The ABS Plastic Extrusion Process: Parameters and Controls

An ABS profile that passes dimensional inspection can still fail in service, and the reasons are usually decided at three points on the line: how dry the pellets were, what the melt actually reached before the die, and how the profile was held and cooled while it was pulled through the sizing station. This article covers industrial profile extrusion of ABS. It does not cover FDM 3D printing with ABS filament, which is a different process with different failure modes and does not appear below.

Short answer. The ABS plastic extrusion process runs in this order: dry the resin → meter it into the feed throat → melt and homogenise it in the plasticising unit → remove volatiles at the vent, where the barrel has one → shape the melt in the die → hold the shape in a sizing station while cooling it → pull it at a haul-off speed synchronised to output → cut to length. Screw type, whether the barrel is vented, and whether sizing is dry vacuum or a water bath are equipment choices, not properties of the material, and they differ between lines that all run ABS successfully. Every step has a window, and every window has two failure directions rather than one. The rest of this article is about where those windows are published, which published numbers disagree with each other, and what evidence to ask an extruder for before you commit to tooling.

Writing "ABS" on a drawing does not specify a material. Acrylonitrile-butadiene-styrene is a family of terpolymers whose flow, softening point and impact behaviour vary widely with rubber content, the acrylonitrile/styrene ratio and molecular weight, all of which the resin producer sets for a particular processing route.

The current international designation system for this family is ISO 19062-1:2015, which classifies ABS moulding and extrusion materials using four designatory properties - Vicat softening temperature, melt mass-flow rate (MFR), Charpy notched impact strength and tensile modulus - alongside information on composition, intended application and/or method of processing, additives, colorants, fillers and reinforcing materials (ISO 19062-1:2015 catalogue entry). The companion part, ISO 19062-2:2019, covers preparation of test specimens and determination of those properties (ISO 19062-2:2019 catalogue entry).

Two points matter when you write this into a specification.

The older designation many drawings still carry, ISO 2580-1:2002, was withdrawn in November 2015 and replaced by ISO 19062-1:2015; ISO 2580-2:2003 was likewise replaced by ISO 19062-2:2019 (ISO 2580-1:2002 catalogue entry). Quoting the withdrawn number is not a small clerical issue, because the current version's designatory flow property is melt mass-flow rate in g/10 min, while the widely quoted MVR is melt volume-flow rate in cm³/10 min. They are different quantities, converting between them requires the melt density at test conditions, and neither number means anything without its test temperature and load. Do not swap the standard number and leave the old property in place.

The second point is what the designation does not do. It classifies; it does not certify that a given grade will run as your section. A grade selected for injection moulding is typically higher-flow than one intended for profile extrusion, and higher flow is a liability when the melt has to hold a section shape between the die exit and the sizing die under its own melt strength. Two grades that both answer to "ABS", and even two grades carrying the same designation, can behave differently on the same line. What to write instead of "ABS": the ISO 19062-1 designation plus the resin producer's grade name, the datasheet values your application depends on with their test conditions, and the producer's stated processing route. If you are still deciding between materials rather than grades, our framework for choosing an extrusion material covers that layer of the decision.

ABS is hygroscopic. Moisture still in the pellet when it reaches the melt does not evaporate quietly; it flashes at the die exit and shows up as splay, silver streaking, surface bubbles or internal voids.

Two published figures circulate for extrusion. One puts as-received moisture at roughly 0.3–0.8 %, with hopper drying around 80–85 °C for about four hours to bring residual moisture below 0.2 %, and warns that over-drying - too long or too hot - causes pellet agglomeration and yellowing (KingStar Mold, a mould and injection moulding supplier). A second gives 80–90 °C for at least 2–4 hours (Jieya, an extruder manufacturer). Both are commercial suppliers, neither states the grade or wall thickness its numbers apply to, and neither is a resin producer's processing guidance. They are working ranges in circulation, not verified settings.

You will also find a stricter threshold, below 0.1 %, quoted widely for ABS; it appears most often in injection moulding material. What we cannot support - and what nobody in the public record has established - is a clean rule that 0.1 % belongs to moulding and 0.2 % is sufficient for extrusion. The two processes impose different shear and residence histories, but that difference has not been shown to license a looser moisture target for extruded parts, and a hidden structural profile is not automatically more tolerant of volatiles than a visible one.

So the residual moisture target is not something to read off a blog table, in either direction. Three things set it: the resin producer's processing guidance for the specific grade, the method by which residual moisture is actually measured (and where and when the sample is taken, including any exposure between dryer and feed throat), and the acceptance criteria of the finished part. Ask for those three, and treat any single industry-wide percentage as a claim that has not been demonstrated.

Screw geometry is the one area where the two published sources agree, and they agree on one number only: both give a compression ratio in the 2.5–3 range. The mould-side source states that for a single screw and adds a length-to-diameter ratio (L/D, screw flighted length divided by screw diameter) of 18–22 for ABS profile work; the extruder-maker source states neither a screw type nor an L/D. A moderate compression ratio and a moderate L/D are not, on their own, evidence that the melt will leave the screw undegraded; screw speed, output rate and residence time all sit outside those ratios.

Barrel temperature is where the public record contradicts itself:

These are not small differences inside one window; they are two different windows. Plausible explanations - different product forms, grades with very different flow, or zone setpoints in one case against a target melt condition in the other - are hypotheses that neither source states and that we have not verified. What can be said is narrower and more useful: neither set of numbers is tied, item by item, to a named grade and a stated set of running conditions, so neither can be adopted as a setting. Averaging them would produce a number that describes nothing, and taking the narrower range because it looks safer is equally unfounded.

The window that governs is the one in your grade's datasheet. Values for ABS glass transition, viscous flow and decomposition onset also circulate from the same commercial source, and they should be read against the actual grade's datasheet rather than treated as fixed boundaries for the material family: ABS is amorphous, so there is no single melting point, and thermal degradation is a function of temperature and time at temperature under the actual shear and atmosphere in the barrel, not a line on a chart. The same source notes that ABS melt is markedly non-Newtonian, with viscosity responding more strongly to shear rate than to temperature - which means a screw turning faster is doing something different from a barrel set hotter, even when a thermocouple reads the same.

The defensible position: set zones from the grade's datasheet processing window, then establish what the melt actually reaches. Where the line carries a melt thermocouple at or near the screw tip, that reading is the variable to control and record. Where it does not, ask how melt condition is established instead, and expect the record to include the measurement location and method, its calibration status, the zone setpoints, the output rate and the screw speed together. A supplier who can only tell you their zone settings is telling you what the controller was asked to do, not what happened to your material.

"Bubbles mean the material wasn't dry" is the reflex, and it is wrong often enough to cost a shift. On a vented barrel, a vacuum port removes residual volatiles from the melt - but only while the port stays open and the vacuum holds. Not every profile line has one, which is the first thing to establish before this section applies to you at all.

Published guidance describes the failure in both directions at the adapter between screw and die: too hot, and the material degrades and discolours with unstable output; too cold, and the vent port blocks, vacuum drops, and the profile surface blisters (KingStar Mold; commercial source). The same source places adapter temperature at or slightly below the metering-section temperature.

The useful discipline is not a fixed order of checks but a way of narrowing. If bubbles appear on a line that was running clean an hour ago, the question is what changed in that hour, and the evidence that answers it is time-stamped: vacuum level trend and adapter zone trend on one side, and on the other the measured residual moisture of what is actually in the hopper now - not only the dryer setpoint log, since a stable setpoint does not prove stable moisture in the pellet. Our PVC extrusion process article covers venting behaviour on rigid PVC, where the volatiles are chemically different and the consequences of losing vacuum are different again; the diagnostic priorities there do not transfer to ABS.

Die Design, Die Swell and Section Balance

Because ABS melt is non-Newtonian and elastic, the extrudate does not leave the die at die dimensions. It swells. A published set of profile-die working values for ABS gives a die swell ratio of roughly 1.01–1.05, a die gap set at about 0.95–1.0 times the target wall thickness, and a polished cavity finish specified by arithmetic mean roughness (Ra) of 0.2 μm or less (KingStar Mold; mould supplier, commercial interest, no external citation, no stated grade or wall-thickness limits). Treat that as a published range to check against your own tooling records, not as a design rule - and note that the compression ratio quoted for a die flow channel is a different quantity from the screw compression ratio above, even though both are written as a bare number.

 

The number that matters more than any of these is the one nobody publishes: the ratio between the thickest and thinnest wall in your section. Melt takes the path of least resistance, so an unbalanced section drains flow into the heavy element at the expense of the thin one. Take an illustrative case of a 3 mm rib alongside a 1 mm web: the imbalance is visible on the drawing, and whether it can be brought into balance is a question for flow-balance analysis, die land and channel adjustment, and a trial - not something that can be settled from the ratio alone in either direction. What is worth saying without qualification is that the review is cheaper at the quotation stage than after the first trial.

Engineers moving a part from rigid PVC to ABS often carry over their calibration assumptions, and this is where that goes wrong - though the difference is in the equipment route a given shop uses, not a rule that ABS must be sized one way and PVC another.

For the dry vacuum route, published practice for ABS profile tooling describes vacuum slots cut densely into the calibrator cavity to hold the profile against the sizing surface, with a calibrator cavity dimensioned at roughly 1.005–1.008 times the finished profile outline to allow for shrinkage, and a calibrator generally shorter than the equivalent PVC profile calibrator. Cooling water is given at 14–18 °C with effective pressure above 0.2 MPa (KingStar Mold; commercial source, values to be checked against line records). The same source quotes a vacuum figure, but without stating whether it is absolute pressure or gauge vacuum, which makes it unusable as written - a vacuum specification that does not say what it is measured against cannot be transferred between lines.

The genuinely useful part is the two-sided failure description:

• Water too warm or pressure too low: the profile leaves the calibrator above its set temperature and deforms downstream.
• Water too cold: quenching locks in residual stress, the profile continues to shrink after the line, and haul-off drag rises to the point of surging.
• Vacuum too low: insufficient hold, the section deforms.
• Vacuum too high: haul-off load rises, the line surges, and pump life shortens.

Two things are worth separating here. The calibrator sets the shape; the cooling stages after it govern how much heat actually leaves the profile, and a line can be correct at one and wrong at the other. And post-extrusion shrinkage - a profile that measures in tolerance at the saw and out of tolerance later - has more than one possible cause: quench-induced residual stress from cooling water that was set to run the line faster is one; insufficient total cooling, orientation from haul-off, and the storage temperature the profile then sees are others. The practical consequence is the same either way: dimensional release should be based on measurement after the profile has stabilised under defined conditions, with the conditioning time and temperature agreed, rather than on the reading taken at the saw.

The table below is a structure for narrowing a cause, not a ranking. We have no frequency data for these defects on ABS profile lines, and neither does any public source we could verify.

The two-directional failure logic behind the cooling and venting rows comes from the published practice cited in the preceding sections; the discriminating-evidence column is an editorial framework for using records you should already have, not a claim about which cause is most common.

Aprofile can be dimensionally perfect and mechanically sound at the saw and still fail in service, because ABS has a specific and well-documented weakness: the polybutadiene rubber phase photo-oxidises.

In a peer-reviewed study of ABS exposed under UVA-351 fluorescent lamps at 50 °C in 360-hour steps - an exposure simulating daylight behind window glass - depth-resolved infrared analysis found the chemical signature of degradation confined to the surface: carbonyl formation and butadiene depletion were substantial at the exposed face, while at depths of about 100 μm and beyond the absorption ratios showed no considerable modification. Scanning electron microscopy found micro-cracking only on the irradiated face (Fiorio et al., Materials 13(1):212, 2020).

The mechanical consequence is out of proportion to that depth, but it is not uniform across properties, and the distinction matters for design. In the same study, measured to ISO 527-1A and ISO 180, ageing produced by far the largest change in strain at break, which collapsed to the point that failure occurred before yield; unnotched impact strength fell sharply as well; notched impact strength fell too, but by a much smaller factor; and tensile modulus rose slightly, by roughly 10 %. That is the classic signature of a thin, brittle, crack-initiating skin on an otherwise intact body - it removes ductility and the ability to absorb an impact on an unnotched face, while a specimen that already contains a notch was never relying on that surface. It is why a part can look sound in cross-section and still snap when flexed.

Two limits on reading this across to your profile: the specimens were injection moulded, not extruded, and the exposure simulated daylight through glass, not direct outdoor exposure. The mechanism transfers; the depth figure and the magnitudes do not, and neither is a constant for ABS in general.

The practical boundary that follows is the one we already apply in selection: unmodified ABS suits low-cost, high-rigidity, indoor applications. Where UV resistance, long outdoor exposure and colour stability are required, ASA is the usual candidate material, and the choice between an ASA grade and an ABS/ASA combination is a selection question to settle with the resin producer's data for the specific service environment, not something this article can settle in the abstract.

When a supplier hands you a weathering report, the question is not how many hours it ran. It is whether the test conditions permit the conclusion being drawn - and for ABS specifically, several common shortcuts do not hold.

A study conducted at the NIST SPHERE facility found the apparent activation energy for yellowing of ABS to be 31 ± 2 kJ/mol, substantially higher than polycarbonate at 21 ± 2, SAN at 19 ± 1 and PBT at 16 ± 4 kJ/mol, with the comparison materials drawn in part from the authors' earlier work in the same programme (Pickett et al., Polymer Degradation and Stability 181 (2020) 109330, NIST copy). Because ABS responds to temperature differently from the materials it is often benchmarked against, a chamber temperature that is "standard" for one material is not neutral for ABS.

Three further findings from the same work change how a report should be read:

• Reciprocity does not hold for ABS colour change. The combined irradiance-response slope was 0.79 ± 0.09, falling to 0.67 at 70 °C. Extrapolating from a high-irradiance test to real-world low-irradiance service therefore under-estimates the degradation rate - the error runs in the unsafe direction. This is measured for colour change, not for mechanical properties.
   
• Lamp choice distorts cross-material comparison. Under a metal halide source, the yellowing rate normalised to each material's own outdoor rate at matched UV dose was 0.71 for ABS against 1.19 for a PC/PBT blend - so that lamp accelerates the two materials relative to reality by factors differing about 1.7-fold. That ratio is a statement about the lamp's spectrum, not about absolute yellowing speed and not about service life.
   
• Very dry chambers exaggerate ABS degradation. ABS showed faster colour shift and gloss loss at relative humidity of 10 % or below, for the specimens tested.

So a weathering report on ABS should state, at minimum: the lamp's spectral power distribution, the chamber temperature, the relative humidity, the irradiance and total UV dose (not only elapsed hours), the specimen formulation and form, the property measured and how - colour, gloss, impact, strain - and the failure threshold being claimed against. Without the first three, the report cannot support a service-life claim; without the rest, it cannot be compared to anything, and none of it substitutes for correlation against real outdoor exposure of the same formulation.

Disclosure on this source: the work was funded by SABIC-HPP as part of a NIST consortium project, two authors are SABIC-HPP employees and one a consultant under SABIC-HPP contract, with the NIST co-author declaring no financial interest. The specimens were 3.2 mm injection-moulded chips containing 3 % coated rutile TiO₂ and formulated with no stabilisers - not extruded profiles. The authors also state that reproducing the effect of rain remained an unsolved problem in accelerated testing at the time of writing, which is a limit on outdoor extrapolation from that work.

The assumption behind most regrind negotiations is that recycled content makes a part worse across the board. The published evidence is more specific than that, and the comparison baseline is the part most often lost.

In the ABS ageing-and-recycling study cited above, mechanically recycling already-aged material - grinding, re-extruding and re-moulding it - increased strain at break and unnotched impact strength relative to that aged material, with a slight reduction in tensile modulus. The authors attribute this to two mechanisms: the brittle photo-oxidised skin is broken up and dispersed through the bulk, and physical ageing is erased by the reprocessing heat history. The finding is recovery from a degraded state, not an improvement over virgin resin, and it did not hold uniformly across the sequence - the study ran three UV ageing steps with two mechanical recycling steps between them, and the second recycling gave a lower strain at break than the first rather than repeating its gains. Notched impact performance recovered far less, and yellowing accumulated across reprocessing. The authors' own framing is that recycled ABS suits parts without demanding requirements for high impact strength, toughness and colour purity (Fiorio et al., 2020).

That turns the purchasing question into a set of screening questions rather than a release. Two of them are quick: does the part carry a notch, a sharp internal corner or a snap feature that concentrates stress, and does it have a colour requirement? A "no" to both narrows the risk but does not clear it, because the study says nothing about the regrind you would actually be buying. The rest of the list is about provenance and control: where the regrind comes from and whether it is in-house or external, how contamination and colour carry-over are excluded, whether lots are traceable and the thermal history is known, at what proportion and in which part of the section it is used, and how the finished profile is verified against the properties the application depends on. Finally, note the limit on the evidence itself: the study used injection-moulded specimens, so it does not convert into a percentage regrind recommendation for extruded profiles.

"Does it have antioxidant?" is a weak question, because the answer rarely predicts behaviour.

In the same study, three formulations were compared across the ageing and recycling sequence: a commercial ABS with no additional stabiliser; the same resin with 4 % of a masterbatch carrying a single hindered-phenol primary antioxidant; and the same resin with 4 % of a masterbatch carrying two phenolic antioxidants plus a phosphite secondary antioxidant. The single-additive formulation was only marginally better than the unmodified resin, while the multi-component package showed markedly higher oxidation onset and peak temperatures and held that advantage through the sequence (Fiorio et al., 2020).

Read that with two qualifications the authors' own data imposes. The base resin is a commercial grade, which does not mean it was additive-free; the comparison is between no additional stabiliser and two specific added systems. And the two masterbatches differ in composition and in total active content at the same 4 % loading, so the result compares two packages as supplied rather than isolating synergy at equal dose. The same work notes that in un-aged material the additives slightly reduced strain at break, behaving as inclusions - stabilisation is a trade, not a free improvement - and it measures thermo-oxidative stability, which is not the same property as UV colour retention. As elsewhere in that study, the specimens were injection moulded.

What to request from a supplier, then, is a description of the stabiliser system - primary antioxidant, secondary antioxidant, and where relevant a UV absorber or hindered amine light stabiliser - matched to the stated service environment, rather than the phrase "contains antioxidant" on a specification sheet.

Each question below has a weak answer that sounds fine and a substantive one. The evidence column lists records a controlled line generally produces; they can be redacted, and an equivalent record that serves the same purpose is a good answer. A supplier who documents differently is not thereby uncontrolled - but should be able to show how the same thing is established.

Two things are worth stating plainly about the numbers in this article.

What is ours and what is not. This article presents no ABS line data. No process capability, energy or first-pass-yield figure appears here for ABS, and none of the process judgements above is offered as our own measured experience on ABS. Figures we have previously published for process capability and specific energy consumption come from our rigid PVC profile lines and are not transferable to ABS - different material, different line, different control problem. Capability data is meaningful per product family and should be requested that way.

Where the published numbers came from. The process parameter ranges quoted above come from two supplier websites, both with a direct commercial interest. Neither ties its numbers, item by item, to a named grade, a wall thickness and a stated set of running conditions, which is what would make them adoptable. They appear here as published working ranges that disagree with each other, which is the honest description. The weathering and recycling findings come from peer-reviewed and government-laboratory work on injection-moulded ABS specimens, not extruded profiles; the mechanisms transfer, the numbers do not.

We have not purchased the full text of the ISO standards; the statements above come from the publicly available catalogue entries and should be read against the purchased standard text.

We extrude custom plastic profiles, including ABS. What we do not normally take on: injection moulding, blow moulding, and sheet, rod or board production. Minimum order quantities and section size limits apply and are stated case by case.

For a technical review, send the section drawing (or a dimensioned sketch), the annual volume, the service environment - indoor or outdoor, temperature range, any chemical or UV exposure - and any colour requirement. A drawing and a service environment produce a useful answer; a request for a unit price without them does not. You can see the ABS sections and finishes we already run on our ABS profile page, or send a section for technical evaluation.