ISO 4406 code up two ranges: real trend or a bad oil sample?

This check covers ISO 4406 codes from automatic particle counters calibrated to ISO 11171 on light, clear lubricating and hydraulic oils sampled on a routine schedule. It does not cover codes from microscope counts, dark or opaque oils that need pore-blockage counting, very clean fluids below scale number 8, or a single sample with no history to compare against.

When should you question a two-range ISO 4406 rise?

Question a two-range ISO 4406 rise whenever one sample jumps two codes in one or more channels and nothing on site explains it, because a single report cannot tell a contamination event from a sampling or counting error. ISO 4406:2021 itself warns about this in clause 4.1: "Particle counts are affected by a variety of factors. These factors include procurement of sample, particle counting accuracy and the sample container (where used), and its cleanliness." (ISO 4406:2021, clause 4.1).

Four situations call for the check before anyone orders a filter cart, a flush or an oil change. First, the rise appears in one sample after a long flat history, with no maintenance, top-up or filter change logged in between. Second, the sample was drawn by a different person, at a different valve, or while the machine was stopped. Third, the same report shows more water than usual, or the oil type is known to foam. Fourth, the rise appears in the 4 and 6 µm(c) channels but the 14 µm(c) channel does not move, or the opposite, and nobody has asked which pattern it is.

In all four situations a two-range rise in an ISO 4406 code is a reason to test the sample before acting on the machine. A confirmed rise is then worth acting on, because a two-range move in a well-sampled system is rarely noise.

What do you need before judging an ISO 4406 code change?

Before judging an ISO 4406 code change, you need the raw particle counts behind each code, the sampling record for each sample, the water result, and the lab's counting method. The code alone is not enough. ISO 4406:2021 assigns each scale number to a band of counts per millilitre, so a code hides where inside the band the count sits (ISO 4406:2021, Table 1).

Collect five items. One, the cumulative counts per mL at 4, 6 and 14 µm(c) for the new sample and at least three earlier samples from the same point. Two, the sampling record: valve or port, machine running or stopped, oil temperature, flushing volume, bottle type and the time between sampling and testing. Three, water content, ideally by Karl Fischer titration under ASTM D6304-25, whose precision statement covers 20 mg/kg to 25,000 mg/kg for Procedure A (ASTM D6304-25, 2025). Four, the counting method: whether the lab counted the neat oil under ISO 11500:2022 or a dilution method such as ASTM D7647-24, and whether the counter is calibrated to ISO 11171, shown by the (c) after each size. Five, the maintenance log for the period between samples: filter changes, top-ups, seal work, flushing and any open-reservoir work.

Without raw counts and the sampling record, the ISO 4406 check cannot go past its first step, and the only safe choice is to resample.

How do you decide whether an ISO 4406 rise is real, step by step?

To decide whether an ISO 4406 rise is real, run five criteria in order: raw count change, sampling point and conditions, water and air interference, the pattern across the three channels, and a confirming resample. Stop at the first criterion that fails, because a later check cannot repair a bad sample. The table below decides between acting on a trend, recounting, and resampling.

The key row is the confirming resample: a two-range ISO 4406 rise that a second, properly drawn sample repeats is a trend, and one that the resample does not repeat is a sample or counting problem, whatever the first report said.

How much does a two-range ISO 4406 rise mean in particle counts?

A two-range ISO 4406 rise means the particle count in that channel grew by a factor of about 2 to 7.7, not exactly four, because each code is a band and the count can sit anywhere inside it. The calculation uses the published band limits. In ISO 4406:2021, scale number 18 covers more than 1,300 and up to 2,500 particles per mL, and scale number 20 covers more than 5,000 and up to 10,000 particles per mL (ISO 4406:2021, Table 1; Maintenance World, 2015). A count that moves from 2,500 to 5,001 per mL changes two codes while growing only about 2 times; a count that moves from 1,301 to 10,000 per mL also changes two codes while growing about 7.7 times (calculated from ISO 4406:2021, Table 1).

A one-range rise spans an even wider spread. Polaris Laboratories shows a 4 µm count moving from 2,485 to 2,510 per mL, which changes the code from 18 to 19 although the count barely moved (Polaris Laboratories, 2018). At the other end, nearly four times as many particles can raise the code by only one (Maintenance World, 2015), for example a count growing from 1,301 to 5,000 per mL (calculated from ISO 4406:2021, Table 1).

The practical rule follows from the table. Read the raw counts and compute the ratio of the new count to the recent counts in the same channel. A ratio near 2 at a band edge is weak evidence, and a ratio of 4 or more is strong evidence that the ISO 4406 particle load really changed.

Why do sources disagree on what one ISO 4406 range step means?

Sources disagree on what one ISO 4406 range step means because some describe the band limits, which double, and others describe actual counts, which need not double. Both are correct about different things. ISO 4406:2021 uses careful wording: "The purpose of this code is to simplify the reporting of particle count data by converting the numbers of particles into broad classes or codes, where an increase in one code generally represents a doubling of the contamination level." (ISO 4406:2021, clause 4.1). Clause 4.3.2 adds that a step ratio "of generally two" was adopted to keep the number of scale numbers within a reasonable limit (ISO 4406:2021, clause 4.3.2).

The Sensemore ISO 4406 calculator page reads the scale band against band: it states that an oil two codes above its target is four times dirtier, which is fair when comparing a result with a specified target band (Sensemore, ISO 4406 calculator). Polaris Laboratories reads the scale count against count and shows a one-code change without doubling and a near-doubling without a code change (Polaris Laboratories, 2018). The band reading also hides movement inside a band: in the Polaris example, a 4 µm count that rises from 1,310 to 2,490 per mL nearly doubles and stays at code 18 (Polaris Laboratories, 2018).

For trend decisions, the count-against-count reading governs. Compare raw counts between samples, and keep the band-against-band reading for setting targets and alarm limits.

Which sampling changes make an ISO 4406 code jump?

The sampling changes that most often make an ISO 4406 code jump are a different sampling point, a stopped or cold machine, a poorly flushed valve or dirty bottle, and a sample left to settle before counting. ISO 4021:1992 prefers extracting samples from a main flowline so that the particulate contaminant in the sample represents the fluid flowing at the sampling point, and allows reservoir sampling only when no suitable sampler is fitted (ISO 4021:1992). As of 2026-10-23, ISO lists ISO 4021:1992 as published and expected to be replaced by ISO/PRF 4021 (ISO 4021:1992, ISO page).

The Sensemore calculator page explains why the point matters: a sample from the bottom of a settled reservoir reads dirty because that is where the dirt went, and a sample downstream of the filter shows how well the filter works rather than what is in the machine (Sensemore, ISO 4406 calculator). Machinery Lubrication recommends a sample bottle at least 2 ISO range codes cleaner than the lowest cleanliness level of interest, and agitation for five minutes in a paint shaker if more than a few minutes pass between sampling and testing (Machinery Lubrication, Oil Analysis 101).

Maintenance itself can also move codes. Turbine Oil Reliability notes that a system flush may temporarily disturb settled debris and worsen the large-particle code before it stabilizes (Turbine Oil Reliability, 2026). Check the maintenance log before calling an ISO 4406 jump a trend.

How do water and air bubbles inflate ISO 4406 particle counts?

Water droplets and air bubbles inflate ISO 4406 particle counts because a light-extinction particle counter sees them as particles, so a wet or aerated sample can report a higher code with no more solid contamination. Machinery Lubrication states that air bubbles and free and emulsified water both appear as particles to optical particle counting (Machinery Lubrication, Oil Analysis 101). ISO 11500:2022, the light-extinction counting procedure, is written for bottle samples of clear, homogeneous, single-phase liquids (ISO 11500:2022), which a sample carrying free water or foam is not.

ASTM D7647-24 addresses this with dilution. Its significance section states that to measure hard particle contamination accurately, it is necessary to negate the counts contributed by small levels of free water, and that its water-masking diluent does this (ASTM D7647-24, section 5.5). It also notes that polydimethylsiloxane defoamant and oxidation by-products can cause falsely high counts as soft particles (ASTM D7647-24, section 5.6). OELCHECK describes its own practice based on ASTM D7647: the sample is homogenized and degassed, propanol makes water droplets invisible, and the diluted sample is counted three times, with the analysis rejected if the counts differ greatly (OELCHECK, particle count).

When an ISO 4406 rise comes with higher water, ask for a dilution recount by ASTM D7647-24 before judging it. OELCHECK's dilution, based on ASTM D7647, uses a solvent mixture of 75% toluene and 25% propanol (OELCHECK, particle count); the lab handles it under its SDS and local solvent rules, not the site.

How many oil samples make an ISO 4406 trend?

None of the standards cited here, including ASTM D7669-20, sets a sample count for an ISO 4406 trend, so treat one high sample as a signal to check and two consistent samples from the same port as the minimum for a confirmed change. ASTM D7669-20 is a guide to practical trend analysis of lubricant condition data, and it leaves limits for specific parameters to other practices or the equipment maker (ASTM D7669-20, 2020). Machinery Lubrication advises that particle count direction alone should not be used as a failure indicator, because the difference between codes can be one particle or several thousand (Jarrett, Machinery Lubrication).

A workable baseline is three or more valid samples from the same port before the jump, so that the raw count spread inside the normal band is known. The confirming resample should be drawn as soon as practical, from the same port, with the machine running at its usual temperature, and before any filtration is added, or the resample cannot test the first result. Machinery Lubrication suggests a cautionary limit at the target or one code above, and an alarm limit two or three codes above the target (Jarrett, Machinery Lubrication).

Trend the three channels separately rather than as one code. The same discipline of fixing the comparison before reading the change applies outside oil analysis, as in deciding whether a rise in customer complaints is product quality or better logging.

Worked ISO 4406 resample example (illustrative)

In this illustrative ISO 4406 resample example, a hydraulic press with a cleanliness target of 18/16/13 reports 20/18/13 after a long flat history. All counts and water values are invented for illustration. The table shows the counts per mL at 4, 6 and 14 µm(c) and the decision each sample supports.

The key row is the resample from the usual line port: it returned 18/16/13, so in this illustrative case the two-range rise came from the reservoir drain sample and its free water, not from the machine. The raw ratio in the new sample was about 2.5 times at 4 µm(c), only a little above the two-fold floor for a two-code rise.

Two findings remain. Water at the reservoir bottom is worth checking at the breather and seals. And the 4 µm(c) count has drifted from 1,900 to 2,300 per mL across valid samples while staying in code 18, which only the raw counts show. In a real case, take band limits from ISO 4406:2021, Table 1.

How should you report an ISO 4406 trend decision?

An ISO 4406 trend decision should be reported as a short record that lets the next reader see which samples count, which do not, and why. Start with the sample identity for each result: machine, sampling port, running state, oil temperature, sampling date, test date and counting method. List the raw counts at 4, 6 and 14 µm(c) next to the codes, so that movement inside a band is visible, and state the ratio of the new count to the baseline in each channel.

Then record each row of the criterion table with its outcome: passed, failed, or not checked. A sample that failed the sampling or interference check stays in the record with its reason, marked invalid, and does not enter the trend line. Do not replace a missing or invalid sample with zero or with the previous value; leave the gap in the trend. Keep open items, such as a dilution recount requested or a resample not yet drawn, apart from conclusions.

If the rise is confirmed, state the action against the cleanliness alarm limits and the channel that drove it. In Lubinpla, a trend analysis task fixes the period, equipment, comparison axis and conversion rules before it looks at results, does not fill missing samples with zero, and interprets only the changes it can trust, in a cited document for your review.

Where does this ISO 4406 trend check not apply?

An ISO 4406 trend check does not apply where the counts themselves are not reliable or not comparable. ISO 4406:2021 notes that reproducibility below scale number 8 is affected by the actual number of particles counted, and that the raw count should be more than 20 particles (ISO 4406:2021, Table 1 note). Very clean fluids therefore show code changes that a resample may not repeat for statistical reasons alone.

Dark and heavy oils fall outside light-extinction counting. ASTM D7647-24 applies to lubricants up to 1,000 mm2/s at 40 °C, is not suitable for samples that are opaque after dilution, and may not suit samples with visible particles (ASTM D7647-24, scope). Machinery Lubrication notes that pore-blockage counters avoid false positives from air, water and dark fluid but have a smaller dynamic range and estimate the size distribution (Machinery Lubrication, Oil Analysis 101). Codes from different counting principles should not share one trend line.

Codes measured before the change to ISO 11171 calibration are not directly comparable either. ISO 4406:2021 reports at 4, 6 and 14 µm(c), where the last two sizes are equivalent to the 5 µm and 15 µm sizes of the first edition (ISO 4406:2021, clause 4.1). Finally, a confirmed trend says the particle load changed; it does not say what the particles are. Ferrography, elemental analysis or patch microscopy answers that question.

Key takeaways

• Read raw counts, not codes; a two-code rise can mean a count that only doubled.

• Compare samples only from the same port, with the machine running at its usual temperature.

• Ask for a dilution recount by ASTM D7647-24 when water or foam is present.

• Trend the 4, 6 and 14 µm(c) channels separately.

• Confirm with a resample from the same port before adding filtration or changing oil.

• Mark invalid samples with their reason and leave the gap; never fill it.