Specifying a Custom Filter: The Data Sheet That Makes a Quote Buildable

A custom filter project rarely fails at the machine. It fails at the enquiry. A buyer sends a part number, a photo and a quantity; the factory quotes something that looks close; the sample arrives and the thread is right but the bypass valve opens at the wrong pressure, or the element is 4 mm too tall for the housing, or the media is rated at 25 microns when the application needs 10. Two months disappear in a loop of re-quotes.

The fix costs nothing: build a specification sheet complete enough that the factory can either quote it or tell you precisely what it cannot do. Below is that sheet, field by field, with the reasoning for each one and the failure it prevents. It is written for buyers sourcing oil, fuel, air, hydraulic and coolant filters for machinery, fleets and industrial equipment from a Chinese filter manufacturer.

Why the data sheet decides the outcome

A filter is a deceptively simple object: a can, a pleated element, a few valves and a gasket. That simplicity is exactly why ambiguity is expensive. Two filters can share a thread, a diameter and a height and still be non-interchangeable, because the difference sits in a spring rate, a media grade or an anti-drain valve that is invisible from the outside.

When the specification is thin, three things happen. The factory quotes its closest existing product, which may be a compromise. The buyer compares prices between suppliers who are actually quoting different products. And the sample approval turns into the real specification exercise, at sample cost and on sample lead time.

Block 1: Identity and application

Start with what the part has to do, not what it looks like. The application determines the performance envelope before any dimension is fixed.

Treat the reference part number as a lead, never as the specification. Cross-references point to products that are claimed to fit; they do not guarantee that the internal construction matches. A part that fits and a part that performs are different claims.

Block 2: The dimensional envelope

Dimensions are where most first samples go wrong, because buyers tend to specify the obvious ones and omit the ones that actually locate the part. Measure the existing filter with a caliper and record the following.

For cartridge elements, add the end-cap construction: which end carries the gasket, whether the caps are metal or polymer, and whether the element is open at both ends or closed at one. These details decide whether the element seals against dirty-side bypass.

Block 3: The connection interface

The interface is the mechanical contract between the filter and the machine. Get it wrong and nothing else matters.

State the thread standard, not just the thread size. A 1-14 thread can be specified to UN or to another standard with a different flank angle; the parts will start onto each other and then bind. If the original is a metric thread on a European machine and the aftermarket equivalent is UN, that is a specification decision with a consequence, not a formatting detail.

Block 4: Media and filtration performance

Media is the part of the filter that does the work, and the part most often described in one word: “paper”. That word covers cellulose, cellulose-synthetic blends, full synthetic glass microfibre and metal mesh, with efficiency curves that differ by a factor of several.

Where the filter must protect an expensive component — a common-rail injection system, a hydraulic servo valve, a turbocharger — the media specification should come from the component supplier’s cleanliness requirement, not from the filter being replaced. Match the filter to what the machine needs, then check that the old filter met it.

Block 5: Valves and internal hardware

Valves are small, cheap, invisible and decisive. A filter with the correct media and the wrong bypass valve will either starve the engine of oil at high load or send unfiltered oil through the circuit at every cold start.

If you do not know the cracking pressure of the original valve, do not estimate it. Send the sample and have the factory measure it, then record the measured value on the approved drawing. That measurement becomes the reference for every future order — the only reliable way to keep a valve setting stable across batches and across suppliers.

Block 6: Pressure, temperature and flow

These three numbers define the service environment. They also decide wall thickness, housing material and seal compound, so they belong on the sheet even when the part looks identical to the one being replaced.

The interaction between rated flow and bypass cracking pressure is where field failures hide. A filter that is correct on paper can run permanently close to its bypass threshold if the flow rating was taken at a lower flow than the machine actually delivers, and the symptom is not a blocked filter — it is an engine that wears faster than it should while every service record looks normal.

Block 7: How the part will be verified

A specification without an acceptance method is an opinion. Decide before the first sample how you will judge it, and put that in the sheet.

• Dimensional report — which dimensions are checked, with what instrument, and against what tolerance.
• Performance tests — efficiency per ISO 16889 or the relevant ISO 4548 part, collapse and burst pressure, valve cracking pressure. Name the standard and the lab.
• Material declaration — media, housing steel grade, seal compound, and any RoHS or REACH statement the destination market requires.
• Traceability — batch code position, and how a delivered filter can be tied back to its production record.
• First-article inspection — a full dimensional and visual report on the first production batch, not just the sample.

Block 8: Marking, packaging and documents

For private label and aftermarket programmes, packaging is part of the product, and it is where a technically correct filter arrives looking like a counterfeit. Specify it with the same precision as the thread.

What to send before the first sample

Before asking for a custom sample, assemble one package. Factories that receive it can usually respond with a feasibility answer and a price in days rather than weeks.

1. Two original filters if available — one to keep, one to cut open.
2. Photographs of the filter from at least four sides, plus the housing it fits into.
3. The completed data sheet from blocks 1 to 8, with unknowns marked as unknown.
4. The machine or engine identification plate, photographed.
5. Your acceptance criteria: the tests you will run and the tolerances you will hold the supplier to.
6. Packaging artwork or a reference sample of how the product should present on shelf.
7. Target quantity and target landed cost, so the factory can propose a route rather than a guess.

Cutting open a competitor’s filter is one of the fastest ways to build the media and valve sections of the sheet. Pleat count, media area, valve type and end-cap construction are all visible in section, and they remove most of the guesswork from block 4 and block 5.

Frequently asked questions

Can a factory work from a sample alone, without a data sheet?

Yes for a straightforward copy, if the sample is representative and the factory is told which characteristics must be held. It is weaker for performance-critical parts, because a copy reproduces what can be measured on a bench and not what the original was designed to achieve. Use the sample to fill the sheet, not to replace it.

How accurate do the dimensions need to be?

Accurate enough to state a tolerance. A dimension without a tolerance is a wish. For most automotive and industrial filters, ±0.5 mm on diameter and ±0.8 mm on height is workable; sealing faces and threads need tighter control and should be called out separately on the drawing.

What if the OEM part number cannot be found?

Then the part number was never the specification. Fall back to measurement and application: machine model, the circuit the filter serves, and the media and valve requirements that follow from it. A factory with cross-reference experience can often identify the family from photographs and dimensions alone.

How long does a custom filter sample usually take?

For a part built on an existing body and element family, sample preparation is commonly a matter of days. Where new tooling, a new end-cap mould or a new media grade is required, the timeline is set by tooling rather than assembly and should be quoted as a separate milestone. Ask which route your part falls into before comparing sample lead times between suppliers.

Should the data sheet include price?

Include the quantity band and the target landed cost, not a single price expectation. A factory that understands the volume and the cost target can propose a construction that meets both; a factory given only a price will meet the price by changing something you did not specify.