Hollow Profile Extrusion: What Actually Controls Wall, Tolerance, and Failure in Plastic Tube and Pipe

Any engineer who has watched a multi-cavity tube drift out of spec knows the hard part of extrusion isn't pushing plastic through a hole. It's controlling what happens inside the void, and inside a void you can't see while the line is running. That cavity is the whole problem in hollow profile extrusion: it's what separates a tube or multi-lumen section from a solid rod, and it's where every hard tolerance and every field failure starts.This is a working guide to how hollow sections are formed, what wall and tolerance you can realistically hold, where the process fails, and how to tell a capable supplier from one that's going to learn on your tooling.

In solid profile extrusion the die does one job: shape a continuous cross-section as the melt leaves the barrel. Introduce a cavity and the whole flow problem changes. A pin or mandrel now sits inside the die, the molten polymer has to split, flow around it, and knit back together downstream. That knit line, the seam weld, is the single most under-discussed weak point in hollow plastic profile extrusion, because it's a region where two melt fronts rejoin under less-than-ideal pressure and temperature, and it never quite has the strength of the surrounding wall.

The void creates a second problem. The mandrel has to hold position while a viscous melt at a few hundred bar pushes against it, and any deflection of that pin shows up directly as uneven wall. So before anyone talks tolerances, understand the real constraint: a hollow section is only as good as the mechanical stability of the mandrel-and-die set forming its inside. If you're starting from the basics of how thermoplastic is melted and shaped, our overview of what plastic extrusion involves covers the upstream fundamentals; here we stay on the hollow-specific failure surface.

Every station constrains the next, and an early handoff that goes wrong is one no downstream correction recovers.Pellets melt and homogenize in the extruder, then the melt enters a crosshead or in-line die where the mandrel forms the bore. The shaped, still-molten tube exits into a vacuum sizing tank, where a calibration sleeve and controlled vacuum pull the outer wall against a reference diameter while the polymer is still soft enough to set its final OD. Cooling follows, the part is measured in-line, then marked and cut to length. The draw-down ratio, how much the melt is stretched between die and sizing, is one of the quieter levers on final wall. The same line geometry also supports a co-extrusion hollow profile, where a second extruder lays a different resin onto the bore or skin: a barrier inner layer, say, or a regrind core under a virgin skin, without changing how the cavity is formed.

A note that explains a lot of equipment choices: high-output pipe and tube lines lean heavily on single-screw extruders, still the workhorse machine type for simple thermoplastic pipe and film because they're simpler, cheaper to run, and energy-efficient on uniform profiles, even as twin-screw grows faster on filled and rigid-PVC compounds (Mordor Intelligence).That's why a shop optimized for continuous tubing usually isn't the same shop optimized for heavily filled compound. For the geometry-and-resin breakdown of how a continuous tube actually takes shape, our piece on how plastic tube extrusion forms hollow products goes station by station. Each downstream station can only work with what the previous one handed it: bore quality is decided at the die, not the sizing tank.

Here's the trap most drawings fall into. Two tubes can share an identical outer diameter and still have completely different wall distributions, one centered, one with the bore pushed off to one side. The OD callout tells you nothing about that. Concentricity does, and it's concentricity, not diameter, that decides whether a pressure tube fails early or a catheter bends the way it should.

 

This is also where concentricity and wall-thickness tolerances in tube extrusion start fighting each other on the drawing. If a print independently dimensions OD tolerance, wall tolerance, and concentricity, those three controls collide: the coaxiality of the inner and outer surfaces already governs how wall can vary relative to OD, so stacking a separate wall tolerance on top often over-constrains the part into something no process can hit economically. That's the variable most suppliers won't raise during quoting, and it's the kind of thing a real DFM review catches before tooling is cut, not after the first sampling run misses.

 

On measurement: the answer to how you even know the wall is even is in-line gauging. On precision medical lines, tubes leaving the extruder typically pass through a ring of four or eight ultrasonic transducers as they enter the quench tank, and inline ultrasonic systems sample on the order of 2,000 measurements per second around the circumference (GaugeAdvisor).Small-bore tubing especially lives or dies by how well that array stays centered. Spot-checking a cut sample at the end tells you what one slice looked like; it does not tell you the wall held for the whole spool.

Numbers, because vague reassurance helps no one. On medical-grade multi-lumen tubing, outer-diameter control around ±0.025 mm is achievable, with each lumen held by its own mandrel pin and its own independently regulated air pressure. That per-lumen pressure control is what keeps one channel from collapsing while another stays round. We treat that figure as a capability statement from our own lines, not a catalog claim, and the honest caveat is that it's geometry- and resin-dependent: a thin-wall, high-modulus resin in a simple round holds far tighter than a soft, multi-cavity section.

The cost argument behind holding a tolerance this tight on multi-lumen tubing runs opposite to intuition. A self-centering die with closed-loop wall control doesn't just improve quality - by automatically optimizing wall and concentricity regardless of line speed, it narrows the wall-thickness tolerance band by about 0.05 mm and cuts material use by up to roughly 2.5% (KraussMaffei). On a line running 24/7, that material saving alone can pay back the better tooling in weeks. But that 2.5% is the equipment maker's figure under continuous production at specific diameters - what your part actually saves depends on resin price, line speed, and how complex the cross-section is, and the payback math only gets real once those are on the table. The point isn't the gadget; it's that tighter process control and lower unit cost are the same investment, not a tradeoff - which is the part a quote sheet never shows.

Bottom line up front: surface fracture and wall drift are shear-and-interface problems, not setpoints you can simply dial out, which is why the fixes have an order.

The defects engineers fear most all live at the die exit and all have specific, diagnosable causes. Sharkskin is a rough, scaly surface: surface melt fracture from excessive shear at the die land as the polymer's skin slips and grabs leaving the die. Push melt velocity higher and it degrades into gross melt fracture, where the whole extrudate distorts. Die drool is the gummy buildup that accumulates at the lip and periodically streaks the part; die lines come from contamination or damage inside the die land; die swell, the extrudate expanding past the die dimension as the polymer relaxes, is what forces sizing to exist at all.

 

There's a real disagreement worth raising, because the textbook framing is incomplete. The common line is that sharkskin is purely a process knob: drop the shear rate or raise the temperature and it goes away. The research disagrees in a way that matters for tooling: studies have found the construction material of the die land itself significantly affects melt fracture in LLDPE, contrary to what capillary-rheometer measurements predict, pointing to polymer-to-metal adhesion failure at the interface as a real driver, not just bulk flow (ResearchGate). So if you've optimized temperature and speed and still fight surface fracture, the die material and land geometry are on the table.

 

Here's how that actually plays out on a line. On a rigid-PVC profile run a couple of years back we chased a sharkskin pattern that wouldn't clear on temperature alone - we lifted the die-land temperature roughly 8 °C and got partial improvement, then dosed in the range of 600 ppm of a fluoropolymer processing aid, and the surface cleaned up inside a single spool while die drool dropped at the same time. That sequence is the rule of thumb we run by: try temperature and shear first, bring in a polymer processing aid at a typical 200–2,000 ppm if that stalls, and only open the die-material conversation when both fail. Dimensional instability over a run is almost always wall drift, which is exactly why the in-line measurement loop from the previous section has to close back to the die, not just alarm.

Most guides pick a lane and never put plastic and metal side by side, even though that's the real decision on a lot of programs. The mechanisms are genuinely different. Hollow profile extrusion in aluminum splits the billet through a porthole or bridge die and re-welds it around the mandrel, which is why aluminum hollow sections carry structural seam welds and why scrap-head lengths on the order of 500–1,000 mm are cut off at startup to clear weak weld zones, as is standard in aluminum hollow extrusion. Plastic holds its cavity differently: a pressurized or suction-fed air path through the mandrel keeps the bore from collapsing through sizing and cooling.

The judgment, not a shrug: for a multi-cavity fluid or wiring channel where per-cavity tolerance and chemical inertness drive the spec, plastic hollow extrusion wins this outright - aluminum can't independently control three small bores.For a structural member carrying mechanical load where wall can be generous and the seam weld is acceptable, aluminum is the right answer, and forcing plastic there is a mistake. The middle case, a rigid, dimensionally stable channel that still needs to be an insulator, is where rigid PVC or PC genuinely competes with aluminum, and where the material decision should be made on cost-per-meter and assembly method, not on reflex. Where exactly that cost-per-meter crossover falls depends on cavity count, wall spec, and run length, and pinning that down for a specific part is the first thing we work out in a DFM review before anyone commits to a die.

Resin follows application, not preference. Rigid PVC dominates building and drainage and accounts for a large share of all PVC resin consumption; PE and PP carry pressure and chemical lines; PC and ABS show up where impact strength, clarity, or dielectric behavior matter. For most custom plastic tubing material decisions the resin is set by service conditions, pressure rating, temperature, what fluid touches the bore, long before aesthetics enter. Plastic pipe demand has grown in the mid-single-digit-percent range annually, with Asia-Pacific leading volume and PVC/PE as the workhorse resins (Grand View Research), which mostly matters here as a reminder that resin supply and tooling availability follow those volumes.

Now the trap buyers fall into most often. The pitch "we use recycled content to save cost" sounds responsible and frequently is - but only under one condition. Controlled, single-material in-plant regrind, blended at a managed ratio, is legitimately usable and we run it. Post-consumer recycled content is a different animal: inconsistent melt flow index and contamination risk mean it typically fails pressure-pipe standards outright, and no blending recipe reliably fixes a feedstock whose viscosity you can't predict spool to spool. If a quote for a pressure-rated part leans on post-consumer recycled resin, that's the line to question. For non-pressure structural sections - channels, trim, enclosures - the calculus loosens considerably, and our extruded plastic channel profiles are exactly the class of part where controlled regrind earns its keep.

This is the part the rest of the article was building toward. The difference between a supplier who delivers and one who iterates on your dime is visible in five questions, and the answers separate capable shops fast.

None of these can be faked in a quote - they're either built into the shop or they aren't. For context on the kind of operation that answers "yes" across the board: we've run custom extrusion since 1998 out of Dongguan, hold ISO 9001, operate 40+ extrusion lines at roughly 2,000 tons annual output, handle sections up to 500 mm wide, and supply UL94-V0 flame-rated material where parts need it.A capable custom plastic tubing manufacturer should put numbers behind every row above without hesitating. For each of these, the right answer depends on your cavity count, wall spec, and annual volume - send the print and we'll answer all five in writing, before you commit to tooling, which is exactly what the DFM review below is for.