The short answer
If you are comparing Atlas Copco air filter specifications, do not start with the micron rating. The four numbers that decide whether a filter actually fits your system are the ISO 8573-1 purity class, the pressure drop at rated flow, the element change interval, and the connection size. The micron number on the front of the data sheet is the one everybody quotes and the one that tells you the least on its own.
That is the whole argument. The rest of this piece is why I check those four things in that order, what went wrong when I didn't, and where this checklist stops being useful.
Why you should trust this checklist (and where it comes from)
I'm the office administrator for a 350-person company. I manage facilities and MRO ordering — roughly $410,000 annually across nine vendors, including two compressed air suppliers. I report to operations on uptime and to finance on invoice compliance, which means I get yelled at from both directions.
In 2023 I bought replacement filter elements for our two rotary screw compressors from a reseller who was about 30% cheaper than the OEM path. The elements fit. They ran. Nine months later one of the compressors threw a high-temperature fault during a production week, and the service technician showed me the differential pressure gauge pinned at the top of its range. Nobody had been watching it. Downtime came to about 11 hours across two shifts. Finance also rejected the reseller's invoice because it listed a generic part description that didn't match our purchase order (handwritten delivery note, no proper line-item invoice). I spent a week cleaning that up.
Take that story with the appropriate grain of salt — I'm not an engineer, and I'm not going to pretend I understand the thermodynamics. What I do understand is the procurement side: which fields on a data sheet I need to verify before I sign, and which fields are marketing.
The four specifications that actually decide fit
1. ISO 8573-1 purity class
This is the specification that should be printed on your requirement sheet before you talk to any vendor. ISO 8573-1:2010 classifies compressed air by three separate parameters: solid particles (by size and count per cubic meter), water (as pressure dew point), and oil (as mg/m³). A filter data sheet that says "0.01 micron" is telling you about one particle size at one efficiency point. It is not telling you the class.
Ask for the class in the three-part format — something like "1.4.1" or "2.4.2" — and check it against what your process actually needs. General plant air and instrument air are not the same requirement, and the filter that satisfies one is often overkill for the other. If the vendor can't state the class, that's a signal about how carefully the filter has been characterized.
2. Pressure drop at rated flow
Most buyers focus on the filtration rating and completely miss the pressure drop at the flow the filter will actually see. Filter elements are rated at a specific flow; run more air through them and the pressure drop climbs, which means your compressor works harder to deliver the same pressure at the point of use. This is the number that quietly shows up on your electricity bill.
Industry guidance commonly cited by CAGI (the Compressed Air and Gas Institute) puts each 2 psi of avoidable pressure drop at roughly 1% of compressor energy — worth confirming against CAGI's current best practices document, since the exact figure depends on your system. On a system running two shifts, that is not a rounding error.
3. Element change interval — and the differential pressure indicator
Every filter data sheet gives a change interval in hours or months. Treat it as a starting condition, not a promise. The real replacement trigger should be differential pressure, which is why I now insist on a differential pressure indicator on anything upstream of a production process. Without one, you're changing elements on a calendar and hoping, or not changing them at all, which is what happened to us.
I should add that the interval figures I've seen from Atlas Copco and from other major compressor OEMs are usually conservative for clean plant air and optimistic for dusty or high-humidity environments. Your actual site conditions are the variable the data sheet can't model.
4. Connection size and rated flow
Boring, and the most common source of returns. Confirm the port size and thread standard, the rated flow in the units your maintenance team uses (m³/min or cfm — pick one and convert, don't assume), and the maximum working pressure. I once ordered a filter with the right filtration class and the wrong port thread because the data sheet listed BSPP and our piping was NPT. That cost us three days and a restocking fee. Note to self: this is now a checkbox on our requisition form.
The counterintuitive part: tighter filtration is not automatically better
It's tempting to think the right move is always the finest available filtration class. But tighter filtration usually means higher pressure drop and shorter element life, and on a system that doesn't need the purity, you're paying twice — once for the element and again in compressor energy. The "just buy the best spec" advice ignores the fact that overspecifying filtration is a real operating cost, not a free upgrade.
What was best practice in 2020 may not apply in 2025 either. Five years ago my default was to order the OEM filter element for every housing without thinking, because third-party compatibility was genuinely hit or miss. That has changed. The better third-party element makers now publish ISO 8573-1 test data, and the market has enough credible options that blanket OEM-only buying is no longer automatic. What hasn't changed is the verification step: whatever you buy, get the class and the pressure drop in writing.
Where this checklist stops working
This approach works for standard plant and instrument air on a system with a stable load. It is weaker in a few situations. If your compressor runs intermittently and the filter sees low flow most of the day, rated-flow pressure drop tells you less than you'd think, and you should talk to someone who can model part-load behavior. If the air feeds a specific process — food contact, breathing air, electronics — don't take my checklist as the requirement; that comes from the process owner or the relevant standard, and I'd get the equipment manufacturer to confirm in writing before ordering.
One more boundary: in markets where the local selling entity is a separate legal company — PT Atlas Copco in Indonesia, for instance, or the equivalent national entity elsewhere — invoicing, warranty terms, and lead times can differ from the regional Atlas Copco catalog. Verify the local entity's part numbers and payment terms before you assume the global data sheet applies as-is. That's less about filtration and more about not eating a rejected expense report.
I'm not 100% sure this checklist covers every configuration — it's built from one company's procurement mistakes, not from an engineering manual. But it has kept our compressors running and our invoices approved for the last two years, which is the standard I'm actually measured against.