BlogApplication notes
How should I set my oil change interval?
- Type
- Application notes
- Published
- 20 November 2026
- Prepared by
- Technical team
- Product group
- Equipment/Service Lubricants
Following the maker's interval by the calendar is the usual approach, but the same oil lasts very differently with heat, water and dirt, so the calendar is early for some machines and late for others. What actually ends an oil's life can be oxidation, depleted additives, water in the oil or built-up particles, and these tend to advance together, each hastening the others.
The maintenance schedule lists an oil change date for each machine, and when the day comes the oil is drained whatever its state. Yet what comes out differs: oil from one machine is clear and barely different from new, while oil drained the same day from an identical unit has already darkened and left sludge on the bottom. Sometimes a filter blocks or a bearing starts to sound different before the date, and the change is brought forward.
Changing too early throws away usable oil, disposal cost and downtime. Changing too late lets acids and sludge corrode parts, while blocked filters and rising viscosity break down the oil film and damage bearings, gears and valves. Setting the interval is therefore as much about equipment reliability as about oil.
A lubricant is a base oil, the bulk of the fluid, plus additives such as antioxidants, antiwear agents and rust inhibitors (Rudnick, 2017). The base oil provides the viscosity that holds a film between metal surfaces, and the additives protect that film and the base oil. The change point is when either of the two can no longer do its job. The turbine-oil monitoring practice states that its test values are only indicative and that equipment type, workload, circuit design and top-up rate must be weighed with them (ASTM International, 2024), and no single change limit fits every machine (Van Rensselar, 2012). The practice for auxiliary equipment oils likewise advises changing on condition and test results rather than fixed service hours (ASTM International, 2023).
When setting a change interval, the technical team sets up the four hypotheses below for what is ending this machine's oil and rules them out one by one. The start is always the same: keep a sample of the new oil as the baseline, then sample from the same point on a regular schedule and trend viscosity, acid number, water and particle count.
Hypothesis 1: oxidised oil
Oxidation goes on wherever oil meets heat and oxygen, so it is the first hypothesis. It runs faster on machines with a blocked cooler or low flow, on machines loaded beyond design, and on those sitting outdoors in summer or near a heat source. A sample from a gearbox on circulating lubrication with months still to go before its change date comes out dark brown with a burnt smell, and the filter, once pulled, has sticky brown film between its pleats.
Oxidation is a chain reaction that keeps itself going. When heat or metal helps strip a hydrogen from an oil molecule, a radical forms, an unstable molecule with an unpaired electron, and it joins with oxygen to become a peroxy radical. The peroxy radical takes a hydrogen from a neighbouring molecule, forming a peroxide called a hydroperoxide and leaving a new radical behind (Rudnick, 2017). A study that aged polyalphaolefin base oil at 140 °C with oxygen confirmed that hydroperoxides break down into peroxy, alkoxy and carbon-centred radicals (Babić et al., 2021).
What remains at the end are acidic compounds such as carboxylic acids and larger, linked molecules. The acids raise the acid number and corrode metal, and the large molecules raise viscosity and drop out as insoluble sludge and varnish, the thin film baked onto surfaces (ASTM International, 2022b). The rate is sensitive to temperature: as a rule of thumb, each 10 °C rise roughly doubles the oxidation rate and halves oil life. It is only an approximation, and in practice contamination and mechanical stress act alongside it (Fitch, n.d.).
Colour and smell are clues; the verdict comes from numbers. Compare kinematic viscosity at 40 °C, the viscosity measured by how fast oil flows under its own weight, with the new oil: as oxidation proceeds, it climbs steadily. Acid number shows how much acid has built up, and infrared spectroscopy, FT-IR, trends oxidation products through the carbonyl band at 1800 to 1670 cm⁻¹. The infrared method does not suit ester-based oils (ASTM International, 2022b). In systems where varnish is a concern, such as turbine oils, membrane patch colorimetry, or MPC, filters insoluble oxidation products onto a patch and trends them as a colour difference, ΔE (ASTM International, 2025c).
When viscosity or acid number begins a steady rise against the new oil, shorten the gap to the next sample, and plan the change once the rise steepens. Before changing, check operating temperature first: if a fouled cooler, low flow or overload has pushed it up, new oil will oxidise just as fast (Fitch, n.d.). Where a varnish indicator is rising, an oil change alone does not lift deposits off surfaces, so consider system flushing or deposit-removal filtration as well.
Hypothesis 2: depleted additives
If viscosity and acid number were normal all along and then turned bad all at once, look at the additives. This shows on machines whose oil has not been changed for a long time, on machines whose heat and load use additives up fast, and on those with so little top-up that almost no fresh additive comes in. Take a compressor whose change has been put off because the last few analyses showed nothing wrong: this time the report comes back showing viscosity and acid number jumping together.
Antioxidants are used up a little each time they break the chain. Primary antioxidants such as hindered phenols and aromatic amines give a hydrogen to peroxy radicals to stop the reaction, and secondary antioxidants decompose hydroperoxides into stable products (Rudnick, 2017). Zinc dialkyldithiophosphate, or ZDDP, forms a protective film on rubbing surfaces to reduce wear and is also known to decompose peroxides (Spikes, 2004). While these remain, the base oil barely changes.
Once the additives are gone, things change. Nothing stops the radicals, the base oil oxidises directly, and viscosity and acid number rise steeply in a short time. This is why oil watched only by viscosity and acid number looks normal until just before its end and then seems to fail suddenly, and why remaining additive is measured on its own (ASTM International, 2022a). In engine oils, the reserve of alkaline additives that neutralise acid is read as the base number, or BN. Used oil contains weakly basic degradation products, so its base number is best measured by the hydrochloric acid titration of ASTM D4739 (ASTM International, 2026).
Remaining additive is measured three ways. Linear sweep voltammetry under ASTM D6971, which reads antioxidants from the current that flows as voltage is raised, gives phenolic and aminic antioxidants in zinc-free turbine oils as a share of the new oil (ASTM International, 2022a). Elemental analysis under ASTM D5185, or ICP, which burns the sample in a hot plasma and reads the light of each element, compares additive elements such as zinc, phosphorus and calcium with the new oil (ASTM International, 2018). For engine oil, watch for the point where falling base number and rising acid number draw close; the change limit must be set from experience for each machine and oil (ASTM International, 2026).
Extending the slope of a steady fall in remaining antioxidant gives an early estimate of when it will run out, and the change is planned ahead of that. The standard advises against judging remaining life from voltammetry alone, though, and recommends reading it with an oxidation stability test and other indicators (ASTM International, 2022a). In large systems part of the oil is sometimes replaced with new to top up additives instead of a full change, but if oxidation products are already high the benefit may not last. Adding aftermarket additives can upset the formulation balance, so consult the oil supplier first.
Hypothesis 3: water in the oil
Water gets in through condensation, coolant leaks, washdown, and breathing, the tank drawing in humid air as it cools. Inside the oil it takes three forms: dissolved water, finely dispersed droplets and free water settled at the bottom. In a season of wide day-to-night temperature swings, droplets form every morning inside the sight window of a gear reducer that stands outdoors, and the oil has lost its clear colour and turned hazy.
Water shortens bearing fatigue life. In full-scale tests on tapered roller bearings with an SAE 20 oil at 25, 100 and 400 ppm water, a clear correlation appeared: the more water, the shorter the fatigue life. The same study found that oils differ in how much water they absorb, and that this difference can affect bearing life (Cantley, 1977).
Water damages the oil as well. Water and metal particles together can sharply speed oxidation (Fitch, n.d.), and some additives are lost, either washed out into the water or split apart by reacting with it, which is hydrolysis (Rudnick, 2017). The standard likewise notes that moisture leads to corrosion, wear, debris and filter plugging (ASTM International, 2025a). In this way water brings forward both hypothesis 1 and hypothesis 2.
Start with appearance and narrow it down with numbers. Hazy oil, water at the tank bottom or a milky emulsion means the oil is already past the amount it can dissolve. Quantify water down to the ppm level by coulometric Karl Fischer titration under ASTM D6304 (ASTM International, 2025a), and keep the crackle test, dropping oil on a hot plate and listening for spitting, for screening only. If sodium, potassium and boron rise together in elemental analysis, suspect coolant, because these elements are common in coolant additives.
Rising water calls for closing its way in before any oil change. Inspect coolers and seals for leaks, shield the points that washdown reaches, and fit desiccant breathers to tank vents. Drain free water from the tank bottom, and reduce dissolved water with a vacuum dehydrator or absorbent filter. Take water targets from the equipment maker's or oil supplier's recommended range, and if coolant is confirmed, the additives may be damaged, so fix the cause and favour a change.
Hypothesis 4: built-up particles
Vents, seals, top-up oil and maintenance work all let particles in, and wear inside the machine makes more. They build up at dusty sites, on tanks with unfiltered vents, and where top-up is poured through a funnel from an open drum. Think of a plant with cement dust in the air, where a fitter pours oil from an open drum through a funnel, and a few weeks later the filter differential pressure alarm sounds earlier than usual.
Hard particles scratch surfaces and age the oil. Particles about the size of the oil film get caught between surfaces, scratch and dent them, and free fresh wear particles in turn, a cycle that feeds itself. Metal particles such as iron and copper also act as oxidation catalysts, hastening hydroperoxide breakdown and radical formation (Rudnick, 2017). That is why clean oil lasts.
Oil cleanliness is expressed as the three-number code of ISO 4406. Each number turns the particle count per mL above one of three size thresholds into a class, and each step up in the code roughly doubles the count (ISO, 2021). So 18/16/13 and 20/18/15 look close as numbers but differ about fourfold in particle count.
The basic measurement is an ISO 4406 code from an automatic particle counter, compared with the target cleanliness (ISO, 2021). In elemental analysis, rising iron, copper and chromium suggest wear, and rising silicon suggests dirt ingress. Elemental analysis reads low for particles larger than a few micrometres, though, so large wear particles need a particle count or ferrography, which gathers iron debris with a magnet for viewing under a microscope (ASTM International, 2018). Shorter filter life is itself a sign of more ingress or of wear starting.
When only the particle count is high, an oil change is the most expensive fix. If viscosity, acid number and additives are normal, so the oil is chemically healthy, filtering it back to cleanliness is the better choice. Check vent filters, seals, and how top-up oil is stored and transferred, and filter new oil before it goes in. If wear elements rise as well, inspect the machine before the oil, and take target cleanliness from the equipment maker's recommendation.
Testing the hypotheses
The four hypotheses are not investigated one by one; they are sorted together from the set of results on each sample. That needs a new-oil baseline first, and a fixed sampling point and interval (ASTM International, 2023). Samples are comparable only when drawn from the same point under the same running conditions, and one result cannot show a trend, so read several in a row and look for changes that leave the normal range (Van Rensselar, 2012).
Look first at what the eye can see. Haze or water points to the water of hypothesis 3 first; deposits point to the oxidation of hypothesis 1 and the particles of hypothesis 4. Viscosity comes next: a rise against the new oil suggests oxidation, and a fall calls for a check on mixing with another oil or fuel. Read chemical degradation from acid number and the infrared oxidation indicator, adding base number for engine oil (ASTM International, 2022b, 2026).
Remaining life is read from antioxidant level and additive elements. If viscosity and acid number are still normal but additives are falling fast, hypothesis 2 is ahead (ASTM International, 2018, 2022a). Contamination shows in Karl Fischer water and the ISO 4406 particle code (ASTM International, 2025a; ISO, 2021), and if wear elements rise too, inspect the machine before the oil (ASTM International, 2018). Once data from similar machines build up, caution and alarm limits can be set statistically under ASTM D7720, though this does not replace limits set by the equipment maker (ASTM International, 2025b).
This is also where the calendar falls short. The four hypotheses progress at different rates by machine and by season. Even on identical units, oil running 10 °C hotter oxidises about twice as fast (Fitch, n.d.), and oil exposed to water or dirt ends sooner. The practice for auxiliary equipment oils aims to change oil in gears, hydraulics, diesel engines, pumps and compressors on condition and test results rather than service hours, cutting operating and maintenance cost while guarding against wear and contamination (ASTM International, 2023), and a gasoline engine study likewise set the drain point from the trend of iron in used oil (Kaleli & Khorramian, 1998).
To move the interval, begin by keeping to the maker's interval while taking two or three samples within it to build a trend. If the oil is still healthy at change time, extend the next interval a step at a time, confirming with the same indicators at each step. If an indicator turns mid-interval, shorten the interval and find which of heat, water or particles is behind it. The decision rests not on one number but on a set: viscosity, acid or base number, additives, water and particles.
When hypotheses overlap, separate what can be reversed from what cannot. Water and particles can be removed by dehydration and filtration, but acids from oxidation and spent additives are hard to undo. The basic call, then: if chemical indicators such as viscosity, acid number and additives near their limits, change the oil; if they are normal and only contamination is high, clean the oil and close the way in. Contamination speeds chemical degradation, so changing oil again and again while contamination continues keeps shortening the interval (Fitch, n.d.; Rudnick, 2017).
Limits and cautions
Change limits from the equipment maker and oil supplier come first, and the standards themselves call indicator values indicative (ASTM International, 2024, 2025b). For synthetic oils such as esters and phosphate esters, some methods such as infrared oxidation analysis do not apply, so agree test methods with the supplier first (ASTM International, 2022b, 2023). Grease is sampled and replaced differently from oil, and its relubrication interval is set separately by bearing type, speed and temperature, so judgements for circulating and splash-lubricated oils do not carry straight over. For equipment under warranty, check the maker's conditions before extending the interval.
When sampling or changing oil, beware of hot oil and residual pressure, and wear protective gloves and eye protection as the product's safety data sheet, the SDS, directs. Dispose of used oil under local rules. A sudden rise in wear elements, coolant in the oil or an unexplained sharp change in viscosity may signal machine failure rather than an oil problem, so refer it to maintenance specialists and an oil analysis laboratory.
Sources
- ASTM International. (2018). ASTM D5185-18, Standard test method for multielement determination of used and unused lubricating oils and base oils by inductively coupled plasma atomic emission spectrometry (ICP-AES). https://store.astm.org/d5185-18.html
- ASTM International. (2022a). ASTM D6971-22, Standard test method for measurement of hindered phenolic and aromatic amine antioxidant content in non-zinc turbine oils by linear sweep voltammetry. https://store.astm.org/d6971-22.html
- ASTM International. (2022b). ASTM D7414-22, Standard test method for condition monitoring of oxidation in in-service petroleum and hydrocarbon based lubricants by trend analysis using Fourier transform infrared (FT-IR) spectrometry. https://store.astm.org/d7414-22.html
- ASTM International. (2023). ASTM D6224-23, Standard practice for in-service monitoring of lubricating oil for auxiliary power plant equipment. https://store.astm.org/d6224-23.html
- ASTM International. (2024). ASTM D4378-24, Standard practice for in-service monitoring of mineral turbine oils for steam, gas, and combined cycle turbines. https://store.astm.org/d4378-24.html
- ASTM International. (2025a). ASTM D6304-25, Standard test method for determination of water in petroleum products, lubricating oils, and additives by coulometric Karl Fischer titration. https://store.astm.org/d6304-25.html
- ASTM International. (2025b). ASTM D7720-21, Standard guide for statistically evaluating measurand alarm limits when using oil analysis to monitor equipment and oil for fitness and contamination. Reapproved 2025. https://store.astm.org/d7720-21r25.html
- ASTM International. (2025c). ASTM D7843-25e1, Standard test method for measurement of lubricant generated insoluble color bodies in in-service turbine oils using membrane patch colorimetry. https://store.astm.org/d7843-25e01.html
- ASTM International. (2026). ASTM D4739-26, Standard test method for base number determination by potentiometric hydrochloric acid titration. https://store.astm.org/d4739-26.html
- Babić, N., Pondaven, S., & Vezin, H. (2021). EPR spin-trapping study of free radical intermediates in polyalphaolefin base oil autoxidation. Polymer Degradation and Stability, 192, 109687. https://hal.univ-lille.fr/hal-03368449
- Cantley, R. E. (1977). The effect of water in lubricating oil on bearing fatigue life. ASLE Transactions, 20(3), 244–248. https://doi.org/10.1080/05698197708982838
- Fitch, B. (n.d.). How heat affects lubricants: Understanding the Arrhenius rate rule. Machinery Lubrication. https://www.machinerylubrication.com/Read/32752/how-heat-affects-lubricants-understanding-the-arrhenius-rate-rule
- International Organization for Standardization [ISO]. (2021). ISO 4406:2021, Hydraulic fluid power — Fluids — Method for coding the level of contamination by solid particles. https://www.iso.org/standard/79716.html
- Kaleli, H., & Khorramian, B. (1998). Used oil analysis and study of oil drain period in gasoline engine (SAE Technical Paper 981448). SAE International. https://doi.org/10.4271/981448
- Rudnick, L. R. (Ed.). (2017). Lubricant additives: Chemistry and applications (3rd ed.). CRC Press. https://www.routledge.com/Lubricant-Additives-Chemistry-and-Applications/Rudnick/p/book/9781498731720
- Spikes, H. (2004). The history and mechanisms of ZDDP. Tribology Letters, 17, 469–489. https://doi.org/10.1023/B:TRIL.0000044495.26882.b5
- Van Rensselar, J. (2012, May). Used oil condemning limits. Tribology & Lubrication Technology. https://www.stle.org/files/TLTArchives/2012/05_May/Cover_Story.aspx