Pipe Wall Tolerance: An Absolute Band, Not a Percentage

The band is written in millimetres

Open a pipe standard and the wall tolerance is not a percentage. It is a pair of numbers in millimetres, or a single number applied to a nominal wall, and it does not scale the way a purchaser expects it to. A 2 mm wall might be permitted plus or minus 0.4 mm. A 10 mm wall gets plus or minus 1.2. A 25 mm wall gets plus or minus 2.7.

Those numbers are not arbitrary. They come from what an extrusion line can hold, and an extrusion line's errors are mostly fixed rather than proportional. Die concentricity is a fixed offset in millimetres. Sizing sleeve wear is a fixed number of microns per thousand metres. Melt temperature variation produces a wall change that depends on the drawdown, not on the wall itself. None of these get bigger because the pipe got bigger.

So the band grows with the wall, but sub-linearly, and the percentage it represents falls. That is the whole story of pipe tolerancing, and almost every argument about a rejected lot comes from one side reading the standard as a percentage and the other side reading it as a band.

Where the numbers in the standard come from

Most pipe standards use one of two shapes. Pressure pipe standards usually give an asymmetric band that never allows the wall to go thin, because thin wall is what fails the hydrostatic test: something like nominal to nominal plus a fixed allowance. Non-pressure and conduit standards usually give a symmetric band, because nothing is being pressurised and the wall only has to be stiff enough.

The fixed allowance in the pressure case is there to absorb the fact that a pipe that measures thin at one point around the circumference will fail there, and the failure is not gradual. A pipe rated at 16 bar with a 10 percent thin patch is not a 14.4 bar pipe; it is a pipe that will burst at the thin patch at some temperature and some time that you cannot predict from the outside.

Table 1 — The band widens with the wall but the percentage narrows, which is why the two readings of the same standard disagree.

Note what happens at the bottom of that table. At 2 mm the band is 0.8 mm wide in total. If the target is 2.0 and the band is plus or minus 0.4, the process has to sit inside a 0.8 mm window all shift, on a product running at thirty metres a minute. That is not generous, it is tight, and it is tight in a way that a percentage reading would hide.

Why a percentage band fails at small diameter

Suppose a buyer writes the specification as plus or minus 10 percent because that is what they remember from a large-diameter job. On a 25 mm wall that is 2.5 mm, which is close to the standard and easy to hold. On a 2 mm wall it is 0.2 mm, which is not.

A 0.2 mm total band on a 2 mm wall means plus or minus 0.1 mm. The ultrasonic gauge used to check it typically has a resolution of 0.01 mm and an accuracy of plus or minus 0.05 mm on a good day, with a couplant and a clean surface. So a third of the permitted band is inside the measurement uncertainty. Two labs will disagree about the same pipe, and both will be right.

The mechanical side is worse. To hold plus or minus 0.1 mm on a 2 mm wall you need the die concentric to better than 5 percent, the vacuum in the calibrator stable to a few millibar, the haul-off to within a fraction of a percent, and the melt temperature to within a degree. A line that does all of that is a line that is being run carefully for a reason, not a line that is running a commodity pipe.

This is why the standards use a band with a floor rather than a percentage. The floor acknowledges that there is a minimum error budget which no amount of care removes, and that below a certain wall the band has to stop narrowing.

Wall thickness is the measurement that decides the reject

In a typical pipe plant, wall is not the only dimensional check but it is the one that generates most of the rejections. Wall and ovality together account for roughly seven of every ten dimensional rejects; length and diameter make up most of the rest.

The reason is that wall is the variable with the most consequences. It sets the pressure rating through the hoop stress equation, it sets the ring stiffness through the second moment of area, it sets the socket fit, and it sets the weight, which is what the customer is actually paying for. A pipe that is 8 percent heavy is a pipe that costs 8 percent more resin and delivers nothing extra, and a pipe that is 8 percent light is a pipe with a claim attached.

Table 2 — Where dimensional rejects actually come from, in the order they should be investigated.

There is a useful discipline in that table. Wall and ovality are process variables; length and diameter are mostly equipment variables. If a lot is failing on wall, the answer is in the line settings. If it is failing on diameter, the answer is usually a worn sleeve, and no amount of process adjustment will fix it.

Ovality is a separate number for a reason

Ovality is the difference between the largest and smallest diameter at a section, usually expressed as a percentage of the nominal diameter. It is specified separately from wall because the two fail independently. A pipe can be perfectly on-wall around its whole circumference and still be visibly out of round, and a pipe can be round and have a wall that varies by 15 percent around the circumference.

The causes are different too. Ovality comes from the pipe being cooled unevenly, from sagging under its own weight while still soft, from being pulled through the haul-off before it is stiff, and from being stacked or coiled while it is still relaxing. Wall variation comes from the die, the drawdown and the melt.

The two interact in one specific way that catches people out. A two-point micrometer measures wall at two opposite points. On an oval pipe, those two points are not representative, and the reading can be comfortably inside tolerance while the thin points, ninety degrees away, are outside it. This is the single most common reason a pipe passes at the plant and fails at the customer's incoming inspection.

Measuring it: the check that settles arguments

The method that ends disputes is a full-circumference ultrasonic scan at three stations along the length, eight points per station, on a pipe that has been conditioned at 23 °C for at least an hour. That gives twenty-four wall readings, a maximum, a minimum, a mean and a spread, and from those you can compute both the wall tolerance and the eccentricity.

Table 3 — Measurement methods, and the specific defect each one is blind to.

Weight per metre deserves a mention because it is the check that most plants actually run, and it is the one that hides the most. It gives you the average wall over the cut length with great precision, and it tells you nothing about whether that average is made of a uniform wall or a wall that is 20 percent thick on one side and 20 percent thin on the other. Both pipes weigh the same. Only one of them passes the hydrostatic test.

What actually moves the wall

If a line is drifting out of wall tolerance, there are five levers and they are worth checking in this order, because the first one accounts for more drift than the other four combined.

1. Drawdown. Wall moves inversely with line speed for a fixed extruder output. A 1 percent error in line speed is a 1 percent error in wall, and line speed is the setting most often changed by hand and least often recorded.
2. Extruder output drift. Screw wear, a partially blocked screen pack or a barrel heater going soft will move output by a few percent over a shift without any alarm.
3. Die concentricity. This shows as a once-per-revolution pattern in the wall, and it is corrected with the die adjustment screws, not with the process.
4. Calibrator vacuum. Too little vacuum and the pipe does not seat against the sleeve, so the wall floats. Too much and the pipe is dragged, which thins it.
5. Melt temperature. A 10 °C rise lowers the melt viscosity, raises output at fixed screw speed and increases drawdown, so the wall falls twice for the same change.

The practical rule is to log line speed and extruder output every hour, not just at start-up, and to plot the wall against the ratio of the two. When the ratio is stable and the wall is not, the problem is downstream in the calibrator or the die. When the ratio itself drifts, the problem is upstream and no die adjustment will help.