COLUMN  ENGINEER’S NOTEBOOK

This article was published in ASHRAE Journal, April 2019. Copyright 2019 ASHRAE. Posted at www.ashrae.org. This article may not be
copied and/or distributed electronically or in paper form without permission of ASHRAE. For more information about ASHRAE Journal, visit
www.ashrae.org.
Steven T. Taylor

VAV System Duct Main

Design
BY STEVEN T. TAYLOR, P.E., FELLOW ASHRAE

There is a dearth of detailed advice in common HVAC design handbooks describ-

ing how to design ductwork for VAV systems. Chapter 21 of ASHRAE Handbook—
Fundamentals, for instance, provides three suggestions for duct sizing friction rates but
does not suggest when to use each and does not distinguish between constant volume
and variable volume systems with respect to design techniques or friction rates. In
this month’s column, I will provide detailed advice and a simple procedure for sizing
“medium-pressure”* duct mains† in VAV systems.
Duct Design Methods This is not a common application, but it demonstrates
Ideally, ductwork should be sized based on life-cycle the impact of duct system effect: the pressure drop of
costs, balancing first costs with life-cycle energy costs, the consecutive elbows is ~40% higher than that of two
with constraints due to noise and space considerations. individual elbows with long straight ducts entering
This is readily done with pipe sizing1 leading to ASHRAE and leaving. Piping, on the other hand, has negligible
Standard 90.12 limits on pipe sizing that vary depending or even below-unity system effects due to consecutive
on whether the system is variable flow or constant flow. fittings.5
But life-cycle costing for duct sizing is simply not practi- Tsal developed a life-cycle cost-based duct design
cal, primarily because: method called the T-method6 in the 1980s, but its
•• There are no standard sizes or standard fittings simplified techniques for calculating both first costs
for ductwork, as there are for piping. This makes cost and energy costs were deemed to be so inaccurate, the
estimating much more difficult. T-method was removed from Chapter 21 in 2013.
•• Duct system effects due to consecutive fittings make Instead, Chapter 21 lists two duct sizing methods:
up a very large portion of fan system pressure drop but •• Equal Friction (EF). Each duct section is sized
they are not readily calculated. In fact, the ASHRAE Duct to have the same friction rate loss, in. w.g. per 100 ft
Fitting Database3 does not include any data on consecu- (Pa/m).
tive fittings and the SMACNA HVAC Systems Duct Design4 •• Static Regain (SR). Each duct section is sized so that
manual has pressure drop data on just one, consecu- Steven T. Taylor, P.E., is a principal of Taylor Engineering in Alameda, Calif. He is a
tive elbows in different planes without turning vanes. member of SSPC 90.1 and GPC 36.

* Theterm “medium pressure” is not strictly defined by ASHRAE or SMACNA but is commonly used to describe ductwork upstream
of VAV boxes. It is typically selected for SMACNA pressure class from 2 in. w.c. to 4 in. w.c. (500 to 1,000 Pa) inclusive with velocities
up to about 4,000 fpm (20 mps). Low pressure systems are typically selected for less than 2 in. pressure (500 Pa) class and less than
2,000 fpm (10 mps) velocity.
† Duct “mains” are defined here as ductwork upstream of the taps to VAV boxes. Toolbox column “VAV Box Duct Design” in the July 2015
ASHRAE Journal addresses sizing of the inlets and outlets to VAV boxes. Branch lines from VAV boxes to outlets are typically sized as
low-pressure ductwork using equal friction method at 0.08 in. w.c. to 0.10 in. w.c. per 100 ft [0.65 to 0.8 Pa/m].

54 A S H R A E J O U R N A L   a s h r a e . o r g   A P R I L 2 0 19
ThisfileislicensedtoStevenTaylor([email protected]).CopyrightASHRAE2019.
 COLUMN  ENGINEER’S NOTEBOOK

the static pressure loss due to friction in that section is to use any of these rates. The chapter includes an exam-
offset by the static pressure regain resulting from a re- ple where the friction rate (0.67 in. w.c. per 100 ft [5.5
duction in duct velocity at the beginning of that section. Pa/m]) is determined by the duct size that maximizes
Neither method has a strong rationale for why it the acoustically allowable velocity (3,000 fpm [15 mps]
should be used to size ducts! in the example); this friction rate is used to size all the
Clearly, there is no intrinsic value to having the same downstream ductwork. During my ~40 years designing
rate of pressure loss due to friction in each duct sec- and peer reviewing duct systems, I have never heard of
tion. In the author’s opinion, the EF method offers only anyone using this approach to size ductwork!
this benefit: it is readily understandable and repeat- The SMACNA HVAC Systems Duct Design manual suggests
able. Two engineers given the same desired friction rate starting with friction rates and velocities in the shaded
will design ducts roughly the same size. But there is no area of Figure 1. This is, of course, a huge range so not
way to relate friction rate to life-cycle costs other than very useful advice. The manual does suggest that a fric-
designing for lower friction rates will almost always tion rate of 0.1 in. w.c. per 100 ft (0.8 Pa/m) is a typical EF
result in higher first costs and lower energy costs. design value for low-pressure ducts.
The SR method results in the static pressure at each
tap to be roughly constant, arguably making the system Friction Rate Reduction Method
self-balancing. But true self-balancing is seldom pos- While ASHRAE and SMACNA Handbooks offer little
sible because of duct size limitations‡ and, moreover, concrete advice on sizing medium-pressure VAV duct
having the same static pressure at taps seldom provides mains, the EDR Advanced VAV System Design Guide7 out-
self-balancing in real systems§ since taps very seldom lines a relatively simple duct sizing technique called the
require the same pressure from that point to supply the Friction Rate Reduction Method. The procedure is as
desired airflow through the tap, through the branch follows:
ductwork, through terminal devices, then through the 1. Starting at the fan discharge, choose the larger duct
final air outlet. The SR method is also more difficult to size from among both of the following design limits:
use than EF because iteration is required; it cannot be a. Maximum velocity (to limit noise). Veloc-
readily done by hand. The SR method generally results ity limits are commonly used as a surrogate for
in larger downstream ducts (and lower energy costs) limiting duct noise generated by the airflow itself
than the EF method when starting with the same initial (as opposed to that generated by the fan). In fact,
duct size. research8 has demonstrated that noise created by
Because of the simplicity of the EF method, it is by far airflow in straight ducts is negligible compared to
the most popular method for sizing all types of HVAC other sources, such as noise generated by fans and
systems, including both constant volume and variable by turbulence at fittings. It is possible that a high
volume. But what is the “right” friction rate for each sys- velocity duct system with smooth fittings may make
tem? Chapter 21 states that friction rates typically range less noise than a low velocity system with abrupt fit-
from 0.05 to 0.20 in. w.c. per 100 ft (0.4 to 1.6 Pa/m) for tings. Nevertheless, limiting velocity to limit noise
low-pressure ductwork and has duct size tables for three is a common practice. Chapter 21 recommends the
rates: 0.08, 0.2, and 0.6 in. w.c. per 100 ft (0.65, 1.6, and velocity limits shown in Table 1.** Velocity limits for
5 Pa/m) friction rates. But it offers no advice about when flat-oval duct are not listed but can be estimated
‡ Round duct diameters are typically limited to multiples of 1 in. (25 mm), and to multiples of 2 in. (51 mm) in large diameter ducts (actual
break point varies by manufacturer). While rectangular duct making machines can readily make any duct dimension, standard practice is
to use multiples of 1 in. (25 mm).
§ One example where SR is useful is the design of stair pressurization fan ductwork. These systems are required by most codes for high
rise buildings to protect exit stairs during a fire. Duct sizing need not consider energy use or noise since these fans operate only for fire
emergencies. Hence, initial velocity can be high, e.g., 4,000 fpm (20 mps), which provides sufficient velocity pressure for substantial
regain to occur at taps for supply air grilles. The taps are also usually identical so the benefit of equal static pressure at each tap that
the SR method provides has the benefit of making the system relatively self-balancing. Using the SR method for this application typically
results in the duct remaining the same size down to the last one or two sections.
** Table 12 from Chapter 21 was extracted from A Practical Guide to Noise and Vibration Control for HVAC Systems, 2nd Edition, 2011,
which includes no references to research to support the velocity limits. They were likely created by the authors based on experience. For
this reason, and because turbulence is the driving factor, not speed, they should not be considered hard limits.

A P R I L 2 0 19   a s h r a e . o r g   A S H R A E J O U R N A L 55
ThisfileislicensedtoStevenTaylor([email protected]).CopyrightASHRAE2019.
COLUMN  ENGINEER’S NOTEBOOK

FIGURE 1   Duct Friction Loss Chart from SMACNA’s HVAC Systems Duct Design, Figure A-1.

USED WITH PERMISSION FROM SMACNA Air Quality cfm at 0.075 lb/ft3

to be between those for rectangular and round but TABLE 1   Recommended Maximum Airflow Velocities to Achieve Specified Acoustic
closer to rectangular due to the flat bottom surface Design Criteria (Table 12 from ASHRAE Handbook—Fundamentals, Chapter 21).
of the flat-oval duct. MAXIMUM AIRFLOW VELOC-
b. Maximum friction rate (to limit fan power). A ITY, FPM
reasonable starting friction rate for VAV systems is NC OR RC RATING IN RECTANGULAR ROUND
0.25 in. to 0.30 in. per 100 ft (2 to 2.5 Pa/m). The ra- DUCT LOCATION ADJOINING OCCUPANCY DUCT DUCT
tionale for this range is shown in the sidebar, Design 1 2 3 4
Friction Rates for VAV Systems, on page 58. In Shaft or Above Solid 45 3,500 5,000
2. At the end of the duct main, choose a minimum Drywall Ceiling
35 2,500 3,500
friction rate, typically 0.10 in. w.c. to 0.15 in. w.c. per 100
ft (0.8 to 1.2 Pa/m) 25 or less 1,500 2,500
3. Decide how many transitions will occur along the Above Suspended 45 2,500 4,500
hydraulically longest duct main (the so-called “index Acoustical Ceiling
35 1,750 3,000
run,” the run with the highest pressure drop that will
25 or less 1,000 2,000
determine the design pressure drop and fan power)
from the fan to the most remote VAV box. Typically, a Duct Within Occupied 45 2,000 3,900
Space
transition should not be made any more frequently than 35 1,450 2,600
every ~20 ft (6 m) since the cost of the transition will
25 or less 950 1,700
generally offset the cost of the sheet metal savings. The

56 A S H R A E J O U R N A L   a s h r a e . o r g   A P R I L 2 0 19
ThisfileislicensedtoStevenTaylor([email protected]).CopyrightASHRAE2019.
COLUMN  ENGINEER’S NOTEBOOK

Design Friction Rates for VAV Systems LCC = C1 f   – 0.2 + C2 f

Some may consider the 0.3 in. per 100 ft [2.5 Pa/m] initial fric- where K and C are constants for a given system.
tion rate to be very high for an energy conserving design. But this LCC is minimized for a given friction rate when the derivative of
design condition represents a reasonable balance between first the LCC with respect to friction rate is zero:
costs (including cost of sheet metal ducts plus the space required ∂LCC
= 0 = –0.2C1 f    – 1.2 + C2
to house them) and energy costs, recognizing that VAV systems ∂f
Now assume that a constant volume system has a minimum
seldom operate at their design capacity.
LCC when the friction rate is 0.1 in. per 100 ft [0.8 Pa/m]. This is
The appropriateness of this friction rate as a design condition
probably the most common design friction rate used for constant
can be demonstrated by a simplified analysis of the life-cycle
volume and low velocity duct systems and that recommended in
costs of a duct distribution system. Assume that the life-cycle
the SMACNA manual. At this rate:
cost (LCC) of the duct system is the sum of first costs (FC)
and life-cycle energy costs (EC, equal to annual energy costs 0 = – 0.2C1(0.1) –1.2 + C2
adjusted by a life-cycle present worth factor), as shown in the 3.2C1 = C2
equation below: The LCC equation can be simplified to:
LCC = FC + EC LCC = C1  f   –0.2 + 3.2C1 f
Assuming a given static pressure class, first costs are roughly Conservatively assume that a variable volume system will have
proportional to duct surface area (area of sheet metal). For round an average annual airflow rate of 60%. With a variable speed
ducts, costs would then be proportional to duct diameter D: drive, the system will use about 30% of the energy used by a con-
stant volume system of the same design size. The LCC equation
FC ∝ D
then becomes:
Assuming that energy costs for a given fan system are propor-
tional to duct friction rate, the friction rate in a standard duct LCC = C1 f  –0.2 + 0.3 × 3.2C1 f
system can be calculated from the following equation from the = C1 f  –0.2 + 0.96C1 f
ASHRAE Handbook—Fundamentals (also used to develop friction Taking the derivative with respect to friction rate and setting to
rate nomographs): zero, it is possible to solve for the friction factor that results in the
f ∝ D  –1.2 V  1.9 lowest LCC:
∂LCC
where D represents the duct hydraulic diameter and V is the = 0 = –0.2C1 f  –1.2 +0.96C1
∂f
velocity. f = (0.21) –083
For a round duct, the velocity for a given airflow rate is inversely = 0.27
proportional to the square of the diameter, so the friction rate var- While this analysis is surely simplistic, it does demonstrate that
ies with diameter as follows: sizing ducts for a higher friction rate for VAV systems than for con-
f ∝ D –5 stant volume systems is technically justified based on life-cycle cost.
If 0.1 in. w.c. per 1 ft [0.8 Pa/m] is the “right” friction rate for constant
Based on the equations above, the life-cycle cost as a function
volume systems, then 0.25 in. to 0.3 in. per 100 ft [2 to 2.5 Pa/m] is
of diameter would be:
“right” for VAV systems. Note that with the Friction Rate Reduction
LCC = FC + EC Method, this rate is only used for the first section of duct, so aver-
= K1D + K2D –5 age friction rates will be less, but still greater than that for constant
and as a function of friction rate as: volume systems.

design is more flexible to accommodate future changes [2.5 less 1.25 = 1.25 Pa/m]) and divide it by the number of
and is more energy efficient with fewer transitions. It is transitions. The result is called the friction rate reduc-
not uncommon to have only three or four major transi- tion factor.
tions along the index run. 5. Size duct along the index run using calculated
4. Take the difference between the maximum friction airflow with diversity†† starting with the maximum
rate as determined in Step 1 (whether determined by the friction rate, then reduce the friction rate at each tran-
friction limit or velocity limit) and the minimum friction sition by the friction rate reduction factor. By design,
rate from Step 2 (e.g., 0.3 less 0.15 = 0.15 in. w.c. per 100 ft the last section will be sized for the minimum friction
†† VAV systems should be sized based on expected loads including diversity. Solar diversity (due to the fact that the sun can only be in
one position at a time) is easily determined with standard load calculation software. Diversity in other load components, such as people,
plug loads, and lighting loads, is a judgment call on the part of the designer, as are the peak design assumptions for these components.

58 A S H R A E J O U R N A L   a s h r a e . o r g   A P R I L 2 0 19
ThisfileislicensedtoStevenTaylor([email protected]).CopyrightASHRAE2019.
COLUMN  ENGINEER’S NOTEBOOK

rate selected in Step 2.

FIGURE 2   Example of duct sizing using the friction rate reduction method from
The method is illustrated in Figure 2, which shows a EDR’s Advanced VAV System Design Guide.
riser diagram of a simple three-story building.
In this example, a maximum friction rate of 0.3 in.
0.3 in.
w.c./100 ft (2.5 Pa/m) was selected with a final rate of
0.15 in. w.c./100 ft (1.23 Pa/m) at the beginning of the last
section. The index run connects to the first floor. Three
transitions exist so the friction rate reduction factor is 0.15 in. 0.2 in.
(0.3 – 0.15)/3 = 0.05 in. w.c./100 ft (0.41 Pa/m). Each sec- 0.25 in.
tion of the run is sized for ever-decreasing friction rates.
The other floors should be sized for the same friction
rate as the duct on the index floor (0.2 in. w.c./100 ft [1.6 0.15 in. 0.2 in.
Pa/m] in this example) primarily for simplicity (typical 0.20 in.
floors will have the same size ducts) and because there
may be operating conditions when those floors are the
index run. 0.15 in.
This technique emulates the static regain method,
resulting in somewhat constant static pressure from system effects and the normal changes that occur as
one end of the duct section to the other, but without design progresses. It is also important to realize that
complex calculations. It is not intended to be pre- precise duct sizing is not necessary for proper opera-
cise, but precision is not possible in most cases due to tion because VAV boxes can adjust for a wide range of
inlet pressures, generally more than what occurs in
medium-pressure systems designed using the Friction
Rate Reduction Method.

Conclusions
The ASHRAE Handbooks and SMACNA manuals
offer no specific advice for how to size VAV system duct
mains upstream of VAV box taps. Fortunately, the EDR’s
Advanced VAV System Design Guide describes a relatively
simple procedure, the Friction Rate Reduction Method,
Advertisement formerly in this space. that should provide a reasonable balance between first
costs and energy costs.

References
1. Taylor, S., McGuire M. 2008. “Sizing pipe using life-cycle
costs.” ASHRAE Journal 10.
2. ASHRAE/IES Standard 90.1-2016, Energy Standard for Buildings
Except Low-Rise Residential Buildings.
3. ASHRAE. 2017. Duct Fitting Database, version 6.00.05.
4. SMACNA. 2006. HVAC Systems Duct Design, 4th edition.
5. Rahmeyer, W.J. 2000. “Pressure loss coefficients for close-
coupled pipe ells.” ASHRAE Transactions 108(1):390-406 (ASHRAE
research project RP-1035).
6. Tsal, R.J., M.S. Adler. 1987. “Evaluation of numerical
methods for ductwork and pipeline optimization.” ASHRAE
Transactions 93(1):17-34.
7. EDR. 2009. Advanced Variable Air Volume (VAV) System Design Guide,
Energy Design Resources, California Public Utilities Commission
8. ASHRAE. 1965. ASHRAE Research Project 37, “Noise
Regeneration in Ducts,” Final Report.

60 A S H R A E J O U R N A L   a s h r a e . o r g   A P R I L 2 0 19
ThisfileislicensedtoStevenTaylor([email protected]).CopyrightASHRAE2019.