Profile Cut Length Accuracy: Error Budget, Encoders and Thermal Contraction

Why cut length is harder than it looks

Cutting a profile to length sounds like the simplest operation on the line. The profile arrives at a known speed, a saw cuts it, the length is set by a number in the controller. What could go wrong is everything downstream of that number.

A profile quoted at 3,000 ±2 mm is a demanding tolerance when you consider that the same profile is 3,000 mm long when it is at 45 °C and 2,988 mm long when it has cooled to 20 °C. That 12 mm difference is six times the tolerance, and it happens after the saw, where the controller cannot see it.

The three cutting methods

Table 1 — Cutting methods. The tolerance column is what the method can hold on a stable line, not what the brochure claims.

Stop-and-cut gives the best length because the profile is stationary during the cut, so there is no synchronisation error. It costs throughput: every cut stops the line, and on a 30-second cycle a 3-second stop is 10% of output. Flying cutters avoid the stop but introduce a synchronisation error between the saw carriage and the profile, which is where the extra millimetre comes from.

Where length error comes from

An error budget is the only way to improve cut length, because the contributions have different causes and different fixes. The table below is a realistic budget for a flying saw on a PVC profile at 6 m/min.

Table 2 — An error budget for a flying saw at 6 m/min. The thermal term is not an error in the same sense — it is predictable, and it should be compensated rather than reduced.

The lesson in the budget is that the controllable errors add up to about ±2 mm while the thermal term is an order of magnitude larger and entirely predictable. Plants that spend money on a better saw before they compensate for contraction are solving the smaller problem.

Thermal contraction after cutting

Every thermoplastic has a coefficient of linear thermal expansion roughly twenty times that of steel. A profile cut at 45 °C and measured at 20 °C is shorter than it was on the saw, by an amount that depends on the material and the temperature difference.

Table 3 — Contraction over a 3 m length cooling from 45 °C to 20 °C. The range reflects the spread between grades; the actual value for a specific compound should be taken from the supplier's data sheet.

Two things follow. First, the cut length must be set longer than the nominal by the expected contraction, or the parts must be measured and accepted at the line temperature. Second, the temperature at the saw must be stable, because a profile that arrives at 40 °C one hour and 55 °C the next will produce two different final lengths from the same setting.

Stabilising the saw temperature is usually cheaper than tightening the saw. A longer cooling run-out, or a spray section before the cutter, brings the profile to a consistent temperature and removes most of the day-to-day length drift.

Encoder and haul-off speed

The cutter's length reference comes from an encoder. If the encoder measures the haul-off belt rather than the profile, every source of slip between belt and profile becomes a length error. Belt slip on a wet profile is not small: 1% slip on 3 m is 30 mm.

The better arrangement is a measuring wheel running directly on the profile, close to the cutter, with a known circumference and a high-resolution encoder. That measures what is actually arriving at the saw rather than what the puller thinks it is delivering.

Table 4 — Length reference options. Moving the encoder closer to the profile always reduces the error, because it removes a transfer step that can slip.

Belt slip also has a second effect: it makes the cut length vary with line speed, because slip increases with tension. If the length error changes when you change speed, the encoder is measuring the wrong thing.

When to cut on the fly and when to stop

Table 5 — Choosing between the two. The decision usually comes down to whether the tolerance is tighter than ±1 mm and whether the line runs long enough for throughput to matter.

A hybrid is common on large profile lines: the haul-off slows rather than stops, and the saw cuts during the slow phase. That gives most of the length accuracy of stop-and-cut with a smaller throughput penalty, at the cost of a more complex control scheme.

Setting a cut-length tolerance

A cut length tolerance should be set from what the assembly needs, then checked against what the process can hold. Quoting ±1 mm on a 4 m PP profile that contracts 13 mm between the saw and the assembly bench is a specification that cannot be met without controlling temperature.

Table 6 — Typical cut length tolerances by application. The tolerance should always be quoted at a stated temperature, otherwise it is not a specification.

Measuring length properly

Length is the easiest dimension to measure badly. Three practices cover most of the error: measure at 23 °C after at least 30 minutes of conditioning, measure on a flat surface with the profile supported over its whole length, and measure the same way every time.

A 3 m profile measured on a bench with 200 mm overhanging each end will read short, because the ends sag. A profile measured immediately off the line will read long, because it has not contracted yet. Neither error is large individually, but together they can consume a tolerance that looked generous on the drawing.