Dr.Vakuum
All applications

Sizing vacuum for thermoforming and MAP

7 min read

On a thermoforming line the vacuum system is rarely the thing that limits package quality. It is very often the thing that limits cycle rate — and it is almost always specified on the wrong number.

This note sets out what residual oxygen actually depends on, why ultimate vacuum carries no useful selection information here, and how to size the pump on the seconds you have rather than the millibar you can reach.

1. Residual oxygen is arithmetic, not a pump specification

Air is 20.9% oxygen. Evacuate a package to an absolute pressure p and, before backfill, the oxygen remaining is simply the same fraction of what is left:

O₂ residual (%) = 20.9 × p / 1013

Evacuation pressure Residual O₂
5 mbar 0.10%
10 mbar 0.21%
20 mbar 0.41%
50 mbar 1.03%
150 mbar 3.09%

That single line explains the whole hierarchy of packaging methods:

  • Gas flushing alone, with no evacuation, leaves 2–5% O₂, and the exact figure drifts with flow rate and sealing speed. It is a flow-through purge, so it never finishes.
  • Evacuation to 5–50 mbar puts you at 0.1–1% O₂ before any gas is admitted at all.
  • Evacuation plus gas flush is what reliably holds below 1%.

Read the table again and note where 150 mbar lands: 3.1%, squarely in the same band as not pulling vacuum at all. This matters when choosing technology, because it is exactly the ultimate pressure of a typical dry claw pump.

Two separate questions. The pump sets oxygen at the instant of sealing. What happens over shelf life is set by the film. Barrier layers such as EVOH lose oxygen-barrier performance as they take up moisture, and a deep-drawn tray is thinnest — hence most permeable — at the corners the forming station stretched most. A perfect vacuum system cannot compensate for a web thinned past its barrier spec.

2. Why ultimate vacuum tells you nothing here

Look at any rotary vane packaging range and you will find the same ultimate pressure — commonly 0.1 mbar — printed against every model from the smallest to one twice its size. Two pumps that differ by a factor of two in the parameter that matters are identical in the parameter buyers ask about.

You also never operate there. The working point is 5–50 mbar, one to two decades above ultimate. Ultimate pressure is a condition indicator — a way of telling whether the oil is still good — not a selection criterion.

The number that actually selects the pump is effective pumping speed at the working pressure, and it is not the number on the badge.

3. Sizing on cycle time

Start from the standard evacuation relation:

t = (V / S) × ln(p₀ / p₁)

Rearranged to size for a cycle:

S = (V / t) × ln(p₀ / p₁)

In units convenient for packaging — V in litres, S in m³/h, t in seconds:

t [s] = 3.6 × (V[L] / S[m³/h]) × ln(p₀/p₁)

A 100 litre die volume pulled from atmosphere to 5 mbar with a 250 m³/h pump: 3.6 × (100/250) × ln(202.6) = 7.6 s. Now apply the two corrections that decide whether you hit that number.

The speed curve — roughly a 25% de-rate

A rotary vane pump delivers about 75% of its nameplate speed at 5 mbar. Nameplate is free-air displacement near atmosphere; the working point is not near atmosphere. A 250 m³/h pump is a ~188 m³/h pump where it does its work.

Conductance — and why it gets worse as the cycle proceeds

Pump and pipework are in series:

1 / S_eff = 1 / S + 1 / C

In this pressure range flow is viscous, and laminar conductance goes as

C ∝ d⁴ · p̄ / l

Two consequences follow, and both are money:

  • Diameter dominates by the fourth power. Going from DN40 to DN50 raises conductance by (50/40)⁴ = 2.44×. It is the cheapest cycle time available, and it is the thing most often got wrong at installation.
  • Conductance falls as you pump down, because it is proportional to mean pressure. A line that is generous at 500 mbar can be the dominant restriction at 5 mbar — which is precisely why the complaint is always "the last two seconds take forever."

A 250 m³/h pump on a line whose conductance is also 250 m³/h gives S_eff = 125 m³/h. You paid for a 250 and installed a 125.

4. The wet-product floor

Water's saturation vapour pressure is about 24 mbar at 20 °C. Until the free moisture on a product is gone, that is the lowest pressure the package will see, whatever the pump is doing.

Below that point you are no longer removing gas, you are evaporating liquid, and evaporation is limited by surface area and heat supply — not by pumping speed. The logarithm stops predicting anything. Doubling the pump does not halve this part of the cycle.

Worse, vapour that condenses inside an oil-sealed pump emulsifies the oil, which loses its sealing film, which raises ultimate pressure, which tempts operators to extend the cycle, which draws in more vapour. Cloudy or milky oil is the tell, and rotor seizure is the end of that road.

Gas ballast is the answer, and the trade is favourable. Admitting a controlled bleed of air raises the ultimate pressure — typically from 0.1 to 0.5 mbar — but §2 established you have ultimate pressure to spare: 0.5 mbar is still an order of magnitude below a 5 mbar working point. You are spending a reserve you were never using in order to prevent a seizure. On wet product, run it continuously.

5. Where a booster earns its place

A Roots booster operates in the 1–100 mbar band, with a speed gradation of typically 2–8× the backing pump. That band is exactly where the backing pump's speed curve has collapsed.

Because every decade of pressure costs the same time at constant speed (ln 10 = 2.303), the last decade is as expensive as the first — but you are crossing it with a de-rated pump on a line whose conductance has also fallen. Multiplying speed there shortens the expensive part of the cycle. This is why a booster beats simply buying a larger single pump on both cycle time and energy: it is not more pump, it is speed placed where the curve failed.

A booster cannot pump against atmosphere and always needs a backing pump.

6. Diagnosing a slow line

Three quite different faults all present as "the pump is getting tired." Separate them in minutes with a gauge and a stopwatch:

  1. Evacuate, isolate, and watch the pressure rise. A steady linear climb is a leak. Compute Q = V·Δp/Δt and check it against the working point; the leak-driven floor should be under 10% of your target pressure.
  2. A rise that flattens near 20–25 mbar is wet product, not a fault.
  3. Failure to reach base at all, with milky oil, is genuine pump degradation.

Rank the leak suspects by wear rate: forming and sealing die seals first — they are heat-cycled every stroke — then the film web at the chamber lip, the knife station, quick-connects, and the pump inlet non-return valve.

One more: a clogged inlet filter is a conductance restriction before it is anything else, so its first symptom is a cycle-time problem. It is routinely misdiagnosed as a worn pump.

Recommended Dr.Vakuum machines

Oil-sealed rotary vane (V and DV Series) remain the default for the evacuation station: high speed in the 5–50 mbar band, good water-vapour tolerance with gas ballast, and a self-cleaning oil circuit.

Dry screw pumps (DSI Series) — oil-free, 0.05 mbar ultimate, with high vapour and particle tolerance. The right answer where oil carry-over is unacceptable or wash-down is daily.

Model Flow (50/60 Hz)
DSI150 150 / 180 m³/h
DSI300 300 / 360 m³/h
DSI630 630 / 756 m³/h

Dry claw pumps (CL Series) belong on the forming side, where 150–200 mbar is sufficient — not on the evacuation station, where their ultimate pressure would leave you at ~3% residual oxygen.

Sizing help. Send us die volume, target pressure, the seconds you have, and the bore and length of the run to the pump, and our engineers will work the de-rated speed and conductance through with you. Talk to an engineer.

See also: the chamber-machine side of the same packaging hall in Vacuum packaging of frozen seafood and meat, and the oil-free marination step upstream in Vacuum tumbling and marination.