is read magnetically or optically. It averages velocity over its own face area, which is useful at a grille, and the common four inch head is generally too big to put inside a duct.

Log-Tchebycheff traverseA pattern of measurement positions that clusters points nearer the duct wall to account for friction slowing the air there. It replaced simple equal area division because it produces a truer average, and the position multipliers are published as tables.

Ak factorA correction constant that converts a measured face velocity at a specific diffuser into an air quantity. Krueger's chief engineer notes it is a ratio, not a physical dimension, that it is only valid for the exact instrument and probe position it was derived with, and that manufacturers largely stopped publishing them.

Green flags vs red flags

Green flags

Accuracy is published as a percentage of reading with a small or absent fixed term, and stated across named velocity bands.
The stated velocity range starts low, in the tens of fpm rather than above a hundred.
A slim telescopic probe is offered if duct traverses are part of the plan, with the probe diameter stated.
Multi point and timed averaging are built into the instrument, not left to you and a notepad.
A calibration document ships in the box, or an NIST traceable version or service is offered by the maker.

Red flags

A single bare percentage with no velocity band and no indication of whether it is reading or full scale.
CFM headlined in the marketing with no mention anywhere that you have to enter the area yourself.
The maker's web page and the manual state accuracy differently, which usually means the web page dropped the fixed term.
A vane head presented for duct work when the head is plainly too wide to fit through a test hole.
No calibration paperwork of any kind mentioned in either the data sheet or the manual.

Who's who: the brands

Testo - German instrument maker whose HVAC line runs from app connected smart probes to hot wire anemometers and capture hoods, and which publishes banded accuracy specs and sells NIST calibration certificates as ordering options.
Extech - Part of Teledyne FLIR, known for broad and affordable test instrument ranges including large vane thermo-anemometers with on meter area entry and optional NIST versions.
Kestrel - US maker of rugged waterproof pocket environmental meters built around a field replaceable impeller that is individually calibrated in the company's own NIST traceable wind tunnel.
Fluke - Best known for electrical test gear, with a smaller indoor air quality line that includes vane anemometers, micromanometers and pitot tubes for duct traversal.
TSI and Alnor - Long standing air flow measurement specialists whose capture hoods and thermal anemometers set much of the balancing convention, and whose HVAC handbook documents the traverse procedures.

How to read a listing without getting fooled

Start at the bottom of the spec table rather than the top. Find the velocity range and note where it starts, because that floor decides whether the meter is usable on returns and low velocity ducts at all. Then find the accuracy line and work out what it is a percentage of. If it says percent of reading, multiply it by a number you actually expect to see. If it says percent of full scale, or if there is a fixed term added on, calculate that figure in absolute fpm and compare it with the same expected reading, because that is your real error. Next look for the volume flow row and see how area gets in: on the meter, in an app, or not at all. Finally check whether the probe geometry suits the job, since a vane diameter or a probe length quietly determines whether the instrument can be used where you meant to use it. Anything the maker declines to publish, treat as unspecified rather than assuming it matches a similar model.

How much should you spend?

Budget meters buy you a display and a velocity number, and the honest ones state a range and a rough percentage without much banding. That is enough to confirm air is moving and to compare one register against another on the same day. Mid range is where on meter area entry, multi point averaging and a properly banded accuracy spec appear, and it is where most people should land, especially if outlet and grille work is the main use. Premium buys the hot wire on a slim telescopic probe, tighter low velocity accuracy, calibration paperwork in the box and calibration services from the maker, and it is worth it only if you are genuinely traversing ducts or need a number you can put in a report. A rugged sealed environmental meter also sits at the premium tier, but you are paying for waterproofing, logging and a dozen weather parameters rather than for HVAC precision. Capture hoods sit above all of this and are a different purchase entirely.

Questions, answered

Can I just multiply the register's face velocity by its outside dimensions?

No, and this is the most common mistake. The grille blocks part of the opening, so the area air actually passes through, the free area, is smaller than the outside dimensions of the register. It is also different from the duct area behind it. Diffuser makers used to publish Ak factors to bridge the gap, but Krueger's chief engineer explains that those factors are specific to the exact instrument and probe position used to derive them and are largely no longer published.

Do I need a hot wire if I only ever check registers at home?

Probably not. At an outlet the useful property is averaging across the face, which is what a wide vane does naturally and a single point hot wire does not. A large vane meter with on meter area entry and a multi point average mode will serve you better and cost less. The hot wire earns its place when you need to get a sensor inside a duct through a small hole, or when velocities drop low enough that a vane will not turn.

How much straight duct do I need before a traverse?

The Alnor handbook says to measure at least 7.5 duct diameters downstream and at least 3 duct diameters upstream of any turn or obstruction, and that traversing as close as 2 diameters downstream and 1 upstream is possible but will hurt accuracy. Fluke's traversal guide quotes 10 equivalent diameters of straight duct ahead of the plane and 3 after. For rectangular ducts, convert to an equivalent diameter first.

Is a calibration certificate worth paying for?

It depends on whether anyone is going to ask. For your own house, no. For work that goes into a commissioning or balancing report, a traceable certificate is often the difference between a number that stands and one that gets queried. Testo ships a calibration protocol with the 425 and sells NIST velocity calibration at three points, Extech offers an AN300-NIST variant, and Kestrel supplies a Certificate of Conformity plus an impeller you can swap yourself to restore factory calibration.

Ready to choose? See our top pick, the alternatives, and the one we'd skip See the Anemometers verdicts