A particle count reports how many particles of each size band are present in a millilitre of oil, and it is expressed as an ISO cleanliness code: two numbers, the first for particles larger than four microns and the second for particles larger than six microns. A code of 18/16 means more particles than a code of 16/14, and the scale is logarithmic, so one step up is roughly double the count. The number that matters is not the absolute figure but where it sits against the cleanliness target for that machine's components. A servo-controlled hydraulic system with a piston pump wants a tighter target than an older gear-pump machine, and the same report can be acceptable on one and a fault on the other. Take samples the same way every time, interpret against a target, and watch the trend rather than the single reading.
What you will notice
| Symptom | What it usually points to |
|---|---|
| Oil looks clean but components keep failing early | Particles below the visible limit do the damage. A count is the only way to see them. |
| Filter elements loading far faster than the schedule | A rising particle count between services means contamination is being generated faster than filtration removes it. |
| New oil added and the count goes up | New oil is not clean oil. It arrives with its own particle count, and it can leave a system dirtier than it found it. |
| Count improving but component wear continuing | The source may have been removed while the residual contamination is still circulating. Cleanliness is a trend, not an event. |
Most likely causes
| Cause | How to confirm it |
|---|---|
| Ingress through a breather or a failed seal | The reservoir breathes every time the oil level changes. A restricted or damaged breather pushes air, and dust, in through the weakest seal instead. |
| Contamination introduced during a repair | Open ports, unplugged hoses and tools laid on a bench are the classic route. A repair that leaves a machine dirty is a repair that fails early. |
| Internal generation from a wearing component | A pump or motor that is starting to wear produces particles continuously. The count rises, filtration removes some, and the rise continues. |
| New oil and top-up containers | Oil decanted into an open can or poured through a dirty funnel arrives contaminated regardless of what the drum label says. |
Reading the report
| Item | What it means in practice |
|---|---|
| First number of the code | Particles larger than four microns per millilitre, expressed on a logarithmic scale. One step is roughly double the count. |
| Second number of the code | Particles larger than six microns. These are the ones that open clearances in pumps and valves. |
| Sample point | Mid-stream from a live sample port is the standard. A bottle dipped in the reservoir reports something else entirely. |
| Trend across samples | Three consecutive readings in the same conditions say more than any single number, because they show whether the system is gaining or losing cleanliness. |
| Water and viscosity on the same report | Water and viscosity often arrive with the count and change the interpretation. Free water makes a figure look better or worse than the system really is. |
How to confirm it
- Take the sample the same way every time
Use the machine's own sample port, with the oil at working temperature and the machine running or immediately stopped as the specification requires. Flush the port first and fill the bottle to the marked line. Sampling from the reservoir with a dip tube produces a number that cannot be compared with anything.
- Record the conditions with the sample
Hours since the last oil change, hours on the machine, whether the system was hot, and what work was done since the last sample. A number without conditions is a number you cannot use.
- Read the code against the target for that machine
Confirm the cleanliness target for the components fitted. A machine with a piston pump and servo control and a machine with an older gear pump do not share a target, and the same report can be comfortably inside one and outside the other.
- Compare against your own trend, not a general figure
The most useful comparison is with the previous three samples from the same machine under the same conditions. A system that is stable at a slightly high figure and a system climbing through the same figure are in different states of health.
- Act on a step change, and investigate before flushing
Two or more ISO steps in one sampling interval is an event: find the source, whether it is a repair, a failed breather or a wearing component. Flushing without finding the source restores the number for a while and nothing else.
- Re-sample after the corrective action
Cleanliness is only confirmed by a second sample. Re-sample after the same number of hours as the original interval, and record the result against the earlier readings so the next decision has history behind it.
What to do about it
Address the source first: breathers, seals, repair cleanliness and the way new oil is transferred into the machine. Filtration removes particles; it does not remove their cause.
Filter to the target rather than to a schedule. Where a count is above target, change the return and pilot filter elements and re-sample on a shorter interval until the trend is flat, then return to the standard schedule.
Use the count alongside the other tests rather than instead of them. A particle count says how dirty, water and viscosity testing say how wet and how thin, and the three together are what justify a change or a repair. Guangzhou Shenyue Machinery Technology Co., Ltd. supplies filters, breathers, seals and hydraulic components from Guangzhou, holds 3,000+ part types on the shelf with 100,000+ part numbers available to order, and states the grade on every line. 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.
Keep the sampling routine on a fixed interval and write the results down. The value of a cleanliness program is in the sequence, not in any single report, and the machine that has three years of trend data is the one where a fault is caught before it becomes a repair.
Parts you may need, by grade
| Part | Grades commonly quoted |
|---|---|
| Return and pilot filter elements | Genuine; OEM-quality; matched aftermarket element to the same rating. |
| Breathers and tank hardware | Genuine; OEM-quality. |
| Seals, O-rings and gaskets | Genuine; OEM-quality; specified material for the application. |
| Test ports, sample valves and fittings | Genuine; OEM-quality; matched aftermarket. |
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
What is a good ISO cleanliness code for an excavator?
It depends on the components. Modern machines with piston pumps and servo control normally carry a tighter target than older machines with gear pumps, and the correct figure is the one the manufacturer specifies for that system. The practical approach is to confirm the target, then compare each sample with your own trend, because a stable figure slightly above target and a climbing figure at the same value are different problems.
Can I judge oil cleanliness by looking at it?
No. Visible contamination starts at around forty microns, and the particles that open clearances in a hydraulic system are a fraction of that. Oil that looks perfectly clean can be two ISO codes too dirty, which is exactly why the count is measured rather than assessed by eye.
Is new oil clean enough to pour straight in?
Not usually. New oil has its own particle count, and oil transferred through an open container or a dirty funnel arrives dirtier still. Where a machine carries a tight cleanliness target, filter the oil as it goes in or have it delivered in a closed transfer arrangement, and sample after the top-up rather than before it.