Every few months a buyer sends me the same message in different words: the spiral machine is cutting short. Sometimes it is 15 mm on a 3 metre length, sometimes 40 mm, and once it was a batch of 300 mm diameter pieces that came out 8 mm under diameter and 20 mm short, which the site fitter discovered only after the hangers were fixed. The machine is rarely the culprit in the way people expect. What has usually happened is that one of three length systems has drifted, and nobody noticed because all three are measured in millimetres and only one of them is on the screen.
I have set up and corrected spiral lines in a dozen shops, and the diagnosis is nearly always the same twenty-minute sequence. This is that sequence, with the numbers that make it useful.
On a spiral tubeformer there are three quantities in play. Confusing them is the root of most arguments.
Strip width is the coil width, fixed by what you bought. It sets the diameter.
Developed circumference is what the strip has to be to close into a tube of that diameter, including the lock seam allowance.
Cut length is what the measuring system decides, and it has nothing to do with the other two except that it is measured on a tube whose diameter those two defined.
For a lock-seam spiral the working relation is simple enough to do on a phone:
W = π × D + seam allowance
where D is the nominal outside diameter and the seam allowance is 25 to 30 mm for a standard four-ply lock, a little more for a heavy lock on 1.0 mm and above. A 1250 mm coil therefore makes a duct of about (1250 − 27) / 3.1416 = 389 mm, which is why 1250 is the most common coil width in the trade: it lands on the 400 mm nominal with a little room. A 1500 mm coil lands near 470 mm, a 1000 mm coil near 310 mm.
Two corrections that people skip. First, use the mean diameter, not the outside diameter, when the wall is thick: on 1.2 mm material at 400 mm the difference is 1.2 mm of circumference, which is small but not nothing. Second, the lock does not consume the allowance symmetrically. The formed seam takes up 2 to 4 mm of the strip depending on roll condition, so a machine that has just had new lock rolls fitted will produce a slightly larger diameter from the same coil than it did last month. That is normal, and it is why diameter should be measured, not assumed, after any roll work.
Six things, and I have met all six in the field.
Coil width tolerance. A slit coil is not exact. Plus or minus 1 mm is normal, and I have measured coils 3 mm under nominal from a supplier who was cutting to the low side to save material. Three millimetres of width is about 1 mm of diameter and, more importantly, it changes the circumference, which changes how the strip tracks.
Forming stretch. The outer fibre of the lock stretches as it is closed. The effect is small on galvanised, larger on aluminium, and largest on stainless where the material work-hardens. A shop that switches from 0.5 mm galvanised to 0.6 mm stainless without re-checking diameter will find the tube comes out 3 to 5 mm larger than expected because the lock does not pull down as tightly.
Encoder wheel slip. This is the big one. The measuring wheel runs on the formed tube or on the strip, and its effective circumference is what converts pulses into millimetres. A 100 mm wheel that has worn 1 mm is 1 percent small, and 1 percent on a 3,000 mm length is 30 mm. I have seen exactly that, with the operator convinced the machine had lost its parameters, when the fix was a 40 dollar wheel.
Cut-off timing. A flying cut-off has to accelerate, cut and return. If the acceleration ramp is timed for a heavier tube than the one running, the knife arrives late and every piece is long. If the tube is lighter than the ramp assumes, pieces are short. This shows up as an error that changes with diameter and line speed, and it is the reason a machine can be perfect on 400 mm and 20 mm out on 200 mm.
Blade kerf and cut quality. A shear cut removes 0 to 1 mm depending on blade clearance. A plasma cut removes 2 to 3 mm of kerf and leaves a bevelled edge that a tape measure reads differently depending on where you place it. If you change from shear to plasma and keep the same length offset, every piece gets shorter by the kerf.
Thermal expansion and coil set. A coil that has been in the sun is longer per wrap than one from a cold store, and a coil with a strong set fights the forming head, which slightly changes the effective feed. This is why the first ten pieces after a coil change are worth measuring.
| What you see | Most likely cause | Check | Fix |
|---|---|---|---|
| Every piece short by the same amount | Cut length offset or kerf change | Measure 5 pieces at 1,500 mm; compare spread with offset | Adjust length offset by the constant error |
| Error grows with length, about 1% | Encoder wheel wear or slip | Wrap the wheel in tape, mark and roll one revolution on the bench | Replace wheel, recalibrate pulses per metre |
| Random scatter of 5 to 15 mm | Strip slipping in the feed, or coil drag | Watch the strip between decoiler and head during a cut | Adjust brake tension, check pinch roll pressure |
| First piece long, rest correct | Acceleration compensation | Run 10 pieces and measure each | Set the ramp parameter for the tube mass |
| Correct on one diameter, off on another | Cut-off timing set for one tube mass | Run 3 diameters at the same line speed | Re-tune ramp, or reduce line speed on small sizes |
| Diameter small, length also short | Coil width under nominal | Measure coil width at both edges, three places | Reject or re-slit the coil, update the width parameter |
| Cut end out of square by 2 to 4 mm | Blade clearance or clamping | Look at the cut face for a bright tear band | Reset clearance to 5% of thickness, check clamp pressure |
Do this with the machine warm, because a cold hydraulic system moves at a different speed and you will chase the number twice.
Set the target to 1,500 mm and run five pieces at normal production speed, not at the slow jog setting. Measure each on the bench with a steel tape against a stop, not with the same tape held by two people in the air. Write the five numbers down. If the spread is under 3 mm and the average is off by a constant, adjust the length offset and you are done. If the spread is over 5 mm, stop adjusting offsets and find the mechanical cause, because no parameter will fix scatter.
Next run three pieces at 3,000 mm. If the error doubles with the length, it is a scale factor, which means the pulses-per-metre calibration or the wheel diameter. If the error stays constant, it is an offset, which means kerf or the knife position.
Then run the same 1,500 mm target at half speed. If the length changes, you have an acceleration compensation problem and the parameter has to be set for the tube mass you actually run, not for the heaviest one the machine can handle.
Write the results in the machine log with the date, the coil batch and the diameter. Two months later, when the same complaint arrives, that page turns a two-day argument into a five-minute check.
Blade clearance on a spiral cut-off should be about 5 percent of material thickness: 0.03 mm on 0.6 mm sheet, 0.05 mm on 1.0 mm. Too tight and the blade chips; too loose and the tube tears, leaving a ragged edge that measures short because the tear pulls material away from the cut line. If the cut face shows a bright band a third of the way through the thickness, clearance is too large.
The clamp matters as much. A tube that rotates or shifts during the cut loses 2 to 5 mm on every piece and the blade life halves. Check the clamp pads for galvanised build-up weekly; a 0.2 mm layer of zinc on the pad is enough to let the tube creep.
Flying cut-off machines also need their return spring or servo tuned so the carriage is back at zero before the next length is complete. When it is not, the machine either waits, which costs cycle time, or cuts short, which costs the batch.
Edge camber is the quiet killer. A coil with 3 mm of camber over 2 metres will make a tube that walks sideways in the forming head, and the operator compensates by steering the strip, which changes the lock overlap and therefore the diameter. Measure camber by laying 2 metres of strip on the floor and pulling a string along one edge. More than 2 mm and the coil is not worth running on a machine you care about.
Telescoped coils do the same thing in a different way: the wrap tension varies, so the strip enters the head at varying tension and the lock is sometimes tight and sometimes loose. If you see a wave pattern in the seam every 300 to 400 mm, look at the coil, not the rolls. A power slitting machine that slits to width in-house will not fix camber, but it will let you control width to 0.3 mm, which removes one variable from the list.
When you move to a new diameter, do not trust the table in the manual. Run 1 metre, cut it, and measure both the diameter at three places and the lock overlap. Overlap should be 8 to 10 mm of visible engagement on a standard lock. Less than 6 mm and the seam will open under pressure; more than 12 mm and the rolls are working too hard, which shows up as a bright burnished line on the outside of the tube.
Then check the tube for straightness over 3 metres. A spiral tube with a consistent bow of more than 3 mm in 3 metres usually means the forming head angle is slightly off for that width, and it will make every elbow fit-up a fight. Correct it before you run the batch, not after.
If you are specifying a new machine rather than fixing an old one, the cut-off arrangement matters more than the headline speed. A machine that cuts with a plasma torch on a 0.6 mm tube will always have a wider length tolerance than one with a flying shear, because the kerf moves with the torch height. The specifications on a spiral duct forming machine usually quote the tolerance for the shear version, so ask which cut-off the number refers to. The difference between a fixed-mould head and a band-type head also changes how much the diameter moves when you change material, which is worth understanding before you buy: the notes on a mould type spiral tubeformer explain the trade-off better than most sales conversations do.
A spiral line is mechanically simple and therefore easy to neglect. Four items cover most of the failures I have seen.
Measuring wheel: check diameter monthly with a caliper, replace below nominal by 0.5 mm, clean the surface weekly because zinc dust packs into the knurl and turns the wheel into a smooth roller.
Lock rolls: check the seam profile every shift on the first piece, and check the roll gap monthly with a feeler gauge. A worn male roll produces a loose lock, which changes both diameter and leak behaviour.
Feed and clamp pads: clean daily, replace when the grooves are gone. This is a 30 dollar part that causes 30 mm errors.
Hydraulic oil and filter: change at the manufacturer's interval, and check the pressure at the cut-off cylinder rather than at the pump. A lazy cylinder is often the real reason the knife arrives late.
None of this is glamorous, and that is the point. The shops that hold 2 mm on a 3 metre spiral cut are not running better machines than everyone else. They are measuring their own output every week and replacing a wheel before it becomes a warranty claim.
What length tolerance should I expect on a spiral line? With a flying shear, a healthy encoder and a stable coil, plus or minus 2 mm on lengths up to 3 metres is realistic. With plasma cut-off, expect plus or minus 4 to 5 mm because of kerf and edge quality, and specify accordingly if the duct is going into a prefabricated module where the length is critical.
Why does the diameter change when I change coil supplier? Width tolerance, yield strength and coating weight all move the lock take-up. Measure the first tube of every coil batch and adjust the width parameter rather than the forming angle. Two coils from different mills at the same nominal width can differ by 2 mm in diameter.
Can I cut spiral duct on a saw instead of a shear? You can, and for thick stainless it is sometimes the better choice, but a saw removes 2 to 3 mm of kerf and leaves burr that has to be dressed before the duct goes into a joint. Add the kerf to the length offset and budget the deburring time; on a 300-piece order it is a real cost.
Do I need to recalibrate after changing material thickness? Check, yes. The pulses-per-metre figure does not change, but the cut-off timing and the lock take-up do. Run three pieces at the new thickness and confirm the length before releasing the batch, especially when moving from 0.5 mm to 0.8 mm or above.