Duct Design 5 — Sizing the Ducts

12 Comments Read/write comments

Duct design schematic diagram showing vents and air flow

To this point in our little series on duct design, we've been calculating intermediate quantitites: available static pressure, total effective length, and friction rate. Today we use all that to find out how big the ducts need to be. We're following the Manual D protocol for duct design, a standard developed by the Air Conditioning Contractors of America (ACCA). Let's jump right in and see how it works.

Sizing the ducts by friction rate

Recall that the rated total external static pressure (TESP) tells us how much resistance we can have across the furnace or air handler when it's delivering the rated air flow. To hit that number, we have to control the resistance of the duct system.

All else being equal, a duct system with a greater total effective length (TEL) has greater resistance. That doesn't mean the total external static pressure is greater, though, because the friction losses in the ducts depend on both the length and the cross-sectional area. That's the unequal part, the knob we use to control the resistance.

If the total effective length is high, we have to increase the duct area. If the length is low, we can use smaller ducts. That's how we ensure the ducts deliver the right amount of air. (Of course, it has to be installed and commissioned, too.)

The friction rate I discussed in part 4 of this series allows us to quantify this process. (It's one of two factors that we have to look at in determining the size. The other is below.) In part 4, I showed an example where the friction rate was 0.073 iwc per 100' of total effective length.

The next step is to use that friction rate and the air flow rate for each duct section in cubic feet per minute (cfm) to find the size necessary to move that amount of air. We do it with software, but duct calculators give the same information.

Here's an example with the new ASHRAE duct size calculator. Our friction rate is 0.073 iwc/100'. Let's say we have a section of ductwork that needs to move 400 cfm. On the Friction Loss/Air Quantity part of the dial, we line up 0.073 with 400 cfm, as seen below.

The ASHRAE duct size calculator showing how to find the duct size for a given friction rate and air flow rate

As you can see, we need a round metal duct that's slightly larger than 10" to do what we want here. For flex installed properly (inner liner pulled tight with no sag or compression), it would be the same size. (See my article on flex duct compression if you don't believe that.)

We don't design for compression, but you can see that if the installer used flex and didn't pull the inner liiner tight, leaving 4% longitudinal compression, you'd need a 12" flex duct rather than 10". If they installed 10" flex duct compressed by 4%, the resistance would be higher, the static pressure would be higher, and the air flow would be lower.

Got that? The process isn't hard. You'd do the same thing for every section of the ductwork, using the same friction rate but putting in the different air flow requirements for each part.

Sizing the ducts by velocity

But just looking at those two sections of the duct calculator aren't the end of the process. We also want to make sure the velocity of the air isn't too high. So we look at the Velocity/Air Quantity section. In my example here, 400 cfm at 0.073 iwc/100' corresponds to a velocity of about 725 feet per minute (fpm). That's fine for supply ducts. To move 400 cfm on the return side in this duct system, we'd need to move to a larger duct.

The ASHRAE duct size calculator showing how to check for velocity with a given air flow rate

In Manual D, Table N3-1 specifies the maximum velocities for supply and return trunks and branches. For supplies, it's 900 fpm. For returns, it's 700 fpm. That's why we'd go up to 12" in this case for a return moving 400 cfm at 0.073 iwc/100'.

When sizing by the friction rate results in too high a velocity, we size by the velocity, which results in a larger duct. But larger ducts also result in less resistance, which means we may get too much air flow in that run. What do we do about that? Install balancing dampers.

In our HVAC design business at Energy Vanguard, we generally don't specify ducts smaller than 4". We do round ducts in one inch increments from 4" to 10" and then every 2" after that, which is why I said we'd use a 12" instead of a 10" duct for a return in that example.

Now we've got the procedure for finding the sizes of all the ducts in a design. I've got only a few topics left to go in this series: laying out the ducts, choosing duct types, and registers and grilles. And then I'll present a case study to show how all this works, from design to installation to commissioning.


Other articles in the Duct Design series:

The Basic Principles of Duct Design, Part 1

Duct Design 2 — Available Static Pressure

Duct Design 3 — Total Effective Length

Duct Design 4 — Calculating Friction Rate


Related Articles

The 2 Primary Causes of Reduced Air Flow in Ducts

How to Install Flex Duct Properly

The Science of Sag - Flex Duct and Air Flow

The Secret to Moving Air Efficiently through Your Duct System


NOTE: Comments are moderated. Your comment will not appear below until approved.


Jun 9 2017 - 10:33am

You should compile these Duct Sizing posts into a PDF format for download. They would be a great help for HVAC Schools.

Jun 9 2017 - 11:36am

Gary, we'll be doing something like that. Our plan is to create online courses with expanded versions of this with homework and videos and more.

Jun 9 2017 - 6:56pm

Allison, as you know, Manual D is not the only way to design ducts and it leaves out important considerations such as duct surface area, as evidenced by the long duct runs in your illustration.

It also depends on dampers to overcome the embedded simplifications in the method. As a result there can be higher static pressures, higher fan motor watt draw, and larger conduction losses.

ACCA Manual D covers one set of procedures that can be followed for proper design of a duct system. It contains the information necessary to:
* determine the maximum allowable duct static pressure to ensure the design airflow can be delivered through the unit and to each room;
* take into account the resistance of evaporator coils,humidifiers, air filters,auxiliary heaters, dampers, etc.
* take into account the performance of the air handler fan under differing static pressures.

Other design methods, such as the Equal Pressure Drop method as applied by Davis Energy Group, are available. Some of these methods do not depend on dampers in every run and can provide a more efficient design.

In the majority of homes currently built, duct systems are designed using rules of thumb and obsolete assumptions about house thermal characteristics(Hawthorne and Reilly 2000). A common rule of thumb is that the registers need to be placed towards the perimeter of the house, either below or above windows. These ideas were based on studies from the 1940s and 1950s that concluded that three factors favored these locations for registers: 1) floor warming due to conduction, 2)counteracting window convection currents, and 3) mixing
capabilities. Buildings designed today differ greatly in thermal characteristics and can thus be designed with registers high on the inside wall.

This system optimization strategy has been proven to work in homes built under the Building America program. Thermal distribution efficiency is greatly improved while retaining comfort conditions in the conditioned rooms.
Shorter ducts provide less conduction area and less restriction to flow which provide opportunities for additional improvements elsewhere.

Jun 10 2017 - 7:49am

John, yes, there are indeed other methods. And yes, the majority of HVAC systems don't get proper design, instead relying on rules of thumb. Regarding the long duct runs you see in the schematic at the top of this article, we do that for a good reason. Yes, extra duct surface area hurts when the ducts are in unconditioned space. No, we don't have to put the supply outlets at the perimeter.

We do third party design and don't always get to work with the HVAC contractor. If we design the system for ducts at the perimeter and the contractor instead keeps them closer in, the design still works. If we design for interior supply vents and the contractor puts them on the perimeter, it may not work so well. Whenever we can, we like to keep the duct runs short.

Thanks, as always, for sharing your wisdom and experience here, John!

Jun 14 2017 - 3:45pm

If the third party design is based on supply around the perimeter, but will still work if the supplies are installed at the interior, would it work better, or could the equipment change if the supplies are at the interior and the duct runs shorter?

I'm thinking particularly of ducted minisplits, which I understand need short runs/low static pressure to work effectively and efficiently; does defaulting to perimeter supplies and thus longer runs rule those out in this approach?

Jun 19 2017 - 6:49pm

Cal, There are many variables in your question... sure longer duct runs in attics will have more heat loss than short runs. This would be a discussion to have when calculating loads.

There is also the question of register placement on the wall or ceiling. If it is just moving the register 8 or 10 feet from the center to the perimiter, while it may have an effect on the load, it is not likely to have an impact on the static pressure/duct sizing. The average 90 degree elbow has 3 times the impact on the friction rate at 20-35 EL (equivalent length) as an 8 foot straight line length extension.  However, moving from wall to ceiling would likely add or remove a fitting, probably have different boot AND quite possibly a different register size. These changes can have significant impact.

Keeping in mind the lower available static pressure available with many ducted mini-splits, making sure there is coordination and agreement on layout and placement is important at the design stage. Having to run an extra 8 or 10 feet of duct is not likely to impact the performance, where adding one or more elbows would be cause to verify acceptable performance.

Jun 19 2017 - 8:05pm

Nothin' to add to what Andy said.

If you're considering a ducted mini, be aware that some brands/models have much lower available static than others. As I recall, Daikin's smallest sizes only have 0.12 IWC available static, which is barely enough for filteration. Mitsubishi's SEZ head can handle 0.20 IWC, and Fujitsu's ARU head goes up to 0.35 IWC. That's a huge difference in terms of how much ducting can be supported by each of these units.

Jun 10 2017 - 1:17am

Adding to John's comment, Manual D is like a cookbook... it's a step-wise approach to duct design with inherent simplifications (i.e., compromises), and the software the implements Manual D imposes even more compromises. For example, in a well designed system, fittings represent most of the pressure drop. However, the most popular duct design software doesn't adjust fitting equivalent lengths for actual velocity as specified in Manual D (see my comment in part 3).

That said, this series is an excellent primer for 'pros' who haven't bothered to read Manual D and rely totally on the software.

BTW, I follow my own version of the equal pressure drop method John mentioned and I totally embrace short duct run-outs. Extending supply run-outs to the perimeter is silly/wasteful in homes built to 21st century energy codes.

Jun 10 2017 - 7:59am

Thanks, David. I agree short run-outs are the way to go but, as I said in my reply to John Proctor, sometimes you have to compromise when you're doing third party design. If I were a builder or HVAC contractor, all my systems would have short run-outs — or no run-outs, as in ductless mini-splits. Most of our designs are for ducts inside the building enclosure, so that mitigates the extra heat gain and heat loss associated with higher duct surface areas.

Jun 10 2017 - 3:57pm

Yup, doing HVAC design without the ability to coordinate with the installer limits our flexibility to do everything as we would want it.

Probably the biggest push-back I get from contractors is my requirement for ductless return whenever possible, using various methods to ensure a very low static return path from every supply. That not only reduces static but saves first cost and blower energy (return side is typically the worst offender in systems with excess static). It also makes it a lot easier to get ducts inside floor trusses. But for some reason, most mechanical contractors can't get their head around systems with ductless or single point return.

In low-load homes with finished basements, I require ductless return. The sensible load in the basement is typically far too small for it's own system, even with mini-split. A ductless or single point return in basement forces air changes between floors even when there's little or no sensible load in the basement. This helps avoid supplemental dehumidification and keeps the basement fresh, and acts to reduce the overall load. To me, it's a real no-brainer.

Jun 10 2017 - 1:30am

Another thought about Manual D...

I've had more than one builder ask me to prepare a Manual D for LEED certification - after the fact! Sigh.

Here's my take on Manual D submittals: Programs and codes that require a Manual D report for compliance are missing the boat. It's just a piece of paper, rarely if ever, followed by field verification that the ducts were actually built to plan, or that the design itself was any good. Far better would be require external static pressure and room-by-room airflow verification. Prove to me that you *build* good duct systems, not what some computer program spits out. I could care less how you got there.

Jun 10 2017 - 8:03am

David, we're in complete agreement here. We also get — and turn down — the occasional request for load calculations and duct designs after the system has already been installed.

Add new comment