A few years back we shipped a spiral tubeformer to a fabricator supplying a pharmaceutical facility. Commissioning went well. Roundness was good, the cut-off was square, the operator could change from 400 to 630 mm in under half an hour. Three weeks later he called me because the site had rejected a batch of 400 mm duct on a leakage test. Not dramatically. Just enough to fail the specified class on the seams, not the joints.
I asked him to send me a strip width measurement. He had been running 1290 mm. For a 400 mm tube with a standard four-ply lock, the correct figure is 1284 mm. Six millimetres too wide does not sound like much until you work out what it does: it pushes the seam overlap deeper into the pocket than the lock was designed for, and the extra material has nowhere to go except outwards. The seam looked fine. It was standing slightly proud, and a proud seam on a spiral tube is a leak path with a very long, very thin cross-section.
That is the whole discipline of spiral duct in one anecdote. The machine is simple. The strip width arithmetic is not, and nobody checks it because it is arithmetic.
A spiral tube is a strip wound helically around a fixed mandrel and locked along the moving edge. The strip enters from a decoiler, passes through a set of forming rolls that curl it to the tube radius, then through a lock-forming head that turns one edge into a pocket and the other into a flange, and finally into a closing head that presses the two together as the tube advances. The tube leaves the machine as a continuous helix, which is why the machine can run a 6 m length as easily as a 1 m one.
What varies between machines is not the concept but three things: how the radius is set, how the lock is closed, and how the tube is cut off.
On the simpler machines the radius comes from a set of interchangeable forming rolls and a mandrel sized to the diameter. On the better ones the rolls are on eccentric shafts and the diameter is set with a handwheel and a scale, which cuts changeover from an hour to fifteen minutes. If you make more than four diameters regularly, that difference is worth more than most of the other features on the spec sheet.
The standard spiral lock is four layers of metal folded together, closed by rollers rather than welded. That is the reason spiral duct is cheap and fast: no welding, no filler, no post-weld cleaning. The trade-off is that the seam is mechanically tight, not metallurgically sealed, and its airtightness depends entirely on how well the four layers are nested.
Flying shear or a stationary cut after a pause. The shear produces a clean end but leaves a small tab where the cut crosses the lock. On ducts that will be flanged, that tab needs to be dressed or it will hold the flange off square by 1 to 2 mm, which then shows up as a gap at the joint on site.
This is where most quality problems start, so it is worth writing down rather than remembering.
The developed width of the strip is the circumference of the tube plus the lock allowance, plus a small correction for the fact that the neutral axis of the bend sits slightly inside the mid-thickness.
For a four-ply spiral lock on galvanized steel in the 0.5 to 1.0 mm range, the allowance works out at 26 to 30 mm depending on the specific roll set. A working formula that has served me well:
Strip width = 3.1416 x (D + t) + 28
Where D is the nominal outside diameter and t is the strip thickness, both in millimetres. The 28 mm is the lock allowance and it is not universal. Check it against the roll drawing for your machine, and if the drawing is missing, measure it: run a strip through the lock head only, close the seam by hand, and measure the pitch of the resulting helix against the circumference.
| Nominal diameter | Strip thickness | Calculated strip width | Standard coil width used | Weight per metre (approx.) |
|---|---|---|---|---|
| 100 mm | 0.5 mm | 344 mm | 345 mm | 1.28 kg/m |
| 150 mm | 0.5 mm | 501 mm | 500 mm | 1.84 kg/m |
| 200 mm | 0.6 mm | 658 mm | 660 mm | 2.95 kg/m |
| 315 mm | 0.6 mm | 1019 mm | 1020 mm | 4.65 kg/m |
| 400 mm | 0.6 mm | 1286 mm | 1285 mm | 5.90 kg/m |
| 500 mm | 0.8 mm | 1600 mm | 1600 mm | 9.85 kg/m |
| 630 mm | 0.8 mm | 2009 mm | 2010 mm | 12.40 kg/m |
| 800 mm | 1.0 mm | 2546 mm | 2545 mm | 19.70 kg/m |
Note how 500 mm and above push past the 1500 or 1600 mm working width of a standard machine. That is not a mistake in the table. Those diameters are made either on a wider machine, or by slitting and welding two strips, or by running the tube in two halves. If your order book has a lot of 630 and 800 mm work, buy the 1600 mm machine and accept that you will be buying wider coil. If it is occasional, buy the 1500 mm machine and sub-contract the big sizes.
The four-ply lock is a continuous helical rib. Structurally, that rib is doing the same job as a stiffener ring, except it is continuous and it is at a shallow angle to the hoop direction. The result is a tube that resists ovalisation far better than its gauge suggests.
Practical numbers from our own test rig, for a 500 mm tube spanning 3 m on simple supports with a 400 Pa internal pressure and a 5 kg point load at mid-span: 0.6 mm spiral deflects about 4.5 mm, 0.8 mm about 2.8 mm, and a 0.8 mm longitudinally seamed tube of the same diameter about 6.2 mm. That last figure surprises people. The longitudinal seam is a straight line of four layers running the length of the tube, and it does nothing for hoop stiffness at all. The spiral seam does, because it wraps.
This is why spiral duct is the default for return air, exhaust and general ventilation in diameters above 300 mm, and why it is a poor choice for high-pressure supply where the fitting count is high and the fittings, not the straight, dominate the pressure loss.
Stainless and aluminium are both awkward on a lock seam. Stainless work-hardens at the fold and cracks; aluminium is soft enough that the four layers gall and the seam will not close cleanly at thicknesses below 0.8 mm. For those materials the usual answer is a welded seam, TIG for stainless, MIG or TIG for aluminium, run continuously along the helix as the tube advances.
A welded spiral tube runs at a third to a half the speed of a locked one, needs a welder who knows what they are doing, and needs post-weld cleaning if the material is 304 or 316. It is also genuinely airtight, which matters in food, pharma and chemical exhaust where a lock seam is simply not acceptable regardless of how well it is formed.
If your work is mostly galvanized HVAC, the locked machine is the right purchase. If you are chasing stainless process duct, look at a spiral tubeformer configured for welding from the start, because retrofitting a weld head onto a lock-only machine is rarely economic.
If the coil was slit from a master coil with a slight edge wave, the strip will curve as it pays off. The tube then comes out as a corkscrew. Operators try to fix this by adjusting the entry guide, which makes it worse. The test is simple: pull 5 m of strip onto the floor and look along it. If it bows more than 10 mm over 5 m, reject the coil. You cannot form a straight tube from a curved strip.
Too loose and the strip balloons and skips; too tight and the strip stretches, which changes the developed width mid-run and makes the seam pitch drift. Set the brake so the strip has a slight, even tension with no sag between the decoiler and the entry rolls, and check it again after the coil has lost half its weight. A coil that was right at 2 tonnes is wrong at 500 kg.
The closing head takes the most load of any station because it is squeezing four layers. When the rolls wear, the seam closes but does not bottom out, and the tube measures correct on diameter while the seam sits proud. Check with a 0.05 mm feeler at three points along a metre of seam. If the feeler enters anywhere, the seam is not closed.
A blade that is 0.3 mm out over the tube diameter gives an end that is visibly off square, and every flange fitted to it inherits the error. Check by standing the tube on a flat plate and measuring the gap under the low side. More than 1 mm on a 500 mm tube is out of tolerance for most specifications.
Suppliers quote changeover in minutes and they usually mean the mechanical part. Real changeover on a 1500 mm machine with fixed roll sets, for one operator working alone: 35 to 50 minutes for a diameter change including the test tube, and about 10 minutes for a gauge change at the same diameter. Machines with eccentric roll adjustment cut the diameter change to 15 to 20 minutes. Neither figure includes fetching the coil, which on a busy shop floor is often the longest part of the job.
If you run small batches, the changeover time is what sets your cost per metre, not the running speed. A machine that runs at 45 m/min but takes an hour to change over is worse than one that runs at 30 m/min and changes in fifteen minutes, for any batch under about 200 m.
Run three tubes at the smallest, middle and largest diameter you will actually sell. For each, measure the diameter at both ends and the middle in two planes, check the seam with a feeler, stand the tube on a plate to check squareness, and run a smoke or pressure test if the specification calls for one. Then change to a different diameter and time the changeover with a stopwatch, with the operator working at normal pace rather than demonstrating.
That last test tells you more about whether the machine suits your order book than the whole specification sheet. For the wider picture on how a spiral line fits alongside rectangular duct equipment, the spiral duct forming machine overview covers the configurations and the auxiliary equipment most shops end up needing.
Not directly. Oval and flat-oval duct need a dedicated line, because the forming rolls have to produce two different radii in the same section. Some shops run a round tube and then flatten it in a press, which works for short lengths and low pressure but produces inconsistent dimensions and should not be used for anything that has to fit a standard fitting.
Most 1500 mm machines bottom out at 80 to 100 mm, and the practical limit is set by the mandrel size rather than the rolls. Below 100 mm the mandrel becomes so slender that it deflects under the closing pressure, and the tube comes out with a taper. If you need 80 mm and below regularly, buy a machine designed for small diameters rather than pushing a large one.
On a well-set machine, the cut-off loss is one tube length of overlap plus the trim at the start of the coil, which typically works out at 0.5 to 1.5 percent of material. If you are seeing more than 3 percent, look at the cut-off and at whether the operator is running the coil down to the last metre instead of planning the batch around the coil weight.
For general ventilation at low pressure, no. Above roughly 500 Pa, or wherever the specification names a leakage class, yes. The sealant goes into the pocket before closing, and the machine needs a small applicator to do it consistently. Applying sealant to a finished seam from the outside is a waste of time and it looks it.