What is technical cleanliness for machined parts, and how do you specify it under ISO 16232 and VDA 19?

How does ISO 16232 measure particles on a part?

A lab rinses the defined surfaces with a test fluid, filters the fluid through a membrane, then weighs the residue, counts and sizes the particles under a microscope, or both. The particles are sorted into lettered size classes by their longest dimension.

The extraction method is chosen to suit the part. Pressure rinsing with a directed jet suits external surfaces and accessible bores; ultrasonic extraction suits small parts that can be immersed; agitation suits closed volumes; internal rinsing suits passages and galleries that a fluid can flow through; air-jet extraction exists for parts that must not get wet. Before routine testing, the extraction is qualified with a declining test, in which the same part is extracted repeatedly and, per ISO 16232, the last extraction must yield no more than 10% of the cumulative total within a maximum of six extractions. If it never declines, the method is not getting the particles off, and the result understates the real contamination.

Two analysis routes follow. Gravimetric analysis weighs everything on the membrane and reports a residual mass per component in milligrams; it is quick and cheap but cannot tell one large chip from a cloud of fine dust. Light-microscope analysis counts and sizes particles and, as industry guidance, separates metallic shiny particles from non-metallic particles and fibers, which is what matters for a hydraulic part where a hard steel chip is far more dangerous than a cotton fiber. Scanning electron microscopy with EDX is used when the material of a particle has to be identified, for example to trace a chip back to a specific alloy in the routing.

Every test also needs a blank value: the same procedure run without the part, to measure what the cabinet, fluid and handling contribute. Per ISO 16232, as in VDA 19.1, the blank should be no more than 10% of the specified or expected cleanliness value; if it is higher, the test cannot distinguish a clean part from a dirty lab.

Counts can be normalized three ways in a Component Cleanliness Code: per component (N), per wetted surface area (A, referenced to an area of a thousand square centimeters) or per wetted volume (V). Normalizing by area is what lets a supplier compare a small fitting with a large housing on the same scale. For a single part number, per component is simpler and is what most drawings use.

How should you choose the limit values?

Start from the largest particle the system can tolerate, and specify that first, separately for metallic and non-metallic particles. A residual-mass limit or a count per size class is a useful second line, not a replacement for the largest-particle limit.

The failure that hurts a hydraulic or fuel system is usually one hard particle bigger than a clearance, not a large total mass of fine dust. A limit written only as a mass, such as a few milligrams per part, can pass a part carrying a single chip large enough to lock a valve. So the useful order is: the maximum allowed size for metallic particles, a separate and usually larger maximum for non-metallic particles, fibers either excluded or limited on their own, and then optionally a mass or a count per size class to track the process over time.

To set the size, look at the tightest gap downstream of the surface in question, the filter rating that protects it, and whether particles from this part can reach it before that filter. A sealing face behind a filter can tolerate much more than a pilot orifice upstream of it. Where a customer standard already exists, use it: hydraulic suppliers often work to ISO 18413 for parts cleanliness, which follows the same extraction-and-analysis logic, and medical device makers commonly quantify extractable residue by the gravimetric method in ASTM F2459-24, with cleaning validation following the manufacturer’s own procedure.

Do not copy a limit from another part because it looked strict. A limit tighter than the system needs forces extra washing stages, cleaner packaging and more expensive testing, and it may be impossible to verify if the blank value of an ordinary lab is close to it.

Which cleanliness callouts cause problems with suppliers?

The most common problems are notes that cannot be measured, limits that cannot be verified, and specifications that confuse part cleanliness with fluid cleanliness. Each one leads to an argument at incoming inspection rather than a clear accept or reject.

• “Free of chips, burrs and contamination.” Every supplier will agree to it and no lab can test it. Keep it as a workmanship note if you like, but add a measurable limit next to it.
• A fluid code on a part drawing. ISO 4406 codes describe particles per milliliter of hydraulic oil. They describe the oil in a running system, not the debris on a dry component, and a supplier cannot report against them without inventing a conversion.
• No tested surface. Rinsing the whole outside of a housing picks up handling dust from faces that never see fluid. Limit the test to the wetted surfaces.
• Burrs that are not loose yet. A cleanliness test only finds what comes off. A burr still attached at a cross-hole intersection passes the test and breaks free in service, so cleanliness and deburring are separate requirements; the edge break guide covers how to call out edges.
• Particles locked in porosity. On machined castings, chips and coolant can sit inside opened pores and come out later. The porosity-after-machining guide explains where that happens.
• Plating or coating after the clean. A part cleaned before finishing has to be cleaned again afterwards if the limit applies to the finished part. Say which state the limit applies to.