Guide to choosing a filter medium

The four questions that decide a medium

Ask them in this order: each rules out candidates the next would not have.

What is the service temperature, and its peak? Not the average: the peak, and how long it lasts. A medium run continuously below its limit holds; the same one, subjected to repeated excursions, ages abruptly. And it is not the medium alone that sets the element’s limit — see below.

What is the chemical nature of what passes through? The fluid, but also condensates, additives, cleaning-in-place products, and the ambient atmosphere. This is the most decisive question, and the one we do not settle in a table — the paragraph that follows says why.

What is the nature of the particle? Its size, but above all its behaviour: abrasive, it wears the medium by friction and shifts the choice towards mechanical strength; sticky or hygroscopic, it plates on and stops releasing, and the answer is then a surface treatment rather than a different material; fibrous, it bridges the openings and clogs faster than its size suggests.

What are the mechanical constraints of the installation? Cleaning method — pulse, reverse air, shaking — admissible differential pressure, vibration, and whether or not there is a support. A medium perfectly suited to the fluid tears if it works without a restrainer basket.

How to read the table above

Three cautions, and they come from a real mistake this project paid for.

The value shown is a maximum limit published by a manufacturer for a given medium. It is not a recommended service temperature, it is not a guarantee, and it is not the limit of the finished element.

A maximum limit is not the lower bound of a range. Written out like that it seems obvious; yet it is the error found twice in our own catalogue during its source audit, where a “maximum operating temperature” line from a media table had been published as the minimum of a service range. We write it here because the same line of the same documentation will be read by others.

A table row can cover several media. Manufacturers’ documents often present a temperature row with several material columns: reading the row without reading the column attributes one medium’s value to another. Every value in the table above carries its provenance mark: open the document, and check the column.

Finally, the limit of the complete element is the lowest among those of the medium, the frame, the adhesives and the gaskets. A medium with plenty of margin may sit in a cartridge whose adhesive gives way well before.

Chemical compatibility cannot be read from a table on a website

We do not publish a medium-by-fluid compatibility matrix, and that is not a gap: it is a decision.

Such a matrix would be wrong by construction. The compatibility of a material with a product depends on concentration, temperature and exposure time taken together — the same material withstands a dilute acid when cold and degrades in a concentrated one when hot. A table cell does not carry those three dimensions. It carries a “compatible” that will be read as a guarantee, and one wrong line leads to choosing a medium that dissolves in a customer’s process.

What we do instead, and what gives a defensible answer: we start from the product safety documentation you send us, cross it with the compatibility table of the medium’s manufacturer — the one who stands behind its own material — and decide case by case, telling you which document we relied on. When the documents do not allow a conclusion, we say that too.

Seven media in the referential carry no value, deliberately

The table leaves cells empty. None is an oversight, and none carries a dash: here is why, material by material.

For polyamide, the values found came from resellers and contradicted each other from one site to the next, by a factor of about two. Taking one of them would have been choosing at random while giving the choice the appearance of a source.

For the metallic materials — stainless steels, zinc-coated steel, brass, bronze — no filter media documentation consulted publishes a temperature limit. That is an expected result and not insufficient searching: these materials are characterised by their grade and their mesh, and their service limit is that of the assembly, not of the wire.

For activated carbon, the only value that came close actually described a polyester substrate with a membrane, not the carbon itself. Quoting it would have attributed one material’s characteristic to another — exactly the defect described above.

The general rule, and it holds beyond this table: the field is absent rather than filled with a plausible value. A dash or a “not stated” would read as missing data to be filled in; an empty cell says that no document attests it, which is the exact information.

What this page does not say

It does not choose your medium. It gives you the order of the questions, the values we can defend document in hand, and the method for handling the one we do not publish. Describe your process and we will walk the rest of the way with you.