Engine oil: grades, synthetic versus mineral, and what the numbers mean

There are two questions you can ask about a bottle of engine oil. The first is whether it is synthetic or mineral. The second is whether it is the right oil for your engine.

The piece this is translated from spends its entire length on the first question and never once answers the second. It never explains what 5W-30 means, never mentions API or ACEA, and never says that the handbook specifies both a viscosity grade and a specification and that both are binding. That ordering is backwards, and the reason is worth stating plainly:

Synthetic versus mineral is a quality question, where the worst case is that you paid too much. Grade and specification are compatibility questions, where the worst case is a blocked particulate filter or a camshaft timing fault.

So this starts with the numbers, and gets to synthetic afterwards — where it belongs, and where the original's central technical claim turns out to be backwards too.

The grade: what 5W-30 actually says

It is two separate measurements with a W between them, and the first thing to know is that W stands for Winter, not weight. These are not "weights" and the numbers are not a scale of thickness you can read off intuitively.

Two consequences people get wrong.

A lower W number is not thinner oil. A 0W-30 and a 5W-30 are the same oil at operating temperature — both are 30-grade at 100 °C. The 0W simply flows better when starting from cold. Since a large share of engine wear happens in the seconds after a cold start, before oil pressure is established everywhere, the lower winter number is doing real work and costs nothing at temperature.

The second number is not "more protection". It is the viscosity the engine was designed around — its bearing clearances, its oil pump's capacity, and on a modern engine its variable valve timing, which uses oil as a hydraulic fluid. Going up a grade because thicker sounds safer is a change to a design parameter. It shows up as slower cam phasing on cold mornings, and it sets codes: P0011 and P0016 are what an engine says when the cam is not where the map expects it, and incorrect engine oil viscosity is on the cause list for both.

The specification, which is the half nobody reads

The grade tells you how thick. It tells you nothing about what the oil is made to do — how much detergent it carries, how it behaves in an exhaust after-treatment system, whether it will hold up to a long drain interval. That is the specification, and it is printed on the same label.

And here is the expensive one, which the original could not have known in 2016 but which decides the purchase today:

If the car has a diesel or petrol particulate filter, it needs a low-SAPS oil — the ACEA C sequences — and putting a full-SAPS A3/B4 oil in it will block the filter.

SAPS is sulphated ash, phosphorus and sulphur. Every engine burns a little oil; that is normal. The additives in a full-SAPS oil leave an ash residue when they burn, and ash is the one thing a particulate filter cannot deal with — soot burns off during regeneration, ash does not. It accumulates until the filter is finished. This is not a slow degradation over the life of the car; the wrong oil can load a filter in months, and a replacement particulate filter is one of the most expensive single parts on a modern diesel.

Note the trap in that: a low-SAPS C3 oil and a full-SAPS A3/B4 oil are commonly both sold as 5W-30. The grade matches and the oil is still wrong. Anyone choosing oil by the number on the front of the bottle will make this mistake, and the original's advice — read the handbook — is right, but it never says that the handbook is specifying two things.

Synthetic, mineral, and what the words are worth

Base oils are classified by how they are made, in five API groups. The original lists five categories too, but they are not these — it mixes marketing terms with base stocks and invents a group for re-refined oil.

"Synthetic" on a label does not reliably mean Group IV. After a 1999 advertising ruling in the United States found that a Group III oil could be marketed as synthetic, the word became a performance claim rather than a statement of chemistry, and most oil sold as synthetic today is Group III. That is not a scandal — a good Group III is an excellent base oil, and it is why synthetic oil stopped being exotic and became ordinary — but it does mean the word on the front tells you less than the approvals on the back.

What synthetics genuinely give you is real and worth paying for on most modern cars: better cold flow, a higher viscosity index so the oil holds its grade across a wider temperature range, more resistance to oxidation, and lower volatility so less of it evaporates and needs topping up.

One thing in the original's chemistry section should simply be disregarded. It lists the elastomer incompatibilities of polyalkylene glycol base oils — against ABS, PVC, polycarbonate, neoprene and so on. PAG is a Group V industrial base oil used in compressors and refrigeration; it is not used in engine oil at all, because it does not mix with hydrocarbon oils. That passage has been imported from an industrial lubrication reference and does not describe anything you can buy for a car.

The film thickness claim, which is the original's central error

This is the load-bearing sentence of the whole piece, and it is wrong:

It says synthetic oils "sit on engine parts with a thinner film", that they have lower oil film thickness, and that they may therefore fail to lubricate the bearings of an ordinary engine properly.

Film thickness in a bearing is set by viscosity at operating temperature. That is what the grade measures, and it is the same number regardless of what the base oil is made from. A synthetic 10W-40 and a mineral 10W-40 are both 40-grade at 100 °C — that is what the grade means, and both must clear the same high-temperature high-shear minimum at 150 °C.

If anything the comparison runs the other way. Synthetics have a higher viscosity index, so they thin out less as temperature climbs. A mineral oil pushed well past normal operating temperature loses grade faster than a synthetic of the same rating — so under the conditions where film thickness is genuinely at risk, the synthetic is holding more of it, not less.

Where the confusion comes from is real, though, and worth naming, because it is why the claim sounds right. Synthetics are usually sold in thinner grades — 0W-20 and 5W-30 rather than 15W-40 — because a high viscosity index is what makes those grades possible at all. So the shelf genuinely does associate "synthetic" with "thin". But that is the grade, not the chemistry. Buy a synthetic 5W-40 and you have a 40-grade oil. The two properties are printed separately on the bottle precisely because they are separate.

Will synthetic wreck an old engine? Three claims, examined

The original's practical conclusion is that you should not put synthetic in an older engine. It gives three reasons, and they are in descending order of merit.

The seals: true in 1985, not true now

The claim is that mineral oil swells seals and prevents leaks, that synthetics lack that property, and that some of them attack seals outright.

The physics behind it is real: PAO does shrink elastomers slightly, on the order of half a percent, which is exactly why early synthetics earned this reputation. But seal compatibility is governed far more by the additive package than by the base stock, and modern synthetic formulations include ester components and dedicated seal-swell agents specifically to supply the swell that PAO does not. Elastomer compatibility is also part of what an API or ACEA licence tests. A licensed modern synthetic is formulated not to do this.

So this is a genuine historical fact reported as a current one — the archive's most common failure mode, and one that costs readers money by steering them away from a better product.

The deposits: a real mechanism, stated far too strongly

The claim is that switching from mineral to synthetic lifts baked deposits off the internals, they circulate, and "after a while the engine fails."

There is a mechanism here, and it is worth respecting: a genuinely neglected, sludged engine can have material dislodged, and if enough of it reaches the oil pickup strainer the result is oil starvation, which does destroy engines. The oil filter is the other thing that can load up.

But "the engine fails" describes the tail, not the case. On an engine with a reasonable service history there is nothing much to dislodge. And the far more common outcome of switching a high-mileage engine to a modern oil is not failure but a leak becoming apparent — where deposits had been sealing a gasket that had already hardened years ago. The oil did not cause that; it stopped hiding it.

The flush: the original's own procedure is the risky part

Having raised the deposit risk, the original prescribes a flush — drain the old oil without changing the filter, fill with flushing oil, idle for twenty minutes, drain, change the filter, fill with synthetic.

Parts of that are sensible. Leaving the old filter in place during the flush so it catches what comes loose is a genuinely good detail, and the note that the flushing product must suit the engine's alloys is a real consideration.

But twenty minutes of idling on a solvent flush is the aggressive version of this procedure, and it is being recommended to exactly the engines least able to take it — old, high-mileage, unknown history, which is the whole premise of the section. That is the population where a flush is most likely to dislodge a sheet of sludge and least likely to have a clean pickup screen waiting for it.

If an engine is clean enough not to need a flush, it does not need one. If it is dirty enough to need one, the safer route is a series of shorter oil-and-filter changes on ordinary oil, letting the detergent package do the work gradually, rather than one hard solvent cycle. Where there is visible sludge under the filler cap, the honest answer is that the flush is a gamble and the sump should come off instead.

The petrol analogy, and why it misleads

The original explains all of this with an analogy: synthetic, semi-synthetic and mineral are "just like super and regular petrol" — putting regular in a car that wants super causes problems, putting super in a car that wants regular does not.

It is a memorable analogy and it is the source of the article's own confusion, because octane is one axis and oil has two.

Octane is a single scalar: resistance to knock, where more is harmless. Oil is specified along two independent axes that do not substitute for each other — viscosity grade, which must match the engine's design, and specification, where newer is generally backwards-compatible. Collapsing them into one produces exactly the reasoning error in the section above: it invites you to think of synthetic as "higher grade oil", and then to worry that higher grade might be too thin, when grade and base stock were never the same property.

The one-line version: a 0W-20 full synthetic is not a "better" oil than a 15W-40 mineral. It is a different viscosity, and in an engine specified for 15W-40 it is the wrong one.

Changing it: the interval, the filter, and the colour

The original's practical advice is largely sound, and this section keeps most of it while fixing the numbers.

On the interval, it says to change the oil about 1,000 km before the stated figure, because the engine spends time idling in traffic that the odometer never counts. The reasoning is right and better than most guides manage — distance is a proxy for engine hours and a poor one in city driving. The prescription is arbitrary, though: 1,000 km is 20% of a 5,000 km interval and 5% of a 20,000 km one. Two corrections:

• Manufacturers already publish the answer. Nearly every handbook carries a severe service schedule alongside the normal one, and short journeys, extended idling, heavy traffic, dust and towing are exactly what it covers. If your driving fits, that column is the interval — not the normal one minus a guess.
• Oil also expires by calendar. Additives deplete and oil oxidises whether the car moves or not, which is why handbooks give a time limit — commonly twelve months — alongside the distance. A car doing 4,000 km a year is on a time-based schedule, and this is the single most-missed interval there is.

On the filter, it says to fit a new one at every change and never skip alternate ones. That is correct and there is nothing to add. A filter that has been left in is a reservoir of the contamination you just paid to remove, and a collapsed or blocked filter sends oil through the bypass valve unfiltered.

On colour, the original is right in a way that deserves emphasis, because it argues against two opposite folk beliefs at once — that good oil should blacken quickly to prove it is working, and that good oil should stay clean to prove it is durable. Both are wrong, and the correct answer is that oil darkens gradually because the detergent and dispersant additives are holding combustion soot in suspension, which is their job. Soot suspended in oil is soot that is not deposited on anything.

The corollary is worth stating outright: you cannot judge oil condition by looking at it. Dark oil is doing its job. The things that actually end an oil's life — additive depletion, fuel dilution, oxidation, viscosity shear — are all invisible on a dipstick.

On additives, the original says to avoid aftermarket thickeners and not to mix oils. The first is right: a finished oil is a balanced formulation, and pouring in a viscosity improver unbalances it while invalidating the specification the engine needs. The second is right as a rule and slightly too strong in practice — topping up with a different brand of the same grade and specification, to get home or to avoid running low, is fine and always better than running the level down. Low oil level is a real fault with real codes behind it; brand mixing is not.

On re-refined oil, the original says never to use it. That needs reframing rather than repeating. Re-refining is now a legitimate industrial process producing Group II and III base stocks that meet the same specifications as virgin ones, and licensed oils containing them are sold by major brands. What the original is really describing is a market problem, not a chemistry one — reprocessed waste oil sold as new, alongside outright counterfeit product. So the test is not the origin of the base stock; it is the licence. An API or ACEA mark and an OEM approval on the label, bought from a seller with something to lose, is the protection. Suspiciously cheap oil in an unfamiliar container is the risk, whatever the base stock.

What this looks like when it goes wrong

Oil is upstream of a lot of the corpus, which is the useful thing about getting it right:

• Pressure. Genuinely low oil pressure and a faulty sensor present identically on the dashboard and are completely different problems — P0520 through P0524 cover the circuit and the condition separately. Treat the light as real until measured.
• Valve timing. P0011 and P0016, as above. Wrong viscosity, degraded oil or a blocked solenoid screen all starve a cam phaser of the pressure it needs.
• Consumption. An engine that uses oil is a different question from one that leaks it, and head gaskets and oil consumption covers telling them apart. Blue smoke and a dropping level with no puddle is consumption.
• Dilution. Oil diluted with fuel thins the oil below its grade without any of it leaving the sump, and short-journey diesels with regenerating filters are where it happens.
• Noise. A rattle on startup that fades is the sound of oil not being where it should be quickly enough — the reason the winter number exists.
• Turbochargers. A turbo is fed by the same oil at far higher shear and temperature, and it is usually the first thing to notice a neglected interval.

And if you are looking at a car you do not own yet, oil is one of the cheap tells: inspecting a used car covers what the filler cap and the dipstick are worth, and — more importantly — why you should not let anyone clean the engine bay before you look at it.

The short version

Open the handbook and find two things: the viscosity grade and the specification. Buy an oil that carries both, from somewhere reputable. If the car has a particulate filter, the specification is not negotiable and the grade alone will not protect you.

After that, synthetic versus mineral is a question about how long the oil lasts and how it behaves at the extremes — worth answering, worth paying for on most modern engines, and nowhere near as important as the two numbers the original never explained.

Video guides

Video, for the parts of this that are easier watched than read. The English-language ones come first.