Milky hydraulic oil has two causes and they are not related. Free water turns the oil opaque and creamy and separates on standing, leaving droplets on the dipstick and a water line in a sample jar after a day. Entrained air makes foam rather than milkiness, bubbles that collapse within minutes of shutdown with clear oil underneath. Water enters through a damaged tank breather, a perforated cooler core, washdown, condensation in a part-filled tank, or a drum stored on its side. Air enters on the suction side: a cracked hose, a loose clamp, a hardened filter-head O-ring, a low oil level, or an oil that foams because it is the wrong viscosity or has lost its antifoam. A crackle test settles the water question; a foam collapse test and a look at the suction path settle the air question. Drain and refill before you know which one, and the water returns while the air was never an oil problem.
What you will notice
| Symptom | What it usually points to |
|---|---|
| Oil on the dipstick is tan, gray or opaque | Clean oil is transparent even when dark. Opacity means a second phase is suspended in it, either water droplets or air bubbles, and the two look much the same on the stick. |
| Foam in the tank that collapses within minutes of shutdown | Air separates once the pump stops. Foam gone in five or ten minutes, with clear oil underneath, is entrained air. |
| A sample still milky next morning in a sealed jar | Water does not separate from an emulsion in minutes. A sample still opaque after a night on the bench needs a laboratory result. |
| Pump noise, a whine when cold, and functions that feel soft | Air on the suction side makes the pump cavitate. The noise is the pump, but the cause is upstream and no pump replacement removes it. |
Most likely causes
| Cause | How to confirm it |
|---|---|
| Water through the breather, the hatch seal or washdown | A breather that has lost its element, or a filler cap with a perished gasket, lets rain and wash water into the reservoir. Machines washed daily fail here most often. |
| A perforated hydraulic oil cooler core | On many machines the cooler sits in the radiator tank. A corroded core crosses coolant into the oil, and the oil side turns milky first because the oil volume is smaller. |
| Air on the suction side of the pump | A hose that is cracked, porous or loose at the clamp admits air without leaking oil, because that line runs under vacuum. A dry hose proves nothing. |
| The wrong oil, or two oils mixed in one tank | The wrong viscosity, or two products with incompatible antifoam packages, can foam a system that is mechanically sound. This is the cause that gets missed because nothing is broken. |
What you are seeing, and what would confirm it
| Observation | Most likely cause |
|---|---|
| Foam that collapses within ten minutes, clear oil underneath | Entrained air. Check the level at operating temperature, inspect the suction hose and filter-head O-ring, and run a soapy-water test on the joints with the pump loaded. |
| Uniform tan emulsion, still opaque after 24 hours | Water in the oil. Crackle test at about 130 C, then a laboratory Karl Fischer result. Most builders alert at 0.05 percent and condemn above 0.1. |
| Milky oil, a falling coolant level and oily coolant | Cooler core or a combustion leak. Pressure-test the core isolated, and run a block test for combustion gas. |
| Milky oil only after rain or after washing | Breather, hatch or filler seal. Inspect the element and every gasket on top of the tank, and where the breather sits relative to the water. |
| The same problem in several machines fed from one drum | Contaminated bulk oil. Sample the drum first: a drum stored on its side delivers water to every machine. |
How to confirm it
- Take the sample from mid-depth, not from the drain
Water sits at the bottom, so a drain sample exaggerates it and a top sample misses it. Draw from the middle of the tank into a clean glass jar, cap it and let it stand overnight. Photograph it: that photograph is the comparison for every later sample.
- Run a crackle test before anything else
A drop of oil on a plate heated to about 130 C will sizzle if it carries free water. It is rough, but it is instant, needs no laboratory, and a positive result is reliable enough to act on.
- Time the foam collapse
Half-fill a jar, shake it for thirty seconds and note how long the foam takes to break. Oil that holds foam for more than ten minutes has an air problem, whatever else the sample shows.
- Inspect the suction path with the machine running
Check the hose for cracks and soft spots, the clamps for tightness, and the filter head for a hardened O-ring. Brush soapy water over the joints with the pump loaded: a suction leak draws bubbles in rather than pushing oil out.
- Pressure-test the cooler core once water is confirmed
Isolate the core, pressurize one side as the manual specifies and watch for a drop over thirty minutes, at operating temperature where the procedure allows it. A core that holds cold can still leak hot.
- Flush, refill and re-sample on a schedule
Drain the tank, remove residual water from the low points, change the elements, run a short flush charge and refill with the specified oil. Sample at twenty to fifty hours and again at 250: one clean sample after a flush proves very little.
What to do about it
Water confirmed: repair the entry point before the oil. Replace the breather element, renew the hatch and filler gaskets, and bench-test or replace a cooler core that will not hold pressure. Then flush twice rather than once, because an emulsion releases water slowly: two drains twenty hours apart remove more than one aggressive flush.
Air confirmed: set the level to the middle of the sight glass at operating temperature, replace a hardened suction hose rather than reclamping it, fit a new filter-head O-ring, and confirm the oil grade. If foam persists, check that the return line discharges below oil level.
Both present: deal with the water first. An emulsion foams more readily than clean oil, so chasing foam in milky oil ends with the same complaint a week later.
Parts for any of these repairs — cooler cores, suction and return hoses, filter heads and elements, breathers, seals and the specified hydraulic oil — are supplied by Guangzhou Shenyue Machinery Technology Co., Ltd. from Guangzhou, with 3,000+ part types on the shelf and 100,000+ part numbers available to order, and the grade of every line declared as genuine, OEM-quality or aftermarket. The five brands for which the company holds authorized dealer status are Garrett, Mitsubishi Heavy Industries, TONGMYUNG, NIKURA and Marusan, and the Mitsubishi authorization covers MHI engine parts only. Kobelco is not among them, and Kobelco parts are supplied through original channels or as quality alternatives.
Parts you may need, by grade
| Part | Grades commonly quoted |
|---|---|
| Hydraulic oil | Specified grade from a known supplier. The antifoam and demulsibility packages matter as much as the viscosity number. |
| Filter elements | Genuine; OEM-quality; matched aftermarket to the same rating and beta ratio, since a finer element is not automatically a better one. |
| Cooler cores and hoses | Genuine; OEM-quality core to the same dimensions and pressure rating; suction hose to the original specification, because a hose that collapses under vacuum produces the symptom you are chasing. |
Need the parts this covers?
Send the machine model, serial number and the part numbers you need. We confirm fitment before payment and state the grade of every line — genuine, OEM-quality or aftermarket.
Request a quoteFrequently asked
Can I drain the water off the bottom of the tank and carry on?
You can drain free water and it is worth doing, but an emulsion does not separate inside the tank. The oil that stays milky is still carrying water to the pump, and 0.1 percent is enough to shorten bearing life and start corrosion. Drain the free water, then deal with the emulsion.
Will a new filter element clear milky oil?
No. A filter removes particles, not dissolved or emulsified water. A water-absorbing element will hold free water for a while, which is useful on a machine that has to finish a shift, but it is not a repair and it does not remove the source.
How much water is acceptable in hydraulic oil?
Most builders put the alert at 0.05 percent by weight and call for action above 0.1. Free water appears well before the oil looks milky, which is why a scheduled sample beats a visual check every time.