Metal Ductwork

Now let's discuss real ductwork; sheet metal ductwork, that is. ¶603.9 of the 2006
IMC (International Mechanical Code) says “All joints, longitudinal and transverse
seams, and connections in ductwork shall be securely fastened and sealed…”; it
further requires any tapes used be UL classified
and marked (note the
emphasis), and any sealants used to be UL classified. In other words, sheet
metal ducts must be airtight.

In fact, it is rare to find airtight metal ducts, and a good project for an experienced
DIY-er is tightening up and reinsulating a deficient duct system. Fred recently
completed such a project at his house, cutting his energy usage nearly 15% and
saving over $300 in the first year's operation. This is not a project to be undertaken
lightly, as there are a lot of hazards in basements and (especially) attics; but as
long as proper
precautions are observed, the job can be completed safely.

Of course, ducts in unconditioned spaces must be insulated to minimize heat
gain in the summer and heat loss in the winter. The 2006 IECC requires most
supply and return ducts outside the building envelope to be insulated to R8.

When you think about insulation values, you need to know the  "
installed value",
because you need to consider
compression of the insulation: Fiberglass
insulation is most effective when there's air "trapped" within the insulation;
overcompressing the insulation reduces the amount of trapped air, thereby
reducing the installed R value.

Here's how to tell if your fiberglass duct insulation is substandard:
  • It looks neat, there's no sagging on the bottom and the top and sides don't
    exhibit "wrinkling".
  • When you push a corner with your finger, there's very little "give".

Paradox: Properly installed duct wrap looks bad; it's baggy, it's wrinkled but it
saves you money every time your system runs,
Sections of metal ducts
are joined with "Slips"
and "Drives". Slips are
applied first, normally
at the top and bottom,
and then the drives
are applied to the
sides and "hammered
over" the slips; here,
the locations were
reversed. The
important thing to
remember is the result
is not airtight.
Here's Fred's DIY,
after removing all
the overcompressed

The joints on this
"transition fitting"  
had not been fully
Here Fred's "rubbing
down" the tape for
good adhesion.

Note the red printing
compliance with

When you see cloth
tape or plain foil
tape there’s a Code
violation; only tape
marked "UL181"
complies with Code.

FYI: When you see
plain tape,
someone's saved

foot of tape.
Here's a "collar" used
to connect a branch
duct to the rectangular
main. This type of
collar is attached by
bending tabs over
inside the duct wall.
Note that it's not tight
at the wall; there was a
good deal of leakage
here, and this is typical
for most of the collars
we see.
Overcompressed (left) and properly installed
Here are two types of

On the left is what
Fred removed; you
can clearly see the
tabs. That collar is
intended for duct
board: The tabs are
much longer than are
necessary for sheet
metal (no wonder they
weren't snug).

On the right is an
"adhesive collar" Fred
used as a
replacement. This has
an adhesive-backed
flange to seal against
After sealing everything that could leak air, we begin to reinsulate the ducts,
the one instance in the trade where "
A Good Job Looks Bad":
the "Duct Wrap" is a little "baggy" and "wrinkled" when properly applied.

Wrinkled Flex - No Good!

Wrinkled Wrap - Excellent!
Here's an adhesive
fitting installation; if
you look closely,
you may see the
screws (used to
secure it) beneath
the tape. Note that
we've applied UL
listed water-based
duct sealant to seal
between the collar
and its flange. It's
not necessary to
tape over the
flange, we were just
on a roll.

An "adjustable
elbow" is attached
to this collar, note
that we've taped
the "gores"
(segments) of the
elbow. The Code
says "All Joints"; we
take that quite

We need to discuss duct fittings. The concern here is the degree of resistance
encountered at a change in direction: Duct systems operate at VERY low
pressure, and a high resistance fitting can have a major impact on airflow; we've
seen cases where improving one fitting corrected substandard flow. It's important
to recognize high resistance fittings and avoid using them. Resistance of fittings
is expressed as Equivalent Length ("EL") of straight duct; a bad fitting might have
an EL of 80 feet, a good one 10 feet. That's quite a difference when you consider
an entire duct system might have less than 100 feet of duct.

90 Degree Elbow

There are several types of elbows available; unfortunately, the easiest ones to
fabricate (the least expensive ones) have the highest resistance to flow. In order of
  1. Square elbow with "Turning Vanes" - EL 10 (The vanes are usually a
    requirement in commercial work, and it's quite rare to find a residential
    contractor who'll go to the expense).
  2. True radius elbow - "Heel" and "Throat" are curved - EL 10 to 20,
    depending on configuration (Your best choice, it's something a residential
    contractor can and will do if asked).
  3. Radius elbow with mitered throat (a 90 degree inside corner) - EL 30 to 45
    (Careful here, this is what you'll get if you don't specify radius heel and
    throat - it takes a few minutes less to fabricate).
  4. No elbow - EL 55 and up.

The images at left and right all depict return duct connections to air movers.
There's a reason for that:
It's the most important fitting in the entire duct system.
Why do we say that, you ask? Because there's a school of thought that says "It's
only a return" in order to save a few of their dollars.
In fact, return ducts are
more important than supply ducts because they're the ones where the high
resistance fittings usually occur.
Radius elbows
BAD (top)
and GOOD (bottom)

There is no fitting -
the installer just cut a
hole in the side of the

This one may look like
a radius elbow, but
guess what?

This contractor added
40 to 50 feet to his
duct system by
providing mitered
throats on these two

There should be a
sign for him here
Last, but by no means least...

ACCA fitting (and we use that word loosely)  #1-F is a special case: The contractor
was so ill-informed, or so focused on reducing cost, that he provided
the worst
type of "Non-fitting"
; he simply ran the supply main over the air mover (installed
in the vertical - "upflow" - position) and "Connected A to B" with straight duct the
same size as the unit outlet. The equivalent length of this THING
Let's stop calling it a "fitting" - there is no way this THING is a fitting)
can be as much as
120 FEET.
Here we see a gas furnace and cooling
coil at the bottom of the image. The
contractor has run the supply main over
the furnace and "Connected A to B".
There is no fitting to assist the air to go
to the right along the main.

In so doing, he has (perhaps
unknowingly) abandoned all hope of
ever delivering rated flow.

The important number here is "H/W":
The height of the horizontal duct divided
be the width of the vertical duct; a
quotient of 0.
42 as seen here has more  
120 feet equivalent length.
As it works out, a
person Fred
considered a friend
provided one of these
"THINGS" when he
replaced Fred's fibrous
glass duct system.

Here's Fred's "FIX"
(He's hoping to
eliminate 60 feet or so;
his H/W is about 0.69
for a length of about
107 feet):
More on Tape

The seams and joints of duct wrap must be stapled, and then sealed with one of
two kinds of tape: UL181 (with the printing - preferred) or a "foil skrim" tape
matching the insulation facing.

There are installers using silver cloth tape (this type of tape
has a rubber-based adhesive that can dry out over time, at
which point the tape can literally fall off) to do the job, and
then stapling THROUGH the tape.

This technique raises a lot of questions:
  1. Do they staple through the tape because they know the tape will eventually
    fall off, and they think the staples will prevent that?
  2. What about the vapor barrier - Do they care that it is punctured hundreds of
  3. What about the condensation that forms on the outer staple surfaces - Do
    they think the staples won't eventually rust, fall apart and allow the seams
    and joints to open up?

Note the multiple layers of tape here: That's always an indication that
something's wrong.

If you have a choice, DO avoid these guys at all costs!

Manual J (It's a book)
tells us "
Duct system
attributes can make
a 100 to 400 SqFt per
ton difference in the
SqFt/Ton value

In other words,
way your duct
system is designed
and installed
dictates how much
cooling you need
Adjustable elbows are
shipped in a straight
configuration; rotating
the "gores" yields any
angle up to 90