There is a kind of situation that leads to arguments at acceptance: the machine runs through a batch, the oil sample goes to the laboratory and the particle count passes; a few days later another sample is taken and the result does not pass. The same machine, the same source of oil, no further treatment in between, and two different conclusions at the end.
More often than not this is not a failure of the filter. It is that some of the particles never travelled through the filter element at all.
Filtration intercepts, it does not collect
Mechanical filtration works by letting the oil pass through the filter medium and holding back the particles that are larger than the filtration accuracy, so that they stay in the element. The filtration accuracy of the ZYD series is ≤1 μm and the cleanliness indicator is NAS class 5 (NAS 1638); both figures describe the oil that has come through the filtration stage, how much was held back and how much is left.
Particles that the oil flow never carried are outside that description. The sediment at the bottom of the tank, the deposit on the tank wall and the debris that collects at the bottom of the filter housing all belong to this group. No matter how strong the filtration is, it cannot act on oil that does not pass through the element.
When the settling happens
When the oil stands still, particles that are denser than the oil settle under gravity. The periods when the machine is stopped, when the oil in the tank is left standing, when the tank is emptied for an element change and when the machine is opened for maintenance are the periods in which sediment gathers most easily. Settling is not a single event; it goes on a little at every standstill.
The next time the machine starts, those particles are lifted again by the oil flow and enter the filtration stage. They are larger than the particles normally carried by the oil, and they block the passages of the medium faster, so a contradictory picture appears: at the start of a run the pressure difference rises quickly, while the filtration accuracy itself is not the problem.

| Where the particles are | Does the element hold them back | How they get into the oil flow |
|---|---|---|
| In the oil that is flowing | Yes, intercepted at the filtration accuracy | They travel with the oil flow anyway |
| In the settled layer at the bottom of the tank | No, they never pass the element | Circulation lifts the oil from the bottom of the tank |
| Attached to the tank wall and to the bottom of the filter housing | No | Disturbed by the flow, and cleaned out during maintenance |
| In oil standing in the pipework and in the storage tank | No | Stirred or drawn off before it enters the machine |
Bringing the sediment back into the flow
Since settling is unavoidable, the answer is not to prevent it but to control under which conditions it is lifted and where it goes once it is.
- Let the oil circulate once at a low flow rate after starting, to lift the sediment and then pass it into the fine filtration stage, so that large particles do not rush straight at the fine elements.
- The five fine elements 150-500-5 are in parallel, so changing one of them does not stop the others, but in a machine with a heavy sediment load this first circulation is still needed.
- Empty the tank before an element change, clear the sediment from the housings and the bottom of the tank before putting the new element back, so that what has settled is not drawn straight into it.
- When the sample point is at the bottom of the tank the sample is more likely to carry sediment, so the sampling point and the operating data have to be read together.
Telling a settled bed from a spent element
The difference is in when the pressure difference appears. A spent element shows a pressure difference that climbs slowly with the volume treated and comes back after the element is changed; sediment that has been lifted shows a sudden rise at one start-up and settles down again after a few circulations. The first calls for an element change, the second for clearing the sediment.
Reading the operating record together with the test results makes the judgement much easier: during which period the pressure difference rose, how long the machine stood before that start-up, and where the sample was taken. That information is worth more than a single pressure reading.
A filter only answers for the oil that flows through it; the particles lying at the bottom of the tank do not pass the element, and no accuracy will hold them back.
