Qinghe Air Filter

Filter efficiency grades: EN 1822, ISO 16890 and MERV compared

Standards

Different filter families carry different efficiency scales: EN 1822 grades HEPA/ULPA by MPPS penetration, ISO 16890 grades HVAC media by particle-size fraction (ePM1/ePM2.5/ePM10), and MERV grades by composite arrestance across particle bands. Read the scale in its own terms rather than treating the labels as interchangeable.

EN 1822 for HEPA and ULPA

EN 1822 defines HEPA and ULPA grades by penetration at the most penetrating particle size. The ladder runs H10–H12 (EPA), H13–H14 (HEPA) and U15–U17 (ULPA). This scale is used for cleanroom terminal filters and critical environments, where the question is how little penetrates at the worst particle size, not how much is arrested on average.

ISO 16890 for HVAC media

ISO 16890 grades HVAC filters by their removal of particle-size fractions — ePM1, ePM2.5, ePM10 — plus a coarse class. It replaced the older EN 779 G/M/F ladder for general ventilation. The value of this scale is that it ties a filter to a real-world particle fraction, which makes it easier to reason about occupant exposure than an abstract average.

MERV for a composite view

MERV, defined by ASHRAE 52.2, assigns a single number across a band of particle sizes, from coarse pre-filters up to high-efficiency grades. It is common in North American specifications. Because each scale measures a different thing, a direct one-to-one translation between MERV, ISO 16890 and EN 1822 is approximate; confirm the governing standard for the region and the equipment before quoting.

From a sourcing perspective

From a sourcing perspective, we recommend reading each efficiency scale in its own terms and confirming the governing standard for the region and the equipment before quoting. Our advice is to treat any one-to-one translation between MERV, ISO 16890 and EN 1822 as approximate, because each scale measures a different thing. In our experience, the practical win is to anchor a specification to the standard the supplier actually tests against, then compare like-for-like rather than chasing an abstract higher number.

Sources