Cut open two oil filters with the same part number and the difference is usually obvious in ten seconds. One contains two valves and a deep pleat pack; the other contains a spring, some paper and a lot of empty space. Both fit the thread.
The two valves that do the important work
The anti-drainback valve is a rubber diaphragm at the inlet holes. When the engine stops, it stops oil draining out of the filter back into the sump. Without it the filter empties overnight, and the next cold start runs the engine on unpressurised bearings for a second or two while the filter refills.
Cheap filters either omit it or use a nitrile diaphragm that hardens after a few heat cycles and stops sealing. Silicone stays flexible far longer and is what a quality filter uses — the difference is visible on a cut-open sample, and it is a fair thing to ask a supplier to specify in writing.
Not every engine needs one. Filters mounted vertically with the opening upward drain less, and some manufacturers omit the valve deliberately. Copying a filter and adding or removing the valve arbitrarily is the problem, not its presence or absence per se.
The bypass valve is a spring-loaded relief. When the filter media is blocked — cold thick oil at start-up, or a filter left in far too long — the valve opens and lets oil past unfiltered. Unfiltered oil is bad; no oil is catastrophic. The valve exists so the engine survives the trade.
Its opening pressure is engine-specific and typically in the region of 0.8–1.5 bar differential. Set too low, the filter bypasses under normal conditions and filters nothing. Set too high, it fails to protect when the media blocks. A supplier who cannot state the opening pressure is not controlling this component.
The media
Three things describe filter media, and only the first appears on most datasheets:
- Micron rating. Usually quoted as nominal — the size at which the filter removes a stated proportion of particles. "20 micron nominal" and "20 micron absolute" describe very different filters, and quoting the nominal figure without saying so is common.
- Efficiency at a stated particle size, expressed as a percentage. This is the more meaningful number, and the one to ask for.
- Dirt holding capacity — how much contaminant the filter absorbs before it blocks. This determines whether the filter survives its service interval, and it correlates strongly with pleat area.
Pleat count and pleat depth are the visible proxy for capacity. A filter with 60 deep pleats holds far more than one with 30 shallow ones in the same can, and you can count them on a cut-open sample without any equipment.
Construction details that fail in service
End cap bonding. The pleat pack is sealed to metal or fibre end caps. If the adhesive fails, oil flows around the media instead of through it and the filter is doing nothing at all — with no external symptom. This is the most consequential invisible defect in the category.
Centre tube strength. The perforated tube inside resists collapse under differential pressure. A thin tube can collapse when cold oil forces its way through, taking the media with it.
Can seam and burst pressure. Oil systems see pressure spikes well above normal running pressure. Filters are burst tested, and a figure should be available from the supplier.
Gasket material and retention. A gasket that swells in hot oil or falls out of its groove during handling causes a leak that the customer notices immediately and blames on the fitter.
A cut-open inspection anyone can run
Take one filter from each batch, cut the can open with a filter cutter, and check:
- Is the anti-drainback valve present, and does the rubber still feel flexible?
- Is there a bypass valve, and does the spring have real tension?
- Count the pleats and measure their depth. Compare against the reference sample kept from the approved first article.
- Pull gently at the end cap seal. It should not separate.
- Check the centre tube for adequate wall thickness and clean perforations.
Keeping a cut-open reference sample from the approved sample batch is the single most useful quality control step in this product category. Every later batch is compared against a physical object rather than against a specification nobody reads.