What actually ends a cartridge's life
A sediment cartridge fails by filling up: captured particles progressively block its pores, resistance to flow rises, and eventually the system downstream is starved of pressure. That resistance is directly measurable as differential pressure (ΔP) — the difference between the gauge before the housing and the gauge after it.
A clean cartridge typically drops 0.1 – 0.3 bar at rated flow. As dirt accumulates the ΔP climbs slowly, then steeply — the curve is a hockey stick, because each blocked pore forces more flow through the remaining ones, loading them faster.
Change sediment cartridges at about 1 bar differential, or at whatever change-out differential the cartridge's own data sheet recommends. Read that carefully: a data sheet's maximum differential is usually a structural figure, the pressure at which the cartridge collapses, and it sits well above the pressure at which the cartridge should be changed. Running to the structural number is running to failure. Past the change-out point flow falls off quickly, and high ΔP drives captured dirt through the media — unloading, which delivers weeks of collected sediment downstream in one slug.
Why the calendar misleads in both directions
Time-based replacement is a proxy, and it errs both ways. On clean municipal water a cartridge replaced every month may go to the bin 10% loaded — money thrown away. After one storm on a river abstraction, a "monthly" cartridge can blind in three days — and until the scheduled change, the system runs starved. The dirt load decides, and the gauges are how the dirt load speaks.
The practical routine costs thirty seconds: read both gauges weekly, log the differential, and change at the threshold. The log is a bonus diagnostic — a ΔP that climbs faster month on month means the raw water is deteriorating or an upstream stage is failing.
The two exceptions where time rules
- Carbon cartridges (GAC and block). Carbon is exhausted by adsorption, which is invisible to a pressure gauge: a fully spent carbon block flows perfectly while removing nothing. Replace carbon on throughput volume, from the data sheet's rated litres for the contaminant it is claimed against. Where no rated volume is published, six months is a common working interval on domestic duty, but it is a habit rather than a specification: it assumes ordinary household throughput and ordinary feed chlorine, and it is wrong in both directions on a busy site or a lightly used one. Verify with a free-chlorine test after the carbon where it protects a membrane.
- Hygiene-limited duty. On potable point-of-use systems, a cartridge that sits wet for many months becomes a biofilm home regardless of dirt load. Around six months is the usual ceiling applied to drinking-water cartridges even at low ΔP, for that reason rather than on any manufacturer's rating — and change any cartridge that has stood unused in a wet housing for weeks.
Reading the change-out like an engineer
The removed cartridge is data:
- Uniformly brown through the depth — correctly rated, correctly loaded; the system is working.
- Surface-caked but clean inside — the cartridge is too fine for the load arriving; add or coarsen the stage in front.
- Looks clean but ΔP was high — suspect fine clay or colloids blinding the fine pores invisibly, or scale/iron precipitating in the media.
- Local dirty patches — flow bypassing, usually a missing or damaged O-ring seating the cartridge.
Rule of thumb: gauges either side of every housing, change sediment cartridges at about 1 bar differential and never at the data sheet's structural maximum, change carbon on its rated volume, and glance at every cartridge you remove — it is a free report on the system's health.