De Laval Cream Separator

Build and test only inside a substantial enclosure. Balance obsessively, run up slowly, and never open the housing while anything is turning.

De Laval specifies gearing to reach six or seven thousand revolutions a minute from a hand crank, skimming sixty litres per hour. Both figures are in the specification.

Leave whole milk in a shallow pan somewhere cool and record how long a cream line takes to form. That number — many hours — is the thing being replaced.

Centripetal acceleration is ω²r. At 7000 rpm on a 50 mm radius that is several thousand g — which is why seconds replace hours.

Machine a bowl that runs true. At these speeds a gram of imbalance is a large cyclic force — balance is a safety requirement, not a refinement.

Support the bowl on a thin shaft rather than a rigid one. De Laval's insight was that a flexible shaft lets the bowl spin about its own centre of MASS above the critical speed, instead of fighting the imbalance.

Run up briskly past the resonance and note the vibration peak then fall away. Lingering at the critical speed is what destroys the machine.

Build a gear train taking a comfortable 60 rpm hand speed to several thousand at the bowl. Ratios above a hundred to one are what make it hand-powered at all.

Introduce milk down the axis. It must enter where the field is weakest and travel outward, or it simply mixes what has already separated.

Fit two outlets at different radii. Dense skim milk collects at the wall; light cream is displaced inward — the LIGHT phase goes to the middle, which surprises people.

Insert nested cones so the milk is split into thin layers. A globule then has millimetres to travel instead of centimetres, and separation time collapses.

Collect both outputs and compare. Good separation leaves skim milk at a fraction of a per cent fat, continuously, with no waiting at all.

Feed faster and separation worsens: residence time in the bowl has dropped below what the globules need to travel. Throughput and purity trade directly.

Find the sludge at the wall — dirt, cells and debris denser than milk. The separator is also a clarifier, and that hygiene benefit mattered as much as the cream.

The patent. US 368,328, "Apparatus for working centrifugal machines", granted 16 August 1887 to Gustaf de Laval of Stockholm, covering the support and hand operation of vertically-mounted centrifugal cream separators. De Laval had built his first continuously operating separator in 1878; the significance of "continuously" is that milk flows in and two streams flow out without the machine stopping, where every earlier method was a batch that had to be set, waited on and skimmed.

The physics is Stokes' law with gravity replaced. A fat globule rises through milk at a rate set by the density difference, the globule's size squared, the fluid viscosity and the acceleration field. Gravity gives 1 g and a cream line in half a day. Spinning at 7000 rpm on a 50 mm radius gives several thousand g, and the same separation happens in seconds. Nothing about the milk changes — only the size of the field driving the separation.

The disc stack is the second idea. Separation time also depends on how FAR a globule must travel, and that distance scales with the depth of the liquid layer. Nesting a stack of conical discs divides the bowl into many thin layers, so each globule has a very short path to the nearest surface, from which it slides inward. This is the same reasoning as a plate settler in water treatment, and it is why a modern separator bowl is full of pressed cones.

De Laval's other contribution is the shaft. A rigid spindle must be balanced almost perfectly or it tears itself apart; a slender flexible one behaves differently. Above its critical speed a flexible shaft lets the rotor spin about its own centre of mass, so residual imbalance is accommodated rather than resisted — a machine that shakes violently at one particular speed and then runs smoothly beyond it. This is now standard rotordynamics, and the same man went on to apply high-speed rotation to the de Laval steam turbine and gave his name to the converging-diverging nozzle used in rocket engines.