The Traverse That Doesn't Add Up
You must verify 10,000 cfm in a 24 in. x 18 in. rectangular supply duct. The only accessible test hole sits about two duct diameters downstream of a 90-degree elbow. You run a 16-point pitot traverse: velocity pressures range from 0.28 to 0.78 in. w.c., with the high readings clustered on the outside of the elbow. The average velocity pressure is 0.50 in. w.c. Your micromanometer reads a barometric pressure of 29.4 in. Hg, and the airstream temperature is 95°F. A previous balancer reported 10,000 cfm from a hot-wire reading taken at the same hole. The fan is belt-driven, running 1,140 rpm with a nameplate maximum of 1,180 rpm at 60 Hz.
1. Is this traverse a valid basis for reporting airflow? What is wrong with it?
Read the worked answer
No. The location fails the standard straight-duct requirement for a pitot traverse - roughly 7.5 duct diameters of straight run upstream and 2.5 downstream. At two diameters past an elbow, the velocity profile is skewed and still rotating; the readings confirm it (a 0.28–0.78 spread with the high side on the outside of the elbow is a textbook distorted profile). Adding points does not repair a bad location.
The prior hot-wire reading at the same hole has the same defect, plus hot-wire sensors are direction-sensitive and read low in swirl. A number repeated twice from a bad location is not corroboration.
What you can say is that the value is indicative only. Before reporting airflow you should find a compliant location (further downstream, or upstream of the elbow), or use a fan-based cross-check.
2. Compute the flow from your data, correct for actual air density, and state where it lands against the tolerance.
Read the worked answer
Standard-density calculation first:
- Area = (24 x 18)/144 = 3.00 ft²
- V = 4005 x √0.50 = 4005 x 0.7071 = 2,832 fpm
- Q = 2,832 x 3.00 = 8,496 cfm
Density correction:
- ρ = 1.325 x Pb / (460 + °F) = 1.325 x 29.4 / 555 = 0.0702 lb/ft³
- V(actual) = 1096 x √(Vp/ρ) = 1096 x √(0.50/0.0702) = 1096 x 2.669 = 2,925 fpm
- Q(actual) = 2,925 x 3.00 = 8,775 cfm
The correction adds about 3.3%. Against 10,000 cfm design, 8,775 cfm is 12.3% low - outside a ±10% tolerance even before considering the location error.
Fan speed check: at 1,140 rpm the fan is 40 rpm below its 1,180 rpm maximum. Applying fan laws, (1180/1140) = 1.035, so a correctly set drive could give roughly 3.5% more flow - about 9,080 cfm, still 9.2% low. That is inside ±10% but only barely, and it is derived from a traverse you do not trust.
3. The previous report says 10,000 cfm. What do you do next, and what goes in your report?
Read the worked answer
Do not simply overwrite their number, and do not simply accept it either - resolve it with independent measurements:
- Fan-side verification: record fan rpm with a tachometer, motor amps against nameplate, and fan static (inlet and outlet) with a pitot or manometer on straight duct at the fan. Compare the measured operating point to the fan curve.
- Static profile: measure duct static at several points between the fan and the test hole to see whether the system resistance matches design.
- Better traverse: request or install a test port at least 7.5 diameters downstream (or take a fan-outlet traverse where the run is straight) and repeat with a proper log-linear point layout.
- Check the drive: confirm the sheave setting and whether the fan is intended to run at 1,180 rpm. A 40 rpm shortfall is a mechanical/commissioning issue for the installing contractor, not something you compensate for in the report.
Report content: the measured values, the density correction used, the actual traverse location and its non-compliance noted as a limitation, the fan rpm and amps, and a clear statement that the airflow is unverified pending a compliant traverse. Flag the previous 10,000 cfm figure as unsupported by the test location. Never present an indicative number as a balanced result.