Blog ·
Water Filter Cartridge Quality Inspection: AQL & Tests | XZH
Cartridge QC for importers: defect classes, AQL sample sizes, on-site hydro, pressure-drop, chlorine and RO rejection tests, plus PO clauses. Get a quote.
Short answer: Generic pre-shipment inspection checklists miss the failures that matter in water filter cartridges, because the defects that cost importers money are inside the cartridge - short-filled media, O-ring bypass, carbon fines, weak end-cap bonds. A cartridge-specific inspection uses zero accepted critical defects, AQL 1.0 for major, 4.0 for minor, samples per SKU lot (200 pieces for a 3,201-10,000 pc lot under Z1.4 Level II), and adds five quick on-site tests: weight against the golden sample (±3%), pressure drop at rated flow, hydrostatic hold on sealed cartridges, a chlorine spot-check and a turbidity check after the labelled flush volume. One inspector man-day (USD 268-399 in Ningbo) covers two to three SKUs.
We manufacture PP, carbon block, UF and RO cartridges in Ningbo and run these checks on our own line every shift. This guide is written from the factory side of the inspection table - including the shortcuts other factories take, so your inspector knows where to look.
Why Generic Pre-Shipment Checklists Miss Water Filter Defects
Search for an inspection checklist and you get the same template used for T-shirts and phone cases: quantity, appearance, dimensions, labels, a drop test. Those checks still matter. But they are built around defects you can see, and a water filter cartridge hides its defects.
Three things make cartridges different:
- The media is sealed inside. A 10-inch GAC cartridge with 260 g of carbon looks identical to one with 320 g. The only external evidence is weight.
- A drop-in cartridge is not a pressure vessel. It sits inside a housing, so a classic “leak test” on the cartridge alone proves nothing. The real failure mode is bypass - water flowing around the media past a badly sized O-ring or a warped end cap - and you only see it in a reference housing.
- Performance defects look perfect. A cartridge with low-iodine carbon or a mislabelled micron rating passes every visual check and fails in your customer’s kitchen three weeks later.
So a cartridge inspection needs a scale, a reference housing, a small flow rig and two cheap test kits alongside the usual AQL visual check. None of it requires a laboratory, and all of it fits in a normal inspection day.
Defect Classification for Water Filter Cartridges
Before any inspection, agree a defect list with the factory in writing. This is the table we recommend for cartridges and inline filters. Adjust the numbers per SKU, but keep the structure.
| Class | Defect | Why it’s in this class |
|---|---|---|
| Critical (0 accepted) | Any leak on a sealed cartridge (inline, quick-change, Everpure-type) during hydrostatic hold | Water damage claim in the field |
| Critical | Bypass in reference housing: O-ring missing, wrong size or cut; end cap not seated | Unfiltered water, no visible symptom |
| Critical | End-cap separation or visible gap around the bond line | Media escapes into the water line |
| Critical | Foreign matter inside (metal, insects, mould), or strong solvent / adhesive odour | Health and brand risk |
| Critical | Wrong model, micron rating or flow arrow on label | Misapplication, compliance exposure |
| Major (AQL 1.0) | Media weight below golden sample -3% | The #1 cost-cutting lever |
| Major | Pressure drop above spec at rated flow | Low flow complaints, early change-outs |
| Major | Turbidity above 1 NTU after the flush volume printed on the label | ”Black water” returns |
| Major | Chlorine spot-check outlet above 0.1 mg/L (carbon SKUs) | Media grade or quantity wrong |
| Major | Dimension out of tolerance affecting fit (e.g. 10” length outside 248-251 mm) | Won’t fit competitor housings |
| Major | Wet-pack RO membrane: pouch damaged or dry | Membrane may be irreversibly damaged |
| Minor (AQL 4.0) | Scuffs, light flash, label bubbles, ink rub | Cosmetic only |
| Minor | Polybag seal crease, carton print smudge | Cosmetic only |
The single most important line is the first Major: media weight against a golden sample. It is where the money is, it is invisible, and it takes ten seconds per piece to check.
Looking for a reliable water filter manufacturer?
Get free samples and a factory-direct quote. MOQ from 500 pcs.
Get a Free Quote →AQL Sample Sizes: How Many Cartridges to Pull
Most inspections use ANSI/ASQ Z1.4 (technically equivalent to ISO 2859-1), General Inspection Level II, single sampling, normal inspection. The numbers below are straight from the Z1.4 master tables. “Ac” is the maximum defects you accept; at “Re” you reject.
| Lot size per SKU | Code | Sample size | Major AQL 1.0 (Ac / Re) | Major AQL 2.5 (Ac / Re) | Minor AQL 4.0 (Ac / Re) |
|---|---|---|---|---|---|
| 501-1,200 | J | 80 | 2 / 3 | 5 / 6 | 7 / 8 |
| 1,201-3,200 | K | 125 | 3 / 4 | 7 / 8 | 10 / 11 |
| 3,201-10,000 | L | 200 | 5 / 6 | 10 / 11 | 14 / 15 |
| 10,001-35,000 | M | 315 | 7 / 8 | 14 / 15 | 21 / 22 |
Three rules we see importers get wrong:
- A lot is one SKU, not one container. A 40ft high-cube typically carries 50,000-80,000 standard 10-inch cartridges across several SKUs. Twelve SKUs of 4,000 pieces is twelve lots of 200 samples each - not one lot of 48,000 with 500 samples.
- Critical means zero. Z1.4 has no “AQL 0” column; you simply state in the PO that one critical defect in the sample rejects the lot.
- Functional tests use a smaller sample. A hydrostatic hold or an RO rejection test takes 10-15 minutes per piece. Run them on Special Level S-2 or S-3 - that is 8-20 pieces for most cartridge lots - and run the visual and weight checks on the full Level II sample.
We recommend AQL 1.0 for major defects instead of the consumer-goods default of 2.5. At a 200-piece sample, 2.5 lets 10 underweight cartridges pass; 1.0 stops at 6. On a product that goes into a drinking water line, the tighter plan is worth it.
On-Site Tests an Inspector Can Actually Run
Everything here uses equipment any competent filter factory already has on its QC bench. Ask for it in the inspection booking; if the factory cannot provide a flow rig and a reference housing, that tells you something about their own in-line QC.
| Test | Applies to | Method | Typical pass limit | Time / piece |
|---|---|---|---|---|
| Weight check | All; critical for GAC, PP, CTO | 0.1 g scale, compare to signed golden sample | Within ±3% of golden sample | 10 s |
| Dimensions | All | Calipers + go/no-go gauge for threads | e.g. 10” slim: length 248-251 mm, OD 63-65 mm | 30 s |
| Bypass check | Drop-in cartridges | Install in reference housing, run at rated flow, check seal areas and outlet clarity | No bypass, O-ring seated 360° | 2 min |
| Hydrostatic hold | Sealed cartridges: inline, quick-change, Everpure-type | Pressurise to 0.8 MPa (116 psi), hold 5 min | No leak, no pressure drop on gauge | 6 min |
| Pressure drop | All | Flow rig at rated flow, e.g. 3.8 L/min (1 GPM) for 10” slim | Per SKU datasheet, e.g. PP 5 µm ≤0.14 bar, GAC ≤0.21 bar, CTO 5 µm ≤0.35 bar | 3 min |
| Fines / turbidity | GAC, CTO, composite | Flush the volume printed on the label, then sample outlet | ≤1 NTU (or clear against white background) | 3 min |
| Chlorine spot-check | Carbon SKUs | Feed ~2 mg/L free chlorine, DPD kit on outlet at rated flow | Outlet <0.1 mg/L | 5 min |
| RO rejection & flow | RO membranes | 250 ppm NaCl, 50 psi, 25 °C, 10 min warm-up, TDS meter on feed and permeate | Rejection ≥96%; flow within ±20% of rated GPD | 15 min |
| UF air-hold | UF modules | Wet fibres, pressurise to 0.2 MPa, isolate 5 min | Decay ≤0.01 MPa | 7 min |
A few notes from our own QC bench:
- The flush-volume limit must match your label. If your packaging tells end users to flush 5 L before use, the inspection limit is “clear after 5 L” - not “clear immediately”. GAC always sheds some fines on first use. Write the test to match the promise.
- The chlorine spot-check is not a capacity test. NSF/ANSI 42 challenges cartridges with 2.0 mg/L chlorine over their full rated volume in an accredited lab. On site you are only checking that a fresh cartridge performs like the golden sample. A CTO block that lets chlorine through at 2 L has the wrong carbon or not enough of it.
- RO test conditions must be written down. The 1812-size residential membranes are rated at 250 ppm NaCl, 50 psi and 77 °F (25 °C). Test at 40 psi and 500 ppm and you will “fail” a good membrane. See our RO membrane sourcing guide for how rejection ratings are defined.
For material-level background on what these numbers mean, see our guides to PP sediment cartridges, CTO carbon block cartridges and UF membrane cartridges.
Where Factories Cut Corners (And How to Catch Each One)
We quote against these shortcuts every week. When a price is 15-25% below ours for the “same” cartridge, the gap is almost always in this table.
| Shortcut | What the factory saves | How your inspector catches it |
|---|---|---|
| Short-filled GAC (e.g. 260 g instead of 320 g in a 10” slim) | 15-20% of media cost | Weigh every sampled piece against the golden sample |
| Light PP melt-blown (e.g. 80 g instead of 120 g for 10” 5 µm) | Resin cost scales with grams | Weight check; light cartridges also compress by hand |
| Low-iodine or coal-based carbon sold as coconut | USD 0.8-1.5 per kg of carbon | Chlorine spot-check, odour, supplier COA with iodine value (coconut typically ≥1,000 mg/g) |
| NBR O-rings instead of EPDM or silicone | A few cents per ring | Material cert, hardness check (typically 70 ±5 Shore A), swell after hydro test |
| Rushed adhesive cure on end caps | Line speed | Tacky bond line, solvent odour, cap lifts under firm hand twist |
| Reworked or returned stock mixed into new lots | Scrap cost | Inconsistent date codes and weights within one carton |
| Nominal micron sold as absolute | Nothing - it’s a labelling claim | Pressure drop far below what a true absolute rating would produce |
None of these shortcuts is visible in a photo. All of them show up on a scale, a flow rig or a DPD kit within an hour. If you are also comparing supplier quotes, our water filter price guide shows where realistic price floors sit for each cartridge type.
When to Inspect: DUPRO, PSI and Loading Supervision
| Inspection | When | Best for | Typical cost (Ningbo) |
|---|---|---|---|
| First article / pre-production | First 10-20 pieces off the line | New SKUs, new moulds, any private-label first order | Often done by the factory with photos + samples couriered to you |
| During production (DUPRO) | 20-30% of order complete | Carbon block, composite, UF - defects decided at moulding and capping | USD 268-399 per man-day |
| Pre-shipment inspection (PSI) | ≥80% packed | Every order | USD 268-500 per man-day |
| Container loading supervision | At loading | Mixed-SKU containers, first orders, retail-ready packaging | Half to one man-day |
For carbon block and composite cartridges, DUPRO is the inspection that saves money. Density and fines are set when the block is moulded; once it is end-capped and bagged, a defect means scrapping the whole cartridge. For simple PP sediment cartridges, a PSI alone is normally enough.
On a container worth USD 40,000-80,000, two inspection days cost about 1-2% of goods value. Our landed cost guide shows how that compares with the other line items in your import budget.
What to Write Into Your PO and QC Specification
An inspector can only fail what the paperwork defines. “Good quality” and “same as sample” are not enforceable. For each SKU, attach a one-page QC spec that states:
- Dimensions and tolerances - length, OD, ID, thread standard (NPT vs BSP), with a go/no-go gauge reference
- Media - type, source (e.g. coconut shell GAC), minimum weight per cartridge, iodine value where relevant
- Pressure drop - maximum at a stated flow rate (e.g. ≤0.21 bar at 3.8 L/min)
- Hydrostatic test - pressure and hold time for sealed cartridges (e.g. 0.8 MPa, 5 min, no leak)
- Chlorine and turbidity limits - test method, feed concentration, flush volume, outlet limit
- RO membranes - rated GPD, test conditions (250 ppm NaCl, 50 psi, 25 °C), minimum rejection 96%, flow tolerance ±20%
- Seals - O-ring material (EPDM or silicone), hardness, food-grade lubricant
- Packaging - sealed polybag, sanitary end caps, carton marks, lot code format on unit and carton, and for RO membranes an intact wet-pack pouch
- Sampling plan - Z1.4 Level II; critical 0, major AQL 1.0, minor AQL 4.0; S-3 for functional tests
- Change control - no change of media supplier, resin, O-ring or adhesive without written approval
- Golden sample - signed and sealed, one kept by each side, replaced every 12 months
This is also the document that makes private-label programmes scale. When you add a second factory later, the QC spec is what lets you prove both are building the same product. If your packaging carries NSF claims, align the spec with the certified build - our certification guide explains which standards apply to which claims.
What a Good Factory Hands You Without Being Asked
From the factory side, here is what we think a buyer should receive with every lot. If a supplier cannot produce these, a third-party inspection becomes your only QC system rather than a verification of theirs.
| Document / record | What it proves |
|---|---|
| Lot-level weight log (sampled every hour) | Media fill was controlled across the whole run, not just the inspected cartons |
| In-line hydrostatic or air-leak records for sealed cartridges | Every piece was tested, not only the inspection sample |
| Carbon and resin COAs per batch | Media grade matches the spec, including iodine value |
| RO / UF bench test summary with test conditions | Performance was measured under the stated conditions |
| Material declarations for food-contact parts | Supports FDA 21 CFR / EU food-contact compliance claims |
| Photos of retained golden sample next to production sample | Both sides are comparing against the same reference |
On our own line, every sealed quick-connect and Everpure-compatible cartridge goes through a pressure test before packing, and media weight is logged hourly. Those records go into the shipment folder with the COAs. A buyer’s inspector then verifies our records instead of starting from zero, which is how an inspection day covers three SKUs instead of one.
After the goods arrive, storage conditions matter too. Carbon cartridges absorb moisture and odours in a humid warehouse - see our shelf life and storage guide. If field complaints do come in, our cartridge troubleshooting guide maps each symptom back to a likely production cause.
Related Guides
- OEM Water Filter Sourcing: 8 Steps from Spec Sheet to First Container
- How to Source Water Filter Cartridges from China OEM Factories
- Coconut Shell Activated Carbon: Why Premium Carbon Filters Better
- NSF 42 vs 53 vs 58 vs 401: Which Certification Do You Need?
- Water Filter Cartridge Logistics: From Ningbo Factory to Global Warehouse
Request a Quote With a QC Spec Attached
Send us your target SKUs and we will return a quotation together with a draft QC specification for each one - media weights, pressure-drop limits, test methods and the AQL plan - so your inspector has something enforceable on day one. MOQ from 500 pcs per SKU, free samples for qualified buyers, and our lot records ship with every order.
Frequently Asked Questions
- What AQL should I use for water filter cartridge inspection?
- Most third-party inspections default to ANSI/ASQ Z1.4 (equivalent to ISO 2859-1), General Inspection Level II, with AQL 2.5 for major and 4.0 for minor defects. That default was written for consumer goods where a major defect is a cosmetic or functional nuisance. For water filter cartridges we recommend zero accepted critical defects (any leak, bypass, end-cap separation, foreign matter or wrong label), AQL 1.0 for major defects such as media underweight or pressure drop out of spec, and AQL 4.0 for minor cosmetic defects. Run the slow functional tests - hydrostatic hold, chlorine spot-check, RO salt rejection - on a smaller special-level sample (S-2 or S-3, typically 8-20 pieces per SKU) rather than the full visual sample.
- How many cartridges should an inspector check in a pre-shipment inspection?
- Sample size depends on the lot size of each SKU, not on the size of the container. Under Z1.4 General Level II, a lot of 1,201-3,200 pieces needs 125 samples, 3,201-10,000 pieces needs 200, and 10,001-35,000 pieces needs 315. A 40ft container with 12 SKUs of 4,000 pieces each is therefore twelve lots of 200 samples, not one lot of 48,000. In practice one inspector man-day covers two to three SKUs with full functional testing, so budget inspection days per SKU, not per container.
- Can a factory-floor inspection verify NSF chlorine or lead reduction claims?
- No. An on-site inspection can verify that the cartridge is built the same way as the certified or approved sample: same media weight, same pressure drop, same chlorine spot-check result, same materials. It cannot replicate an NSF/ANSI 42 or 53 capacity test, which runs thousands of litres of spiked challenge water through the cartridge in an accredited lab. What a spot-check does catch is the most common production failure - a cartridge that passes chlorine at 2 L but has a fraction of the carbon it should have. Use the on-site test to confirm consistency and the lab report to support the claim.
- What are the most common defects in water filter cartridges from China?
- In our experience the defects that reach importers are rarely cosmetic. The top five are: short-filled GAC or light PP melt-blown cartridges (the cheapest way to cut cost, invisible from the outside); end-cap bonding failures from rushed adhesive cure; O-ring bypass from wrong O-ring size, hardness or material; carbon fines that do not clear after the flush volume printed on the label; and mislabelled micron rating or flow direction. All five can be caught on site with a scale, a reference housing, a flow rig and a turbidity check - none of them require a laboratory.
- How much does a third-party pre-shipment inspection for water filters cost?
- In the Ningbo and Yangtze River Delta area, independent inspection companies typically charge USD 268-399 per man-day, and the multinational labs USD 300-500 per man-day, usually with no travel surcharge because the area is a major manufacturing hub. One man-day covers roughly two to three cartridge SKUs with full AQL visual inspection plus functional tests. On a container worth USD 40,000-80,000, a two-day inspection is about 1-2% of goods value, which is less than the cost of re-labelling a single pallet in your own warehouse.
- When should I inspect carbon block cartridges - during production or before shipment?
- For carbon block (CTO) and composite cartridges, book a during-production inspection when 20-30% of the order is complete, in addition to the pre-shipment inspection. Block density, fines and end-cap bonding are decided at the moulding and capping stage, and once a block is end-capped and bagged a defect means scrapping the whole cartridge. Catching a density or fines problem on the first 20% of output costs a line adjustment; catching it at 100% packed costs the order. Simple PP sediment cartridges can usually go straight to a pre-shipment inspection.
- What should I write in the purchase order to make QC enforceable?
- Write numbers, not adjectives. For each SKU the PO or attached QC spec should state: dimensions with tolerances, media type and minimum weight, pressure drop limit at a stated flow rate, hydrostatic test pressure and duration for sealed cartridges, chlorine spot-check method and limit where applicable, RO rejection and flow at stated test conditions, O-ring material and hardness, label and packaging requirements, the AQL plan, and a clause that no material, supplier or BOM change is allowed without written approval. Also require a signed golden sample kept by both sides. An inspector can only fail what the paperwork defines.
Looking for OEM Water Filter Cartridges?
Get a factory-direct quote from Ningbo XZH — 15+ years OEM experience, 50+ export markets, MOQ from 500 pcs.
Get a Free Quote →Get Our Full Product Catalog
100+ SKUs with specs, MOQ, and factory-direct wholesale pricing — everything covered in this article and more.