Duct Junction Loss Coefficients: Which Velocity Should You Use?
AI-generated editorial illustration; not a verified installation or construction detail.
A coefficient needs a reference section
A junction calculation can look tidy and still be wrong by a factor of nearly three. The cause may be a valid loss coefficient multiplied by the velocity pressure from the wrong duct section.
The example below produces a branch total-pressure loss of 21.6 Pa. Substituting the common-duct velocity gives 60 Pa. Nothing about the fitting changed; only the reference used in the calculation changed. This is a useful error to investigate before adding fan-pressure margin.
Read the subscripts before entering the number
ASHRAE’s Duct Design chapter defines a local loss coefficient as total-pressure loss divided by velocity pressure at a specified section. With SI units, velocity V = Q/A, velocity pressure pv = ρV²/2, and fitting loss Δpt = C × pv. Use the actual internal duct area for V, rather than an equivalent circular area used for friction sizing.
For a dividing junction, label the upstream common section c, the outgoing branch b, and the outgoing straight-through section s. In the chapter’s branch-referenced convention, Cb multiplies pv,b; Cs multiplies pv,s.
Read the source diagram and notes every time. ASHRAE’s separate fitting supplement also contains common-section-referenced junction coefficients labelled Cc,b and Cc,s. The extra subscript matters: a coefficient for the branch path does not automatically use branch velocity pressure.
Worked example: one inlet, two outlet paths
Consider an illustrative dividing junction carrying steady, low-speed air at an assumed uniform density of 1.20 kg/m³. Ignore leakage, elevation effects and temperature change. All dimensions below are clear internal dimensions. The chosen velocities demonstrate the arithmetic; they are not project design recommendations.
Common section c: 1,200 L/s = 1.20 m³/s through 400 × 300 mm. Area Ac = 0.120 m², so Vc = 1.20/0.120 = 10.0 m/s.
Branch section b: 360 L/s = 0.360 m³/s through 300 × 200 mm. Area Ab = 0.060 m², so Vb = 0.360/0.060 = 6.0 m/s.
Straight-through section s: 840 L/s = 0.840 m³/s through 400 × 300 mm. Area As = 0.120 m², so Vs = 0.840/0.120 = 7.0 m/s. Continuity checks: 0.360 + 0.840 = 1.20 m³/s.
The three velocity pressures are therefore pv,c = 0.5 × 1.20 × 10² = 60.0 Pa; pv,b = 0.5 × 1.20 × 6² = 21.6 Pa; and pv,s = 0.5 × 1.20 × 7² = 29.4 Pa.
Assume Cb = 1.00 referenced to b and Cs = 0.20 referenced to s. These are deliberately illustrative coefficients, not ASHRAE table values or verified data for a fabricated tee. A project calculation must replace them with applicable fitting data.
The common-to-branch loss is Δpt,c→b = 1.00 × 21.6 = 21.6 Pa. The common-to-straight loss is Δpt,c→s = 0.20 × 29.4 = 5.88 Pa.
Keep two path entries. Air continuing straight does not also traverse the branch exit. Adding 21.6 and 5.88 Pa as one series junction loss would produce an incorrect 27.48 Pa entry. Each route receives its own junction contribution before its downstream losses are added.
Original teaching example at ρ = 1.20 kg/m³. Coefficients are illustrative, not verified values for a selected fitting. Branch and straight losses belong to separate airflow paths.
Convert the coefficient to check the spreadsheet
Suppose a worksheet stores all junction coefficients on the common-section basis. Preserve the same physical pressure loss by changing the coefficient as well as its reference:
Cc,b = Cb × (Vb/Vc)² = 1.00 × (6/10)² = 0.36. Then Δpt,c→b = 0.36 × 60 = 21.6 Pa.
Likewise, Cc,s = 0.20 × (7/10)² = 0.098. Multiplying 0.098 × 60 gives the same 5.88 Pa straight-path loss.
The incorrect branch calculation, 1.00 × 60 = 60 Pa, is 38.4 Pa too high, or approximately 178% above the correct illustrative result. Checking both coefficient bases is a compact spreadsheet test: they must return the same pressure loss, apart from rounding.
Static pressure can rise across this junction
McGill AirFlow’s fundamentals guide explains pressure changes using total pressure = static pressure + velocity pressure. Applying that relationship to our example shows why a static-pressure reading alone cannot establish the fitting’s total-pressure loss.
Assume the common-section static pressure is 100 Pa gauge. Its total pressure is 100 + 60 = 160 Pa. After the branch-path loss, branch total pressure is 160 − 21.6 = 138.4 Pa.
Branch static pressure is therefore 138.4 − 21.6 = 116.8 Pa. Static pressure has risen by 16.8 Pa while total pressure has fallen by 21.6 Pa. The reduction in velocity pressure accounts for both quantities.
Do not subtract that static rise again from the calculated total-pressure loss. Doing so would count the velocity conversion twice. These values refer to the sections represented by the coefficient definition, not arbitrary pressure taps beside the turbulent takeoff.
A negative coefficient needs interpretation
For converging junctions, ASHRAE describes energy transfer between unequal-velocity streams during mixing. A path associated with the slower incoming stream can show a negative total-pressure loss coefficient. That sign describes an exchange between streams; the junction has not become a fan.
Verify direction, reference section and the companion path before accepting a negative value. Do not replace it with zero merely because a worksheet expects every fitting entry to be positive.
Turn the teaching example into a reviewable calculation
The ASHRAE Duct Fitting Database supplies fitting outlines, coefficient tables and calculated pressure losses using entered geometry and airflow. Use the applicable source to establish the real coefficient, retaining its fitting identifier and assumptions.
For this example, the lookup ratios would include Qb/Qc = 0.30, Ab/Ac = 0.50 and As/Ac = 1.00. Confirm which ratios the selected table actually requests; some use velocity ratios or different area combinations.
A practical review checklist:
Draw flow arrows and mark c, b and s on the calculation sketch.
Record internal areas, airflow units and the density used.
Store each coefficient with its source, path and reference section.
Recalculate pressure loss on a second reference basis as a cross-check.
Confirm that each airflow path receives only its own junction loss.
Revisit the lookup when branch flow, geometry or operating direction changes.
Where this method needs more evidence
A correct reference velocity cannot rescue an unsuitable fitting coefficient. McGill’s guide discusses how closely spaced fittings interact; isolated-fitting data may not describe the installed arrangement. Check nearby elbows, dampers and transitions against applicable test data or a suitable engineering assessment.
This example excludes straight-duct resistance, equipment losses and fan-system effects. It establishes two junction entries, not a complete fan duty. Its most useful deliverable is a traceable calculation row: fitting, route, coefficient, reference area, reference flow, density and resulting total-pressure loss.
Related learning resource
For a broader study sequence, Nexora’s HVAC Duct Design and Sizing Handbook is a 25-chapter PDF covering airflow, duct sizing, pressure-loss calculations and fan ESP. Review its contents alongside the original fitting data needed for your project.
Sources and further reading
Engineering education only. Verify project conditions, applicable requirements and manufacturer data; use qualified professional judgment before design, selection, operation or maintenance decisions.
Comments