Compressed Air for EPP and EPS Moulding Lines: ISO 8573-1 Purity, Demand and Distribution

2026-10-22Industry News22 min read

Quick answer

Specify compressed air for an EPP or EPS moulding line: ISO 8573-1 purity classes, how much air a press really uses, receiver sizing and distribution.

Quick answer: A bead foam press uses compressed air for three quite different jobs — moving cylinders and actuators, pushing beads through filling guns, and blowing parts and water off the tool — and each job wants a different purity. The practical way to specify this is ISO 8573-1, which grades compressed air on three independent contaminants written as [particles : water : oil]. Most general plant pneumatics on an EPP or EPS line sit around [3:4:3] to [2:4:2]; instrument air feeding control valves and positioners normally goes tighter, around [2:3:1] or [3:4:1]; and any air that meets raw bead or a part surface should be treated as a contact utility, not as shop air. Size the system on peak demand during fill and ejection rather than on the average, add a receiver so short peaks do not force you to buy a bigger compressor, and treat every leak as a pressure problem that costs money every hour the plant runs.

EPP/EPS shape moulding machine on the production floor, the main consumer of compressed air on a bead foam line — Every pneumatic function on the press — clamping, ejection, bead fill and blow-off — draws from one plant air system. Specifying that system properly is cheaper than oversizing it to hide the problems.

Key takeaways

Write the air spec as three numbers, not one. ISO 8573-1 grades solid particles, water and oil independently, as [particles : water : oil]. "Clean dry air" is not a specification.
Separate contact air from actuation air. Air that touches bead or a part surface belongs to the same contamination conversation as cleanliness-critical packaging; air that only moves a cylinder does not.
Pressure dew point, not atmospheric dew point. Water class is defined at line pressure. A "−40 °C dew point" without the word pressure means nothing.
A refrigerated dryer lands you at water class 4 (+3 °C pressure dew point). Anything tighter needs a desiccant dryer, and anything tighter than that costs purge air you also have to pay for.
Size on peak, not average. Fill and ejection are short, sharp peaks. A receiver absorbs them; a bigger compressor is an expensive way to do the same job.
Every bar of unnecessary pressure is paid for every hour. Running the header 1 bar higher than the highest consumer needs buys nothing and costs roughly 7% in compressor energy for that step.
Poor drainage design ruins good dryers. Dead-end runs, liquid take-offs and low points without drains put condensate back into the air after it has been dried.
Compressed air is not vacuum. It is a separate machine, a separate pipe network and a separate maintenance regime from the vacuum system that evacuates the mould.

Where compressed air is actually used on a bead foam line

Before specifying a compressor or a filter train, it is worth listing what the air does. On most shape moulding installations the users fall into four groups, and they do not all need the same quality.

Main compressed air consumers on an EPP/EPS shape moulding line, with the practical quality concern for each
Use	What the air does	Demand shape	Main quality concern
Cylinders and actuators	Clamp and stripping movement, core pulls, ejector actuation, hopper and loader valves, gun feed valves	Intermittent, synced to the machine cycle	Water washing out lubricant and causing sticky valves in cold start-up; scale scoring seals
Bead conveying and filling guns	Carries pre-expanded bead along hose into the cavity, then assists the crack-fill cycle	Sharp, short peaks at every fill	Liquid water making bead clump; oil aerosol contaminating bead surfaces before fusion
Ejection and blow-off	Parts and condensate blown clear of the tool and the take-out area	Short, high flow, every cycle	Any carryover lands directly on the part surface and the tool cavity
Instrument and control air	Pneumatic control valves, positioners, safety circuits, sensing lines	Small but continuous	The most sensitive user — oil and water here cause drift and stiction rather than obvious failures

Compressed air and vacuum are two separate systems. Vacuum runs below atmospheric pressure and exists on this process mainly to pull condensate out of the steam chest and hold the part during ejection; compressed air runs above it and moves cylinders and bead. They share nothing but a plant floor, and they fail in opposite directions. Sizing, pump selection and leak testing for the vacuum side are covered separately in our article on vacuum systems for EPP and EPS moulding — this article is the other half of the utility picture.

Specifying purity: how ISO 8573-1 works

ISO 8573-1 classifies compressed air purity across three contaminant categories. Each is graded on its own scale, so a single reagent never implies the others. The standard and its current edition, ISO 8573-1:2010, expresses this as three class numbers in a fixed order — [particles : water : oil].

A specification written as ISO 8573-1:2010 [2:4:2] therefore means class 2 for solid particles, class 4 for water and class 2 for total oil. Lower numbers are cleaner, but the scales are not identical for the three categories, so "class 2" does not mean the same degree of strictness across all three.

ISO 8573-1 purity classes. Solid particles are counted per cubic metre by size band; water is a pressure dew point; total oil includes liquid oil, aerosol and vapour
Class	Particles — 0.1–0.5 µm	Particles — 0.5–1 µm	Particles — 1–5 µm	Water — pressure dew point	Total oil
0	Specified by the user or supplier, and stricter than class 1			As specified, stricter than class 1	As specified, stricter than class 1
1	≤ 20,000 /m³	≤ 400 /m³	≤ 10 /m³	≤ −70 °C	≤ 0.01 mg/m³
2	≤ 400,000 /m³	≤ 6,000 /m³	≤ 100 /m³	≤ −40 °C	≤ 0.1 mg/m³
3	not specified	≤ 90,000 /m³	≤ 1,000 /m³	≤ −20 °C	≤ 1 mg/m³
4	not specified	not specified	≤ 10,000 /m³	≤ +3 °C	≤ 5 mg/m³
5	not specified	not specified	≤ 100,000 /m³	≤ +7 °C	—
6	Mass-concentration basis rather than particle count			≤ +10 °C	—

Class 0 does not mean zero. Class 0 carries no published numeric limit. It means the limit has been agreed between user and supplier and is stricter than class 1 — which means a specification that says "class 0" without stating the contaminant, the measurable limit, the test method and the sampling point is incomplete. If you need it, write all four.

Dryer choice is really a water class choice

The water digit is the one that drives most of the capital cost, and in practice it narrows to a single decision:

Refrigerated dryer — typically delivers around +3 °C pressure dew point, which is water class 4. For indoor lines with no cold pipe runs, this is usually sufficient.
Desiccant (adsorption) dryer — commonly delivers −20 °C or −40 °C pressure dew point, i.e. water class 3 or 2. Needed when lines run outdoors, in unheated buildings, or where any condensation at the point of use is unacceptable.
Membrane or deliquescent options trade purge losses and consumables against dew point, and suit small point-of-use flows rather than a whole header.

The cost of going tighter is not only purchase price. Heat-reactivated desiccant dryers also consume purge air — air that has already been compressed, dried and paid for — and that demand has to be added to the compressor sizing. This is the same trap as buying oversized filtration: every extra stage also adds pressure drop, and every bar of pressure drop has to be generated.

The units that catch people out

Two conventions in air specification produce most of the avoidable disputes:

Pressure dew point versus atmospheric dew point. ISO 8573-1 water classes are pressure dew points — the temperature at which condensation begins at line pressure. The same air relaxed to atmosphere has a considerably lower dew point. Always state which one you mean.
Free air delivery versus compressed volume. Compressor output ("free air delivery") is quoted as the volume the compressor draws in at ambient conditions, not the compressed volume delivered. Any comparison that mixes the two is wrong by roughly the compression ratio.

Which air touches your product

The most useful reframing of compressed air on a foam line is to stop treating it as one utility and start treating it as contact air and power air.

Assigning ISO 8573-1 targets by application on an EPP or EPS line. These are starting points for discussion, not requirements imposed by the standard
Application	Typical starting specification	Reasoning	Worth tightening when
General plant actuation	[3:4:3] to [2:4:2]	Refrigerated dryer plus particulate and coalescing filtration; adequate for cylinders in a heated plant	Valves begin sticking seasonally, or the header runs through unheated space
Instrument and control air	[3:4:1] to [2:3:1]	Oil and particle grades matter more than extreme dryness, because contamination causes drift rather than visible failure	Safety-related valves, positioners and sensing lines are on the header
Bead conveying and fill air	[2:4:1] upward	This air meets raw bead before fusion, so oil and water here become part of the material	Surface defects appear that the defect sequence does not explain
Air contacting finished parts	Site-specific; often [2:3:1] or tighter	Ejection blow-off and handling air deposit whatever they carry onto the part	Parts go into electronics, medical or food-adjacent packaging

The last row matters more than it looks. If air used for ejection blow-off carries droplets, those droplets land on the tool cavity and the part surface, and the symptom resembles a process defect rather than a utility problem. That misdiagnosis is expensive because it sends the team hunting through the moulding cycle instead of checking the dryer. When you are working through unexplained surface contamination, it is worth resolving whether the cause is air quality before going deeper into the EPP and EPS moulding defect troubleshooting sequence.

Electronics packaging has an additional reason to care. If you are moulding static-dissipative packaging, oil and moisture films on the part surface can interfere with the consistency of dissipative behaviour and with subsequent printing or labelling. The design side of that discussion is covered in our guide to ESD-safe EPP and EPS packaging moulds; the utility side is purely about keeping contact air clean.

How much air does a press actually use

Most budgets and most compressor quotes start with a guess. A cylinder-by-cylinder estimate takes about an hour and is almost always lower than the guess, because the largest named air users turn out to be intermittent.

Calculating one cylinder

A pneumatic cylinder consumes air to fill whichever side is being supplied, at line pressure. Converting that to free air — the number that matters against compressor output — simply multiplies by the compression ratio.

Free air per stroke = stroke volume × (gauge pressure + atmospheric pressure) ÷ atmospheric pressure
For a double-acting cylinder, add the return stroke, which uses the smaller annular area.

Worked example. A clamp cylinder with a 100 mm bore, a 32 mm rod and a 200 mm stroke, running at 6 bar gauge.

Bore area = π/4 × 0.1² = 0.00785 m² → stroke volume = 0.00785 × 0.2 = 1.57 L
Rod area = π/4 × 0.032² = 0.00080 m² → annular return area 0.00705 m² → return volume = 1.41 L
Compression ratio at 6 bar gauge = (6 + 1.013) ÷ 1.013 = 6.92
Out stroke = 1.57 × 6.92 = 10.9 L free air | Return = 1.41 × 6.92 = 9.8 L free air
Round trip ≈ 20.7 L free air per machine cycle

At a 90-second cycle that is roughly 40 cycles an hour, or about 0.83 m³/h for that one cylinder. Multiply by the number of cylinders, then add blow-off and fill. The headline number is usually modest; the shape is the problem.

The peak is what sizes the compressor

Fill air and ejection blow-off are not gentle. They are short bursts, they often overlap, and they tend to happen at the same point in every cycle. If several presses share a header, the peaks from independent machines will occasionally coincide, and no one plans for it.

Fill assistance and gun air — brief, high flow, adds directly to the compressed demand at every fill.
Ejection blow-off — short, often the single largest instantaneous draw on the machine.
Actuation — distributed around the cycle, predictable from the cylinder table above.
Continuous background load — desiccant purge, instrument air, dry-running losses and the leaks nobody has found yet.

Sizing on average demand is the most common and most expensive error. Average demand produces a compressor that cannot hold pressure through the fill peak. The usual consequence is a decision to raise header pressure, which "solves" the pressure dip by paying more energy every hour of every day for the rest of the plant's life.

Receiver sizing: buying seconds instead of horsepower

A receiver is a buffer. If a peak lasts long enough to dip the header, storing air near the point of use is far cheaper than installing a compressor that can cover the peak outright. The sizing relationship is straightforward when everything is expressed in consistent units:

V = Q_net × t × p_atm ÷ (p₁ − p₂)
V receiver volume | Q_net shortfall = peak demand − compressor output, in free air terms per second | t duration of the peak in seconds | p₁, p₂ receiver pressures before and after the peak, absolute

Worked example. During fill, demand peaks at 160 L/s of free air for 12 seconds while the compressor delivers 100 L/s. The shortfall is 60 L/s. The receiver is allowed to fall from 7.0 to 6.5 bar gauge — that is 8.01 to 7.51 bar absolute, a difference of 0.5 bar.

V = 60 × 12 × 1.013 ÷ 0.5 = 1,459 L ≈ a 1.5 m³ receiver

The arithmetic says something useful beyond the number: the tighter you hold the allowable pressure band, the larger the vessel has to be. Allowing a half-bar swing is usually invisible to the process and dramatically cheaper than demanding a tenth.

Distribution: getting specified air to the point of use

A plant can dry its air properly and still deliver water to the machine, because almost all distribution problems are drainage problems in disguise.

Take-offs from the top of the header. Liquid water travels along the bottom of a horizontal pipe. Drawing supply from the top of the main leaves most of it behind; drawing from the bottom collects it.
Drop legs and drains at every low point and every drop. A drop leg plus an automatic drain is cheaper than a filter and does not need elements.
Slope horizontal runs in the direction of flow, with the drain at the far low point, so gravity does the work continuously rather than at shut-down.
Size pipe on velocity and pressure drop, not on connection size. Undersized pipe is a permanent energy tax; the pressure lost in the line has to be generated at the compressor forever.
Close the loop with a ring main where the layout allows it. A ring reduces drop in a long run and lets a section be isolated without stopping the plant.
Filter near the point of use, not only at the compressor. Rust and scale generated inside the pipe downstream of the dryer arrives after the treatment train.
Keep hands off the wrong materials. Distribution pipe should be pressure-rated and suited to the duty; whatever you choose, do not improvise with pipe intended for other services.

Filtration stages, in the order they belong. A typical train runs a particulate pre-filter, then a coalescing filter for oil aerosol and fine liquid, then the dryer, then dust removal after the dryer, with an activated-carbon stage added only where oil vapour has to be removed. Activated carbon is not a substitute for a coalescer — putting vapour removal in front of liquid removal just saturates the expensive element quickly.

Leaks and pressure: the two cheapest energy projects

Compressed air is expensive per unit of useful work, and the losses are nearly invisible, which makes them durable. Two actions reliably pay back faster than any equipment purchase.

Fix leaks

In poorly maintained plants it is common to see a large share of generated air never doing useful work — figures in the region of a fifth to a third of output are frequently quoted for unmanaged systems, with well-run plants holding well below that. The cheap test is simply to observe pressure decay: with all consumers off and the compressor isolated, watch how fast the header falls.

Listen with an ultrasonic detector rather than by ear — most leaks are inaudible over plant noise.
Tag, log and repair in batches; a leak list that is not tracked becomes background.
Concentrate first on couplings, drain valves, hose ends and quick connectors, which fail far more often than pipe.
Treat every failed automatic drain as a leak as well as a water problem — an open drain is both.

Stop paying for pressure you do not use

Set the header to what the most demanding consumer actually requires, then reduce everything else at its own regulator rather than throttling at the compressor. A widely used planning estimate puts the saving at roughly 7% of compressor energy per bar of discharge pressure removed, which makes an unnecessary bar one of the most expensive habits a plant can keep.

Compressed air is not a cleaning tool. Blowing swarf or dust off clothing, benches or the machine with a shop line can drive particles into skin and eyes, and regulators restrict it accordingly — in the United States, OSHA 1910.242(b) limits the pressure at the nozzle to below 30 psi (roughly 2 bar) for cleaning purposes, even then requiring chip guarding and personal protective equipment. Specify dedicated low-pressure cleaning nozzles and put restrictions in place so the high-pressure line cannot be misused.

A note on lifecycle cost

Compressors are unusual among plant purchases in that the purchase price is the smaller number. Across a typical service life, the electricity consumed dominates everything — capital, maintenance and filtration combined. Two consequences follow for anyone specifying a system for a new line.

First, an efficiency argument usually beats a capital-cost argument over any horizon worth planning. Variable-speed drive, correct sizing, lower operating pressure and repaired leaks all look like operating detail and are actually the main cost levers. Second, tighter purity than the process needs is a permanent tax: deeper drying costs purge air, finer filtration costs pressure drop, and both are paid for every hour rather than once. Pick the class from the application, not from a catalogue.

Eight steps to specifying air for a new or retrofitted foam line

List every air user: cylinders with bore, rod and stroke, fill air, blow-off, instrument air, dryer purge.
Convert each to free air per cycle, then to per-hour average and per-second peak. Keep the two separate.
Assign an ISO 8573-1 three-digit class to each user. Group them — usually into plant air, instrument air and contact air.
Choose drying from the water digit. Refrigerated for about +3 °C pressure dew point; desiccant only where the application or the pipe routing genuinely needs it.
Size the compressor against average demand and the receiver against peak shortfall. Do not size the compressor against the peak.
Set the header to the highest genuine requirement and regulate down locally. Do not buy pressure dips.
Design the pipe for drainage: top take-offs, drop legs, automatic drains at low points, slope in the direction of flow.
Commission with measurements — dew point at the point of use, pressure drop across each filter element, and a leak-down test with results recorded as the baseline.

Keep the utility work out of the cycle-time work. A utility upgrade and a cycle-time project both improve output per hour, and doing them together makes it impossible to attribute the gain. Do cycle reduction as its own exercise following the sequence in where EPP and EPS cycle time is really spent, and keep air improvements separate so each can be verified.

Frequently asked questions

What does an ISO 8573-1 specification like [2:4:2] actually mean?

It is three independent purity classes written in the fixed order particles, water, oil. In this example that means class 2 for solid particles, class 4 for water (a pressure dew point of +3 °C or lower) and class 2 for total oil (up to 0.1 mg/m³, including liquid oil, aerosol and vapour). The digits do not imply each other, so class 2 particles says nothing about water or oil performance. The current edition is ISO 8573-1:2010.

Is a refrigerated dryer enough for an EPP or EPS moulding line?

For most indoor installations, yes. A refrigerated dryer typically delivers around +3 °C pressure dew point, which corresponds to water class 4, and that is sufficient for general actuation air in a heated plant. You need to move to a desiccant dryer — water class 3 or 2 — when lines run outdoors or through unheated space where further cooling would produce condensation, or where liquid water at any point of use would contaminate the bead or the part.

Why does ISO 8573-1 use pressure dew point instead of normal dew point?

Because the practical question is whether condensation will form inside the pipe, and that depends on pressure. Water condenses more readily at higher partial pressures, so the same parcel of air shows a higher dew point at line pressure than it does after relaxing to atmosphere. A specification that simply says "minus 40 degrees" without stating pressure dew point is ambiguous and will not survive a dispute.

What air quality should I specify for air that touches the bead or the finished part?

Treat it as a contact utility rather than as shop air. Bead conveying and fill air should be held much tighter on both water and oil than general actuation air, because anything it carries becomes part of the material before fusion. Air used for ejection blow-off deposits carryover directly onto the part surface and the tool cavity. A common starting point for the instrument and control circuit is around [3:4:1] to [2:3:1], going tighter only where the application demands it.

Class 0 is described as the cleanest air, so should I specify it?

Only if you can write the number down. Class 0 has no published numeric limit in the standard; it means the limit is agreed between user and supplier and is stricter than class 1. A complete class 0 requirement states the contaminant, the measurable limit, the test method and the sampling point. Writing "class 0" on its own specifies nothing, and it is one of the three independent digits — it does not set the other two.

How do I estimate how much air one pneumatic cylinder uses?

Convert the volume it consumes on each stroke into free air. Multiply the piston area by the stroke to get the volume at line pressure, then multiply by the compression ratio, which is gauge pressure plus atmospheric pressure divided by atmospheric pressure. A 100 mm bore cylinder with a 200 mm stroke at 6 bar gauge uses roughly 10.9 litres of free air on the out stroke, and about 9.8 litres on the return for a 32 mm rod, so around 20.7 litres per full cycle.

Should the compressor be sized for peak air demand?

No. Size the compressor for average demand and use a receiver to cover short peaks. The receiver volume needed equals the peak shortfall multiplied by the peak duration and atmospheric pressure, divided by the pressure drop you are prepared to allow in the receiver. Because that last term is in the denominator, tolerating a half-bar swing instead of a tenth dramatically reduces the vessel size required, and the difference is invisible to the process.

How much does running the header at unnecessarily high pressure cost?

A widely used planning estimate puts it at roughly 7 percent of compressor energy per bar of discharge pressure removed, which makes an unnecessary bar a permanent rather than one-off cost. Set the header to what the most demanding legitimate consumer requires and reduce every other circuit at its own local regulator. Raising header pressure to hide a peak-siting problem is the most expensive way to avoid buying a receiver.

Why does my air still contain water at the machine when the dryer is working?

Almost always because of how the pipe is arranged rather than how the air was dried. Liquid water travels along the bottom of horizontal runs and collects in low points, so supply take-offs should come from the top of the header, horizontal lines should slope in the direction of flow, and every low point and drop needs a drop leg with an automatic drain. Rust and scale generated downstream of the dryer also arrive after the treatment train, which is why point-of-use filtration matters.

How do I find compressed air leaks cheaply?

Start with a leak-down test: isolate the compressor, shut all consumers and watch how fast the header pressure falls. Then locate individual leaks with an ultrasonic detector rather than by ear, since most are inaudible over plant noise. Concentrate on couplings, quick connectors, hose ends and drain valves, which fail far more often than pipe, and log each leak with tag and repair date so the list actually closes out.

In-Process Quality Control for EPP and EPS Moulding: SPC, Part Weight and Traceability

2026-10-21

How to Cut EPS Foam: Tools, Techniques and Tips

2026-10-20

Vacuum Systems for EPP and EPS Moulding: Pump Type, Sizing and Leak Control

2026-10-19

Need a custom EPP / EPS mold for this application?

Send us your drawing, target density and annual volume — our engineers reply with a tooling proposal, lead time and quotation within 24 hours.

Get a quotation