Airflow Measurement and CFM Verification

Purpose

This standing instruction produces a system airflow figure in CFM that carries its own basis: which method produced it, what that method assumes, and what it was cross-checked against.

Capacity, coil temperature, temperature split and half the charge diagnostics all read differently at 300 CFM per ton than at 400. A shop that guesses airflow from a static reading, or reports whichever of two methods gave the friendlier answer, condemns ductwork that was fine and replaces blowers that were doing their job.

Safety actions that gate this work

• The temperature-rise method requires the furnace to fire for the whole measurement, so a personal CO monitor is on your body before the burner lights and stays on until you leave. A cracked exchanger or a spilling vent shows up as a rising reading long before anyone feels it.
• If you smell gas at the meter while clocking it, everyone leaves the building immediately, nothing electrical is switched, no phone is used inside, and the gas utility is called from outside. Do not go back in to shut anything off.
• Supply plenum metal near a firing heat exchanger will burn a hand. Use a probe with a handle, work through a drilled port rather than an open seam, and never push a probe into a hole that opens on the flue path.
• Register readings at ceilings put you on a ladder with a hood in both hands. Set the feet level, keep your belt buckle inside the rails, and have the hood handed up rather than carried while climbing.

Scope

Covers total system airflow verification on residential and small light-commercial forced-air equipment: gas furnaces, electric and heat pump air handlers, and single-zone rooftop units.

Does not cover static pressure, which the static pressure diagnostic SOP owns and which this procedure requires as an input. Per-room delivered air belongs to the register and grille balancing visit SOP. Refrigerant charge verification is a separate SOP and must not be attempted until airflow is verified.

Choosing the method, and what each number is based on

The constant 1.08 in the first row is derived for standard air at about 0.075 lb per cubic foot, near sea level. At roughly 5,000 ft the density is about 15 percent lower and the constant falls with it, to about 0.92, so using the sea-level constant there understates the computed airflow by roughly the same 15 percent.

Roles and handoffs

Procedure

1. Establish the target airflow before measuring anything. Acceptance: a target CFM with its source written down, taken from the equipment's own data for the installed coil and mode rather than a rule of thumb, and a note that heating and cooling airflow can differ on the same unit. Wrong looks like 400 CFM per ton applied to a heat pump whose data calls for more. Stop rule: no target from the data means you record the measurement and stop short of a verdict. Hazard: none physical; stand clear of the cabinet while you read the plate.

2. Take the static map first, on this visit. Acceptance: TESP, the rated ESP and every component drop from the static pressure diagnostic SOP, dated today. Wrong looks like an airflow number with no pressure behind it, which cannot say why the airflow is what it is. Stop rule: a filter over the 0.20 in w.c. shop gate gets corrected and the static repeated before airflow is measured at all. Hazard: as that SOP states, power off at the disconnect for any drilling; and before drilling, identify what lines the plenum, because on a building constructed no later than 1980 duct wrap, tape or joint compound of unknown type is presumed asbestos-containing under 29 CFR 1926.1101 for construction work or 1910.1001 in general industry and is not drilled at all, while lined duct that is drilled gets collection at the bit.

3. Pick the method against the table and write down which one and why. Acceptance: the method named on the ticket with the equipment condition that selected it, plus that method's basis line. Wrong looks like running two methods and reporting only the friendlier one. Stop rule: if the only available method is the blower table, say so and label the figure a prediction rather than a measurement. Hazard: none physical; a decision made at the truck.

4. On the temperature-rise route, clock the gas meter for the real input. Acceptance: every other gas appliance off, a timed revolution of the smallest dial, the heating value obtained from the utility rather than assumed, and the input compared against the nameplate. Stop rule: an input more than about 10 percent off the plate is a gas pressure, orifice or derate problem and is corrected before any airflow is computed from it. Hazard: you are at a live meter with an appliance firing, so the CO monitor is on you, and a gas odor means everyone leaves with nothing switched and the call made outside.

5. Place the temperature probes where the reading is not contaminated. Acceptance: return probe upstream of any bypass, supply probe out of line of sight of the heat exchanger and past the first turn, both stable within 1 degree F over 2 minutes after at least 5 minutes of steady firing. Wrong looks like a rise above the nameplate range with no other airflow symptom, which is the signature of radiant heat striking the probe. Stop rule: relocate and re-read before diagnosing low airflow. Hazard: the plenum is hot; work through a port, not an open seam.

6. Compute the airflow and print the constant with its density basis. Acceptance: CFM to the nearest 10, with the clocked input, the steady-state efficiency from manufacturer data rather than the AFUE, the measured rise and the constant all shown. Wrong looks like AFUE used as efficiency, which understates output and therefore airflow. Stop rule: above roughly 2,000 ft use the density-corrected constant and say which one you used. Hazard: the equipment is still firing, so the CO monitor stays on.

7. Cross-check with a second, independent method and reconcile. Acceptance: a second CFM value by a different method, both reported, agreeing within a shop band of 10 percent of the higher figure. Wrong looks like a single number with nothing to contradict it. Stop rule: a disagreement beyond the band means neither value is reported alone; the usual cause is a leaky filter slot under the flow plate or a probe still reading radiant heat. Hazard: hood work at ceiling registers is ladder work, so set the feet level and keep your buckle inside the rails.

8. Return to service and verify the protection you disturbed. Acceptance: flow plate out, filter back with the arrow correct, slot cover refitted, panels on, thermostat restored, the interlock proven by opening the door on a running blower, a final temperature rise inside the nameplate range with the cabinet closed, and every appliance shut off at step 4 restored and confirmed operating, each logged by name, any standing pilot relit per its own lighting instructions with the space ventilated and no ignition source introduced until the odor check is clean. Stop rule: if the high limit cycled at any point, establish why before treating the limit as the fault, and never jumper it. Hazard: this step re-fires the appliance with the family present, so the panel goes on before the burner lights and the CO monitor stays on until you are out the door.

The record this produces

An airflow line on the ticket that a later reader can re-derive:

• Target CFM, the document it came from, and the mode it applies to
• Method used and the one-line basis for that method
• Raw inputs: clocked input and heating value, efficiency and its source, return and supply temperatures, probe locations
• Computed CFM, CFM per ton, and the percentage of target
• The second method, its value, and the percentage difference between the two
• TESP and component drops from the same visit
• Disposition: accept, or the specific restriction that has to move first

The charge tech reads this before touching a gauge, because superheat and subcooling read differently on a starved coil. Six months later, a callback tech can tell whether airflow changed or somebody just used a different method.

Worked pass: 3-ton system on an 80,000 BTU/hr nameplate furnace, complaint of weak cooling

Step 1: the equipment data calls for 1,200 CFM in cooling with the installed coil, which is 400 CFM per ton for 3 tons. That is the target, sourced from the manufacturer's data rather than a rule of thumb.

Step 2: static map from earlier the same visit, TESP 0.78 in w.c. against 0.50 rated, filter clean at 0.11.

Step 4: clocked at 45 seconds for one cubic foot, so 3,600 divided by 45 gives 80 cubic feet per hour, and at the utility's stated 1,030 BTU per cubic foot that is 82,400 BTU/hr. That is about 3 percent over the 80,000 plate, inside the 10 percent gate, so the procedure continues on the clocked figure.

Step 5 FAILS. First reading: return 70 F, supply 138 F, a rise of 68 F against a nameplate range of 35 to 65 F. Nothing else looks like severe restriction and the static map does not support a rise that high. Stop rule taken: the supply probe is in line of sight of the heat exchanger. Relocated past the first turn and re-read after 6 minutes of firing: return 70 F, supply 122 F, a rise of 52 F, stable within 1 degree over 2 minutes.

Steady-state efficiency from the manufacturer's data is 80 percent, so output is 82,400 times 0.80, which is 65,920 BTU/hr. On the bad reading, 1.08 times 68 is 73.44, and 65,920 divided by 73.44 is about 898 CFM, or 299 CFM per ton. On the corrected reading, 1.08 times 52 is 56.16, and 65,920 divided by 56.16 is about 1,174 CFM, or 391 CFM per ton. The contaminated probe would have reported 75 percent of target and sold a duct renovation; the real figure is 1,174 divided by 1,200, or 98 percent.

Step 6: the constant used is 1.08 at an elevation under 2,000 ft, so no density correction applies, and the ticket says so.

Step 7: a calibrated flow plate in the filter slot reads 1,120 CFM. The two differ by 54 CFM, and 54 divided by 1,174 is about 4.6 percent, inside the 10 percent band. Both are reported. The lower plate figure is consistent with the plate adding resistance of its own, which is stated rather than explained away.

Step 8: plate out, filter back, doors on, interlock proven by opening the door on a running blower and watching it stop, final rise 51 F closed up, inside the nameplate range. The range and the water heater turned off at step 4 are restored and confirmed lit, both logged by name. The limit did not cycle at any point.

References

• Equipment nameplate and the manufacturer's installation instructions for rated input, steady-state efficiency, nameplate temperature-rise range, and required airflow by mode
• The gas utility for the local heating value in BTU per cubic foot, which varies and must not be assumed
• ANSI/ACCA 5 QI for airflow verification at installation and ACCA Manual S for the airflow the selected equipment was matched at, in the editions your jurisdiction or utility program has adopted
• Flow plate and anemometer manufacturer documentation for the calibration basis and reference pressure of the instrument
• See related: the static pressure diagnostic SOP, the supply and return temperature split verification SOP, and the register and grille balancing visit SOP