Last month a contractor in Riyadh asked me why his supplier's double-wall duct was costing almost double the plain galvanized spiral, and whether he could get away with wrapping single-wall himself. Fair question. I have been around double-wall lines for years now, and the honest answer is: sometimes you can wrap it, and sometimes wrapping costs you more than the machine-made product. The difference shows up in joints, not in the straight runs.
Double-wall, sometimes called pre-insulated spiral duct, is basically two concentric spiral tubes with insulation between them. The inner shell can be solid galvanized or perforated for sound absorption, the fill is usually 25 to 50 mm glass wool at 12 to 16 kg per cubic meter, and the outer skin is a light-gauge spiral tube, often 0.5 mm pre-painted or galvanized. On a proper line, both shells come off the same spiral tube forming machine in one pass sequence, which is why the concentricity is good and the assembly is fast.
Field-wrapped single wall is cheaper on paper. But every elbow, every takeoff, every joint needs manual cutting and taping, and a lazy wrap leaves thermal bridges at every hanger. In a 40-degree summer that is money leaking out of the building for twenty years.
On a standard spiral line, one machine runs one strip into one tube. On a double-wall line, the sequence matters. Typically the inner shell runs first, then the insulation feeding station lays the glass wool blanket over the moving strip or over the formed inner tube, then the outer skin locks over it. The outer skin is deliberately formed a size or two larger so it slides over the insulated inner shell without tearing the blanket.
The details that separate a good double-wall tube from a rattling one:
Most double-wall outer skins are corrugated. The corrugation adds hoop stiffness so you can drop the outer skin to 0.5 mm without it denting when a technician steps on it in a ceiling void. Corrugation pitch is usually 6 to 12 mm with a depth of 1 to 2 mm. If the corrugation rolls are worn, the skin loses stiffness and you get oil-canning on the finished duct. Check the rolls every 40,000 to 50,000 meters of production, roughly.
Glass wool blanket comes in compressed rolls and expands as it feeds. If the feed tension is too high, the blanket stretches thin and you lose R-value in spots. Too low and it bunches at the lock seam. On the lines I have commissioned, the sweet spot is a light drag brake plus a guide comb that keeps the blanket centered. You hear the difference: a bunched blanket makes the outer skin sit proud on one side.
If the outer skin is not concentric with the inner shell, the insulation thickness varies around the circumference. Two millimeters of eccentricity on a 50 mm fill means one side has 48 mm and the other 52. Not fatal, but it stacks up over a duct run and hangers will chew through the thin side. Good lines keep the outer skin oversize consistent, often 25 to 35 mm larger in diameter than the inner shell depending on fill thickness.
This is the part buyers get wrong most often. The straight duct is easy. The joints decide whether the system holds its thermal and acoustic performance. Here is a comparison of the joint options I see in real projects:
| Joint type | Air leakage | Insulation continuity | Install speed | Relative cost |
|---|---|---|---|---|
| Male/female sliding coupling with gasket | Good, if gasket is intact | Gap of 40-60 mm at each joint, needs collar cover | Fast, push-fit | Low |
| Draw-band coupling | Moderate, sealant dependent | Small gap, easy to close with band collar | Fast | Low |
| Flanged joint (angle ring or std flange) | Excellent with gasket and bolts | Best, flange covers full joint | Slower, bolt work | Medium |
| Clamp-together with inner spigot | Good | Insulation must be stepped back | Very fast | Medium |
| Field-wrapped butt joint | Poor if rushed | Depends entirely on workmanship | Slow | Lowest |
For pressure classes up to 500 Pa class A or so, sliding couplings with a rubber gasket are standard and fine. Above that, or on exposed architectural work where leakage testing is specified, flanged joints pay for themselves. A proper round flange forming line makes this painless, but plenty of shops buy flanges loose and it still works.
The one mistake I see constantly: the inner shell and outer skin are joined separately, but nobody seals the annular gap where insulation shows. Warm humid air sneaks into that gap and condenses on the inner shell. In a chilled-water AHU system you get wet insulation and eventually mold complaints. A simple insulated joint collar solves it for pennies.
Sliding joints let the fitter push one section deep into the next, which crushes the insulation at the overlap and creates a hard thermal bridge. Specify and stamp a stop mark, or better, fit an internal stop bead during production. Costs nothing on the line, saves arguments on site.
When a branch is cut into double-wall duct, the inner liner frays if it is cut with a coarse tooth or a hot torch. Frayed glass fibers downstream of the takeoff end up in occupied spaces, and facility managers notice. Cut the liner cold with a sharp knife, deburr the inner shell, and cap the cut fibers with a collar or mastic ring.
A standard band hanger cinched tight flattens the outer skin and squeezes the insulation to nearly nothing at the contact line. Use wide-band hangers, minimum 40 mm, or insulated hanger inserts. On one hospital job in Southeast Asia the contractor reused 20 mm band hangers from a single-wall job, and the thermographic survey showed a dotted line of heat gain along every hanger. They replaced all of them. Nobody was happy.
Push-fit couplings ship with gaskets installed, but handling knocks some out. A thirty-second check before assembly catches the missing ones. After assembly it takes a ladder, a test report failure, and a callback.
When you are comparing suppliers, ask for these specifics instead of a generic catalog sheet:
Inner shell gauge and whether it is solid or perforated. Fill density and thickness, and whether the density is measured on the finished duct or just the incoming blanket. Outer skin gauge and coating. Joint type per diameter, and whether gaskets are EPDM or cheap foam. Concentricity tolerance, typically within 2 mm. Length of duct sections, because 3-meter sticks mean 33 percent more joints than 6-meter sticks on the same run.
For acceptance on site, the two tests worth insisting on: a leakage test per SMACNA or DW/144 class on a representative section, and a simple thermal check with an IR camera after the first week of operation. The camera costs nothing to run and catches every crushed hanger and missing collar.
If you are a duct shop already running a spiral tubeformer, adding double-wall means an insulation feed station, a second forming pass or a second smaller machine for the outer skin, and joint collars tooling. The machine investment is real but modest compared to a full line. The commercial upside is that pre-insulated duct sells at a premium and the customers who buy it, hospitals, data centers, hotels in hot climates, are exactly the customers who also buy big and reorder.
What I tell smaller shops: start with the joint system. Buy or tool up for gasketed couplings and flanges first, run double-wall as a semi-manual assembly with purchased outer skin if volume is low, and automate the insulation feed only when orders justify it. The spiral tubeformer you already own does most of the work either way.
25 mm glass wool at 12 to 16 kg/m3 covers most comfort air conditioning up to around 50 mm equivalent thermal requirement. Middle East projects often ask for 50 mm fill for exposed runs. Confirm with the project spec, because thickness drives outer skin size and joint collar size, and changing later means retooling.
Not fully. Rotary brush cleaning works on solid-liner double wall but the brushes cannot pass most standard couplings, so access doors are needed every few meters. Perforated liner duct should not be brush cleaned at all, since the brushes pull fibers out of the fill. This is why some hospital specs limit perforated liner to non-return runs.
Material is typically 60 to 100 percent more per meter, but installed cost gap narrows to 20 to 40 percent because fitting and wrapping labor disappears. On long straight runs with few fittings the gap narrows further. Break-even depends almost entirely on your local labor rate.
With a good EPDM gasket and a stop bead, sliding couplings handle up to around 750 Pa positive reliably. Beyond that, or for negative pressure above 500 Pa where collapse risk matters, move to flanged joints with bolted gaskets. Leakage class targets in the spec, not habit, should make this call.