Foam Is Only the Air You Can See

Why the fine bubbles stay

Whether a bubble reaches the surface before the oil reaches the suction depends on how fast it rises and how long it has. For a fine bubble, rise speed goes with the square of its diameter and inversely with the oil's viscosity. Halve the diameter and it rises four times slower; cool the oil and it slows again.

A typical ISO VG 320 gear oil sheet lists 320 mm²/s at 40 °C and 25 at 100 °C. The ASTM D341 viscosity and temperature relation carries that to near 1,300 mm²/s at 20 °C and near 110 at 60 °C.

Treat the bubble as a rigid sphere, and the arithmetic is short: a 0.5 mm bubble climbs about 30 mm in five minutes at 20 °C, and about 370 mm at 60 °C.

The same arithmetic explains the morning. Oil tends to foam when it is cold, and cold oil holds its fine air longest. As the tank warms, the large bubbles leave, the surface layer collapses and the glass clears. The fine fraction lingers, and while the machine runs, every pass through the gears, pump and return line can make more.

Five minutes in the tank

Now give it the time a tank gives it. A 200 litre console tank circulating 40 litres a minute turns over every five minutes, and at 40 °C a 0.2 mm bubble in the gear oil needs more than an hour to rise 300 mm from return to suction.

So it goes round again, and the air release standard spells out what that can do: oil pressure that cannot be held and incomplete films in bearings and gears. That is the twitching needle.

One five-minute pass clears nothing much under half a millimetre, even from warm gear oil. A typical ISO VG 46 hydraulic oil does no better at 20 °C, and at 60 °C it still keeps everything under about 0.2 mm.

Where the air gets in

Usually it is the machine. Air typically enters a hydraulic circuit through the suction line when seals and fittings are not tight. The suction side runs below atmospheric pressure, so a loose joint there pulls air in without letting oil out; it stays dry and nobody sees it.

Returns and drains that end above the oil beat air in with every litre. One pump maker's reservoir guidance asks for drain and bleed lines ending at about half the fluid level, the suction inlet at least eight pipe diameters below the surface, and baffles that give the bubbles time to rise.

Level and cold starts decide where the foam goes. A sump filled to the top of the glass leaves the foam nowhere to go but the breather. One gear maker's fault table files oil at the breather under frequent cold starts or a high oil level, and the remedy it lists is mechanical: an expansion tank.

The oil can change too. In Flender's trend table, cloudy oil points to water or other foreign fluids, and a falling silicon value to an anti-foaming additive used up. The same manual forbids mixing oil types or makers. Each is a question for the laboratory.

Air corrupts two readings you already trust

The level. Flender's manual allows it to be read only at standstill, with the oil foam-free, after 5 to 10 minutes shut down. The rule is not fussiness: a foaming glass gives a false level, and a top-up made against it can overfill the sump, which feeds the breather and the foam in turn.

The particle count. The same manual has the laboratory thin the sample so that no air bubbles are counted. A light-extinction counter registers whatever blocks its beam, and the counting method, ISO 11500, is written for single-phase liquids; a bubble is a second phase. Before a jumped cleanliness code sends a filter cart out, ask whether that sample was degassed.

At the next stop, with a stopwatch

At the next planned stop, with the machine isolated under the site procedure, collect five things.

• The sample, taken running before the stop or within the manual's window after it, with the new-oil reference, noting that the oil was aerated.
• Oil temperature at the stop.
• Foam band height and the haze below it at the stop, then at 1, 5 and 10 minutes.
• Standstill level at 10 minutes, logged with the temperature.
• Every return and drain against that level, the baffle, the suction depth and any disturbed suction joint.

Foam that collapses within minutes while the haze stays tells you the fine fraction is large. Trust the level only against the mark and the last standstill reading. A level taken while running carries the air and, on a circulating system, misses the oil still out in the lines.

Two tests, two questions

Foam and air release are separate properties, and a good result on one says nothing about the other. ASTM D892 blows air through the oil at 24 °C and 93.5 °C and rates the foam it builds and the foam left after settling. ASTM D3427 times how long entrained air takes to fall to 0.2 percent by volume.

Some data sheets carry only one: the gear oil sheet above lists foam figures and no air release line at all. The air release method may not rank oils where residence times are short and gas contents high, so read both results as a change against the new-oil reference, not as a pass mark.

Ask for both, with the test temperature on the report, plus viscosity at 40 °C against the nominal grade, water by Karl Fischer and the element trend, silicon included.

Air path or changed oil

If foam and air release match the new-oil reference, the oil's air handling has not changed, and the air path you found on the walk is the job.

The creamy band was the easy part. The air that never reached the surface is what the pump has been running on since the cold start, and a stopwatch at the next shutdown will tell you more about it than another can of additive.